OUP user menu

Characterization of carotenoid biosynthetic genes in the ascomycete Gibberella zeae

Jian-Ming Jin, Jungkwan Lee, Yin-Won Lee
DOI: http://dx.doi.org/10.1111/j.1574-6968.2009.01854.x 197-202 First published online: 1 January 2010


Carotenoids are a structurally diverse class of terpenoid pigments that are synthesized by many microorganisms and plants. In this study, we identified five putative carotenoid biosynthetic genes from the ascomycete Gibberella zeae (GzCarB, GzCarO, GzCarRA, GzCarT, and GzCarX). HPLC showed that the fungus produces two carotenoids: neurosporaxanthin and torulene. We deleted the five genes individually to determine their functions. GzCarB, GzCarRA, and GzCarT were required for neurosporaxanthin biosynthesis, but the deletion of GzCarX or GzCarOgzcarX or ΔgzcarO) failed to alter the production of neurosporaxanthin or torulene. ΔgzcarRA and ΔgzcarB did not produce neurosporaxanthin or torulene. ΔgzcarB led to the accumulation of phytoene, which is an intermediate in carotenoid biosynthesis, but ΔgzcarRA did not. ΔgzcarT produced torulene but not neurosporaxanthin. Based on these functional studies and similarities to carotenoid biosynthesis genes in other fungi, we deduced the functions of the three genes and propose the carotenoid biosynthetic pathway of G. zeae.

  • carotenoid
  • Gibberella zeae
  • neurosporaxanthin
  • phytoene
  • torulene


Carotenoids are important natural terpenoid pigments produced by many microorganisms and plants, but not animals. Of the >700 natural carotenoids that have been identified, most are C40 terpenoids that vary in the number of conjugated double bonds, end-group structures, and oxygen-containing functional groups (Britton et al., 2004). The interesting properties and human health benefits of carotenoids have received much attention. Carotenoids exhibit significant anticarcinogenic and antioxidant activity, and play an important role in preventing chronic disease (Landrum & Bone, 2001).

Carotenoids are derived from the isoprenoid biosynthetic pathway (Umeno et al., 2005). In fungi, the first isoprene, isopentenyl pyrophosphate, is synthesized from mevalonic acid, with acetyl-CoA as a starting precursor. Short-chain-length prenyltransferase then synthesizes geranyl pyrophosphate, farnesyl pyrophosphate, or geranylgeranyl pyrophosphate (GGPP). GGPP is the immediate precursor of C40-carotenoids. Phytoene synthase catalyzes the condensation of two GGPP molecules into phytoene. Phytoene dehydrogenase catalyzes a desaturation process of four consecutive steps from phytoene to lycopene as a final product. Finally, lycopene is cyclized by lycopene cyclase to produce β-carotene (Sandmann, 2002; Sieiro et al., 2003).

Filamentous ascomycetes, such as Neurospora crassa and Fusarium fujikuroi, produce the carotene-derived pigment neurosporaxanthin, a C35 acidic apo-carotenoid (Avalos & Cerdà-Olmedo, 1987; Schmidhauser et al., 1990). Phytoene is first synthesized by the bifunctional enzyme phytoene synthase/lycopene cyclase Al-2 in N. crassa and by CarRA in F. fujikuroi (Arrach et al., 2002; Linnemannstöns et al., 2002). The phytoene dehydrogenases Al-1 and CarB of N. crassa and F. fujikuroi, respectively, introduce up to five double bonds into phytoene, yielding 3,4-dihydrolycopene as an intermediate step in the formation of torulene (Hausmann & Sandmann, 2000; Linnemannstöns et al., 2002). Lycopene cyclase synthesizes torulene from 3,4-dihydrolycopene. Lycopene cyclase and phytoene synthase activity are present in one fungal protein (Arrach et al., 2001). Torulene is then converted into β-apo-4-carotenal by the torulene-cleaving oxygenase Cao-2 in N. crassa (Saelices et al., 2007) or CarT in F. fujikuroi (Prado-Cabrero et al., 2007a). Finally, β-apo-4′-carotenal is oxidized to neurosporaxanthin by the aldehyde dehydrogenase Ylo-1 in N. crassa (Estrada et al., 2008a, b).

Gibberella zeae (anamorph: Fusarium graminearum) causes head blight of small grains and produces mycotoxins such as zearalenone and trichothecenes (Leslie & Summerell, 2006). The complete genome of G. zeae has been sequenced (http://www.broad.mit.edu/annotation/fungi/fusarium/), enabling functional studies of numerous genes via reverse genetics. From the genome database, we identified five putative genes related to carotenoid biosynthesis and characterized three of them using both targeted gene deletion and chemical analyses.

Materials and methods

Strains and culture conditions

Strain GZ03643, provided by Robert Bowden (USDA-ARS, Manhattan, KS), was used as the wild-type G. zeae strain. A GZ03643-derived PKS12-deleted mutant (Δpks12) (Kim et al., 2005) was used to generate double mutants of PKS12 and carotenoid biosynthesis genes. For DNA isolation, the fungal strains were grown in 50 mL complete medium (CM; Leslie & Summerell, 2006).

Nucleic acid manipulations, PCR primers, and targeted gene deletion

Fungal genomic DNA was extracted as described previously (Leslie & Summerell, 2006). Standard procedures were used for restriction endonuclease digestion, gel blotting, and 32P labeling of probes (Sambrook et al., 2001). DNA fragments used as probes were amplified from GZ03643 genomic DNA using appropriate primer pairs (Supporting Information, Table S1). PCR reactions were performed as described previously (Kim et al., 2005). The candidate carotenoid biosynthetic genes were deleted using the double-joint PCR method (Yu et al., 2004). Fungal transformation was performed as described previously (Kim et al., 2005).

Preparation of carotenoids

For pigment production, fungal strains were grown on CM for 7 days at 25 °C under cool-white fluorescent lights, after which the cultures were harvested, dried in a ventilated hood, ground in a blender, and then extracted with acetone. The acetone extracts were applied to an Al2O3 column (Duksan Pure Chemicals, Ansan, Korea) and eluted with petroleum ether (30–60 °C), chloroform, and chloroform : methanol (3 : 1 v/v). The carotenoids were purified using C18 reserve-phase silica-gel chromatography (Merck, Darmstadt, Germany), with neurosporaxanthin purified from Δpks12 mutant, torulene from the ΔgzcarT/pks12 double mutant, and phytoene from the ΔgzcarB/pks12 double mutant. Retinal was obtained from Sigma-Aldrich (St. Louis, MO).

HPLC analysis

The fungal strains were grown on CM for 4 days at 25 °C under cool-white fluorescent lights. Then, 2 g of each culture was extracted with acetone, applied to a 0.3 g silica gel column (Merck), and eluted with chloroform : methanol (3 : 1 v/v). The elution was dried and dissolved in 5 mL chloroform. The resulting carotenoids were analyzed using an HP 1100 HPLC system (Hewlett Packard, Palo Alto, CA) and Symmetry C18 column (4.6 × 250 mm; Waters, Milford, MA). Absorption was measured at 298 nm for phytoene, 386 nm for retinal, and 462 nm for neurosporaxanthin and torulene. The mobile phase was acetonitrile : methanol : chloroform (47 : 47 : 6 v/v/v) at a flow rate of 1 mL min−1.

Genetic complementation via outcrossing

To test the genetic linkage between GzCarB or GzCarRA and carotenoid production, we fertilized the MAT1-2 deletion strain Δmat1-2 with ΔgzcarB/pks12 or ΔgzcarRA/pks12, as described previously (Lee et al., 2003). The Δmat1-2 strain carries the wild-type alleles GzCARB, GzCARRA, and PKS12. Each outcross was performed in triplicate on separate carrot agar plates, with 20–30 single ascospores randomly isolated from each plate 10 days after sexual induction. The genotype of each progeny was determined using PCR with specific primer pairs: GZCARB-5for/GEN-R and GZCARRA-5for/GEN-R primers were used to amplify the GzCARB and GzCARRA loci, respectively, and the presence of the PKS12 locus was determined using P12-5′f/HygB-r primers designed previously (Kim et al., 2005). Each progeny was grown on CM for 7 days, after which pigmentation was compared with that of its genotype.


Molecular organization of the carotenoid biosynthesis gene cluster

Four genes (FGSG_03064.3–FGSG_03067.3) were located at 9.2 kb of the putative gene cluster on supercontig 2 of the F. graminearum genome (Fig. 1a). The organization of the gene cluster was very similar to that of the cluster containing four genes related to carotenoid biosynthesis in F. fujikuroi (Thewes et al., 2005). The gene cluster included a gene coding for an opsin-like protein (FGSG_03064.3), two genes predicted to code for torulene biosynthesis enzymes (FGSG_03065.3 and FGSG_03066.3), and a gene coding for a putative carotenoid cleaving oxygenase (FGSG_03067.3). FGSG_03064.3 and FGSG_03067.3 proteins showed 73% sequence identity to opsin (CarO) and carotenoid oxygenase (CarX) of F. fujikuroi. FGSG_03065.3 and FGSG_03066.3 proteins exhibited 92% and 81% identity, respectively, to the phytoene dehydrogenase (CarB) and bifunctional enzyme (CarRA) of F. fujikuroi. In addition, the predicted protein FGSG_02625.3 shared 82% identity with torulene oxygenase (CarT) of F. fujikuroi. Based on these similarities, the five G. zeae genes FGSG_03064.3–FGSG_003067.3 and FGSG_02625.3 were designated as GzCarO, GzCarB, GzCarRA, GzCarX, and GzCarT, respectively.


Functional analysis of the carotenoid biosynthetic genes. (a) Organization of the carotenoid biosynthetic genes in Gibberella zeae. Numbers and arrows indicate the ORFs and transcriptional direction, respectively. (b) Deletion strategy. The target gene in the wild-type strain GZ03643 (WT) was replaced with geneticin-resistant gene (gen) through homologous recombinations. (c) Southern hybridization to confirm gene replacement events. Lane 1, wild-type strain GZ03643; lane 2, targeted-gene deletion strain originated from the wild type; lane 3, targeted-gene deletion strain originated from Δpks12; lane 4, strain that carries ectopic transformation event. Sizes of DNA are indicated on the right of the blot.

Targeted deletions of the carotenoid biosynthesis genes in G. zeae

We deleted the five genes individually via targeted mutagenesis (Fig. 1b). Southern blot analysis was performed on genomic DNA from the wild-type strain and genR transformants. Size variations of hybridized bands between the deletion and wild-type strains suggested that each gene has been replaced with the gen cassette (Fig. 1c). All deletion mutants did not show any noticeable phenotype changes on sexual and asexual development, mycelia growth, and zearalenone production.

Carotenoid production

As PKS12 is responsible for the biosynthesis of the pigment aurofusarin, Δpks12 was used to observe the carotenoids. The double-deletion mutants ΔgzcarX/pks12, ΔgzcarO/pks12, and ΔgzcarT/pks12 produced orange pigments, as did Δpks12 single mutants. The color of ΔgzcaRA/pks12 and ΔgzcarB/pks12 was white (Fig. 2). The carotenoid components of the deletion mutants were analyzed using HPLC (Fig. 3). Peaks were identified by comparing both retention times and peak absorption spectra with those of authentic substances. GZ03643 and Δpks12 produced two main carotenoid pigments: neurosporaxanthin and torulene; phytoene and retinal were not detected. The profiles of ΔgzcarX and ΔgzcarO mutants were the same as those of GZ03643 and Δpks12. Neither the ΔgzcarRA nor ΔgzcarB mutant produced neurosporaxanthin or torulene, but phytoene was detected in the ΔgzcarB mutant. ΔgzcarT mutant produced torulene but not neurosporaxanthin (Fig. 3).


Pigmentation of transgenic strains. Photographs were taken 7 days after inoculation on CM.


HPLC analysis of carotenoids produced by the wild-type and transgenic strains. Samples were analyzed at either 298 or 462 nm. Retention times for neurosporaxanthin (1), torulene (2), and phytoene (3) are 7.3, 43.0, and 62.2 min, respectively.

Genetic complementation by outcrossing

We isolated 69 and 64 ascospores from the outcrosses between Δmat1-2 and ΔgzcarB/pks12 and between Δmat1-2 and ΔgzcarRA/pks12, respectively. Segregations between PKS12 and GzCARB or GzCARRA loci fit a 1 : 1 : 1 : 1 ratio (Table S2). The genotypes of the progeny were consistent with the expected phenotypes: all progeny carrying the gzcarB/pks12 or gzcarRA/pks12 genotype were white, whereas all progeny carrying GzCARB/pks12 or GzCARRA/pks12 exhibited an orange pigment, thus confirming the genetic linkage between GzCARB and GzCARRA and carotenoid production.


Carotenoids, the most ubiquitous natural pigments produced by numerous fungi and plants, have been studied extensively because of their biological importance. However, the production and biosynthetic pathway of carotenoids in the ascomycete fungus G. zeae has not been studied. This is the first report identifying carotenoids produced by the fungus and characterizing carotenoid biosynthesis genes in the fungus.

GzCarRA exhibits high sequence similarity to CarRA of F. fujikuroi (Linnemannstöns et al., 2002) and Al-2 of N. crassa (Arrach et al., 2002). These genes encode a bifunctional enzyme with both phytoene synthase and carotene cyclase activity. Our study showed that ΔgzcarRA does not produce phytoene, suggesting that GzCarRA is required for phytoene synthesis, and the high sequence similarity between GzCarRA and CarRA suggests that GzCarRA also has cartotene cyclase activity.

GzCarB is highly similar to the CarB gene of F. fujikuroi (Linnemannstöns et al., 2002), and Al-1 of N. crassa (Schmidhauser et al., 1990). Al-1 synthesizes 3,4-didehydrolycopene by introducing double bonds to the phytoene substrate via phytofluene, ɛ-carotene, neurosporene, and lycopene. The major products of this enzyme are 3,4-didehydrolycopene and lycopene. γ-Carotene is not the substrate of Al-1, suggesting that torulene is synthesized from 3,4-didehydrolycopene (Hausmann & Sandmann, 2000). In our study, ΔgzcarB accumulated phytoene, indicating that GzCarB also plays a role in the dehydrogenation of phytoene. Therefore, we deduced that GzCarB is a phytoene dehydrogenase that catalyzes the formation of 3,4-didehydrolycopene and lycopene (Fig. 4).


Proposed carotenoid biosynthetic pathway in Gibberella zeae.

GzCarX and GzCarO show high similarity to carotenoid cleavage oxygenase CarX (Thewes et al., 2005) and opsin-like protein CarO (Prado et al., 2004), respectively, from F. fujikuroi. CarX expressed in Escherichia coli synthesizes retinal from β-carotene, γ-carotene, β-apo-8′-carotenal, and torulene, indicating that the function of CarX is in retinal biosynthesis (Prado-Cabrero et al., 2007b). Opsins are a class of retinal-binding proteins with seven transmembrane helical domains. In this study, G. zeae did not produce retinal and neither ΔgzcarX nor ΔgzcarO affected neurosporaxanthin and torulene production, suggesting that both genes are not functional in the fungus.

GzCarT is highly similar to CarT in F. fujikuroi. CarT functions as a torulene oxygenase, given its catalysis of the conversion of torulene into β-apo-4′-carotenal in vitro and the accumulation of torulene by the CarT null mutant of F. fujikuroi (Prado-Cabrero et al., 2007a). As expected, ΔgzcarT also accumulated torulene and produced no neurosporaxanthin, suggesting that GzCarT is a torulene oxygenase.

Based on our results, we propose the following neurosporaxanthin biosynthetic pathway in G. zeae (Fig. 4). Torulene is first synthesized by GzCarRA and GzCarB. The colorless carotenoid phytoene is synthesized from two molecules of GGPP by GzCarRA. GzCarRA is a bifunctional enzyme that contains two domains: one catalyzing phytoene synthesis and the other catalyzing the formation of β-ionone rings. The phytoene dehydrogenase GzCarB catalyzes the synthesis of 3,4-dihydrolycopene from phytoene by the stepwise introduction of up to five conjugated double bonds. Then 3,4-dihydrolycopene is converted to torulene by GzCarRA. Torulene is subsequently converted into β-apo-4′-carotenal by the torulene-cleaving oxygenase GzCarT. Finally, β-apo-4′-carotenal is oxidized to neurosporaxanthin by an aldehyde dehydrogenase (Fig. 4).

In conclusion, we identified carotenoids produced by G. zeae and characterized three G. zeae genes that are related to carotenoid biosynthesis. Two of the three genes are contained in a putative carotenoid biosynthetic gene cluster, but the third is not linked to the cluster. All three genes are required for neurosporaxanthin production. Based on these results, we propose a carotenoid biosynthetic pathway in G. zeae. In addition, the Δpks12 strain can be used to easily differentiate carotenoid production, which highlights G. zeae as a system for further carotenoid studies, including identification of other genes required for carotenoid biosynthesis and regulation of carotenoid production.

Supporting Information

Additional Supporting Information may be found in the online version of this article:

Table S1. Primers used in this study.

Table S2. Genetic complementation by outcrossing.

Please note: Wiley-Blackwell is not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.


This work was supported by the Crop Functional Genomics Center of the 21st Century Frontier Research Program funded by the Korean Ministry of Education, Science and Technology (CG1411), and the National Research Foundation of Korea (NRF) grant funded by the Korea government (MEST) (2009-0063350).


  • Editor: Richard Staples


View Abstract