To develop a microbial cell factory, the selection of the promoter elements required for driving heterologous protein expression is crucial. At present, only very few promoters have been identified in non-conventional yeast and their metabolic regulation is not fully elucidated. Engineered promoter elements are an integral part of synthetic biology and metabolic engineering and are used to enhance the titers of homologous and heterologous proteins. Eukaryotic core promoter elements are the crucial region for the binding of transcription factors and other regulatory factors involved in transcriptional control. So the development of suitably engineered promoters in organisms with poorly defined genetic tools is extremely important for making these organisms as industrial production hosts for fuels and chemicals. The synthetic promoter approach has been developed quickly in S. cerevisiae compared to other non-conventional yeast. However, recently a large number of genome sequences of non-conventional yeasts are available (e.g., Y. lipolytica) and have boosted both basic and applied research to understand the genotypic and phenotypic features and further development of metabolic engineering tools for biofuel production. Most of the promoter engineering work has been concentrated on upstream regulatory sequences (Hartner et al., 2008; Xuan et al., 2009), engineering core promoter elements (Blazeck et al., 2012), and/or by creating random mutations in core promoter elements (Berg et al., 2013), and this will lead to tight and tuneable control over a metabolic regulatory pathways (Teo and Chang, 2014).

Along with the core promoter elements, upstream elements such as binding sequences for transcription factors can also be modified to increase the strength of the promoters. For example, TetR protein, which is a well-studied and extensively used protein in bacterial molecular biology, retains its binding to DNA control elements in yeast and sensitive to synthetic inducers, such as anhydrotetracycline or doxycycline. So promoter with TetR-binding sites has been used to control different genes in different yeasts (Blount et al., 2012b). Still, non-conventional yeast promoters are poorly characterised, and findings from S. cerevisiae can be applicable to the most the highly efficient non-conventional yeast systems also because most of the yeast promoters are capable of cross recognition (Van Ooyen et al., 2006).

The commercial K. lactis uses inducible promoter PLAC4, which is a lactose inducible promoter for recombinant protein production, and the proteins are produced into the culture fluid, which makes the recombinant protein purification easy. Highly efficient protein secretion machinery and the high biomass attained in submerged culture condition makes K. lactis an attractive cell factory for heterologous protein production (Van den Berg et al., 1990; Gellissen and Hollenberg, 1997). A hybrid promoter approach was established in K. lactis by combining core promoter elements of Trichoderma reesei cellobiohydrolase1 (cbh1) to the β-galactosidase (lac4) promoter of K. lactis for increasing recombinant protein production in the yeast (Madhavan and Sukumaran, 2014). This may open up the possibilities of manipulation of core promoter elements across classes of the organisms by introducing important core promoter sequences to create highly active synthetic promoters.

Pichia pastoris is a widely used protein production platform for enzymes and pharmaceuticals (Vogl et al., 2013). Highly active methanol-inducible promoter, alcohol oxidase 1 gene (pAOX1), is the most widely used and tightly controlled promoter in P. pastoris. This has been well studied in terms of promoter control elements and different transcription factor. Random mutated library of pAOX1 has been constructed and resulted in less glucose repression and increased activity compared to wild pAOX1 (Berg et al., 2013). The constitutive pGAP (glyceraldehyde-3-phosphate dehydrogenase promoter) served as a scaffold, and this resulted in the development of a series of promoters with a range of activity from 0.01- to 19.6-fold of wild-type pGAP activity (Qin et al., 2011). Blazeck et al. (2012) reported that the combination of core promoter elements with synthetic upstream activator sequence could increase the expression level.

The field of synthetic biology witnessed a transformation by the advent of CRISPR–Cas (Clustered Regularly Interspaced Short Palindromic Repeat) system in several host organisms (Cong et al., 2013). CRISPR-mediated RNA-guided control of gene using nuclease-deficient Cas9 (dCas9) was reported to control and regulate the expression of genes inside the yeast, with synthetic single stranded-guide RNAs (sgRNAs) targeting the genes to silence on a genome. This technique helped to create a series of simultaneous targeted integrations and double- stranded breaks in S. cerevisiae rapidly and efficiently (Bao et al., 2015). The dCas9 was fused to Mxi1, a protein that involved in the attraction of histone deacetylase Sin3p homolog, a component of gene-silencing complex in yeast. Further, the dCas9 along with sgRNAs repressed the TEF promoter of S. cerevisiae about 10-fold, Mxi1 (fusion protein) could downregulate the promoter by 53-fold (Gilbert et al., 2013). This technique is also been applied to an industrially important non-conventional K. lactis strain, where muconic acid production was improved by incorporating six genes at three targeted loci and helped in the introduction of a pathway for the production of muconic acid precursor (Horwitz et al., 2015).

Industrial biofuel production often requires strains that can withstand several stress conditions like extreme pH, high temperature, osmotolerance, shearing forces, organic acids, and other inhibitors. All these properties are complex traits and are encoded by several genes at different loci. Due to the complexity of yeast genome, normal genetic analysis methods, therefore, fail to characterize the underlying genetic regulatory network, and efforts to optimize one of these traits classically depend on adaptive laboratory evolution or other controlled breeding strategies. The phenotypic and genotypic characterization of non-conventional yeast species deserve special attention and will help to identify strains and species with novel and/or improved industrially important properties, and the advent of next-generation sequencing at low cost and in a comparably short time have now opened the way for the use of advanced genome analysis tools like quantitative trait loci analysis to identify, at least under some conditions, all causative genes for a certain trait (Bloom et al., 2013).

Tsai et al. (2015) reviewed about the requirement of genetic circuits in yeast synthetic biology. The timely expression of genes in metabolic pathway could minimize the extra energy and nutrient resources and maximize the production capacity. This emphasizes the need for genetic circuits which include multiple regulatory elements arranged to create different logical gates. The combination of these modules resulted in a complicated network (Elowitz and Leibler, 2000) and genetic switches (Gardner et al., 2000). Recently, Blount et al. (2012a) reviewed the status of Yeast genetic circuits. The genetic circuits in S. cerevisiae are far behind the E. coli. Blount et al. (2012b) also proposed three important principles to create genetic circuits in yeast: (1) the circuit elements should not be related to the host physiologic elements to retain orthogonality, (2) the circuit elements needs to be tuneable and inducible, and (3) self-contained groups should be built to obtain modularity. In order to construct genetic circuits, technologies that create fast prototyping, evaluation, and optimization of metabolic pathways in host species are essential. The current key research area of synthetic biology is how to reduce the time required for making genetic circuits. The finding of new metabolic pathways and its combination can be efficiently addressed by better pathway assembly techniques, like ligation-free assembly (Li and Elledge, 2007; Vroom and Wang, 2008) and BioBricks (Shetty et al., 2008). Another key area of research of synthetic biology is the synthesis of reusable constructs with predictable behavior. The construction of a different array of controllable and tunable expression system is very useful in metabolic engineering. A genetic switch that responds to environmental cues and leads to further downstream regulation is also advantageous for biofuel production (Dueber et al., 2007).

For example, incorporation of FadR-binding sites in GAL1 promoter region, after RNA polymerase binding region and before the transcription start site, represses the promoter activity. But the presence of fatty acid, like myristic acid, changes the conformation of FadR, reduces the affinity to DNA-binding sequences, and deregulates the expression of genes (Blount et al., 2012a). Teo and Chang (2014) incorporated several upstream activator elements which in turn detect copper and phosphate starvation, and then downstream gene will only be activated by the depletion of both fatty acid and copper or fatty acid and phosphate. This synthetic device is useful for sensing the presence of fatty acid and can be included into pathways with the production or depletion of fatty acids. The regulatory responses of yeast to external stimuli have also been engineered for synthetic biology applications. For example, mitogen-activated kinase cascade have genetically engineered to rewire the novel regulatory networks (Kiel and Serrano, 2011).

Extensive attempts have been done to increase the range of substrate utilization capabilities of S. cerevisiae through metabolic engineering. A schematic diagram for different routes of substrate assimilation is illustrated in the Figure 2. Efficient conversion of carbon substrates, such as arabinose, xylose, and cellobiose, has been materialized through the introduction of hydrolytic enzymes and transporter genes in yeast. Engineered strains of S. cerevisiae capable of utilizing starch and xylose have been reported (Wei et al., 2013). But all these recombinant strains perform well at a fermentation temperature of 30°C but not at the higher temperatures. But most of the hydrolytic enzymes for lignocellulosic biomass function optimally at 50°C. Majority of the non-conventional yeasts are thermotolerant and ferment sugars at high temperature, but the range of substrate utilization is less (Katahira et al., 2004). Metabolic engineering of these non-conventional yeasts is essential to widen the range of substrate utilization. However, certain non-conventional yeasts utilize a variety of carbon substrates, but they do not possess other important characteristics (ethanol tolerance, inhibitor tolerance, etc.).

Pichia stipitis are the ascomycetous yeast that can ferment xylose to ethanol at nearly maximum yield without any by-product formation. Few metabolic engineering approaches have been established in P. stipitis to resolve its crabtree effect, low ethanol tolerance, etc. Although P. stipitis is an efficient biofuel producer due to its natural capability for xylose fermentation, further improvement in production host is hampered by a lack of genetic engineering tools. For further improvement of this strain, RNA sequencing was done in search of highly active constitutive promoters. Several promoters like TEF1, Xy11p, ADH1p, and ADH2p have been exploited in S. stipitis for efficient gene expression (Cho and Jeffries, 1999). Few transcriptomic approaches were carried out in S. stipitis under glucose/xylose inducing conditions. Recent transcriptomic study under lignocellulosic biomass inducing condition and oxygen limited condition resulted in the mapping of several important genes (Bullard et al., 2010). Synthetic biology applications in S. stipitis are very less at this moment, and most of them involve the genetic transfer of its efficient biosynthetic genes to S. cerevisiae to make it able to utilize pentose sugars, although the full genome sequencing (Jeffries and Van Vleet, 2009) and the more efficient genetic transformation systems (e.g., plasmid vectors and a loxP/Cre recombination system) and drug resistance markers (Laplaza et al., 2006) might enable better room for genetic modification of this industrial strain.

The thermotolerant yeast H. polymorpha cannot utilize starch and xylose and its ferment glucose and xylose to ethanol at high temperature. For the efficient utilization of xylose and starch, amylolytic and xylanolytic enzymes were heterologously expressed. Genes encoding α amylase and glucoamylase, SWA2 and GAM1, from the yeast Schwanniomyces occidentalis, encoding α-amylase and glucoamylase, were transferred in to H. polymorpha under the well characterized constitutive promoter of the H. polymorpha glyceraldehyde-3-phosphate dehydrogenase gene (Voronovsky et al., 2009). Since the organism is highly thermotolerant, engineering the substrate utilization strategies will help to improve the industrial production.

Kluyveromyces marxianus KY3 is a highly efficient hexose-fermenting yeast and has many other advantages like thermotolerance, high cell density, temperature and pH tolerance, high secretion of heterologous proteins, and efficient substrate utilization (Yanase et al., 2010). The xylose utilization of K. marxianus is weak. To alleviate the xylose utilization, this strain was engineered for the conversion of xylose to xylitol and ethanol at the higher temperature. This was done by replacing native xylose reductase gene of the K. marxianus strain with xylose reductase gene from P. stipitis. This modified strain with both NADPH and NADH as the coenzyme could produce 55 g/L ethanol and 32 g/L xylitol (Zhang et al., 2013). Other engineering approaches include the protein engineering for altering the cofactor requirement of xylose reductase of Candida tenuis (Petschacher and Nidetzky, 2008) and optimized expression levels of Candida shehatae xylose reductase, Candida tropicalis xylose dehydrogenase, and P. pastoris xylose kinase by combinatorial transcriptional engineering (Tsai et al., 2015). Recently, evolutionary adaptation technique has been used to enhance the utilization of xylose in K. marxianus (Sharma et al., 2016).

Many microorganisms possess biomass-degrading enzymes that can efficiently degrade lignocellulose materials, but their fermentative production of ethanol is less. Recently, Arora et al. (2015a,b) reviewed the importance of highly efficient cellulosomes for the production of biofuels from lignocellulosic biomass. Several attempts have been done to convert enzymatically efficient hosts into ethanol producing host to serve as the cell factory for ethanol production. But still, an efficient ethanol fermenter needs to be developed. To develop yeast for CBP bioethanol production, a synthetic biology technique, called “promoter-based gene assembly and simultaneous overexpression” (PGASO), that can simultaneously insert and express multiple genes into yeast. K. marxianus has a number of advantages, such as heat and toxin tolerance, over the model organisms K. lactis and S. cerevisiae. To formulate an efficient cellulose cocktail, a filter-paper-activity assay for selecting heterologous cellulolytic enzymes was developed in this study and used to select five cellulase degrading genes (two cellobiohydrolases, two endo-β-1,4-glucanases and one beta-glucosidase genes) from different fungi. A fungal cellodextrin transporter gene was selected to transport cellodextrin into the cytoplasm. These six multiple genes were assembled into the genome of the host using PGASO technology. Experimental results indicated that the developed strain KR7 contains five recombinantly expressed heterologous cellulase genes and that strain could transform crystalline cellulose into ethanol (Chang et al., 2013).

Despite the engineering of metabolic networks in yeast for biofuel production, yeast could not achieve the high level of biofuel production because of the toxic nature of products. The toxicity is mainly due to the hydrophobicity of the accumulated products inside the membrane, and this will lead to membrane disruption by the inhibition of ATP-generating pumps and conformation changes in the proteins that maintain the fluidity (Jeffries and Jin, 2000; Dunlop, 2011). Several genetic manipulations elicit enhanced tolerance against several advanced biofuels. The success of all these attempts was not up to the mark.

Even though most of the non-conventional yeast species can naturally fight with most of the inhibitors, such as nitroaromatics, aromatics, halogenated organo- phosphates, metals, and alkanes (Zinjarde et al., 2014), still, there is an urgent need for engineering some non-conventional, like Y. lipolytica, against fermentation inhibitors present in biomass hydrolyzates, like acids and phenolics. The inhibitor tolerance is a quantitative trait and is determined by several complex genes. Hence large-scale analysis of gene expression, different expression cassettes construction, and optimization will be essential to impart inhibitor tolerance to yeast. Y. lipolytica has been engineered to express recombinant laccases (Madzak et al., 2005), which could improve its detoxification capacities. Several studies reported the strains of non-conventional yeast species, such as Schizo saccharomycespombe, Pichia kudriavzevii, D. bruxellensis, Torulaspora delbrueckii, and Wickerhamomyces anomala, with promising fermentative features and superior ethanol tolerance levels than that of S. cerevisiae (Mukherjee et al., 2014; Ruyters et al., 2015). D. bruxellensis has been reported as one of the excellent yeast in terms of product tolerance and production. D. bruxellensis and S. cerevisiae possesses an almost similar molecular mechanism for this trait (Piskur et al., 2006). Kwon et al. (2011) reported that P. kudriavzevii is extremely tolerant to 3 g/L furfural and also tolerate acetic acid of up to 10 g/L (Oberoi et al., 2012) and formic acid up to 2 g/L (Dandi et al., 2013). The genetic engineering tools for this yeast are limited, and genome sequence has recently been reported by Chan et al. (2012).

Yarrowia lipolytica is a well-studied oleaginous organism and extensively used for industrial biofuel production, and it has been served as a model organism for biofuel research, especially for fatty acid-derived fuels (Beopoulos et al., 2009; Tai and Stephanopoulos, 2013; Blazeck et al., 2014; Zhou et al., 2016). Several metabolic engineering tools are available for Y. lipolytica (Juretzek et al., 2001; Madzak, 2015). The completely annotated genome is available (Dujon et al., 2004), and its metabolism and regulation are also studied in detail (Pan and Hua, 2012). System and synthetic biology approaches have been established in this organism (Morin et al., 2011; Pomraning et al., 2015). Several metabolic engineering studies strengthen the lipid production in this organism. Different target genes were found for overexpression from the metabolic pathway and manipulated to increase the fatty acid accumulation. For example, inhibiting beta-oxidation, by targeted deletion of the six POX genes or the MFE gene (Dulermo and Nicaud, 2011) and overexpression of enzymes leading to TAG production (DGA2) (Beopoulos et al., 2012) and GPD1 (Dulermo and Nicaud, 2011), increased the lipid content. An improved Y. lipolytica strain could produce a very high lipid yield and lipid titers of ~55 g/L under optimized conditions. This further proved the economic feasibility of Y. lipolytica for biofuel production (Qiao et al., 2015). Y. lipolytica is not able to utilize cellulose and starch. Wei et al. (2014) modified the strain by incorporating cellulases for the conversion of cellulosic substrates. In addition, two alpha amylases—starch-degrading enzyme—have been expressed in this host (Park et al., 1997; Celinska et al., 2015).

Studies revealed that the use of intron-containing translation elongation factor-1a (TEF) promoter is capable of increasing the oil production Y. lipolytica compared to intron-less TEF promoter. They have exploited this expression system for the overexpression of diacylglycerol acyl transferase (DGA1), the final key enzyme of the triglyceride (TAG) synthesis pathway, which resulted in a fourfold enhancement in lipid yield compared to wild-type, to a lipid yield of 33.8% of DCW. They also proved that the overexpression of acetyl-CoA carboxylase (ACC1), the first key enzyme of fatty acid biosynthesis, increased lipid content twofold over control, or 17.9% lipid content. The co-expression of ACC1 and DGA1 also improved the production of oil content (Tai and Stephanopoulos, 2013). Ledesma-Amaro et al. (2015) performed a strain engineering approach to obtain a consolidated bioprocess for the direct production of lipids from raw starch. Further, they proved that lipid production from starch can be enhanced by both metabolic engineering and culture condition optimization. Blazeck et al. (2014) rewired Y. lipolytica native metabolism for increased lipid titers, by coupling combinatorial multiplexing of lipogenesis targets with phenotypic induction. The tri-level metabolic control results in accumulation of 90% lipid content and 60-fold improvement over wild-type strain. Sheng and Feng (2015) reviewed the metabolic engineering of non-conventional yeast strains to improve the production of fatty acid-derived biofuels, and they discussed the bottlenecks that limit the productivity of biofuels and suggested the appropriate strategies to overcome the current bottlenecks. Some of the non-conventional yeast system currently employed for biofuel production is listed in Table 1.