Metabolic engineering of non-conventional yeasts for construction of the advanced producers of biofuels and high-value chemicals

Highlights

• •Non-conventional yeasts are promising producers of biofuels and high-value chemicals.
• •Non-conventional yeasts could be used to produce biofuels from lignocellulosic sugars and from crude glycerol as well riboflavin from glucose or cheese whey.

Abstract

Non-conventional yeasts, i.e. yeasts different from Saccharomyces cerevisiae, represent heterogenous group of unicellular fungi consisting of near 1500 species. Some of these species have interesting and sometimes unique properties like ability to grow on methanol, n-alkanes, ferment pentose sugars xylose and l-arabinose, grow at high temperatures (50°С and more), overproduce riboflavin (vitamin B2) and others. These unique properties are important for development of basic science; moreover, some of them possess also significant applied interest for elaboration of new biotechnologies. Current paper represents review of the recent own results and of those of other authors in the field of non-conventional yeast study for construction of the advanced producers of biofuels (ethanol, isobutanol) from lignocellulosic sugars glucose and xylose or crude glycerol (Ogataea polymorpha, Magnusiomyces magnusii) and vitamin B2 (riboflavin) from glucose and cheese whey (Candida famata).

Introduction

Non-conventional yeasts represent heterogenous group of unicellular fungi different from that of Saccharomyces cerevisiae (or Saccharomyces genus in general). Some others add to conventional yeasts also Schizosaccharomyces pombe. Any definition means that non-conventional yeasts represent diverse group of yeast organisms consisting around 1500 species. Some species of non-conventional yeasts are characterized by unique properties. They include ability to grow on methanol, n-alkanes, efficient alcoholic fermentation of most abundant pentose sugar, xylose, thermotolerance, riboflavin oversynthesis, production of large amounts of lipids and carotenoids and others. Several species of non-conventional yeasts, like O. polymorpha and C. famata, belong to favorite microorganisms used in the laboratory of the author for studying alcoholic fermentation, riboflavin synthesis, construction of the advanced strains using metabolic engineering and related research. In particular, we studied high-temperature xylose alcoholic fermentation and construction of the advanced ethanol producers from xylose and glycerol in O. polymorpha; production of isobutanol in multinuclear yeast Magnusiomyces magnusii and finally on construction of the advanced producers of riboflavin and flavin nucleotides in the flavinogenic yeast C. famata.

Ethanol production from xylose in the thermotolerant yeast Ogataea polymorpha

Xylose is the most abundant pentose sugar in our planet and being the second most abundant sugar after glucose. Xylose is the most abundant sugar of hemicelluloses. Lignocellulosic crop residues comprise more than half of the world's agricultural plant biomass [1] and significant fractions of the total can be recovered without competing with other uses [2,3]. Hemicelluloses are heteropolymers consisting residues of different sugars, including xylose, L-arabinose, glucose, mannose, galactose, rhamnose, glucuronic and galacturonic acids. Of different hemicelluloses, xylan, arabinoxylan, glucuronoxylan and xyloglucan, all containing large portion of xylose residues, are the most common [4]. Xylose content varies from 5.3% in spruce or pine to 35.3% in corn cobs [5]. Xylose constitutes about 17% of the total dry weight in woody angiosperms and ranges up to 31% in herbaceous angiosperms [6]. Depending on the substrate and reaction conditions, dilute acid pretreatments of lignocellulosic residues can recover 80–95% of the xylose from the feedstock [7]. High abundance of xylose in nature led to intriguing possibility to convert it and in general xylose-containing feedstocks like lignocellulose to useful substances and fuel ethanol producing by yeast is considered as the most important bioproduct.

There are many xylose-metabolizing yeasts described to date. Genes for xylose assimilation appear to be widespread in the ascomycete yeasts, including unexpectedly, several yeasts lacking the ability to grow on this sugar [8]. The best-known xylose-utilizing yeasts are Scheffersomyces (Pichia) stipitis, Scheffersomyces (Candida) shehatae, Pachysolen tannophilus, Spathaspora passalidarum, Kluyveromyces marxianus and Ogataea polymorpha [9], [10], [11], [12], [13]. Most yeast researchers work with recombinant Saccharomyces cerevisiae based on the strong argument that in spite this species does not grow on xylose at all, it has long history of industrial use for ethanol production from conventional feedstocks like glucose and sucrose. However, the best engineered strains of S. cerevisiae in certain parameters of xylose fermentation still are inferior to the wild-type of the native xylose fermenting yeasts, e.g. S. stipitis. For example, ethanol productivity of the recombinant S. cerevisiae was lower than in S. stipitis on xylose as the sole carbon source, mainly due to formation of xylitol and glycerol by baker's yeast [14]. Still all known wild-type and recombinant yeast strains have serious drawbacks which prevent their immediate industrial application.

The author works in the field of xylose alcoholic fermentation using O. polymorpha. This organism (earlier known as Hansenula polymorpha and Pichia angusta) belongs to methanol-utilizing (methylotrophic) yeasts and is a popular organism for basic and applied research. It is used as model organism for studying methanol metabolism, peroxisome homeostasis [15], [16], [17]. It is also one of the most thermotolerant yeasts known with maximal growth temperature 50 °C and higher and is used to study mechanisms of thermotolerance [18,19]. In the field of applied biotechnology, O. polymorpha is popular for synthesis heterologous proteins, glutathione and penicillin as well as for ethanol production from xylose and crude glycerol [13,[20], [21], [22], [23], [24], [25], [26], [27], [28]]. O. polymorpha is industrial organism with GRAS status and is used for production of hepatitis B surface antigen, anticoagulants, somatotropin, insulin, interferon, phytase, hexose oxidase and triacylglycerol lipase [28]. Methods of molecular genetics are well developed for O. polymorpha [29], [30], [31], [32], [33], [34], [35] and genome of this species is sequenced, publicly available and annotated (see also: [8]; https://mycocosm.jgi.doe.gov/Hanpo2/Hanpo2.home.html).

First article on xylose alcoholic fermentation was published in 2003 by the author's team [11]. It was found that this organism grows well on xylose and glucose and ferment them to ethanol. It was also found that fermentation of both sugars is activated during limited aeration and that amount of ethanol accumulated from xylose is near 100 times lower than that form glucose as accumulated only 0.4 g of ethanol/L in xylose medium though efficient xylose fermentation is prerequisite for development of the feasible process of lignocellulose conversion to fuel ethanol [36]. Starch and xylan are not utilized and fermented by O. polymorpha, however, strains fermenting starch and xylan were successfully constructed by expression of heterologous genes coding for secretory α-amylase and glucoamylase coding by the SWA2 and GAM1 genes from the starch-utilizing yeast Schwanniomyces occidentalis (resulted in starch fermentation) as well as XYN2 of Trichoderma reesei and xlnD of Aspergillus niger that code for secretory endoxylanase and secretory β-xylosidase resulted in xylan fermentation [37]. Ethanol titers, especially that from xylan, were low, however, work was done on the wild-type strain of O. polymorpha with low efficiency of xylose fermentation. Nevertheless, the mentioned result is the proof of concept that O. polymorpha is capable to ferment not only monosaccharides but also biopolymers. In other words, the consolidate bioprocesses (CBP), a promising technology for ethanol production directly from polysaccharides, is thus applicable to O. polymorpha.

Why alcoholic fermentation of xylose in the thermotolerant yeast O. polymorpha at elevated temperatures is so important? This is because high temperatures alcoholic fermentation has several important advantages (see also [28]: (i) contamination should be low as only minority of microorganisms are thermotolerant; (ii) the higher fermentation temperature, the lower costs of fermentation broth heating to temperature of distillation; (iii) fermentation at high temperature goes faster than that at 30 °C; (iv) fermentation at elevated temperatures avoids expensive bioreactor cooling required for mesophilic microorganisms; (v) in the case substrates for fermentation are polymers (starch or pretreated lignocellulose) simultaneous saccharification and fermentation (SSF) and simultaneous saccharification and co-fermentation (SSCF i.e. co-consumption of both glucose and xylose) which combine enzymatic hydrolysis of polymer with subsequent fermentation of liberated sugars in the same vessel, is a promising process. As hydrolyzing enzymes (amylases, cellulases, hemicellulases) optimally work at around 50 °C, only thermotolerant microorganisms could be used in SSF process [38], [39], [40], [41], [42]. O. polymorpha corresponds to these criteria in the case ethanol yield and productivity are sufficiently high and accumulation of byproducts is low.

In O. polymorpha, fermentation was active at high temperatures, up to 45 °C–48 °C. As elevated fermentation temperature has several important advantages, several successful attempts have been made to further improve thermotolerance of this yeast [19]. Though O. polymorpha grows satisfactorily well at 50 °C it ferments very poorly at this border temperature. The increase in the intracellular level of trehalose in O. polymorpha due to deletions of acid trehalase gene ATH1, resulted in a 6-fold higher ethanol production of xylose fermentation at 50 °C. The overexpression of the heat shock proteins Hsp16 and Hsp104, have also led to three to six times improved ethanol production at 50 °C [19]. Very important trait for ethanol producers is high ethanol tolerance. O. polymorpha appears to be more tolerant to ethanol than S. stipitis still being more sensitive than S. cerevisiae [11,43]. Overexpression of the own ETT1 gene (a homolog of S. cerevisiae MPE1 gene), significantly increased the O. polymorpha ethanol tolerance, resulting in 10- and 3-fold improvements in the growth on agar and in liquid media with ethanol, respectively. The ethanol tolerance of O. polymorpha was also enhanced by overexpression of S. cerevisiae MPR1, which codes for acetyltransferase [44]. Still in standard experiments, growth and fermentation of O. polymorpha were conducted at 45 °C.

The reasons of low ethanol production from xylose remain to be known only partially. I suggest that this depends on low expression of the enzymes responsible for xylose metabolism and on inefficient action of regulatory factors which have not been characterized. It is known that the first enzyme of xylose metabolism in yeasts, xylose reductase (gene XYL1) prefers NADPH whereas the second enzyme, xylitol dehydrogenase which oxidizes xylitol to xylulose (gene XYL2), uses only NAD. This pathway is known as oxidoreductive one. This leads to cofactor imbalance when NADH produced in the second reaction cannot be reoxidized in the first reaction leading to lack of NAD regeneration and accumulation of xylitol instead of ethanol [45,46]. Contrary, xylose-utilizing bacteria use isomerase pathway in which xylose is directly isomerized to xylulose. Xylose isomerase does not use any cofactors so no disbalance problem is appeared. Improvements of xylose fermentation in yeasts like xylose-utilizing recombinant strains of S. cerevisiae was based either on introduction of bacterial gene coding for xylose isomerase and establishing thus isomerase pathway or on protein engineering of xylose reductase or xylitol dehydrogenase (oxidoreductive pathway) to eliminate cofactor disbalance problem [13]. To improve xylose fermentation in O. polymorpha, we used both approaches. To improve oxidoreductive pathway, site-specific mutagenesis of XYL1 gene leading to decreasing affinity of xylose reductase to NADPH and maintaining affinity to NADH unchanged was done. Resulted strains were characterized by increase in ethanol production from xylose [47]. To establish isomerase pathway in O. polymorpha, first the deletion mutants xyl1∆ defective in xylose reductase and double deletion mutant xyl2A∆ xyl2B∆ defective in two paralogs of xylitol dehydrogenase which displayed 50–60 times drop in specific activity of this enzymes, were constructed. They failed to grow on xylose as sole carbon and energy source. Gene xylA from Escherichia coli and Streptomyces coelicolor were introduced into such triple mutants. Transformants grew on xylose as sole carbon source as small colonies and displayed activity of xylose isomerase, however, accumulated even less ethanol from xylose relative to the wild-type strain. Large colonies on xylose were isolated from such transformants which characterized by increase in xylulokinase activity [48]. Such last transformants however, accumulated the same amount of ethanol as the wild-type strain, therefore in further work on strain development, the strain with engineered xylose reductase (oxidoreductive pathway) was used [47]. However, the work of Ukrainian group on establishing isomerase pathway in O. polymorpha [48] was the first positive example of successful expression of E. coli xylA gene in yeasts as many attempts to achieve that in S. cerevisiae failed [49], [50], [51]. In independent research, it was found that overexpression of PDC1 gene coding for pyruvate decarboxylase, activates ethanol production form xylose in O. polymorpha [52]. Thus, it was found that enzymes of first steps of xylose metabolism (xylose reductase, xylitol dehydrogenase, xylulokinase) and that of alcoholic fermentation (pyruvate decarboxylase) are important for ethanol production from xylose [47,48,52]. The scheme of xylose and glucose metabolism in yeasts, including O. polymorpha, is presented in Fig. 1.

Fig. 1.

Scheme of glucose and xylose metabolism and fermentation in yeasts.

Most attention was paid on strain development. i.e. on construction of O. polymorpha producing elevated amounts of ethanol form xylose. For this, methods of metabolic engineering and classical random selection were used. At the beginning, the best parental strain for further development was chosen. It was found that the wild-type strain NCYC495 consumed significant amounts of accumulated ethanol, however, one earlier isolated mutant 2Eth- defective in ethanol utilization retained most of the accumulated ethanol. It was used for further work. To further increase ethanol synthesis from xylose, the engineered gene XYL1m (which products has low affinity to NADPH) and native genes XYL2, XYL3 and PDC1 in the strain 2Eth- was overexpressed. It was found that the highest amount of ethanol was produced by transformants with overexpressed genes XYL1m, XYL2 and XYL3 (7.4 g/L) whereas additional overexpression of PDC1 dropped for unknown reason ethanol accumulation to 5.0 g/L [53]. The further improved strain was selected among the mutants resistant to anticancer drug, inhibitor of glycolysis 3-bromopyruvate. It was found that 70% of resistant mutants synthesized enhanced amounts of ethanol from xylose and practically did not accumulate xylitol. The best strain accumulated 9.8 g of ethanol/L and was used for further development [53]. To investigate the molecular aspects of 3-bromopyruvate resistance leading to activation of ethanol production from xylose, insertion mutagenesis was applied. One insertion mutant was isolated which overproduced ethanol from xylose but not from glucose. Insertion cassette disrupted gene ATG13 involved in autophagy initiation. Deletion mutant in this gene had similar phenotype. However, autophagy was normal in insertion mutant (though defective in deletant). Insertion did not impaired HORMA domain compulsory for autophagy. It was concluded that ATG13 is involved both in autophagy and regulation of xylose fermentation though the last process is independent on autophagy [54]. The molecular nature of mutation(s) in 3-bromopyruvate-resistant strain obtained by conventional mutagenesis and used in strain development, was not analyzed. However, in the best isolated strain, almost 10-fold increase in expression of TAL1 gene coding for cytosolic transaldolase and almost 47-fold increase in expression of RPE1 gene coding for ribulose-5-phosphate epimerase was observed [55]. Further selection was based on knock out of homolog of S, cerevisiae gene CAT8 coding for transcription factor involved in gluconeogenesis. It was suggested that xylose is partially gluconeogenic substrate and that defect in this process will direct more carbon from xylose toward ethanol instead of hexoses. Indeed, cat8∆ mutants isolated either from wild-type strain or from advanced ethanol producer from xylose was characterized by further increase in ethanol synthesis from xylose but not from glucose [55]. For further strain development, attention was paid on peroxisomal enzymes of pentose phosphate pathway, namely transketolase (gene DAS1) and transaldolase (gene TAL2). It was found that deletion of each of these genes impaired xylose fermentation having no influence on fermentation of glucose whereas overexpression of these genes activated ethanol production from xylose [56]. It was also shown that defects in peroxisome biogenesis due to deletions of PEX3 and PEX6 genes strongly inhibits xylose alcoholic fermentation. Overexpression of both genes DAS1 and TAL2 simultaneously in the best for that time ethanol producer from xylose (based on the strain cat8∆) further increased ethanol production from xylose to 16 g of ethanol/L at 45 °C [56]. Further increase in ethanol production form xylose was achieved by selection for mutants forming colonies on plates with L-arabinose as sole carbon source ([28], unpublished observation). L-Arabinose-growing strains were characterized by further elevated ethanol production from xylose (to 20 g/L and more) and from L-arabinose (still very low, near 0.1 g/L) at 45 °C. Achieved ethanol titer from xylose is highest at so high temperature, however should be further elevated to reach the highest level obtained for mesophilic yeasts.

It was also found on the important role of sugar transporters in xylose metabolism and fermentation. Own hexose transporter Hxt1 of O. polymorpha was engineered to eliminate glucose inhibition and endocytosis. Resulted strain utilized simultaneously glucose and xylose from their mixture, in contrast to parental strain, and accumulated elevated amounts of ethanol in the mixture of glucose (70%) and xylose (30%) [57].

Several additional genes were found to be important for xylose and in some cases also glucose, fermentation. Increase in ethanol production from xylose was observed after overexpression of genes TKL1 (cytosolic transketolase), TAL1 (cytosolic transaldolase) [56], HXS1 (hexose sensor), AZF1 (transcription factor) [58] and deletions of genes MIG1, HAP4A, HAP4B (transcription factors) [59]. Overexpression of HXS1 and AZF1 genes also activated glucose alcoholic fermentation. Scheme of major genetic manipulations developed and applied in author's work on xylose alcoholic fermentation is presented in Fig. 2. It would be important to manipulate with these genes on background of the best available ethanol producer from xylose isolated as grooving strains on L-arabinose. Additionally, methods of random selection should be used based on involvements of selection approaches developed recently in the author's lab [60]. The methods use selection for resistance to 2-deoxyglucose, glucosamine, trehalose, glyoxylic acid, oxythiamine and other growth inhibitors. It was suggested that correspondingly engineered strains of O. polymorpha will be competitive regarding ethanol titers with mesophilic yeasts, however it will occur at temperatures 45 °C and higher. Further steps in strain development will also include: (i) identification of the new regulatory genes participating in xylose fermentation and their modifications (ii) overexpression of formerly identified genes in the advanced xylose-fermenting strains of O. polymorpha: HXS1, AZF1, TKL1, TAL1; (iii) expression of fungal or bacterial phosphoketolase pathway of xylose metabolism; (iv) activation of L-arabinose metabolism and fermentation due to activation of the corresponding structural genes and identification of the regulatory genes. I also plan to study utilization of all sugars from real industrial lignocellulosic hydrolysates and select mutants resistant to inhibitors in hydrolysates if necessary. Summarizing, the author's lab studied main features of alcoholic fermentation of the most abundant pentose, xylose, identified genes involved in regulation of this process and, using combination of methods of metabolic engineering and classical random selection, enhanced ethanol production form xylose at least 50-folds. Isolated strains could be searched for use in SSCF process of simultaneous saccharification and co-fermentation of glucose and xylose and to serve as good starting material for further strain and process development. I also suggest that in future, the engineered strains of O. polymorpha could be implemented in real industrial process of the 2nd generation ethanol production from lignocellulose.

Fig. 2.

Scheme of engineered metabolic and regulatory pathways in O. polymorpha during construction of the advanced ethanol producer from xylose.

Ethanol production from crude glycerol by O. polymorpha

During last decades, biodiesel became very popular liquid fuel. Biodiesel represents methyl (or ethyl) esters of fatty acids obtained by transesterification from plant oils. Glycerol appeared to be the major byproduct of biodiesel production accounting for 10% of the produced biodiesel [61]. Crude glycerol contains methanol, salts of sodium or potassium, heavy metals and other impurities and its purification is quite expensive. Availability of huge amounts of crude glycerol rises question of its bioconversion to high-value products. Bacteria and yeasts are considered for this purpose. Most popular products of crude glycerol bioconversion are 1,2-propanediol, 1,3-propanediol, 2,3-butanediol, dihydroxyacetone, organic and amino acids [62], [63], [64], [65]. One of the most promising products which can be obtained from crude glycerol is ethanol [66]. Being cheap, crude glycerol is very competitive with sugars as substrate for ethanol production. Among yeasts, S. cerevisiae is not too promising organism as poorly grows on glycerol [67]. Much higher yields of ethanol from glycerol were achieved using yeast Pachysolen tannophilus [68]. However, non-methylotrophic yeasts are unable to utilize methanol available in crude glycerol therefore methanol-utilizing yeasts, like O. polymorpha, have clear advantages for ethanol production from crude glycerol. It was found that the wild-type strain of O. polymorpha produces low amounts of ethanol form glycerol whereas mentioned above strain of this species with overexpressed PDC1 coding for pyruvate decarboxylase [52] showed 3.5 times increase in ethanol production [69]. Additional overexpression of ADH1 coding for alcohol dehydrogenase gave further increase for more than 30%, i.e. overall 4.6 times increase relative to the wild-type strain of O. polymorpha. The ethanol yield was near 0.36 g of ethanol/g of consumed glycerol. Ethanol concentration reached 5.0 g/L at 45 °C and only 4.3 g/L at 37 °C [69]. To further improve ethanol synthesis from glycerol, genes coding for initial reactions of glycerol metabolism were overexpressed. O. polymorpha contains simultaneously two pathways of glycerol metabolism, kinase (glycerol kinase Gut1 and glycerol-3-phosphate dehydrogenase Gpd1) and oxidoreductive (glycerol dehydrogenase Gcy1 and dihydroxyacetone kinase Dhk1) [70,71]. Author's team overexpressed separately genes of both pathways and in both cases received increase in ethanol accumulation, up to 10.7 g/L with ethanol yield 0.13 g/g [72]. Additionally, in one of resulted strains with overexpression of genes PDC1, ADH1, GCY1 and DAK1, heterologous gene FPS1 from Komagataella phaffii was overexpressed, which is known to encode glycerol facilitator which is responsible for glycerol transport. Resulted strain of O. polymorpha accumulated 1.08 times more ethanol. Using crude glycerol as substrate, results were much more modest. The best strain with overexpression of all 5 genes (own genes PDC1, ADH1, GCY1, DAK1 and heterologous gene FPS1 from K. phaffii) accumulated 3.55 g of ethanol/L which is less than that from pure glycerol, however, 18 times more than was produced by the wild-type strain (0.2 g/L) [72]. Obtained results are inferior to that obtained on alternative yeast producer P. tannophilus [68], however, O. polymorpha ferments faster at high temperatures and additionally, it can utilize methanol available as contamination in crude glycerol. There are many lines for further strain development which include simultaneous overexpression of the enzymes of kinase and oxidoreductive pathways as well as random selection with antimetabolites which have been developed and applied for ethanol overproducing strains from xylose and glucose [53,60]. All these factors allow to predict that O. polymorpha could be the very promising organism for development of industrial process of ethanol production from crude glycerol.

Isobutanol overproduction in the multinuclear yeast Magnusiomyces magnusii

Currently, near 120 billion liters of bioethanol is produced annually with over 100 billion liters used as fuel (https://knect365.com/energy/article/c07f7fba-48fa-464f-9f21-12f913fc67f7/world-ethanol-production-to-expand-steadily-in-2019). However, ethanol has serious drawbacks as the fuel. It needs expensive dehydration step and dehydrated (absolute) ethanol absorbs moisture from the air which leads to corrosion of car engine and gas station parts. Ethanol contains relatively low energy relative to gasoline and can be blended with gasoline only to 15%. Moreover, ethanol cannot be blended with diesel. At the same time, higher alcohols like butanol and isobutanol are free of these drawbacks: they are not hygroscopic, contain much higher energy density and can be blended with gasoline and diesel fuel in any ratio [73,74]. Higher alcohols also have other uses in chemical synthesis, and as solvents as well as additives to perfumes (http://www.biobutanol.com/Butanol-Isomers-isobutanol,-n-butanol,‑tert-butanol.html). Yeasts produce isobutanol as a component of fusel alcohols, however, its concentration is near 1000 times less than that of ethanol. During last decade several publications reported the construction of more active isobutanol-producing yeasts. In S. cerevisiae, overexpression of the genes in valine biosynthesis from pyruvate (ILV2, ILV5, ILV3) of native mitochondrial or cytosolic localization of the products and those on the overexpression of aldehyde dehydrogenase ARO10 led to the accumulation of up to 600 mg of isobutanol per L [75], [76], [77]. Substantial increase in isobutanol production was achieved by the elimination of the alternative pathway for pyruvate utilization by the deletion of LPD1 gene of pyruvate dehydrogenase complex, overexpression of genes coding for pyruvate carboxylase, malate dehydrogenase, and malic enzyme and activation of isobutanol biosynthetic pathway which led to accumulation of near 2.1 g of isobutanol per L [78,79]. In the non-conventional yeast K. phafii, overexpression of the endogenous enzymes of valine biosynthesis and Ehrlich pathway led to an accumulation of 2.2 g of isobutanol per L which is the highest titer described for yeast [80].

During testing collection of non-conventional yeasts for production of higher alcohols, author's team paid attention on the multinuclear yeast Magnusiomyces (Endomyces, Diplodascus) magnusii, which accumulated 440 mg of isobutanol/L. Methods of molecular genetics have not been developed for M. magnusii and its genome has not been sequenced. In the author's lab, the methods of M. magnusii transformation were developed using dominant markers of resistance to zeocin and nourseothricin, expressed the heterologous gene ILV2 from S. cerevisiae in M. magnusii, cloned own strong constitutive promoter TEF1. To improve synthesis of isobutanol, ILV2 and ARO10 genes of S. cerevisiae in M. magnusii under control of TEF1 promoter were overexpressed [81]. Additionally, the method of selection for resistance to acetolactate synthase inhibitor bispyribac was applied. Application of both molecular cloning and classical selection approaches led to strains with elevated synthesis of isobutanol (620 mg/L) and incubation of such producer with 2-oxovalerate further elevated isobutanol accumulation to 760 mg/L. It was also found that the wild-type strain of M. magnusii accumulates even higher amounts of n-butanol (820 mg/L) during incubation with 2-oxovalerate. These data clearly show that M. magnusii has great potential as producer of higher alcohols isobutanol and n-butanol [81]. Further progress in this direction depends on sequencing and annotation of M. magnusii genome and adaptation of several modern molecular methods like CRISPR-Cas9 genome editing.

Riboflavin and flavin nucleotide oversynthesis in the flavinogenic yeast Candida famata

Riboflavin (vitamin B₂) is an important water-soluble vitamin which serves as the biosynthetic precursor of flavin coenzymes, flavin mononucleotide (FMN) and flavin adenine dinucleotide (FAD) which act as coenzymes of numerous enzymes, mostly of oxidoreductases [82]. Riboflavin has numerous applications with market value estimated to exceed 9 billion US dollars in 2021 (https://www.mordorintelligence.com/industry-reports/riboflavin-market). It is primarily used as an animal feed additive, in food industry as a yellow colorant and in medicine as the component of multivitamin mixtures and as drug for treatment of some diseases [83,84]. Currently, riboflavin is produced biotechnologically using engineered strains of the bacterium, Bacillus subtilis, and the filamentous fungus, Ashbya (Eremothecium) gossypii. There is the group of so named flavinogenic yeasts which overproduce riboflavin under iron starvation [82]. Candida famata (teleomorph Debaryomyces subglobosus) represents one of the most efficient riboflavin producers among this group. In addition to C. famata, this group includes Meyerozyma (Pichia) guilliermondii, Schwanniomyces occidentalis, the pathogenic yeast Candida albicans and some others [85]. The exact reasons of riboflavin oversynthesis under iron limitation by flavinogenic yeasts are not known. It was hypothesized that overproduced and secreted under iron deficiency riboflavin non-enzymatically reduces practically insoluble Fe³⁺ cation to much better soluble Fe²⁺cation and thus maintains cell growth due to supplying with iron [82]. The other peculiarity of C. famata is its high osmotolerance, as it grows in the presence of high concentrations of NaCl, up to 2.5 M [85].

The strain C. famata dep8 isolated by conventional mutagenesis, belongs to the most active riboflavin producers known and was used for industrial production of riboflavin; however, due to genetic instability, its industrial use was terminated [82]. Still the use of yeast as riboflavin producer has numerous advantages relative to bacteria (absence of phage lysis, growth to dense biomass) and filamentous fungi (riboflavin synthesis during growth phase, unicellular organism which does not plug pipelines), so construction of the yeast competitive producer free of available drawback (genetic instability) would be of great importance.

Known earlier industrial riboflavin overproducer C. famata dep8 was isolated using conventional mutagenesis. However, it was genetically unstable reverting to riboflavin non-overproducing strains. Author's group has found that reversions invariably occur in gene SEF1 and that introduction into strain dep8 additional copy of gene SEF1 from the relative flavinogenic yeast Debaryomyces hansenii (with sequenced genome) greatly stabilized riboflavin production. Introduction of the additional copy of SEF1 gene also substantially increased riboflavin production [86]. It was decided to construct the genetically stable competitive yeast riboflavin producer using combination of the methods of molecular genetics and classical selection. First, the riboflavin-overproducing strain AF-4 was isolated from the wild-type strain C. famata VKM Y-9 in six consecutive steps of conventional mutagenesis. It included selection for (i) resistance to the riboflavin structural analog 7-methyl-8-trifluoromethyl-10-(1′-d-ribityl)isoalloxazine; (ii) to 8-azaguanine; (iii) to 6-azauracil; (iv) to 6-diazo-5-oxo-l-norleucine, which inhibits purine biosynthesis; (v) to the natural nucleoside guanosine which unexpectedly appeared to inhibit growth of C. famata; (vi) selection for brightly yellow colonies on the plates with high pH (6.8). The most flavinogenic mutant, designated AF-4, was selected, which accumulated about 680 mg riboflavin/L in iron sufficient medium, whereas the wild-type strain under these conditions accumulated only 2 to 3 mg riboflavin/L [86]. It was found that strain AF-4 is, in contrast to the strain dep8, genetically stable and does not revert to riboflavin non-overproducing revertants.

To start molecular genetics approaches, the corresponding tools were developed for C. famata. Methods of genetic transformation were developed for C. famata based on dominant resistance markers and on marker LEU2 leading to prototrophy of leucine auxotrophs. To this aim, the C. famata ARS elements were cloned, resulting in the construction of replicative plasmids providing effective transformation of this yeast. The methods of spheroplast transformation and electrotransformation were optimized, reaching transformation frequencies of 10⁴ and 10⁵ transformants/mg plasmid DNA, respectively [87]. Several dominant markers for transformation have also been developed. Such work was hampered by the fact that C. famata belongs to CUG yeast clade which decode this codon as serine rather than leucine [88], so standard genes coding for antibiotic resistance could not be used without codon adjusting. Nevertheless, the ble gene from Staphylococcus aureus (which does not contain a CUG codon) conferring resistance to the antibiotic phleomycin; modified D. hansenii IMH3 and ARO4 genes conferring resistance to mycophenolic acid and fluorophenylalanine, respectively, and BSD gene from Aspergillus terreus coding resistance to blasticidin, were successfully used as dominant markers [35,86,89,90]. Methods of insertion mutagenesis have been developed for C. famata and genes SEF1, MET2 have been tagged which disruptions led to inability to overproduce riboflavin [89]. Besides, a targeted gene deletion system in C. famata has been developed, using LEU2 to knock out the gene SEF1 which is an important tool for molecular research [85]. Structural genes of riboflavin synthesis have been cloned in C. famata. First, collection of C. famata riboflavin auxotrophs was generated by conventional mutagenesis. Biochemical identification of the mutants was carried out by identification of accumulated intermediates of riboflavin synthesis in the culture medium. Finally, the genes RIB1 (coding for GTP cyclohydrolase II), RIB2 (specific reductase), RIB5 (dimethylribityllumazine synthase), RIB6 (dihydroxybutanone phosphate synthase) and RIB7 (riboflavin synthase) were cloned via functional complementation as DNA fragments from a gene library of the corresponding mutations of strains defective in riboflavin synthesis [91], [92], [93]. The scheme of riboflavin synthesis in yeasts is presented in Fig. 3.

Fig. 3.

Scheme of riboflavin and flavin coenzyme biosynthesis in yeasts.

A reporter system for the yeast C. famata was developed. The system includes the Kluyveromyces lactis LAC4 gene, encoding β-galactosidase as a reporter gene, and the C. famata mutant lac4, as a recipient strain. Analysis of different promoters from C. famata and D. hansenii showed that the promoter TEF1 (translation elongation factor 1A) was found to be the strongest one among those tested [94]. Methods of CRISPR-Cas9 for genome editing were adapted for C. famata [95]. However, genome sequence of C. famata wild-type strain VKM Y-9 and riboflavin overproducing strain AF-4 remains unknown which hampers further strain development. At the same time, genome sequence of the former industrial strain C. famata dep8 is publicly available (https://mycocosm.jgi.doe.gov/Debsub1/Debsub1.info.html). Its comparison to genome sequences of the strains VKM T-9 and AF-4 is of great interest and hopefully will be possible in the near future.

Author's work on C. famata was concentrated in two directions: (i) identification of genes involved in regulation of riboflavin synthesis or limiting riboflavin production: (ii) construction of the genetically stable efficient and competitive overproducers of riboflavin and flavin nucleotides. Works in the first direction have been started by identification of the mentioned before gene SEF1 coding for putative transcription activator [89]. It was found that its knock out or point mutations led to inability to overproduce riboflavin whereas its overexpression led to genetic stability and increase in riboflavin production [86]. It was also found that SEF1 promoters from flavinogenic yeasts (C. famata, C. albicans) as well as from C. tropicalis are sufficient to restore the riboflavin oversynthesis in transformants of C. famata sef1Δ mutant whereas SEF1 promoters from other non-flavinogenic yeasts (S. stipitis, S. cerevisiae) did not [96]. Other regulatory gene identified in C. famata was SFU1. In the pathogenic flavinogenic yeast C. albicans, the GATA family transcription factor Sfu1 acts due to physical binding Sfu1 with Sef1 and sequestration of Sef1 to the cytoplasm, where it is unable to perform transcriptional activation of iron acquisition genes [97]. In C. famata, sfu1Δ mutant was characterized by elevated riboflavin synthesis, in contrast to that of sef1 deletion or points mutants which suggest on the role of SFU1 as negative factor in regulation of riboflavin synthesis [96]. It was known earlier that the disruption of the VMA1 coding for vacuolar H⁺-ATPase caused riboflavin oversynthesis in the filamentous fungus A. gossypii and the flavinogenic yeast M. guilliermondi [98,99]. Knock out of VMA1 homolog in the wild-type strain of C. famata also led to riboflavin oversynthesis which could be interesting for construction of the advanced riboflavin producers after VMA1 knock out in the advanced riboflavin producer [96], still the mechanisms of vacuolar ATPase action in regulation of riboflavin synthesis are not known.

Additional factor involved in regulation of riboflavin synthesis appeared to be the gene RFE1 (homolog of mammal BCRP) coding for riboflavin excretion (efflux) from the yeast cell [100]. It is known that mammary gland contains riboflavin pump responsible for efflux of this vitamin to the milk [101]. Several BRCP homologs in the related to C. famata flavinogenic yeast with sequenced genome D. hansenii were identified. Closest homolog (29% identities, 49% positives) designated as RFE1 (RiboFlavin Excretase) was cloned and expressed under control of strong constitutive promoter TEF1 in the riboflavin overproducing strain of C. famata [90,100]. It was suggested that overexpression of riboflavin excretase will activate efflux of riboflavin and thus overall synthesis of this vitamin. Expressed D, hansenii Rfe1 protein was localized in the C. famata cell membranes but not in the nucleus. The constructed strain BRP/RFE1 showed an increase in riboflavin production relative to the parental strain accumulation after 120 h of flask cultivation near 1.7 g/L [100]. It is interesting to note that intracellular riboflavin concentration the BRP/RFE1 transformant was 1.5-fold decreased as compared to that in the parental strain which proves active riboflavin excretion from the cells in the RFE1 transformants. In bacteria, the role of riboflavin in regulation of riboflavin synthesis gene expression is well documented whereas in yeasts it was believed that regulation of riboflavin biosynthesis genes is triggered by iron ions [82]. Data on RFE1 transformant properties in C. famata suggest that riboflavin also could be potentially involved in regulation of the expression of the riboflavin pathway structural genes in flavinogenic yeasts. The corresponding mechanisms should be studied in future works.

Independent approaches to increase riboflavin synthesis in C. famata were done by overexpression of the structural genes of riboflavin synthesis RIB1, RIB6, RIB7 [86,90,102] and activation of the synthesis of purine precursor of riboflavin, GTP, due to overexpression of the engineered heterologous from D, hansenii genes PRS3 (coding for phosphoribosyl pyrophosphate synthetase) and ADE4 (coding for phosphoribosyl amidotransferase). Engineering these genes was done to avoid the feedback inhibition of both enzymes by purine nucleotides. Overexpression of the engineered PRS3 and ADE4 genes in the advanced riboflavin producer BRP (see below) led to increase of GTP pool and 2-fold enhanced riboflavin production [103]. Thus, there is substantial arsenal of tools for construction of the high productive riboflavin overproducer in C. famata.

Genetically stable riboflavin overproducing strains of C, famata have been constructed on background of isolated by conventional mutagenesis non-reverting strain AF-4 [86]. First, one or two additional copies of transcription activator gene SEF1 along with overexpression of IMH3 (IMP dehydrogenase) ortholog from D. hansenii together with native C. famata genes RIB1 (GTP cyclohydrolase II) and RIB7 (riboflavin synthase) were introduced in AF-4 strain [86]. Engineered strain accumulated more at least 4 times more riboflavin relative the parental strain. Further increase in riboflavin production was achieved in the strain with overexpression of genes SEF1, RIB1 and RIB7 designated as BRP (Best Riboflavin Producer). Medium for strain cultivation was optimized and fed-batch cultivation of the producer in the bioreactor was carried out. Maximal titer of riboflavin in a 7 L laboratory bioreactor during fed-batch fermentation in the optimized medium was 16.4 g/L which is close to the highest numbers obtained with other riboflavin producers [82,90]. The scheme of construction of genetically stable riboflavin overproducing strains is presented in Fig. 4. Later on, even more efficient riboflavin producers have been constructed, however, they were tested only in shake flasks. Such improvements were obtained after introduction in BRP strain of D. hansenii RFE1 gene coding for riboflavin excretase [100] and activation of purine biosynthesis de novo after overexpression of engineered D. hansenii genes PRS3 and ADE4 [103]. Last strain was designated as BRPI (Best Riboflavin Producer Improved). It served as parental strain for overexpression gene RIB6. Resulted strain, designated as PRPI/RIB6 is apparently the most active riboflavin producer constructed by us [102]. It contains overexpression of genes SEF1, PRS3, ADE4, RIB1, RIB6 and RIB7 and accumulates during flask cultivation by 13–28% more riboflavin relative to the parental strain and the yield of this vitamin exceeded 400 mg/g dry weight [102]. Further improvement of BRPI/RIB6 riboflavin producer is definitely possible by additional (i) further overexpression of SEF1 transcription activator; (ii) overexpression of RFE1 gene; (iii) knock out of VMA1 gene; (iv) activation of the precursor of riboflavin synthesis, ribulose-5-phosphate due to overexpression of GND1 gene coding for 6-phosphogluconate dehydrogenase. Resulted strains should be superior to best known riboflavin producers.

Fig. 4.

Scheme of metabolic engineering of C. famata for construction of the advanced riboflavin producers (left). Right – kinetics of riboflavin production in the laboratory bioreactor.

In all previous experiments, riboflavin producers were cultivated in glucose medium. However, C. famata grows on lactose and displays β-galactosidase activity [94]. Recently, it was found that C. famata also grows and overproduces riboflavin on whey supplemented only with nitrogen source like ammonium sulfate [104]. Riboflavin synthesis on whey was activated in strains which expressed transcription activator gene SEF1 under control of lactose-induced promoter LAC4 of β-galactosidase gene. However, the best results were obtained with mentioned above strain BRPI/RIB6 [102] which produced riboflavin on whey with titer exceeding 2.5 g/L and yield higher than 300 mg/g dry weight which is close to the yield obtained on glucose. These data support the feasibility of using cheese whey, an abundant waste of dairy industry, in synthesis of a high-value added product, riboflavin [104].

Riboflavin biosynthetic derivatives, flavin coenzymes FMN and FAD, are also valuable substances with numerous applications, mostly in food industry and medicine. E.g., FMN has an advantage for using in food industry as yellow colorant for soft drinks and yogurts because of its much higher solubility as compared to riboflavin [105]. Till recently, there were no microbial producers of FMN synthesizing this nucleotide de novo. C. famata secretes only trace amounts of FMN. To construct FMN overproducer, gene FMN1 of D. hansenii coding for riboflavin kinase, was overexpressed under control of strong constitutive promoter TEF1 in C. famata riboflavin overproducing strain AF-4. Resulted transformants were characterized by 30-fold increase in the riboflavin kinase activity and 400-fold increase in FMN production which led to 200–250 mg of FMN/L under the high-density cultivation condition [106]. Optimization of the medium content and cultivation conditions further increased FMN accumulation [107]. To construct FAD overproducer, gene D. hansenii FAD1 coding for FAD synthetase, was cloned and overexpressed in constructed before FMN overproducer. Obtained transformants displayed 7–15 times elevated activity of FAD synthetase and FAD accumulation in the medium. Optimization of the medium composition and cultivation conditions further increased FAD accumulation for 65-fold relative to the parental strain. Cultivation of the recombinant strain in the optimized condition for 40 h resulted in accumulation of 450 mg of FAD/L [108]. Maximal achieved yields of FMN and FAD are relatively modest due to use of the weak riboflavin producer AF-4 as the parental strain. Recently, advanced riboflavin producers have been used for construction of the more efficient FMN overproducer. Obtained results are very promising as accumulation of near 540 mg of FMN/L in cheese whey supplemented with nitrogen source was obtained in shake-flask experiments [109]. Thus, the flavinogenic yeast C. famata could be the rich source of riboflavin and flavin nucleotides FMN and FAD. Further strain development still is necessary.

Concluding remarks

Presented materials clearly show the huge potential of non-conventional yeasts in production of biofuels and high-value chemicals from standard carbon sources but also from wastes like lignocellulose and cheese whey. Thus, thermotolerant yeast O. polymorpha appeared to be promising organism for construction of the advanced ethanol producers from major pentose sugar of lignocellulose, xylose, and from glycerol. Metabolic engineering and random selection of O. polymorpha elevated ethanol accumulation from xylose and glycerol 50 and 20 times from xylose and glycerol, respectively, at elevated temperature 45 °C. Efficient approach has been developed and applied for construction of the strains which simultaneously utilize and ferment glucose and xylose. In the flavinogenic yeast C. famata, methods of molecular genetics have been developed and efficient producers of riboflavin and flavin coenzymes flavin mononucleotide and flavin adenine dinucleotide have been constructed. Further improvement of the available producers is possible to achieve in the near future.

Funding

The works in these field have been supported by grants and National Research Foundation of Ukraine grant 2020.01/0090, NAS of Ukraine (grant 0120U10163) and of National Science centre of Poland (grants Opus UMO-2018/29/B/NZ1/01–497, Opus 2020/37/B/NZ1/02232, Opus 2021/41/B/NZ1/01224).

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

• Data will be made available on request.