Keywords

8.1 Introduction

Carotenoids are natural fat-soluble pigments widely represented in nature, where they have very diverse functions (Britton et al. 1998, 2004). Carotenoids are universally present in photosynthetic species, where they play essential roles associated to photosynthesis. Their functions are better known in higher plants, where they are responsible of the yellow, orange or reddish colors of many fruits and flowers. However, their most important roles are related with light harvesting and photoprotection of the photosynthetic machinery (Domonkos et al. 2013), and they are present in considerable amounts in green tissues, where their colors are masked by chlorophyll. These pigments are also found in some animals to provide their characteristic colors, as some birds and fishes. Irrespective of their sporadic coloring functions, carotenoids are ubiquitous in animals, which get them through their food to produce retinoids, as the visual chromophore retinal (vitamin A) and the signaling molecule retinoic acid (Blomhoff and Blomhoff 2006). However, with some outstanding exception represented by certain aphids (Moran and Jarvik 2010), animals lack the genes required for their synthesis and get carotenoids through the diet. The carotenoids are also produced by non-photosynthetic microorganisms and are found in many archaea, bacteria, and fungi (Sieiro et al. 2003; Avalos and Cerdá-Olmedo 2004; Avalos et al. 2014a; Sandmann and Misawa 2002).

From another point of view, several carotenoids are valuable biotechnological compounds, with diverse commercial applications in cosmetics and food industry, and some fungi are used as industrial carotenoid producers (Avalos and Cerdá-Olmedo 2004). Carotenoids are used, e.g., to provide color to egg yolks in poultry, to flesh of some fish, or to crustaceans’ shells in aquaculture. Different epidemiological and clinical studies have revealed counteracting effects of the consumption of carotenoids on several chronic illnesses, resulting in reduced risks of heart disease, cancer, eye disease, and other health disorders (Krinsky and Johnson 2005; Stahl and Sies 2005; Rao and Rao 2007), with a positive impact specially in older people (Woodside et al. 2015). Their health-promoting properties, attributed to a large extent to their antioxidant properties (Stahl and Sies 2003; Tapiero et al. 2004), are leading to an expanding interest in the biotechnological production of carotenoids, which are increasingly used as food and feed additives or as vitamin dietary supplements.

The carotenoids are terpenoid compounds containing an aliphatic polyene chain usually composed of eight isoprene units. Terpenoids, also known as isoprenoids, are a vast family of organic chemicals that derive from the sequential condensation of 5-carbon (C5) isoprene units. Their heterogeneity comes from variations in the number of units and in subsequent chemical modifications, which gives rise to a large structural diversification. Thus, terpenoids cover from small volatile compounds, as geraniol or limonene, to essential cell components as sterols, dolichols, or ubiquinone. The C5 precursor in the condensation steps is isopentenyl pyrophosphate (IPP). Depending on the organism or the cell compartment, IPP may be synthesized through two different biochemical pathways. The first one proceeds from acetyl-CoA through hydroxymethylglutaryl coenzyme A (HMG-CoA) and mevalonate (Fig. 8.1), and it is known as mevalonate pathway. The second one, known as methylerythritol pathway, was discovered later in bacteria (Rohmer et al. 1993) and involves the condensation of hydroxyethyl-thiamin and glyceraldehyde 3-phosphate, via d-1-deoxyxylulose 1-phosphate. The plants have both biosynthetic pathways, but the available information in fungi, based on labeling of carotenoids upon addition of radioactive mevalonic acid in different species, indicate the origin of fungal carotenoids through the mevalonate pathway.

Fig. 8.1
figure 1

First steps of the mevalonate terpenoid pathway. The branches for sterol and carotenoids biosynthesis are indicated

The early steps of terpenoid biosynthesis are shown in Fig. 8.1. The first condensation step, between IPP and its isomer dimethylallyl diphosphate (DMAPP), results in the production of the C10 monoterpene geranyl pyrophosphate (GPP). Subsequent IPP additions give rise to the C15 sesquiterpene C15 farnesyl pyrophosphate (FPP) and the C20 diterpene geranylgeranyl pyrophosphate (GGPP). FPP is the source of sterols, whose connection with cholesterol biosynthesis has attracted considerably the attention of researchers (DeBose-Boyd 2008). Similarly, this early part of the pathway is particularly relevant because of the use of FPP and GGPP in posttranslational modifications of proteins associated to cancer (Zahra Bathaie et al. 2016). Carotenoids are tetraterpenoids derived from phytoene, a colorless precursor generated by a head-to-head condensation of two GGPP molecules (Fig. 8.2), a reaction achieved by the enzyme phytoene synthase. The reaction is very similar to the condensation of two FPP units to produce squalene, the precursor of sterols.

Fig. 8.2
figure 2

β-carotene biosynthesis from GGPP. Enzymes responsible for each reaction in Mucorales are indicated close to the chemical modification from each enzymatic step (gray arrowhead)

In all the carotenoid pathways, the phytoene hydrocarbon backbone is object of several desaturations, which generate a chain of conjugated double bonds that confers the ability to absorb visible light, typically in the blue region of the spectrum. This chromophore provides the typical yellow, orange, or reddish pigmentations to the different carotenoids, depending on their specific absorption spectra. In the fungus Phycomyces blakesleeanus, phytoene is synthesized in a cis configuration, which is isomerized to its trans isomer in the first desaturation step (Fraser and Bramley 1994). Typically, carotenoids contain one or two cycled ends, known as ionone rings, resulting from a cyclase enzymatic activity. The rings may be α or β depending on the position of a C=C double bond, with the β-ionone ring being predominant in fungi. Additionally, the carotenoids may suffer other enzymatic reactions, explaining their wide structural diversity in nature (Britton et al. 2004). In the absence of oxygen in the molecule, the carotenoids are termed carotenes. However, the carotenoid biosynthetic pathways frequently include oxidative steps, which result in the production of xanthophylls.

Carotenoid biosynthesis is frequent but not universal in fungi. Thus, two well-known fungal models, the yeast Saccharomyces cerevisiae and the mold Aspergillus nidulans, do not produce carotenoids. However, carotenoid production is present in many fungi, either unicellular of filamentous, and the genetics and biochemistry of their biosynthetic pathways have been subjected to detailed attention in some model systems, which are described in the next sections.

8.2 Carotenoid Biosynthesis in Filamentous Fungi

Carotenoid biosynthesis has been investigated in different filamentous fungi, but some of them stand out for the amount of available information. These are the mucorales (a phylum of Mucoromycotina, formerly known as zygomycetes) Phycomyces blakesleeanus, Mucor circinelloides, and Blakeslea trispora for β-carotene biosynthesis and the ascomycetes Neurospora crassa and Fusarium fujikuroi for neurosporaxanthin biosynthesis. In the next sections, we update the available information on the production of these carotenoids, with special attention to the biotechnological production in the case of β-carotene.

8.2.1 β-Carotene Biosynthesis in Mucorales

The yellowish β-carotene, one of the most ubiquitous carotenes in nature, is the one usually found in mucorales fungi, as shown in P. blakesleeanus (Cerdá-Olmedo 1987), M. circinelloides (Navarro et al. 1995; Fraser et al. 1996), and B. trispora (Lampila et al. 1985). β-carotene production has been also found in filamentous fungi from other taxonomic groups, as the basidiomycetes Sclerotium rolfsii (Georgiou et al. 2001b) and Sclerotinia sclerotiorum (Georgiou et al. 2001a), the ascomycetes Aspergillus giganteus (El-Jack et al. 1988), Cercospora nicotianae (Daub and Payne 1989), and Penicillium sp. (Han et al. 2005), as well as in the “imperfect fungus” Aschersonia aleyroides (van Eijk et al. 1979). As cited in later sections, β-carotene is also present in variable amounts as a secondary product or as an intermediary molecule of the carotenoid pathway in other fungi.

β-carotene biosynthesis requires four desaturations on the phytoene backbone, resulting in the reddish intermediate lycopene, and the β-cyclization of both ends of the molecule (Fig. 8.2). Because of the symmetry of phytoene, the four desaturations are actually two pairs, one corresponding to internal positions and another to external positions. In photosynthetic organisms two different enzymes, typically known as phytoene and ζ-carotene desaturases, achieve each couple of desaturations (Domonkos et al. 2013). However, a single enzyme carries out the four desaturations in fungi. Similarly, separate genes encode the phytoene synthase and the cyclase enzymes in photosynthetic species and non-photosynthetic bacteria, but both enzymatic activities reside in a single polypeptide in fungi. Therefore, only two fungal genes are needed to make β-carotene from GGPP, one encoding a bifunctional phytoene synthase/lycopene cyclase and another encoding a desaturase. The desaturase gene has been described in C. nicotianae (Ehrenshaft and Daub 1994) and in the mucorales P. blakesleeanus (Ruiz-Hidalgo et al. 1997a), M. circinelloides (Velayos et al. 2000a), and B. trispora (Rodríguez-Saiz et al. 2004), where it is known as gene carB. In the three species of the mucorales group investigated, the phytoene synthase/lycopene cyclase gene, called carRA or carRP, is linked in the genome with the gene carB and divergently transcribed from a common upstream region (Velayos et al. 2000b; Arrach et al. 2001; Rodríguez-Saiz et al. 2004), forming a single regulatory unit (Fig. 8.3).

Fig. 8.3
figure 3

(a) Genomic organization of the carRA and carB genes in M. circinelloides (Mc, data from strain CBS 277.49, NCBI reference AMYB01000001), B. trispora (Bt, strain NRRL2457, NCBI reference AY176663), and P. blakesleeanus (Pb, NRRL 1555, NCBI reference NW_017265138). (b) Schematic representation of the alignment of their sequences with MUSCLE (Edgar 2004) and visualization of alignment’s conservation with the CLC sequence viewer software (Version 7.7.1; CLC bio Inc., Massachusetts, USA). Darker lines indicate most conserved sequences. Percentages in the genes under the conservation graphic correspond to the lowest value of the identity matrix between the three sequences of the alignment. In the case of carRA, the percentages for the carR (left) and carA (right) domains are indicated

The functions of these genes have been supported by extensive genetic and biochemical analyses that started with the genetic characterization of their mutant phenotypes, resulting in the alteration of the wild-type yellow pigmentation. Thus, the mutants of the gene carB are albino and accumulate phytoene, while the mutants of the gene carRA are either albino and without phytoene, in the case of the alteration of the phytoene synthase domain, or reddish because of the accumulation of lycopene in the case of the loss of the lycopene cyclase domain. Such mutants have been described in P. blakesleeanus (Cerdá-Olmedo 1985, 1987), M. circinelloides (Navarro et al. 1995; Velayos et al. 1997), and B. trispora (Mehta and Cerdá-Olmedo 1995; Mehta et al. 2003). The first substrate of the desaturase is 15-cis-phytoene, as shows the accumulation of this isomeric form in the carB mutants of P. blakesleeanus (Goodwin 1980), and purified desaturase enzyme converts it to all-trans desaturated carotene products (Fraser and Bramley 1994). In P. blakesleeanus, all the phytoene-accumulating mutants belong to a single complementation group (Ootaki et al. 1973), which defines genetically the gene carB, and the heterokaryons carrying wild-type and carB mutant nuclei accumulate partially desaturated intermediates, as phytofluene, ζ-carotene, and neurosporene. Their amounts, in relation to the proportion of wild-type and carB mutant nuclei, are consistent with the operation of the enzyme as a four-unit complex (De la Guardia et al. 1971). The analysis of the carotenes produced by leaky mutants of the carB gene with different catalytic efficiencies for the different steps, in one case leading to a significant overaccumulation of ζ-carotene, provided further evidence to the participation of a single desaturase in the four reactions (Bejarano et al. 1987).

In relation to carRA/carRP, the bifunctional role of this gene was confirmed in M. circinelloides through the phenotypic effects of partial carRP deletions (Velayos et al. 2000b) and by expression in E. coli strains engineered to produce carotene but missing phytoene synthase or lycopene cyclase activities (Sanz et al. 2011). Similar conclusions were reached in P. blakesleeanus with the correspondence between albino and lycopene-accumulating mutants of the carRA gene, called carA and carR, with mutations affecting the phytoene synthase or the lycopene cyclase in the same gene, respectively (Arrach et al. 2001). These molecular data corroborated former genetic interpretations of mutations in the carR and carA domains of the carRA gene, which led to propose the cleavage of the CarRA polypeptide in separate CarR and CarA enzymes (Torres-Martínez et al. 1980). This hypothesis was later supported by the finding of a conserved proteolytic cleavage site between both protein domains in P. blakesleeanus (Arrach et al. 2001) and independent 40-kDa CarR and 30-kDa CarA polypeptides by western blot analyses in B. trispora (Breitenbach et al. 2012).

The carB or carR mutant phenotypes may be reproduced by specific chemical block of the enzymatic activities. In P. blakesleeanus, cinnamic alcohol, thymol, and diphenylamine inhibit the phytoene desaturase (Bejarano and Cerdá-Olmedo 1989), while CPTA (2-[4-chlorophenylthio]-triethylamine), imidazole, nicotine, and other nitrogenated compounds obstruct the lycopene cyclase (Elahi et al. 1973a, b). Different lines of evidence, based on quantitative analysis of heterokaryons between carRA and carB mutants (De la Guardia et al. 1971; Aragón et al. 1976), the effect of specific enzyme inhibitors on carotene proportions (Candau et al. 1991b), and complementation analysis between null and leaky carB mutants (Sanz et al. 2002), strongly support the organization of the enzymes as a complex consisting of four desaturases and two cyclases. In the complex, the 4 + 2 monomers carry out sequentially the six reactions, and all the desaturase and cyclase units are able to act on any of their corresponding substrates, which are transferred internally. However, this scenario may differ between different mucorales, as indicates the finding in B. trispora of two different lycopene cyclases, one similar to the CarR enzyme of P. blakesleeanus, and a second one insensitive to inhibitors and involved only in the cyclization of lycopene into γ-carotene (Mehta and Cerdá-Olmedo 1999).

There is very limited information on the physical location of carotene synthesis in fungal cells. In P. blakesleeanus, it occurs in a distinct cell compartment than sterol biosynthesis despite the coincidence in the first enzymatic steps. This is evidenced by the different specific radioactivity found in ergosterol and β-carotene in the wild type and in a carS β-carotene-overproducing mutant, mentioned in the next section, when 14C-labeled mevalonate was added to the culture medium (Bejarano and Cerdá-Olmedo 1992). Therefore, both pathways are physically separated and use independent precursor pools. A similar experimental approach also identified independent syntheses for ubiquinone, triacylglycerols, and β-carotene in P. blakesleeanus and B. trispora (Kuzina et al. 2006). Taking together, the available data are consistent with control mechanisms to locate the enzymes for the synthesis of different terpenoids, including those of carotenogenesis, in appropriate cell compartments.

In different zygomycetes it is well established that β-carotene is not an end product but the source of a large array of apocarotenoid derivatives, which comprise at least C18 trisporoids, C15 cyclofarnesoids, and C7 methylhexanoids (Barrero et al. 2011). C18 trisporoids include a family of chemicals, the trisporic acids (Austin et al. 1969, 1970), which play the role of sexual hormones in the life cycle of these fungi (see next sections). β-carotene degradation starts with the activity of carotenoid oxygenases (CCDs), a family of enzymes catalyzing the oxidative cleavage of specific carotenoid substrates to produce apocarotenoids (Ahrazem et al. 2016). The first CCD enzymes in this fungal group where identified in Rhizopus oryzae (Tsp3 and Tsp4) and B. trispora (Tsp3), their names coming from their presumed participation in the production of trisporic acids (Burmester et al. 2007). The apocarotenoids found in P. blakesleeanus are consistent with the sequential cleavage of β-carotene at two different internal positions (Polaino et al. 2010) (Fig. 8.4), the first cleavage carried out by the CCD enzyme CarS, ortholog of Tsp3 (Medina et al. 2011; Tagua et al. 2012). The gene carS was formerly interpreted as regulatory (see next sections) because of the accumulation of large amounts of β-carotene resulting from its mutation. The second cleavage reaction, achieved on the CarS product β-apo-12′-carotenal, is carried out in P. blakesleeanus by the CCD AcaA (Medina et al. 2011). The genome of this fungus contains a third predicted CCD enzyme, but its function has not yet been elucidated. Other genes and enzymes for trisporoid metabolism are currently under investigation (Avalos et al. 2014a), but they are out of the scope of this review.

Fig. 8.4
figure 4

Cleavage reactions of β-carotene in P. blakesleeanus. Dotted gray lines in β-carotene and β-apo-12′-carotenal indicate cleavage sites for the CCDs CarS and AcaA, respectively

8.2.2 Regulation of β-Carotene Biosynthesis in Mucorales

In the three mucorales that have received more attention, P. blakesleeanus, M. circinelloides, and B. trispora, the synthesis of β-carotene is stimulated by light and, depending on the strains, by the interaction with a strain of the opposite sex. As described above, the three species have the same β-carotene biosynthetic pathway mediated by the products of a similar set of divergently transcribed genes. However, the coding sequences of the carB or carRA/carRP genes are more similar in the three species than their regulatory intergenic sequences (Fig. 8.3), and there are differences in their responses to external signals.

Surface cultures of wild-type P. blakesleeanus grown in the light contain about tenfold more β-carotene than those grown in the dark (Bergman et al. 1973; Bejarano et al. 1991; Salgado et al. 1991), and a comparable induction is found in M. circinelloides (Navarro et al. 1995; Velayos et al. 1997). Such stimulation was not initially described for B. trispora (Sutter 1970), but when a dark-grown culture of this fungus is illuminated, its β-carotene content increases (Quiles-Rosillo et al. 2005). The regulation by light in the mucorales is also associated to the different developmental stages exhibited by their surface cultures, which include the formation of asexual fruiting bodies called sporangiophores. This process has been investigated in special detail in P. blakesleeanus, which is able to form two classes of sporangiophores. Light represses the formation of small ones, called microphores, and stimulates the formation of large ones, macrophores, in a process known as photomorphogenesis (Corrochano and Cerdá-Olmedo 1992). Illumination of this fungus during 12-h periods at different growth times showed that carotenogenesis is able to respond to light only during a specific developmental stage, coinciding with the cessation of growth and the start of sporangiophore formation (Bejarano et al. 1991). These authors also found two levels of light sensitivity of photocarotenogenesis in this fungus, with a minor increase of β-carotene content after a weak illumination and a much stronger response observed only with a more intense illumination.

Regulation by light of the genes carB and carRA/carRP is achieved at transcription level. The magnitude of the photoinduction, referred to the mRNA content in the dark, is particularly high in M. circinelloides, where 10 s of blue light are sufficient to produce about a 100-fold increase of the transcript levels during the first 20 min after light exposure (Velayos et al. 2000a, b). Photoinduction is also exhibited in M. circinelloides in the gene carG, coding for the GGPP synthase (Velayos et al. 2003) providing the substrate for CarRA phytoene synthase activity. A similar carRA/carB response is exhibited by B. trispora, but in this case the photoinduction is ephemeral, and the mRNA content returns to dark levels just 40 min after the light pulse. This feature may explain the lack of increase in carotenoid content under continuous illumination in this species. In contrast, the photoinduction rate is not so strong in P. blakesleeanus, probably because of the presence of higher mRNA amounts in the dark. Thus, the exposure to light in this fungus results in less than a tenfold increase in carB (Ruiz-Hidalgo et al. 1997b; Blasco et al. 2001) and carRA mRNA levels (Almeida and Cerdá-Olmedo 2008; Sanz et al. 2010). The time course of mRNA accumulation shows a clear biphasic kinetics for both genes in P. blakesleeanus (Sanz et al. 2010), which could be related with the two different levels of light sensitivity formerly found for photocarotenogenesis in this fungus (Bejarano et al. 1991).

The analysis of mutants affected in photoinduction has been a valuable tool to identify the regulatory proteins responsible for the control by light. Mutagenic searches in P. blakesleeanus led to identify several mutants with reduced photocarotenogenesis: madA, madB, picA, picB (López-Díaz and Cerdá-Olmedo 1980), carC (Revuelta and Eslava 1983), and pimA (Flores et al. 1998). The madA, madB, and pimA mutants are also affected in photomorphogenesis, and they were actually identified because of this developmental alteration. Photocarotenogenesis is also modified by the presence of inhibitors of protein phosphatases and kinases (Tsolakis et al. 1999), suggesting that at least some of these genes could participate in a signal transduction mechanism involving phosphorylation events. Mutants with reduced carotene content in the light have also been described in M. circinelloides (Navarro et al. 1995; Velayos et al. 1997).

While the mutants of the genes madA and madB are only partially affected in photocarotenogenesis, the double mutant madA madB is mostly insensitive (Jayaram et al. 1980; López-Díaz and Cerdá-Olmedo 1980). A combination of genetic and genomic data led to identify the gene madA in the P. blakesleeanus genome (Idnurm et al. 2006), encoding a protein similar to the flavin photoreceptor White Collar-1 (WC-1) of N. crassa, responsible of photocarotenogenesis in this fungus (described in a later section). Accordingly, the action spectrum of carotenogenesis in P. blakesleeanus is consistent with the mediation of flavin photoreceptors (Bejarano et al. 1991). In N. crassa, WC-1 interacts with a smaller protein, WC-2, to form a photoactive complex. In a similar way, MadB is orthologous of WC-2 and it interacts with MadA (Sanz et al. 2009). Unexpectedly, the analysis of the genome of P. blakesleeanus revealed two additional WC-1-like proteins, named WcoA and WcoB, and three WC-2-like proteins, WctB, WctC, and WctD (Corrochano and Garre 2010), pointing to a remarkable complexity of the photosensory systems in this fungus. The action spectra for the two sensitivity levels of photocarotenogenesis, mentioned above, are consistent with the participation of flavin photoreceptors, but are not totally coincident (Bejarano et al. 1991), indicating differences in their respective photoreception systems. This could be explained by the participation of other WC-1-like photoreceptor or the combination of MadA with a different WC-2-like protein. Regrettably, at present there is no methodology available to obtain knockout transformants in P. blakesleeanus, and therefore the functions of the proteins WcoA, WcoB, WctB, WctC, and WctD remain to be elucidated. Yet, such methodology is available in M. circinelloides (Torres-Martínez et al. 2012), which contains a similar set of White Collar-1 genes, called mcwc-1a, mcwc-1b, and mcwc-1c. In contrast to P. blakesleeanus, the absence of the MadA counterpart, MCWC-1a, does not affect photocarotenogenesis but prevents sporangiophore phototropism. On the other hand, photocarotenogenesis is lost in the mcwc-1c null mutant while no phenotypic alteration is apparent in the mcwc-1b null strain (Silva et al. 2006). The functions of the MCWC-1a and MCWC-1c proteins are interconnected: unlike the mcwc-1a and mcwc-1b genes, expression of mcwc-1c is strongly induced by light, but this induction requires a functional mcwc-1a gene. Moreover, carB and carRP photoinduction is severely reduced in the absence of mcwc-1a (Silva et al. 2006).

8.2.3 β-Carotene Overproduction in Mucorales

The mutagenesis screenings in mucorales allow the isolation of carotenoid-overproducing mutants, which stand out because of their deeper pigmentation. In P. blakesleeanus such mutants are affected in three genes: carS (Murillo and Cerdá-Olmedo 1976), already mentioned because of its role in β-carotene cleavage (Medina et al. 2011; Tagua et al. 2012), carD (Salgado et al. 1989), and carF (Mehta et al. 1997). Depending on the mutated gene and the allele, the carotene content in the dark raises from 10- to 100-fold that of the wild type. This increase is not accompanied by concomitant increments in carRA and carB transcripts (Almeida and Cerdá-Olmedo 2008), but at least in the case of the carS and carD mutants, they exhibit higher capacities of incorporation of mevalonic acid into carotene in their cell extracts (Salgado et al. 1991), suggesting posttranscriptional regulatory effects. The carotene increase is less pronounced, about three- to fivefold, in carotene-overproducing mutants of B. trispora (Mehta and Cerdá-Olmedo 1995; Mehta et al. 2003) and in some mutants of M. circinelloides (Navarro et al. 1995; Fraser et al. 1996). However, the null mutants of the gene crgA of this species reach levels as high as 80-fold the content of the wild type (Navarro et al. 2001), reminding the phenotype of the carS mutants of P. blakesleeanus.

Despite their similarities, the molecular basis for carotene overproduction is very different in the mutants of the gene carS of P. blakesleeanus and those of crgA of M. circinelloides. As already indicated, CarS is a β-carotene-cleaving oxygenase leading to the formation of apocarotenoid derivatives that are not visually detected because of their lack of pigmentation. Therefore, β-carotene overaccumulation in the carS mutants is explained by a block of its CarS-mediated degradation. This leads to reconsider a former regulatory hypothesis, according to which the carS mutants were affected in the detection of β-carotene in a feedback control mechanism (Bejarano et al. 1988). However, the occurrence of a feedback regulation through the detection of an apocarotenoid product is not yet discarded. Moreover, a mutagenesis screening from a carS mutant led to identify albino mutants with a second mutation in the carS gene (Salgado and Cerdá-Olmedo 1992), whose molecular basis remains to be explained.

The CrgA protein of M. circinelloides, which has no sequence relation with CarS, consists of two amino-terminal RING finger (RF) domains, two glutamine-rich regions, a LON protease domain, and a carboxy-terminal isoprenylation site (Navarro et al. 2000). The RF domains have been described to interact with E3 ligase-type proteins that mediate ubiquitylation of target proteins, usually labeling them for their degradation. At least one of the RF domains is essential for its regulatory function in carotenogenesis, suggesting that CrgA may function as an E3 ubiquitin ligase (Lorca-Pascual et al. 2004). β-carotene overproduction in the crgA mutants is due to a strong increase in the carRA and carB transcript levels, pointing to CrgA as a negative regulator (Navarro et al. 2001; Lorca-Pascual et al. 2004), a function in which it seems to play a role in the ubiquitylation-independent degradation of MCWC-1b (Silva et al. 2008). CrgA also regulates vegetative growth and sporulation in M. circinelloides, suggesting a wider regulatory function for this protein (Quiles-Rosillo et al. 2003; Murcia-Flores et al. 2007). However, despite the carotene overaccumulation, the crgA mutants maintain the ability to respond to light with a further increase in the carotene content and in the transcription of at least the carB gene (Navarro et al. 2001). The lack of efficient transformation procedures hinders the study of the function of crgA orthologs in P. blakesleeanus and B. trispora, but the crgA gene of B. trispora is able to complement the crgA mutation in M. circinelloides, indicating conserved functions for this protein in both species (Quiles-Rosillo et al. 2005).

As found with the carS mutants, the loss of either phytoene desaturase in carB mutants or of lycopene cyclase in carR mutants (R domain of gene carRA) leads to the accumulation of large amounts of phytoene or lycopene (Ootaki et al. 1973), respectively, and similar results are obtained by their chemical inactivation (Bejarano and Cerdá-Olmedo 1989; Candau et al. 1991b). This high carotene production was interpreted as a clue for the occurrence of feedback or end product regulation in this species. The increases are more variable, depending on the strain or culture conditions, in similar mutants of M. circinelloides (Navarro et al. 1995; Velayos et al. 1997) and only minor in those of B. trispora (Mehta and Cerdá-Olmedo 1995). Despite their large carotene content, light is still able to stimulate phytoene accumulation in the carB mutants of P. blakesleeanus as well as that of β-carotene in some carS mutants (Bejarano et al. 1991). However, light is mostly ineffective on carA or carR mutants, some of them carrying leaky mutations, suggesting a regulatory role for the CarRA polypeptide before its proteolytic cleavage. No photoinduction could be detected either in double carA carS mutants, with a carotene content halfway between those of the two separate mutants (Bejarano et al. 1991). Other observations point to further functions for the CarA protein, facilitating substrate transfer between enzymatic aggregates in the cyclization reactions (Murillo et al. 1981) or participating in the light sensitivity of photomorphogenesis (Corrochano and Cerdá-Olmedo 1990).

β-carotene accumulation in P. blakesleeanus is stimulated by the presence of different chemicals, distributed in at least two families of compounds acting through independent mechanisms (Bejarano et al. 1988). The first family consists of apocarotenoids with a terminal β-ring in the molecule, such as retinol (vitamin A) or β-ionone (Eslava et al. 1974), which predictably exert their action competing with β-carotene in the binding to an enzyme or regulatory protein. In the presence of vitamin A, β-carotene concentration of the wild type rises to levels comparable to those of phytoene or lycopene in the carB or carR mutants in the absence of the inducer, but it is still able to increase significantly the carotene content of these mutants (Eslava et al. 1974). As already mentioned, the carA mutants are insensitive to light, but they exhibit a patent β-carotene increase in the presence of vitamin A. Such increase is similar to that exhibited by the carA carS double mutant, which hardly responds to vitamin A. According to the available information, it was proposed a regulatory model in which β-carotene interacts with a carA/carS complex to downregulate the pathway (Bejarano et al. 1988), but such regulatory mechanism has not been demonstrated experimentally.

The second family of chemicals consists of a series of aromatic compounds (Cerdá-Olmedo and Hüttermann 1986), with veratrol and dimethyl phthalate among the most efficient inducers. Their mechanism of action is independent from that of vitamin A, and therefore their effects are additive (Bejarano et al. 1988). The desaturase inhibitor diphenylamine is a biphenolic compound, and, not surprisingly, some of the phenolic activators, e.g., cinnamic alcohol, also interfere with the desaturase activity (Bejarano and Cerdá-Olmedo 1989). The mechanism of action of the phenolic activators in P. blakesleeanus has not been unraveled.

When two strains of P. blakesleeanus of opposite sex are grown together, the cultures (known as mated cultures) exhibit a fivefold increase in their β-carotene levels compared to the respective single cultures (Govind and Cerdá-Olmedo 1986). This stimulation is visualized as a pigmented band when the mycelia of the two strains meet on an agar surface (see, e.g., Tagua et al. 2012) and does not require a physical contact (Burgeff 1924), indicating the mediation of diffusible products. The stimulations of the synthesis of β-carotene and the formation of sexual structures in mated cultures are probably independent, as suggest the opposite effect of acetate on both processes (Kuzina and Cerdá-Olmedo 2006). The sexual stimulation of carotene biosynthesis is the result of a strong induction in the expression of the carRA and carB genes (Almeida and Cerdá-Olmedo 2008) and in the corresponding enzymatic activities (Salgado et al. 1991).

The upregulation of carotenogenesis by mating is due to the production of trisporic acids, sexual hormones already mentioned in Sect. 8.2.1 whose synthesis requires the collaboration of the two partners (Schachtschabel et al. 2008) and triggers their sexual differentiation (Sutter 1987). At least in B. trispora, either the sexual interaction or the addition of trisporic acids stimulates the expression of the tsp3 gene, needed for the production of these hormones (Burmester et al. 2007). Actually, addition of trisporic acids to single cultures reproduces the carotenoid stimulation in P. blakesleeanus (Govind and Cerdá-Olmedo 1986) or in other related species (Sahadevan et al. 2013), and in P. blakesleeanus it is additive to the inducing effects of dimethyl phthalate, light, or carS mutation (Govind and Cerdá-Olmedo 1986). The sexual stimulation is more efficient in sexual heterokaryons, i.e., those holding nuclei of the strains of the opposite sex in the same cytoplasm. Such heterokaryons are unstable because of random variations in the nuclear proportions but may be stabilized through the introduction of recessive lethal mutations in the partner nuclei (Murillo et al. 1978). The β-carotene content in surface cultures of these heterokaryons increases considerably if the participating nuclei hold carS mutations, reaching up to 2.5% of the total dry mass in media containing cheap industrial subproducts (Cerdá-Olmedo 1989). The carotene levels are not far from this value in partial sexual diploids with a carF mutation (Mehta and Cerdá-Olmedo 2001) and may reach about 3% of mycelial dry mass in a carS carF double mutant (Mehta et al. 1997).

Production of β-carotene has not reached so high levels in M. circinelloides, but in contrast to P. blakesleeanus or B. trispora, this species is amenable to genetic transformation, making possible gene-engineered manipulations (Torres-Martínez et al. 2012). This methodology allowed the generation of strains with increased copy numbers of genes for early steps of the terpenoid pathway, resulting in improvements of up to fivefold in β-carotene production (Csernetics et al. 2011) or with modifications of the carotenoid pathway through the expression of heterologous genes for β-carotene oxidizing enzymes, leading to the accumulation of variable amounts of xanthophylls as β-cryptoxanthin, zeaxanthin, echinenone, canthaxanthin, or astaxanthin (Papp et al. 2006, 2013; Csernetics et al. 2011, 2015). The molecular procedures for these engineered modifications in M. circinelloides have been recently reviewed (Barredo 2012).

8.2.4 B. trispora as an Industrial Carotene Source

The global carotenoid market has kept growing in the recent years, and β-carotene is the most commercialized product (Kirti et al. 2014). β-carotene may be obtained by chemical synthesis or by biotechnological production, with B. trispora and the alga Dunaliella salina as the preferred microbial sources (Ribeiro et al. 2011). In the case of B. trispora, the basis of its biotechnological application is the high β-carotene yields resulting from the sexual interaction in submerged mated cultures, consistent with increased carRA/carB expression (Schmidt et al. 2005) and CarRA/CarB enzymatic levels (Breitenbach et al. 2012). The sexual stimulation is accompanied by many other physiological and developmental changes and affects the expression of many other genes (Kuzina et al. 2008). As a different approach, addition of trisporic acids to single cultures reveals that these hormones act as global metabolic regulators, producing changes in total protein patterns and in metabolism of fatty acids, amino acids, and carbohydrates (Sun et al. 2012). At industrial level, β-carotene production by B. trispora is based on fermentation technology, which implies large-scale growth of strains of opposite sex in submerged conditions and feasible subsequent industrial steps, such as adequate biomass separation and carotenoid extraction and purifications procedures.

A major point in the fermentation procedure is the use of suitable mating partners. Since they may differ in their germination and growth capabilities, the use of adequate proportions of the spores used to start the cultures is an important aspect in the process (Böhme et al. 2006). The production has been improved trough different experimental approaches. Thus, the comparison of the β-carotene levels after growth with different carbon and nitrogen sources, trace elements, pH, or inoculum sizes, as well as the result of combining varying concentrations of different media components, allowed a 42% yield increase (Choudhari and Singhal 2008). The production increases considerably through the use of β-carotene-overproducing mutants. B. trispora spores are multinucleate, which hinders the isolation of recessive mutants, but efficient mutagenesis protocols are available based on exposure to the potent chemical mutagen N-methyl-N′-nitro-N-nitrosoguanidine (Cerdá-Olmedo and Mehta 2012). A color screening of B. trispora colonies derived from spores surviving to this mutagenic agent allowed the identification of “superyellow” mutants, with β-carotene content reaching up to sixfold higher levels than those of the original strain (Mehta and Cerdá-Olmedo 1995). This yield may be improved further in a new round of mutagenesis from a superyellow mutant (example in Mehta et al. 2003), a typical strategy in industrial strain improvements that may be repeated many times. The β-carotene level increases in the superyellow mutants are not as high as those produced by sexual stimulation between wild-type strains, but the use of such mutants in mated cultures results in very high β-carotene productions (Mehta et al. 2003). Recently, alternative screening protocols and novel mutagenesis methods, as N+ ion implantation (Wang et al. 2013) and atmospheric and room temperature plasma (Qiang et al. 2014), proved also efficient to obtain carotene-overproducing strains.

Because of the industrial use of B. trispora, chemical activation of its β-carotene production has received intense attention. The aromatic chemicals that stimulate carotenogenesis in P. blakesleeanus are not effective in this fungus, and those of the retinoid group are less effective. Thus, addition of retinol to the medium results in only a twofold to threefold increase of β-carotene production (Choudhari et al. 2008), compared to a 20-fold increase in P. blakesleeanus (Eslava et al. 1974; Bejarano et al. 1988). Earlier studies found minor stimulatory effects with a large diversity of chemicals, including certain monoterpenes, as α-pinene (Cederberg and Neujahr 1969), and different nitrogenated compounds, as succinimide, pyridazyne, isonicotinoylhydrazine, or some pyridine derivatives (Ninet et al. 1969). More recent reports described new activators at appropriate concentrations and proposed putative inducing mechanisms. Addition of sorbitan monolaurate surfactant span-20 duplicates β-carotene production presumably through the shortening of the hyphal length and improvement of dispersed growth (Choudhari et al. 2008). The presence of ketoconazole, an inhibitor of the enzyme converting lanosterol to ergosterol, results in a threefold increase of β-carotene content, possibly because of compensatory effects in early steps of terpenoid biosynthesis (Tang et al. 2008). Addition of arachidonic acid at the appropriate moment during the fermentation process results in a 1.7-fold rise in the β-carotene content, which correlates with comparable increases in the mRNA levels of the hmgR, carRA, and carB genes (Hu et al. 2012) and with metabolic changes leading to increased glycolysis and fatty acid biosyntheses (Hu et al. 2013a). More predictable stimulatory effects are those produced by addition of terpenoid precursors, as mevalonic acid, isopentenyl alcohol, dimethylallyl alcohol, or geraniol (Shi et al. 2012).

Many observations point to a positive correlation between oxidative stress and β-carotene production in the B. trispora cultures. The fungus protects itself against reactive oxygen species with specific detoxifying enzymes, among which play a major role the catalases and superoxide dismutases (SOD). Addition of H2O2 results in a rise in β-carotene accumulation (Jeong et al. 1999), which is accompanied by a higher catalase activity (Wang et al. 2014). On the other hand, chemical inhibition of SOD activity increases the β-carotene content (Gessler et al. 2002). A similar effect is obtained increasing the oxidative stress by raising the dissolved oxygen concentrations in the medium, either directly (Nanou and Roukas 2011) or with the presence of n-hexadecane (Liu and Wu 2006a), n-hexane, n-dodecane (Xu et al. 2007), or butylated hydroxytoluene (Nanou and Roukas 2010). In the cases investigated, SOD and catalase activities are simultaneously increased in parallel to β-carotene concentrations. These enzyme activities also raise upon addition of liquid paraffin to the medium, which enhance oxygen concentration and β-carotene production (Hu et al. 2013b). However, not all agents producing oxidative stress induce β-carotene accumulation in this fungus, as showed the lack of induction upon addition of iron ions (Nanou and Roukas 2013). The endogenous and exogenous factors producing oxidative stress in B. trispora and the effects on the fungus under fermentative conditions have been recently reviewed (Roukas 2016).

Considerable efforts have been dedicated to develop novel medium compositions with aim of reducing costs and increasing β-carotene yields. A detailed scrutiny on the effect of a diversity of media components showed improved productions in the presence of corn steep liquor, oleic and linoleic acids, kerosene, and the already mentioned butylated hydroxytoluene (Mantzouridou et al. 2002). The positive impact of oils in the production has been confirmed in other reports, such as those based on the use of crude olive or soybean oils (Mantzouridou et al. 2006) or waste cooking oil (Nanou and Roukas 2016). Supplementation of the medium with industrial glycerol, obtained either from soap manufacturing or biodiesel production industries, allows up to a tenfold increase in β-carotene levels (Mantzouridou et al. 2008). Very promising results have been also obtained using agro-food wastes rich in carbohydrates and mineral salts, as beet molasses (Goksungur et al. 2004), cheese whey (Varzakakou and Roukas 2010), cabbage, and peach peels or watermelon husks (Papaioannou and Liakopoulou-Kyriakides 2012). The studies on the use of cheese whey provide a good example of a complete survey on the effect of different variables in the search of the optimal production conditions, either in flasks (Varzakakou and Roukas 2010) or in fermentation reactors (Varzakakou et al. 2011; Roukas et al. 2015). These and other reports are based on the behavior of wild-type strains, but they might be efficient also with β-carotene-overproducing mutants.

The conditions developed for the production of β-carotene are also applicable for the production of lycopene through the block of the cyclase activity, which is usually achieved by chemical inhibition (Mantzouridou and Tsimidou 2008). Thus, lycopene is produced in high yields by mated cultures in large-scale fermentors in the presence of imidazole or pyridine (López-Nieto et al. 2004). Examples on the use of other inhibitors from this group are frequent in the literature, as reported, e.g., for 2-methyl imidazole (Pegklidou et al. 2008), piperidine, creatinine (Liu et al. 2012), or nicotine (Shi et al. 2012). Other chemicals used to block the cyclase activity are some amines, as CFTA (Hsu et al. 1972) and other amine-derived compounds (Hsu et al. 1974; Wang et al. 2016). A more efficient approach consists in the use of mutants devoid of cyclase activity. Such mutants are independent of chemical treatments and allow very high lycopene productions, especially in the form of heterokaryons with nuclei of both sexes carrying additional overproducing mutations (Mehta et al. 2003).

8.2.5 Neurosporaxanthin Biosynthesis in Neurospora and Fusarium

Neurosporaxanthin is a C35 carboxylic apocarotenoid initially discovered in the fungus N. crassa. Early biochemical analyses of cultures from this fungus revealed a complex carotenoid mixture, consisting of neutral carotenes and xanthophylls (Zalokar 1954). This included an acidic carotenoid that was subsequently isolated, characterized, and named neurosporaxanthin (Zalokar 1957), whose chemical structure was determined as β-apo-4′-carotenoic acid (Aasen and Jensen 1965). This acidic xanthophyll was later found also in other fungi, as Arthrobotrys oligospora (Valadon and Cooke 1963), a mutant of Verticillium albo-atrum (Valadon and Mummery 1969), Podospora anserina (Strobel et al. 2009), and different Fusarium species (Avalos and Estrada 2010). Analyses of the carotenoid mixtures in N. crassa (Harding et al. 1969; Mitzka and Rau 1977), F. aquaeductuum (Bindl et al. 1970), and F. fujikuroi (Avalos and Cerdá-Olmedo 1987) are consistent with a biosynthetic pathway that shares the first steps with that of β-carotene but differs in the occurrence of an additional desaturation, a single cyclization, and two final oxidative steps (Fig. 8.5). The substrate of the cyclization may differ between different species. The identification of β-zeacarotene, γ-carotene, lycopene, and 3,4-didehydrolycopene in N. crassa indicates that the cyclization may occur on substrates with either three, four, or five desaturations, while 3,4-didehydrolycopene has not been described in F. fujikuroi or F. aquaeductuum. In some cases, neurosporaxanthin is subject of esterification reactions in the terminal carboxylic group; thus, it is detected as a methyl ester in Verticillium agaricinum (Valadon and Mummery 1977) and as a glycosyl ester in a marine Fusarium species (Sakaki et al. 2002a).

Fig. 8.5
figure 5

Neurosporaxanthin and β-carotene biosynthesis in N. crassa and Fusarium sp. Enzymes responsible for each reaction (AL-1, AL-2, CAO-2, and YLO-1 for N. crassa, and CarRA, CarB, CarT, CarD, and CarX for Fusarium sp.) are indicated close to the chemical modification produced in each enzymatic step (gray arrowhead). Dotted gray lines in torulene and β-carotene indicate cleavage sites for the corresponding CCD enzymes. Cleavage of β-carotene has been only found in Fusarium sp. Reactions and chemical structures up to ζ-carotene are as shown in Fig. 8.2

The genes and enzymatic activities responsible for all the steps needed for neurosporaxanthin biosynthesis are known in N. crassa (Avalos and Corrochano 2013) and F. fujikuroi (Avalos et al. 2014b). The first genes of the pathway were identified in N. crassa by means of the genetic analysis of albino mutants, called in order of identification al-1, al-2, and al-3 (Huang 1964). The gene al-3 encodes the prenyl transferase responsible for the synthesis of the phytoene precursor GGPP from FPP (Nelson et al. 1989; Carattoli et al. 1991), as demonstrated the lack of this activity in al-3 mutants (Spurgeon et al. 1979) or its capacity to replace a bacterial GGPP synthase in E. coli (Sandmann et al. 1993). This reaction is needed also for the synthesis of other essential terpenoids derived from GGPP, e.g., ubiquinone, dolichols or prenylated proteins, and therefore only leaky mutations are isolated for this gene (Barbato et al. 1996).

The synthesis of phytoene from GGPP is achieved in N. crassa by AL-2, whose gene was identified by its capacity to complement the al-2 mutation (Schmidhauser et al. 1994). The activity of AL-2 was inferred from the albino phenotype of the al-2 mutant and the sequence similarity of its carboxy-end region with phytoene synthase from bacteria or plants. A similar approach led to clone al-1, encoding the phytoene desaturase (Schmidhauser et al. 1990), with sequence similarity to bacterial desaturases and able to complement a mutant of Rhodobacter capsulatus lacking the desaturase activity (Bartley et al. 1990). As found for CarB in mucorales, the carotenoids produced by AL-1 in R. capsulatus indicated that this enzyme is able to catalyze four desaturations. Later in vitro and in vivo studies demonstrated that AL-1 acts as homomultimeric complex able to catalyze the five desaturations needed to convert phytoene to 3,4-didehydrolycopene or torulene in the neurosporaxanthin biosynthetic pathway (Hausmann and Sandmann 2000).

Likewise, as described for the CarRA/CarRP counterparts in mucorales, AL-2 is a bifunctional protein whose cyclase activity generates the β-ionone ring of torulene. The cyclase function of AL-2, initially disregarded because of the lack of mutants specifically affected in this enzymatic activity, was later confirmed by complementation when these mutants became available (Arrach et al. 2002). Moreover, enzymatic assays expressing AL-2 in appropriate E. coli strains not only demonstrated the cyclase function of the protein but also its preference to produce a single cyclization (Sandmann et al. 2006). The knowledge of the al-1 and al-2 sequences facilitated the cloning of its orthologs in other fungi, including carB and carRA in F. fujikuroi (Fernández-Martín et al. 2000; Linnemannstöns et al. 2002). The ability of CarB to carry out the five desaturations was evidenced by the isolation of a carB mutant specifically affected in the fifth desaturation (Prado-Cabrero et al. 2009). In the case of AL-2/CarRA, in contrast to what was observed in mucorales, the bifunctional protein is probably not separated in two independent polypeptides, as suggests the difficult identification of mutants affected in the cyclase activity and the finding in N. crassa of a mutant of the phytoene synthase domain that is also affected in the cyclase activity (Díaz-Sánchez et al. 2011a).

Based on the assumption that a CCD enzyme achieves the torulene cleavage step in neurosporaxanthin biosynthesis, the genomes of F. fujikuroi and N. crassa were searched for CCD-encoding genes, leading to the identification of two candidates in each species. Torulene-accumulating mutants formerly identified in F. fujikuroi (Avalos and Cerdá-Olmedo 1987) facilitated the assignation of torulene cleavage activity to one of them, that was called carT (Prado-Cabrero et al. 2007). Similar mutants were also characterized in N. crassa, allowing the identification of the carT ortholog, cao-2 (Saelices et al. 2007). In both species, the complementation of torulene-accumulating mutants and the analysis of their enzymatic activity in vivo and in vitro established the function of the CarT/CAO-2 proteins as CCD enzymes cleaving torulene to produce C35 β-apo-4′-carotenal. In the case of CAO-2, the enzyme was not active on γ-carotene, indicating the need of all the previous desaturations. The function of carT, together to those of carRA and carB, was also confirmed by targeted deletion experiments in Fusarium graminearum, also known as Gibberella zeae (Jin et al. 2010).

The last reaction of the pathway, the oxidation of the terminal aldehyde group to a carboxylic group, is achieved by the aldehyde dehydrogenase YLO-1 in N. crassa (Estrada et al. 2008b) and its ortholog CarD in F. fujikuroi (Díaz-Sánchez et al. 2011b). The name of the ylo-1 mutant comes from its characteristic yellow pigmentation resulting from the accumulation of a complex mixture of carotenoids, in which neurosporaxanthin is missing (Goldie and Subden 1973). Later chemical analyses showed the presence of unusual carotenoids (Estrada et al. 2008b; Sandmann et al. 2008), possibly derived from the β-apo-4′-carotenal predictably accumulated in the absence of YLO-1 activity. However, the young colonies of the carD mutant of F. fujikuroi exhibit an orange pigmentation, as expected from β-apo-4′-carotenal accumulation, but they turn progressively to yellow as they get older due to its conversion to β-apo-4′-carotenol and, possibly, also to derived fatty acid esters (Díaz-Sánchez et al. 2011b).

Despite the predominance of neurosporaxanthin in N. crassa and F. fujikuroi, presumably due to the preference of AL-2 and CarRA to introduce a single cyclization in the pathway, the carotenoid analyses in both species detect minor amounts of β-carotene (Mitzka and Rau 1977), indicating that these enzymes are able to carry out a second cyclization. In the case of N. crassa, in vitro studies showed that γ-carotene is not accepted as a substrate by AL-1 (Hausmann and Sandmann 2000), and therefore if the cyclization is produced on lycopene instead of on 3,4-didehydrolycopene, the resulting γ-carotene may be only converted to β-carotene (Fig. 8.5). The occurrence of a second cyclization might be more frequent than expected in F. fujikuroi and other Fusarium species, where the genes carRA and carB are linked in a gene cluster with gene carX, encoding a β-carotene-cleaving CCD enzyme (Thewes et al. 2005), and with carO, encoding a rhodopsin (Prado et al. 2004). CarO is a photoactive proton pump that plays a role in conidia germination (García-Martínez et al. 2015). The product of β-carotene cleavage is retinal, used as chromophore by the CarO rhodopsin and presumably by a second rhodopsin in these fungi, called OpsA (Estrada and Avalos 2009). Retinal has its maximal absorption in the UV-A region of the spectrum and therefore does not provide pigmentation. There are no reports on retinal levels in these fungi, hindering any estimation of the proportion of carotene substrates deviated to the β-carotene branch. N. crassa has also a gene for a photoactive rhodopsin (Bieszke et al. 1999a, b), but despite having a second CCD gene, no retinal-forming activity has been found so far in this fungus (Díaz-Sánchez et al. 2013).

Information on the physical location of neurosporaxanthin biosynthesis in the fungal cells is very scarce. F. fujikuroi is well known for the production of gibberellins, a group of growth promoting plant hormones (Tudzynski 2005). Gibberellins are terpenoids that share with carotenoids their origin from GGPP, but meanwhile the carotenoids are accumulated in the cell, the gibberellins are excreted to the medium. Experiments with F. fujikuroi cultures in the presence of 14C-labeled mevalonate, as mentioned before for P. blakesleeanus (Bejarano and Cerdá-Olmedo 1992), indicated different cellular locations and independent substrate pools for the synthesis of carotenoids, gibberellins, and sterols (Domenech et al. 1996). In the case of carotenoids, the majority of the enzymes are associated to membranes, making difficult their purification for in vitro studies (Bramley 1985). Consequently, the carotenogenic enzymes of N. crassa are partially solubilized by detergent treatments, and their activities are improved by lipid addition (Mitzka-Schnabel 1985). In the same fungus, carotenoids have been found associated to the external membrane of mitochondria (Neupert and Ludwig 1971), suggesting that they are synthesized in this organelle, but they have been found also in lipid globules and in membranes of the endoplasmic reticulum (Mitzka-Schnabel and Rau 1980). Regarding the association of the carotenogenic enzymatic machinery to membranes, it is particularly relevant that the aldehyde dehydrogenases YLO-1 and CarD include a terminal transmembrane domain (Estrada et al. 2008b; Díaz-Sánchez et al. 2011b), very infrequent in this protein family.

8.2.6 Regulation of Neurosporaxanthin Biosynthesis

As already mentioned for β-carotene production by P. blakesleeanus and M. circinelloides, the synthesis of neurosporaxanthin is stimulated by light in N. crassa (Avalos and Corrochano 2013) and Fusarium sp. (Avalos and Estrada 2010). This photoresponse has been investigated in especial detail in N. crassa, currently a leading model in fungal photobiology (Chen et al. 2010). When grown in the dark, the submerged cultures of this fungus are albino and contain phytoene, but its exposure to light results in the rapid accumulation of colored carotenoids, reaching a maximum in about 6 h (Rau et al. 1968) or 12 h (Zalokar 1954), depending on the culture conditions. Light dosage experiments showed that 1 min of light (Zalokar 1955) or even less (Zalokar 1955; Schrott 1980) is sufficient to cause a detectable response, but after the light pulse, the mycelia become temporarily insensitive to a second light exposure (Schrott 1981). Fluence response experiments showed a biphasic curve, with a first step resulting from just some seconds of exposure at the highest light intensity and a second one needing at least 16 min of illumination (Zalokar 1955; Schrott 1980). In contrast, at least 1 h is needed for a significant response in F. aquaeductuum (Bindl et al. 1970) and F. fujikuroi (Avalos and Schrott 1990).

N. crassa photocarotenogenesis is very sensitivity to temperature. The carotenoids accumulate in much higher amounts when dark-grown mycelia are illuminated and incubated at low temperature (Harding 1974) than at high temperature, with an optimal response between 6 and 12 °C. The effect of low temperature is not only on total carotenoid production but also on enzymatic efficiency, as indicates the accumulation of a higher proportion of neurosporaxanthin, a feature that facilitated the characterization of mutants affected in the last reactions of the pathway (Saelices et al. 2007; Estrada et al. 2008b). Additionally, the analysis of intermediary carotenoids showed an unexpected change in the order of the enzymatic steps depending on the temperature of illumination (Fig. 8.6), with the cyclization of apo-4′-lycopenoic acid as the last reaction at low temperature (Estrada et al. 2008a). Therefore, 3,4-didehydrolycopene may be recognized as a substrate not only by the cyclase activity of AL-2 but also by the cleaving oxygenase CAO-2, with the latter prevailing under cold conditions. Such effect has not been described in Fusarium, and, at least in F. aquaeductuum, the synthesis of carotenoids is less efficient at lower temperatures (Rau 1962).

Fig. 8.6
figure 6

Alternative pathways for late steps of neurosporaxanthin biosynthesis in N. crassa. The reactions shaded in gray are predominant under cold conditions. The dotted gray lines in 3,4-didehydrolycopene and torulene indicate cleavage sites for the CAO-2 CCD enzyme

Action spectrum of photocarotenogenesis in N. crassa reveals an optimal efficiency at light wavelengths ranging from 440 to 490 nm (Zalokar 1955), with peaks at about 450 and 480 nm (de Fabo et al. 1976). As already mentioned for P. blakesleeanus, the spectrum shape is consistent with the participation of flavin photoreceptors. This hypothesis was corroborated by the reduced photoinduction exhibited by flavin deficient mutants (Paietta and Sargent 1981), although such deficiency was not restored by external supplementation of riboflavin or some analogs (Paietta and Sargent 1983). Conclusive confirmation was attained with the identification of the responsible photoreceptor, which turned out to be a heterodimeric complex called White Collar (WC), consisting of the flavin photoreceptor WC-1 and its partner WC-2. Thus, the wc-1 and wc-2 mutants are unable to photoinduce carotenogenesis, a defect that was confirmed at the level of phytoene synthase activity (Harding and Turner 1981), carotenoid accumulation (Degli-Innocenti and Russo 1984), and mRNA levels for structural genes (Saelices et al. 2007). Upon its activation by light, the WC complex acts directly on the promoters of the target genes (He and Liu 2005), presumably binding to upstream regulatory elements as those found in the al-3 promoter (Carattoli et al. 1994). As a result, there is an increase in the mRNA levels of the structural genes al-1 (Schmidhauser et al. 1990), al-2 (Schmidhauser et al. 1994), al-3 (Baima et al. 1991), and cao-2 (Saelices et al. 2007). The response is very fast, reaching top levels after 15–30 min of illumination and decreasing afterward. However, the gene for the last step of the pathway, ylo-1, is not affected by light (Estrada et al. 2008b).

The regulation by light is basically conserved in Fusarium, as indicates the similar action spectrum described for F. aquaeductuum (Rau 1967) and the identification of the orthologous wc-1 genes wcoA in F. fujikuroi (Estrada and Avalos 2008) and wc1 in F. oxysporum (Ruiz-Roldán et al. 2008). However, the WC heterodimer is not the only photoreceptor involved in photocarotenogenesis in Fusarium, since the mutants deprived of either WcoA or Wc1 are still able to accumulate carotenoids in response to light. Recently, it was shown that photocarotenogenesis in F. fujikuroi is carried out in two stages, a fast one mediated by WcoA and a slower one in which participates another flavin photoreceptor, the DASH cryptochrome CryD (Castrillo and Avalos 2015). The significant carotenoid accumulation in the WcoA mutants under constant illumination may be attributed to CryD, whose photoactivity has been experimentally demonstrated (Castrillo et al. 2015). Interestingly, the mutation of either wcoA or cryD also affects the production of other secondary metabolites (Estrada and Avalos 2008; Castrillo et al. 2013), including in the dark in the case of wcoA, indicating other functions for these photoreceptors in addition to the control of carotenogenesis.

The photoinduction of carotenogenesis in Fusarium reminds that of N. crassa, but the maximal levels are reached in this case approximately after 1 h of light, as observed in F. fujikuroi (Prado et al. 2004; Prado-Cabrero et al. 2007), F. oxysporum (Rodríguez-Ortiz et al. 2012), and F. verticillioides (Ádám et al. 2011). In the same way that ylo-1 of N. crassa, carD is not affected by light (Díaz-Sánchez et al. 2011b), as seems also essentially unaffected the GGP synthase gene ggs-1 (Mende et al. 1997), in this case in contrast to its N. crassa ortholog al-3 (Baima et al. 1991). Photoinduction of carotenogenesis is also affected by other regulatory mechanisms, as indicates the lower photoresponse exhibited by the mutants of the adenylate cyclase gene acyA in F. fujikuroi (García-Martínez et al. 2012) or the mating-type MAT1-2-1 gene in F. verticillioides (Ádám et al. 2011).

The decrease in the mRNA levels for the structural genes that follows the photoinduction peak even if illumination persists, a process known as photoadaptation, has been investigated in detail in N. crassa, where it is achieved though the deactivation of the WC complex by hyperphosphorylation (He and Liu 2005). In this process it plays a central role a small flavin photoreceptor, known as VIVID or VVD (Shrode et al. 2001; Schwerdtfeger and Linden 2003). In the vvd mutants, the decrease in the mRNA levels of the al genes after their rapid light induction is less pronounced, leading to a higher carotenoid accumulation (Youssar et al. 2005; Navarro-Sampedro et al. 2008). Other proteins presumably collaborate with VVD in photoadaptation, but although new mutants affected in this process have been identified (Navarro-Sampedro et al. 2008), the molecular functions of the affected genes remain to be elucidated.

In addition to light, other regulatory factors control neurosporaxanthin biosynthesis in N. crassa and Fusarium. In N. crassa the conidia formed by aerial mycelia are orange because of the accumulation of carotenoids, and the synthesis is coupled to the process of conidiation in a light–independent manner. Thus, when conidiation is induced in the wild type in the dark by transferring submerged-grown mycelia to air, there is an increase in al-1 and al-2 transcript levels about 16 h after their transfer, coinciding with the formation of the conidia (Li and Schmidhauser 1995). This induction is also observed under continuous illumination, when al-1 and al-2 transcript levels are attenuated by photoadaptation, and it is absent in regulatory mutants defective in conidiation, such as fluffy or fluffyoid. However, the wc mutants produce pigmented conidia irrespective of illumination. The separation of the stimulatory effects of light by the WC complex and of conidiation by FLUFFY or FLUFFYOID is clearly shown when the mycelium is simultaneously exposed to light and air. In this case there are two different inductions, a rapid and transient one produced by light and a late one after 16 h, which coincides with the formation of the major constrictions during the conidiation process. In the case of gene al-3, the separate regulations by light and conidiation are also proved by the synthesis of specific transcripts for each inducing condition from different transcription start sites: with a size of 1.6 kb in the first case that requires a functional WC complex and 2.2 kb in the second, dependent of a functional FLUFFY protein (Arpaia et al. 1995). The complexity of the regulation of this gene is also shown by the finding of more than one translation start site (Vittorioso et al. 1994).

Different observations indicate that the cAMP signaling pathway, involved in other biological processes, participates in the regulation of neurosporaxanthin biosynthesis in N. crassa and F. fujikuroi. During the lag phase preceding carotenoid photoinduction, there is a transient increase of the cAMP levels in N. crassa, and the cr-1 (crisp-1) mutants, defective in adenylyl-cyclase activity, contain more carotenoids in the dark (Kritsky et al. 1982). A similar phenotype is exhibited by the mutants of the adenylate cyclase gene acyA of F. fujikuroi in the dark (García-Martínez et al. 2012). Moreover, carotenoid photoinduction in N. crassa is reduced by exogenous cAMP addition (Harding 1973), and artificial upregulation of the gene gna-1, coding for a Gα component of a heterotrimeric G complex, exhibited developmental alterations and contained less carotenoids than the control strain (Yang and Borkovich 1999). As a later observation, the lack of histidine kinase DCC-1 increases conidia formation and carotenoid production in N. crassa (Barba-Ostria et al. 2011), but this effect is reversed by cAMP addition. Taken together, these observations indicate an inverse correlation between cAMP levels and carotenoid production in these fungi.

Neurosporaxanthin biosynthesis in N. crassa and F. fujikuroi is stimulated by nitrogen starvation. In N. crassa, the mRNA levels for the genes al-1 and al-2 in the dark are higher when the wild type is grown in a medium with 2 mM NH4Cl (nitrogen limitation) than with 50 mM NH4Cl (nitrogen excess) (Sokolovsky et al. 1992). The same pattern was found in wc mutants, indicating that the effect of nitrogen is independent of the regulation by light. Likewise, nitrogen starvation stimulates carotenoid biosynthesis in F. fujikuroi in the dark, as showed experiments with either immobilized mycelia (Garbayo et al. 2003) or shake cultures (Rodríguez-Ortiz et al. 2009). In both cases, the transfer from high nitrogen to low nitrogen conditions resulted in an increased carotenoid production, corroborated in the case of the shake cultures with an increase in carRA and carB mRNA levels. As in N. crassa, the induction by nitrogen starvation in F. fujikuroi is independent of the induction by light, as shown by their additive effects when both stimulating conditions are simultaneously present (Rodríguez-Ortiz et al. 2009).

Neurosporaxanthin-overproducing mutants, called carS, have been described in F. fujikuroi (Avalos and Cerdá-Olmedo 1987; Rodríguez-Ortiz et al. 2013) and F. oxysporum (Rodríguez-Ortiz et al. 2012). Their higher carotenoid content is consistent with the finding of an increased carotenogenic activity of their cell extracts in vitro (Avalos et al. 1988) and higher mRNA levels for the structural genes carRA, carB (Prado et al. 2004; Thewes et al. 2005), carT (Prado-Cabrero et al. 2007), and carD (Díaz-Sánchez et al. 2011b). Except for the case of carD, the mRNA levels for these genes in the dark are as high as those in the wild type after 1 h illumination; however, the carS mutants are still able to respond to light either in F. fujikuroi (Prado-Cabrero et al. 2007) or in F. oxysporum (Rodríguez-Ortiz et al. 2012). The gene carS encodes a protein with RF and LON domains, with clear similarities with the domain features of CrgA of M. circinelloides despite a considerable sequence divergence (Rodríguez-Ortiz et al. 2012, 2013), and the carS mutant phenotype coincides with that of the crgA mutants. The carS mutation not only affects the synthesis of carotenoids but also the production of other secondary metabolites, such as gibberellins or bikaverin (Candau et al. 1991a; Rodríguez-Ortiz et al. 2009). The mechanism of action of the CarS protein in Fusarium is currently object of detailed investigation.

In contrast to Fusarium, no carS-like mutants are known in N. crassa. Two mutants of this fungus, ccb-1 and ccb-2, have an increased carotenoid content in the dark (Linden et al. 1997), but their carotenoid levels are quite low compared to those of the carS strains of Fusarium (Avalos and Cerdá-Olmedo 1987; Rodríguez-Ortiz et al. 2012). The ccb-1 and cbb-2 mutants are also affected in hyphal morphology, more severely in the case of ccb-1, which is the only one exhibiting an increase in al-1 and al-2 mRNA levels. Other carotenoid-overproducing mutants described in this fungus, as vvd (Shrode et al. 2001) and ovc (Harding et al. 1984), exhibit this phenotype only under light. In the latter case, the ovc phenotype is accompanied by osmotic sensitivity (Youssar et al. 2005), and it is due to a large deletion covering 21 genes (Youssar and Avalos 2007), none of them with a predictable relation with the regulation of carotenogenesis. Carotenoid production may be also augmented in N. crassa by genetic engineering, as reveals the visible increase resulting from the expression of the catalytic domain of 3-hydroxy-3-methylglutaryl coenzyme A reductase from S. cerevisiae under control of a strong promoter (Wang and Keasling 2002).

8.3 Carotenoid Biosynthesis in Yeasts

The color of many pigmented yeasts, frequently in the phylum of basidiomycetes, is due to the production of carotenoids. β-carotene biosynthesis has been described in some species, as Rhodosporidium sp. (de Miguel et al. 1997), Sporidiobolus pararoseus (Han et al. 2012), Ustilago maydis (Estrada et al. 2010), and other Ustilago species (Will et al. 1984, 1985). However, yeasts usually produce xanthophylls. To this group belong canthaxanthin, found in Cantharellus cinnabarinus (Haxo 1950), plectaniaxanthin in Dioszegia sp. (Madhour et al. 2005), and phillipsiaxanthin, a lycopene derivative with hydroxyl and keto groups, in the ascomycete Phillipsia carminea (Arpin and Liaaen-Jensen 1967). The structures of these xanthophylls are depicted in Fig. 8.7. However, the most investigated cases correspond to the productions of astaxanthin by Xanthophyllomyces dendrorhous, and torularhodin by Rhodotorula sp., summarized in the next sections. The biotechnological production of carotenoids by yeast has been object of a recent review (Mata-Gómez et al. 2014).

Fig. 8.7
figure 7

Chemical structures of some yeast carotenoids

8.3.1 Astaxanthin Biosynthesis in Xanthophyllomyces

Astaxanthin is a red xanthophyll of economic importance because of its benefiting effects on human health (Ambati et al. 2014). The high antioxidant activity and health-promoting properties of this xanthophyll have led to an increasing use in the nutraceutical market, where it is sold as an encapsulated product. Astaxanthin is also widely used as animal feed additive, especially as a source of pigmentation for some fishes and crustaceans in aquaculture industries, as those of trouts, salmons, or shrimps (Higuera-Ciapara et al. 2006). Astaxanthin is produced by different microorganisms, including algae, bacteria, and fungi (Alcaino et al. 2014), and it is industrially obtained from the alga Haemematococcus pluvialis (Ambati et al. 2014) and the yeast X. dendrorhous (Rodríguez-Sáiz et al. 2010), also known as Phaffia rhodozyma (the anamorph form) (Johnson 2003).

Astaxanthin is synthesized from β-carotene through the introduction of keto and hydroxyl groups in the β-rings of the molecule (Fig. 8.8). Therefore, the steps up to β-carotene coincide with those described in former sections for β-carotene-producing fungi (Fig. 8.2). All the genes specifically involved in astaxanthin biosynthesis are known in X. dendrorhous, as well as some involved in early steps of the terpenoid pathway, as the one coding for the enzyme that isomerizes IPP to DMAPP (Kajiwara et al. 1997), the FPP synthase gene fpp (Alcaíno et al. 2014), and the GGPP synthase gene crtE (Niklitschek et al. 2008). The phytoene synthase/lycopene cyclase is encoded by the gene crtYB, the first one in fungi where the double function of the enzyme was discovered (Verdoes et al. 1999a). Between the synthase and cyclase reactions carried out by CrtYB, the four desaturations are achieved by the desaturase encoded by the gene crtI (Verdoes et al. 1999b). Albino mutants accumulating phytoene or no carotenes, affected in crtI and presumably in crtYB, are readily obtained by mutagenesis (Girard et al. 1994), but no mutants affected in the cyclase activity have been described. However, such activity is blocked in the presence of nicotine, resulting in the accumulation of lycopene (Ducrey Sanpietro and Kula 1998). On the other hand, diphenylamine blocks the desaturase activity and, at a lower concentration, the activity of CrtS-CrtR on β-carotene, described below.

Fig. 8.8
figure 8

Astaxanthin biosynthesis in X. dendrorhous. In the reactions achieved by CrtS, the presumed participation of CrtR is omitted. Reactions and chemical structures up to lycopene and 3-4-didehydrolycopene are as shown in Figs. 8.2 and 8.5. Enzymes responsible for each reaction are indicated close to the chemical modification produced in each enzymatic step (gray arrowhead). The uncycled end of torulene and derivatives is represented in a folded form, as drawn in original works (Visser et al. 2003; Martín et al. 2008). HDCO 3-hydroxy-3′4′-didehydro-β,ψ-carotene-4-one, DCD 3,3′-dihydroxy-β,ψ-carotene-4,4′dione

The conversion of β-carotene to astaxanthin might result in a diversity of intermediates depending in the order of introduction of the keto and hydroxyl groups. However, early chemical analysis of the carotenoids in this yeast revealed the presence of significant amounts of echinenone, 3-hydroxy-echinenone and phoenicoxanthin (Andrewes et al. 1976), suggesting the order of reactions displayed in Fig. 8.8. The 3-hydroxylations and 4-ketolations are achieved by a single enzyme of the cytochrome P450 monooxygenase family. The responsible gene was simultaneously cloned by two different research groups, that named it crtS (Álvarez et al. 2006) and asy (Ojima et al. 2006), referred to crtS hereafter. The achievement of both types of oxidations by a single enzyme differs from what was found in astaxanthin-producing algae and bacteria, where independent ketolases and hydroxylases are involved in the conversion of β-carotene to astaxanthin (Fraser et al. 1997). However, the expression of crtS in a heterologous system, as M. circinelloides, results only in the production of hydroxylated derivatives, indicating the participation in X. dendrorhous of at least an additional protein (Martín et al. 2008). This turned out to be a cytochrome P450 reductase, encoded by the gene crtR, that provides the electrons for the oxygenation reactions achieved by CrtS, as indicates the lack of astaxanthin in crtR deletion mutants (Alcaíno et al. 2008). When CrtS is expressed in a metabolically engineered S. cerevisiae strain (see next section), it is only active when co-expressed with crtR; despite this yeast having its own cytochrome P450 reductase (Ukibe et al. 2009). This indicates a specific interaction between CrtS and CrtR for proper activity, a conclusion that has been supported by protein modeling and molecular dynamics simulations (Alcaíno et al. 2012).

The chemical analyses of the carotenoids accumulated in X. dendrorhous identified a novel xanthophyll, 3-hydroxy-3′4′-didehydro-β,ψ-carotene-4-one (HDCO, Fig. 8.8), that implies the occurrence of a side biosynthetic branch of monocycled carotenoids in this yeast. More detailed chemical analyses found also torulene and postulated the origin of HDCO through an additional desaturation on γ-carotene and the introduction of keto and hydroxyl groups at the only β-ionone ring of the molecule (An et al. 1999). Thus, as already described for N. crassa and Fusarium sp., this side pathway in X. dendrorhous shows the capacity of the CrtI desaturase to carry out a fifth desaturation that impedes the introduction of the second cyclization in the molecule. Moreover, the detection of minor amounts of the HDCO derivative 3,3′-dihydroxy-β,ψ-carotene-4,4′dione (DCD) indicates that the CrtS-CrtR enzymatic machinery is also able to recognize as a substrate the uncycled end of HDCO (Fig. 8.8).

8.3.2 Biotechnological Astaxanthin Production

The yeast X. dendrorhous (P. rhodozyma) is the only fungus used for industrial astaxanthin production. The yields may vary considerably depending on the strain and the growth conditions (Schmidt et al. 2011), and the biotechnological use of this yeast implies the improvement of the production through the isolation of astaxanthin-overproducing strains and the development of more efficient culture conditions (see, e.g., the study described by Meyer et al. 1993). Despite this fungus is diploid (Kucsera et al. 1998), deep-pigmented mutants producing more carotenoids are easily detected after adequate mutagenesis treatments (Visser et al. 2003), allowing increases in astaxanthin content ranging from 2–3-fold (Fang and Cheng 1992; Stachowiak 2013; Barbachano-Torres et al. 2014) to 10–15-fold (Miao et al. 2010). In addition, different enrichment methods have been successfully used to obtain mutants with higher astaxanthin production, as those based on antymicin resistance (An et al. 1989), β-ionone inhibition (Lewis et al. 1990), flow cytometry sorting (An et al. 1991; Brehm-Stecher and Johnson 2012), photosensitization (An 1997), DPA resistance (Chumpolkulwong et al. 1997), or low-dose gamma irradiation (Sun et al. 2004). Incubation under continuous illumination has no relevant effect on the accumulation of carotenoids in this yeast (Johnson and Lewis 1979), or it even has a detrimental effect (An and Johnson 1990), but it results in increase of the carotenoid content in antimycin-treated cells (An and Johnson 1990), in some astaxanthin-overproducing mutants (Meyer and Du Preez 1994), and in an industrial strain (de la Fuente et al. 2010).

X. dendrorhous is amenable to genetic transformation (Martínez et al. 1998), allowing to obtain strains with enhanced astaxanthin production through engineered metabolic alterations (Visser et al. 2003). In this way, different yield improvements were obtained through the increased expression of early genes of the terpenoid pathway (Hara et al. 2014), the GGPP synthase crtE gene (Breitenbach et al. 2011), the bifunctional crtYB gene (Ledetzky et al. 2014), and the crtS gene (Chi et al. 2015) or through the deletion of genes related to ergosterol biosynthesis (Loto et al. 2012; Yamamoto et al. 2016), thus releasing the repression of HMG-CoA reductase. A similar stimulation was obtained by the targeted deletion of mig1, encoding the orthologous global regulator of catabolite repression, indicating that glucose availability plays a regulatory role in carotenoid biosynthesis in X. dendrorhous (Alcaíno et al. 2016). Actually, glucose exerts a negative effect on astaxanthin biosynthesis (Yamane et al. 1997). Recently, the introduction of several vectors to an overproducing mutant to obtain the combined increased expression of crtYB and crtS (Gassel et al. 2013), and later with the addition of a truncated version of the HMG-CoA reductase gene hmg1 from S. cerevisiae and the endogenous crtE gene (Gassel et al. 2014), allowed to reach carotenoid levels comparable to the extremely high production obtained with the alga H. pluvialis. However, the modified expression of carotenogenic genes may eventually result in undesired effects for astaxanthin production. Thus, overexpression of crtYB or crtI may lead to increased proportions of β-carotene and echinone or monocyclic carotenoids, respectively (Verdoes et al. 2003).

As described for B. trispora, oxidative stress affects the synthesis of astaxanthin in X. dendrorhous. A positive correlation has been found between oxygen levels in the medium and astaxanthin biosynthesis (Yamane et al. 1997). Photochemical generation of singlet oxygen with rose bengal or alpha-terthienyl results in a higher astaxanthin accumulation, but H2O2 or the peroxyl radical-generating agent t-butylhydroperoxide decreases the content of astaxanthin, although it is compensated by a higher β-carotene accumulation (Schroeder and Johnson 1995a). Other authors, however, found the opposite effect for H2O2, its addition increasing the astaxanthin proportion at expenses of that of β-carotene (Liu and Wu 2006b). Moreover, the presence of H2O2 under continuous culture renewal conditions that allow higher astaxanthin yields compared to standard batch cultures leads to a further improvement in the production (Liu and Wu 2007). The increased astaxanthin production in the presence of oxidative agents must be related with the antioxidant properties of this xanthophyll, as indicates the positive correlation between carotenoid content and survival of the yeast in media supplemented with H2O2 or under chemical conditions generating superoxide (Schroeder and Johnson 1993) or singlet oxygen (Schroeder and Johnson 1995b). The proteomes in cells at different culture stages are consistent with a relation of astaxanthin production with the defense against reactive oxygen species generated during metabolism (Martinez-Moya et al. 2011).

The composition of the medium exerts a strong influence on astaxanthin production in X. dendrorhous cultures (Johnson and Lewis 1979). The use of a non-fermentable carbon source, succinate, instead of the fermentable glucose results in a duplication of the carotenoid level (Wozniak et al. 2011), which has been associated to an increased acetyl-CoA availability and cellular respiration rate and presumably also to enhanced oxidative stress (Martinez-Moya et al. 2015). A significant increase in total carotenoid content is also obtained with low nitrogen or phosphate concentrations and with citrate addition, but changes in astaxanthin are not so marked (Flores-Cotera et al. 2001). An increase in carotenoid content is also produced in different genetic backgrounds by addition of ethanol, possibly also related with enhanced oxidative stress, but the proportion of astaxanthin was not determined (Gu et al. 1997). In this case, increased HMG-CoA reductase was detected in the presence of ethanol, which may be the result of a regulatory effect. In support to this observation, addition of mevalonic acid, the product of this enzyme, enhances the accumulation of astaxanthin (Calo et al. 1995). In other cases, the chemical basis of the inducing agents remains to be identified. The presence of a fungal contaminant, identified as Epicoccum nigrum, in Petri dishes with X. dendrorhous colonies, or the incubation of the yeast in media with filtered extracts from E. nigrum cultures, results in increased carotenoid production, suggesting that the contaminant fungus produces a stimulatory metabolite (Echavarri-Erasun and Johnson 2004). Similarly, unidentified fungal elicitors from diverse fungi (Wang et al. 2006) or certain plant extracts (Kim et al. 2007) have inducing effects.

Considering the variations in astaxanthin content depending on the culture conditions and media composition, many efforts have been addressed to improve the culture parameters for optimal production (see, e.g., Ramírez et al. 2001; Liu et al. 2008). The changes in basic fermentation parameters, as pH, temperature, inoculum, and carbon or nitrogen concentrations, usually done on batch cultures because of experimental amenability, may lead to tenfold variations in carotene production (Ramírez et al. 2001). As already mentioned for β-carotene production with B. trispora, the use of cheap substrates contributes to increase the profitability of astaxanthin production as an industrial activity, and many different substrates proved useful for industrial growth of X. dendrorhous. The alternative nutrient sources successfully used include corn wet-milling coproducts (Hayman et al. 1995), peat hydrolysates (Acheampong and Martin 1995; Vázquez and Martin 1998), sugarcane juice (Fontana et al. 1996), wood hemicellulosic hydrolysates (Parajo et al. 1998), beet blackstrap molasses (An et al. 2001), raw coconut milk (Domínguez-Bocanegra and Torres-Muñoz 2004), mustard wastes (Tinoi et al. 2006), pineapple juice (Jirasripongpun et al. 2008), sugarcane bagasse and barley straw hydroxylates (Montanti et al. 2011), mussel-processing wastewater (Amado and Vázquez 2015), and others (Frengova and Beshkova 2009). Although behavior of this yeast may be similar in aerated fermenters or in batch cultures (Acheampong and Martin 1995), the growth conditions in industrial fermenters require to optimize other parameters. A representative example is found in the search for the optimal pH and dilution rate combination using a response surface methodology (Vázquez and Martin 1998).

Contrary to the many efforts dedicated to develop the industrial astaxanthin production by X. dendrorhous, the regulation of the expression of the carotenogenic genes has received little attention. The amplification of the cDNA products for the genes crtI and crtYB revealed the occurrence of alternative splicing events, leading to the formation of predictably nonfunctional proteins (Lodato et al. 2003). Interestingly, the proportion of correctly spliced mRNAs decreased with aging for the gene crtI, suggesting a possible regulatory mechanism of CrtI levels mediated through transcript maturation. Some expression studies have focused the attention on the relation between enhanced carotenoid biosynthetic activity and mRNA levels for the structural genes. In a comparison between a wild type and two carotenoid-overproducing mutants, no correlation was found in RT-PCR analyses between transcript levels for the genes idi, crtE, crtYB, and crtI and carotenoid production (Lodato et al. 2004). A later detailed study of the kinetics of growth, carotenoid content, and mRNA levels for the same genes, in this case extended to crtS, revealed only minor differences between the wild type and one mutant (Lodato et al. 2007). In both strains, the transcript levels for the genes idi and crtE were similar at either the exponential or the stationary phase, while those for crtYB, crtI, and crtS reached maximal levels at the end of the exponential phase and decreased afterward. On the other hand, RT-qPCR analyses of the expression of the whole set of genes involved in astaxanthin biosynthesis found increased mRNA levels for crtE and crtR in an overproducing mutant, but against expected, the pattern was the contrary for crtYB and crtS in advanced culture stages (Castelblanco-Matiz et al. 2015). Even so, in this strain crtYB and crtS contained mutations in their coding sequences, which might result in a higher enzymatic activity. However, other overproducing mutants may have a different molecular explanation, as indicates the much higher levels for genes crtE, crtYB (called in this case pbs), crtI, and crtS (called here ast) in a particular mutant strain compared to the wild type, while genes for earlier steps in the terpenoid pathway were basically unaffected (Miao et al. 2011). The overproducing phenotype in this case was also related with a decrease in ergosterol and fatty acids production.

Other reports have described the effects on gene expression of culture conditions leading to a higher carotenoid content. The inducing effect of succinate on astaxanthin biosynthesis is accompanied by some changes in the expression patterns of the carotenogenic genes, with a delay in the maximal mRNA levels in succinate for crtS and for the alternative spliced versions of crtYB and crtI, although such changes seem insufficient to explain the differences in carotenoid production (Wozniak et al. 2011). However, the inducing effect of ethanol may be explained by a transient but significant increase in crtYB, crtI, and crtS transcript levels, while the opposite pattern was found upon addition of glucose (Marcoleta et al. 2011). The stimulating effect of oxygen on astaxanthin biosynthesis may be at least partially explained by an effect on gene expression, as indicates the positive correlation between dissolved oxygen and mRNA levels for crtE, crtYB, crtI, and crtS (Wu et al. 2011).

8.3.3 Biosynthesis of Torularhodin and Other Xanthophylls

Torularhodin is well known because of its production by the yeasts of the genus Rhodotorula (Moliné et al. 2012), but it has been also found in other yeasts, as those of the genera Peniophora, Cystofilobasidium, Rhodosporidium, Sporobolomyces, and Sporidiobolus (Buzzini et al. 2007; Iurkov et al. 2008). Torularhodin has a chemical structure similar to that of neurosporaxanthin, but with 40 carbon atoms instead of 35 (Sperstad et al. 2006). A first biosynthetic pathway for this xanthophyll was early proposed for Rhodotorula glutinis, based on the intermediates accumulated by their cultures under unfavorable biosynthetic conditions, as low temperature or the presence of methylheptenone and β-ionone vapors, and from former chemical studies (Simpson et al. 1964). The sequence of reactions to produce torularhodin (Fig. 8.9) coincides with that of neurosporaxanthin biosynthesis up to torulene (see Fig. 8.5). However, instead of the cleaving reaction characteristic of this pathway, in R. glutinis, torulene is object of an oxidation reaction to generate an aldehyde group, which is oxidized further to generate the final torularhodin carboxylic group. Chemical studies in Cystofilobasidium demonstrated the occurrence not only of the aldehyde intermediate, torularhodinaldehyde, but also of 16′-hydroxytorulene, indicating three successive oxidation steps from torulene to the final product (Herz et al. 2007), as depicted in Fig. 8.9. The chemical analysis of the carotenoids produced by 13 different Rhodotorula species showed considerable differences in their concentrations and the presence of variable proportions of β-carotene in addition to torularhodin and its precursors γ-carotene and torulene (Buzzini et al. 2007). Therefore, as found in the neurosporaxanthin pathway, the γ-carotene intermediate in torularhodin biosynthesis may be subject of a second cyclization instead of the fifth desaturation (Fig. 8.9).

Fig. 8.9
figure 9

Proposed torularhodin biosynthetic pathway. Reactions and chemical structures up to β-zeacarotene are as shown in Figs. 8.2 and 8.5. Reactions between torulene and torularhodin are as proposed by Herz et al. (2007)

In contrast to the synthesis of astaxanthin or other fungal carotenoids described in former sections, no reports are available on the identification or characterization of the genes involved in torularhodin biosynthesis in the producing yeasts. However, many reports have been dedicated to the improvement of the production of this xanthophyll in Rhodotorula species. Mutagenesis methods, based in the exposure to either chemical mutagens (Frengova et al. 2004b) or UV-B radiation (Moliné et al. 2012), have been successfully employed to obtain torularhodin-overproducing mutants. Illumination has a retarding effect on growth of R. glutinis, but blue light stimulates significantly the accumulation of torularhodin biosynthesis (Sakaki et al. 2000, 2001), an effect that is more pronounced in an overproducing mutant (Sakaki et al. 2000, 2001). Actually, torularhodin exerts a photoprotective role, as indicates the positive correlation between the presence of this xanthophyll and the resistance to UV-B in strains of Rhodotorula mucilaginosa with different carotenoid content (Moliné et al. 2010). In another report, illumination of the yeast in a fermentor resulted in a duplication of the carotenoid content, but in this case β-carotene was predominant (68%) while torularhodin content only reached 21.5% (Zhang et al. 2014). In contrast, oxidative stress artificially induced by addition of chemicals generating reactive oxygen species, as methyl viologen or methylene blue, increased both the carotenoid content and the torularhodin proportion in R. glutinis (Sakaki et al. 2002b). The protective role of torularhodin against oxidative stress was formerly observed in R. mucilaginosa, in which the absence of colored carotenoids produced by addition of diphenylamine resulted in cells more sensitive to duroquinone or to hyperoxia than those with a normal carotenoid content (Moore et al. 1989).

Because of its characteristics, torularhodin is a xanthophyll with potential applications in food and cosmetic industries (Hernández-Almanza et al. 2014a; Zoz et al. 2015). Recent reports describe different efforts to develop industrial culture conditions allowing improved carotenoid production with Rhodotorula species or with other yeasts, usually containing torularhodin mixed with other carotenoids, as torulene and β-carotene. In a classical optimization work, the effect of initial pH, temperature, aeration rate, carbon (glucose, molasses, sucrose, and whey lactose), nitrogen concentration, as well as cotton seed oil and Tween 80 as potentially activating agents, was investigated in batch cultures of strains of R. mucilaginosa (Aksu and Eren 2005) and R. glutinis (Aksu and Eren 2007). As already described for β-carotene and astaxanthin productions, research efforts with Rhodotorula species have been frequently addressed to check the potential use of low-cost substrates for carotenoid production, although the torularhodin proportion in the carotenoid mixtures was frequently disregarded. The cheap substrates tested include grape must, glucose syrup, beet molasses, soybean or maize flour extracts (Buzzini and Martini 1999), hydrolyzed mung bean waste flour (Tinoi et al. 2005), crude glycerol from biodiesel plants (Saenge et al. 2011), and brewery effluents (Schneider et al. 2013). In a different approach, some Rhodotorula species were cocultivated with yogurt starter bacteria in media with whey ultrafiltrates (Frengova et al. 1994, 2004a). In some cases β-carotene is particularly abundant, leading to different reports on the potential use of Rhodotorula strains for its biotechnological production, e.g., in stirred or airlift tanks (Yen and Chang 2015), in solid-state fermentation (Hernández-Almanza et al. 2014b), or to obtain β-carotene-enriched biomass with different processed waste substrates (Marova et al. 2012).

In addition to Rhodotorula, other torularhodin-producing yeasts have been considered as alternative biotechnological carotenoid sources. Glycerol is particularly efficient for torularhodin biosynthesis by Sporobolomyces ruberrimus (Razani et al. 2007), even using raw glycerol from biodiesel production (Cardoso et al. 2016). Basic fermentation parameters and their effects on carotenoid production were also investigated in Sporidiobolus pararoseus (Valduga et al. 2009), which was more recently optimized in a cheap glycerol-based medium (Valduga et al. 2014), but the proportion of torularhodin in the total carotenoid mixtures was not determined in these studies. Carob pulp syrup and sugarcane molasses proved efficient for the production of carotenoids by Rhodosporidium toruloides (Freitas et al. 2014). A particularly innovative strategy to enhance the carotenoid production by this species was the expression of the Pdr10 multidrug transporter from S. cerevisiae, adapted for the codon usage of R. toruloides and the culture of this yeast in two-phase media containing oil (Lee et al. 2016). Unexpectedly, a higher proportion of torulene was exported when pdr10 was expressed, indicating that torularhodin export was less efficient. Torularhodin may be also obtained from Sporobolomyces salmonicolor biomass, as a potential side product in the use of this yeast for exopolysaccharides production (Dimitrova et al. 2013).

8.3.4 Use of Yeasts for Heterologous Carotenoid Production

Heterologous expression in non-carotenogenic microorganisms, as E. coli and some yeasts, has been a powerful tool in functional studies of carotenoid biosynthetic genes and in the development of new carotenoid producing systems (Schmidt-Dannert 2000; Sandmann 2002). In the case of yeasts, the experimental amenability derived from its extensive use in basic research and in brewing and fermentative industries has made S. cerevisiae a favorite system for heterologous expression. The first report on the use of S. cerevisiae for heterologous carotenoid production consisted in the introduction of a plasmid with the Erwinia uredovora genes crtE, crtB, and crtI (coding for GGPP synthase, phytoene synthase, and phytoene desaturase, respectively) under control of yeast promoters. The expression of these foreign genes led to lycopene production, which was mostly replaced by β-carotene if the lycopene cyclase crtY was also included (Yamano et al. 1994). Later, the overexpression of crtI and crtYB from X. dendrorhous in S. cerevisiae allowed the accumulation of variable amounts of β-carotene and their precursors. The production was enhanced by increasing substrate supply through the expression of either the homologous GGPP synthase gene or its ortholog from X. dendrorhous crtE and a truncated version of the S. cerevisiae HMG-CoA reductase gene hmg1 (Verwaal et al. 2007), mentioned in a former section. Improved production was also attained by expressing in parallel to ctrI and crtYB from X. dendrorhous the mevalonate kinase gene mvaK1 from Staphylococcus aureus (Lange and Steinbüchel 2011), by codon optimization for S. cerevisiae of the crtI and crtYB genes (Li et al. 2013) or simply incubating at 20 °C instead of its usual growth temperature of 30 °C (Shi et al. 2014).

The biosynthesis of β-carotene in S. cerevisiae has been subsequently altered in different ways. Production of lycopene was also attained with X. dendrorhous genes through the elimination of the cyclase activity of the crtYB gene of X. dendrorhous, co-expressed with crtE and crtI (Xie et al. 2015). In this case, the carotene levels were raised further improving the catalytic activity of CrtE by directed evolution and varying the copy number of the crt genes. In other study, lycopene production was particularly efficient by combining the crtE (GGPP synthase) and crtB (phytoene synthase) genes from the bacterium Pantoea agglomerans with the carB gene from B. trispora (called BtCrtI in this study), accompanied by other host alterations (Xie et al. 2015). As an innovative approach, the crtI, crtE, and crtYB coding sequences were also expressed in S. cerevisiae as a single polycistronic transcript separated by the T2A sequence of the Thosea asigna virus, resulting in the subsequent separation in independent polypeptides during mRNA translation (Beekwilder et al. 2014); the additional expression of the β-carotene cleavage dioxygenase RiCCD1 from raspberry in these S. cerevisiae cells allowed the production of β-ionone. This apocarotenoid was also produced by separate expression in S. cerevisiae of the crtI, crtE, and crtYB genes in parallel with the Petunia hybrida PhCCD1 gene (López et al. 2015), while the expression of algal β-carotene ketolase genes in similar S. cerevisiae β-carotene producing strains allowed the accumulation of significant amounts of canthaxanthin (Chang et al. 2015).

The production of carotenoids in S. cerevisiae, presumably at expenses of the deviation of substrates normally used for sterols biosynthesis, results in stressed cells. This conclusion is supported by the specific induction in the carotenoid-producing cells of genes of the pleiotropic drug resistance response (PDR), involved in secretion of toxic compounds (Verwaal et al. 2010). Actually, the addition of an appropriate solvent leads to secretion of carotenoids, indicating that the cells are unable to secrete it in the usual hydrophilic environment of the laboratory cultures. Accordingly, the synthesis of different xanthophyll mixtures, obtained through the expression of crtI and crtYB in combination with crtS and crtR from X. dendrorhous or with the bacterial ketolase and hydroxylase genes crtW and crtZ, results in a reduced growth compared to the wild type, although these xanthophyll-producing strains exhibit a better growth capacity in the presence of 1.7 mM H2O2 (Ukibe et al. 2009). In contrast, the canthaxanthin-producing strains obtained by expression of algal β-carotene ketolase genes exhibit a slightly faster growth than the control strain in the absence of artificially induced oxidative stress (Chang et al. 2015).

Besides S. cerevisiae, other biotechnological yeasts lacking carotenoid biosynthesis have been used for heterologous carotenoid production. The sequences of the crtE, crtB, and crtI genes from E. uredovora were modified for optimal codon usage and expressed in Candida utilis under control of constitutive promoters from this yeast, leading to lycopene production (Miura et al. 1998), accumulated in higher levels if the cells were engineered for increased HMGCoA reductase activity and reduced squalene synthesis (Shimada et al. 1998). Lycopene was efficiently converted to β-carotene in the same yeast by the additional expression of the cyclase gene crtY from E. uredovora and to astaxanthin if the β-carotene ketolase (crtW) and hydroxylase (crtZ) genes from Agrobacterium aurantiacum were also expressed (Miura et al. 1998). Another yeast successfully used for carotenoid synthesis is Pichia pastoris, known as industrial producer of heterologous proteins. Two plasmid combinations, one with the already mentioned crtE, crtB, and crtI genes from E. uredovora and another carrying these genes together with the cyclase crtL gene from the plant Ficus carica, in all cases under control of yeast promoters, led, respectively, to notable lycopene and β-carotene productions (Araya-Garay et al. 2012b), while the supplementary expression of crtW and crtZ genes from A. aurantiacum added astaxanthin to the produced carotenoids (Araya-Garay et al. 2012a).

8.4 Conclusions

Because of their easy growth and manipulation, the fungi have been a major choice by the researchers as a tool to investigate the molecular basis of carotenoid biosynthesis and its regulation in microorganisms and by biotechnologists as a source for carotenoids demanded by the market. Usually, depending on their specific features, different species have been used for one purpose or the other. P. blakesleeanus, and later also M. circinelloides, have been extensively used to learn about the biochemistry and the genetics of β-carotene production, while N. crassa and F. fujikuroi have been the major sources of information on neurosporaxanthin biosynthesis. In the case of β-carotene, most of the efforts on its biotechnological production have been dedicated to B. trispora, which was benefited from the scientific background generated with P. blakesleeanus and M. circinelloides. In contrast, the possible biotechnological interest of neurosporaxanthin has not received attention. Regarding the yeasts, X. dendrorhous (P. rhodozyma) has been the major object of attention for astaxanthin production, both from the scientific and biotechnological points of view, while genetics and biochemistry of the carotenoids pathways for the production of other xanthophylls have received very limited attention. Torularhodin production is a notable exception, but the efforts on this xanthophyll were mostly focused on its biotechnological production. The concentration of the research on very few species has allowed a deep knowledge on the molecular basis of carotenoid biosynthesis of the cases investigated, and the high conservation between the ortholog genes of the carotenoid pathways facilitates the prediction for the capacity to produce carotenoids by organisms for which the genome sequences become available. Nevertheless, the low number of fungi investigated in detail allows anticipating that novel fungal carotenoids or more suitable carotenoid producers could be available in the future with more extensive screenings, widening the biotechnological potential of the fungi.