Báo cáo khoa học: Deviation of the neurosporaxanthin pathway towards b-carotene biosynthesis in Fusarium fujikuroi by a point mutation in the phytoene desaturase gene

Deviation of the neurosporaxanthin pathway towards b-carotene biosynthesis in Fusarium fujikuroi by a point mutation in the phytoene desaturase gene Alfonso Prado-Cabrero1, Patrick Schaub2, Violeta Dı´az-Sa´ nchez1, Alejandro F. Estrada1, Salim Al-Babili2 and Javier Avalos1

1 Departamento de Gene´ tica, Universidad de Sevilla, Spain 2 Albert-Ludwigs University of Freiburg, Faculty of Biology, Germany

Keywords carB; carotenogenesis; carotenoid overproducing mutant; filamentous fungi; PDS enzyme

Correspondence J. Avalos, Departamento de Gene´ tica, Universidad de Sevilla, Apartado 1095, E-41080 Sevilla, Spain Fax: +34 95 455 7104 Tel: +34 95 455 7110 E-mail: avalos@us.es

(Received 12 May 2009, revised 12 June 2009, accepted 22 June 2009)

doi:10.1111/j.1742-4658.2009.07164.x

Carotenoids are widespread terpenoid pigments with applications in the food and feed industries. Upon illumination, the gibberellin-producing fun- gus Fusarium fujikuroi (Gibberella fujikuroi mating population C) develops an orange pigmentation caused by an accumulation of the carboxylic apoc- arotenoid neurosporaxanthin. The synthesis of this xanthophyll includes five desaturation steps presumed to be catalysed by the carB-encoded phy- toene desaturase. In this study, we identified a yellow mutant (SF21) by mutagenesis of a carotenoid-overproducing strain. HPLC analyses indi- cated a specific impairment in the ability of SF21-CarB to perform the fifth desaturation, as implied by the accumulation of c-carotene and b-carotene, which arise through four-step desaturation. Sequencing of the SF21 carB allele revealed a single mutation resulting in an exchange of a residue con- served in other five-step desaturases. Targeted carB allele replacement proved that this single mutation is the cause of the SF21 carotenoid pat- tern. In support, expression of SF21 CarB in engineered carotene-produc- ing Escherichia coli strains demonstrated its reduced ability to catalyse the fifth desaturation step on both monocyclic and acyclic substrates. Further mutagenesis of SF21 led to the isolation of two mutants, SF73 and SF98, showing low desaturase activities, which mediated only two desaturation steps, resulting in accumulation of the intermediate f-carotene at low levels. Both strains contained an additional mutation affecting a CarB domain tentatively associated with carotenoid binding. SF21 exhibited higher carot- enoid amounts than its precursor strain or the SF73 and SF98 mutants, although carotenogenic mRNA levels were similar in the four strains.

Introduction

Carotenoids are terpenoid pigments widely distributed in nature, produced by all photosynthetic organisms [1] and many nonphotosynthetic microorganisms, such as bacteria and fungi [2,3]. In plants and algae, carote- noids play essential roles as accessory pigments in pho- tosynthesis [4], and provide red, orange, or yellow

colours to many fruits and flowers. Animals lack the ability to synthesize carotenoids and rely on their diet to produce the vision chromophore retinal [5] or the vertebrate morphogen retinoic acid [6]. Carotenoids are also beneficial for human health as protective agents against oxidative stress, cancer, sight degenera-

Abbreviations PDS, phytotene desaturase; PPO, protoporphyrinogen IX oxidase.

FEBS Journal 276 (2009) 4582–4597 ª 2009 The Authors Journal compilation ª 2009 FEBS

4582

A. Prado-Cabrero et al.

Alteration of Fusarium phytoene desaturase

but

[10],

tion syndromes and cardiovascular diseases [7]. In addition, carotenoids are responsible for the pigmenta- tion of some birds, insects, fish and crustaceans.

other and to the hydroxyneurosporene dehydrogenase of Rhodobacter sphaeroides show low sequence similarity to other bacterial counterparts like the Pantoea phytoene desaturase CrtI. The low sequence conservation suggests a convergent evolution of both groups, further substantiated by their different sensitivities to chemical inhibitors [9]. Other PDS- related enzymes act as isomerases [14], e.g. the plant and cyanobacterial prolycopene isomerase CrtISO [15], the animal all-trans-retinol:all- or as saturases, e.g. trans-13,14-dihydroretinol saturase RetSat [16].

Most naturally occurring carotenoids share a typical chemical structure derived from the C40 polyene chain of the colourless precursor phytoene, a carotene syn- thesized by the enzyme phytoene synthase through the condensation of two geranylgeranyl pyrophosphate molecules (Fig. 1). Carotenoid biosynthetic pathways proceed through the sequential introduction of conju- gated double bonds in the phytoene backbone to yield increasingly desaturated molecules absorbing visible light. Desaturation steps are usually followed by cycli- zation reactions catalysed by carotene cyclases. The generated end-rings may be further modified by differ- ent oxidases introducing oxygen-containing functional groups. Carotenoids are divided into carotenes consist- ing of hydrocarbons and their oxygenated derivatives the xanthophylls [8].

cyclase CarRA [21],

synthase ⁄ carotene

Desaturation steps are achieved by a group of enzymes, with phytoene desaturases (PDSs) as their most representative members. PDS enzymes differ in the number of introduced double bonds, which range from two to five [9]. Some PDS-related enzymes desat- urate substrates other than phytoene, e.g. hydroxyneu- rosporene [10], dehydrosqualene [11] or f-carotene [12]. Plants, algae and cyanobacteria employ two enzymes, PDS and f-carotene desaturase, to perform the four desaturation reactions required for lycopene formation [13]. These enzymes are evolutionarily related to each

Many fungal species are useful tools for the produc- tion of secondary metabolites and the analysis of their biosyntheses. One example is the ascomycete Fusari- um fujikuroi (Gibberella fujikuroi MP-C), known for its ability to produce gibberellins [17], growth-promoting plant hormones with agricultural applications. Upon illumination, F. fujikuroi develops an orange pigmenta- tion caused by the accumulation of neurosporaxanthin [18], a carboxylic apocarotenoid originally found in the fungus Neurospora crassa [19]. Neurosporaxanthin is produced from phytoene through five desaturations, an end-cyclization, an oxidative cleavage reaction and a final oxidation step (Fig. 1). This pathway is medi- ated by the PDS CarB [20,21], the bifunctional phyto- the ene carotenoid cleaving oxygenase CarT [22] and finally by the presumed aldehyde dehydrogenase CarD, which is currently under investigation. F. fujikuroi also accumu- lates minor amounts of b-carotene [18] resulting from

Fig. 1. Carotenoid and retinal biosynthesis in Fusarium fujikuroi. The pathway involves CarRA, CarB, the cleaving oxygenases CarX and CarT, and a postulated dehydrogenase CarD. Desaturations introduced by the CarB enzyme are circled. The grey arrow indi- cates the reaction affected in the SF21 mutant. Reactions under-represented in this strain are shaded.

FEBS Journal 276 (2009) 4582–4597 ª 2009 The Authors Journal compilation ª 2009 FEBS

4583

A. Prado-Cabrero et al.

Alteration of Fusarium phytoene desaturase

strains. Finally, a hypothesis

producing E. coli is proposed to explain the structural basis of the effect of the mutation.

Results

Isolation and phenotypic analysis of a yellow mutant

end-cyclization of the intermediate c-carotene, cataly- sed by CarRA (Fig. 1). b-Carotene is the substrate for CarX, a second carotenoid-cleaving oxygenase, which produces retinal [23,24]. Expression of the identified car genes is stimulated by light and derepressed in the dark in carotenoid-overproducing mutants, generically called carS [22,23,25]. The mutated regulatory gene(s) responsible for the carS phenotype remains to be iden- tified.

[31],

The pale pigmentation of wild-type F. fujikuroi hinders the identification of colour mutants with alterations in the carotenoid pattern. Such mutants are easily identi- fied in deeply orange-pigmented strains like the carS carotenoid-overproducing mutants [18]. A screening for colour mutants was performed after chemical mutagenesis of the carS strain SF4, a descendent of the nitrate reductase-deficient mutant SF1 (Table 1), not affected in carotenoid biosynthesis and formerly used as a recipient strain for transformation experi- ments [25]. This search led to the identification of a mutant with a striking yellow colour (Fig. 2). This subcultured from single conidia and mutant was denominated as SF21.

the same end-product,

As in F. fujikuroi, a single desaturase gene has been found in other carotenogenic fungi: the ascomycetes N. crassa (al-1) [26] and Cercospora nicotianae (pdh1) [27], the zygomycetes Phycomyces blakesleeanus, Mu- cor circinelloides and Blakeslea trispora (carB) [28–30] and the basidiomycete Xhanthophyllomyces dendror- hous (crtI) formerly Phaffia rhodozyma. These enzymes, more similar to those of carotenogenic bacte- ria than to desaturases of photosynthetic organisms, are presumably responsible for all desaturation steps in the corresponding carotenoid pathways. The ability to carry out four desaturations was first inferred from genetic approaches for the CarB PDS from P. blakes- leeanus [32], and later confirmed by heterologous expression in Escherichia coli [28]. A similar heterolo- gous approach demonstrated the ability of CrtI from X. dendrorhous to catalyse the four desaturations from phytoene to lycopene [31] and of AL-1 from N. crassa to achieve the five desaturations from phytoene to 3,4- didehydrolycopene [33]. The carotenoid pathway of N. crassa coincides with that of F. fujikuroi in the syn- thesis of the apocarotenoid neurosporaxanthin, but both fungi differ in the order of the reactions. Whereas in N. crassa the five desatu- rations are performed first and followed by cyclization reaction as a later step [34], in F. fujikuroi the cycliza- tion reaction precedes the fourth and fifth desatura- tion steps, as indicated by the absence of lycopene in different and the occurrence of b-zeacarotene strains [18].

The carotenoids produced by SF21 were analysed by spectrophotometry, TLC and HPLC (Fig. 2). As the indicated by their colours, UV ⁄ Vis spectra of SF21 carotenoid samples differed from that of its ancestor strain SF4 in shape and maximal absorption. TLC analyses revealed that most of the carotenoids accumulated by SF4 were highly polar, pointing to the predominant component. neurosporaxanthin as The minor neutral fraction contained torulene and traces of other carotene intermediates. Parallel separa- tion of the SF21 crude carotenoid samples revealed two predominant bands corresponding to c-carotene and b-carotene. In contrast to SF4, no torulene could be detected, and the neurosporaxanthin band was much paler. HPLC analyses of the neutral carotenoid fractions from both strains confirmed the predomi- nance of torulene in SF4 and the accumulation of large amounts of c-carotene and b-carotene in SF21 (Fig. 2). The amount of phytoene was low in both strains, but was significantly higher in SF21 than in SF4.

Quantification of the carotenoid contents in mycelial samples from light- or dark-grown cultures showed similar results, except for a higher neurosporaxanthin content in the illuminated SF4 samples (Fig. 2). As expected, its parental strain SF1 produced moderate amounts of neurosporaxanthin only in the light. The carotenoid concentration was at least threefold higher in SF21 than in SF4, with neurosporaxanthin repre- senting < 10% of the total carotene.

The accumulation of phytoene in carB mutants [20] and the lack of strains blocked in a later desaturation step indicated that CarB is responsible for all five desaturations. Here, we provide conclusive evidence for the ability of the CarB enzyme to carry out the five desaturation reactions and to discriminate between different carotenoid substrates. We have isolated and characterized a F. fujikuroi carB mutant impaired in the catalysis of the fifth desaturation (i.e. that con- verting c-carotene to torulene), but fully able to cata- lyse the preceding four desaturation steps. The effect the mutation was confirmed by targeted allele of replacement and comparing the activity of wild-type and altered CarB enzymes in different carotenoid-

FEBS Journal 276 (2009) 4582–4597 ª 2009 The Authors Journal compilation ª 2009 FEBS

4584

A. Prado-Cabrero et al.

Alteration of Fusarium phytoene desaturase

Table 1. Fusarium fujikuroi strains used in this study. Only the relevant transformant is included. For clarity, wild-type carB alleles (carB+) are also indicated. NG, N-methyl-N¢-nitro-N-nitrosoguanidine.

Strain

Genotypea,b

Origin

Colour in the dark

FKMC1995 SF1

carB+ niaD4 carB+

White White

SF4 SF21 SF73 SF98 T5 SF191 SF214

niaD4 carS35 carB+ niaD4 carS35 carB36 niaD4 carS35 carB37 niaD4 carS35 carB38 niaD4 carB+ ⁄ carB36 hygR niaD4 carS63 carB+ ⁄ carB36 hygR niaD4 carS63 carB36

FKMC1995, spontaneous ClO3K resistance SF1, NG mutagenesis SF4, NG mutagenesis SF21, NG mutagenesis SF21, NG mutagenesis SF1, transformation with pB21H T5, NG mutagenesis SF191, spontaneous

Orange Yellow Greenish Pale greenish White Orange Yellow

plasmid loss

SF215

niaD4 carS63 carB36

SF191, spontaneous

Yellow

plasmid loss

SF216

niaD4 carS63 carB+

SF191, spontaneous

Orange

plasmid loss

a carS mutations are tentatively assigned to a single hypothetical carS gene. b carB37 and carB38 alleles include also the carB36 mutation.

Identification of a mutation in the SF21 carB allele

Replacement of the wild-type carB allele by carB36

and the

c-carotene

required for

torulene

i.e. the The carotenoid pattern of the mutant SF21, subsequent accumulation of deviation of the pathway to b-carotene, suggested impaired c-carotene to torulene desaturation activity. This may be the result of an altered CarB if all five synthesis are desaturations sole F. fujikuroi PDS enzyme catalysed by this (Fig. 1). To test this hypothesis, we cloned and sequenced the carB alleles from strains SF21 and wild-type FKMC1995.

identical

to that of

IMI58289 except

The carB gene was formerly cloned from the wild- type F. fujikuroi strain IMI58289 [21] and its sequence was deposited in the EMBL database (accession num- ber AJ426418). The carB sequence from FKMC1995 for a was C196 fi T transition which does not affect the encoded protein sequence. The predicted CarB protein shared a similar structural organization with other PDS and PDS-related enzymes of different origins (Fig. 3), including the characteristic N-terminal dinu- cleotide-binding domain [35,36].

Compared with carB from FKMC1995, the SF21 carB allele, designated here as carB36, showed a single point mutation, a C608 fi T transition, resulting in a Pro170 fi Leu substitution. The corresponding resi- located in a predicted a-helix-rich region due is from the presumed carotene binding far (Fig. 3) domain harbouring the mutations formerly identified in three P. blakesleaanus carB mutants [37].

The generation of mutant SF21 from wild-type FKMC1995 includes two chemical mutagenesis steps (Table 1), presumably resulting in further random mutations in addition to that found in the SF21 carB36 allele. To check if this allele is sufficient to pro- duce the deviation of the pathway to b-carotene in a wild-type carotenogenesis background, a two-step strategy was used to replace the carB allele of strain SF1 with carB36 (Fig. 4A,C). Ten transformants were isolated after transformation of the SF1 strain with a plasmid carrying carB36. In five of them, Southern blot analyses showed the incorporation of a single copy of the plasmid at the carB locus (Fig. 4A,B). Three of these strains were checked for carotenoid content. Compared with the wild-type, the three trans- formants contained approximately twofold more carot- enoids upon illumination, but exhibited similar carotenoid compositions. One of these transformants (T5, indicated by an asterisk in Fig. 4B) was chosen for further investigation. T5 conidia were grown on Petri dishes to search for mutant colonies, expected at low frequency from spontaneous plasmid loss by homologous recombination (Fig. 4C). Transfer of indi- vidual colonies to selective medium showed variable frequencies of hygromycin sensitive strains, usually > 1%. However, all the strains tested were orange and contained the wild-type carB allele, suggesting preferential recombination through the same DNA segment that led to the plasmid integration. No yellow

FEBS Journal 276 (2009) 4582–4597 ª 2009 The Authors Journal compilation ª 2009 FEBS

4585

A. Prado-Cabrero et al.

Alteration of Fusarium phytoene desaturase

Fig. 2. SF21 phenotype. Representative colonies of SF4 and SF21 strains grown in the dark at 22 (cid:2)C on DGasn agar. TLC and HPLC analy- ses of carotenoid samples from 9-day-old mycelia of both strains grown under the same conditions. UV ⁄ Vis spectra (350–550 nm) and maxi- mal absorbance wavelengths (nm) of accumulated carotenoids are shown in the insets. Below: quantitative analyses of the carotenoids produced by SF1, SF4 and SF21. A scheme of the pathway is presented on the left. Phytoene, phytofluene, f-carotene, b-zeacarotene, c-car- otene, b-carotene, torulene and neurosporaxanthin are abbreviated as P, Pf, f, b-z, c, b, T and Nx, respectively. The identities of the interme- diates are depicted by colour. Surfaces are proportional to amounts, indicated in lgÆg)1 dry mass. The data show average and standard deviation (outer semicircles) from three independent determinations. Left and right semicircles correspond to cultures grown in the dark and under continuous light, respectively. SF1 contained only trace amounts of carotenoids in the dark. SF4 contained low amounts of phytoene, c-carotene and b-carotene, represented as approximate calculations. Circles missing in the SF4 and SF21 schemes correspond to undetected carotenoids.

colonies were detected after visual inspection of at least 120 Petri dishes with (cid:2) 250–500 colonies, proba- bly because of the difficult identification of this pheno- type in the pale pigmented background of T5.

(Fig. 4A). Hygromycin-sensitive strains were obtained from SF191 and checked by PCR for the loss of the carB36 allele. One of them, called SF216, harboured a single wild-type carB allele (PCR test not shown) and had a lower carotenoid content (Fig. 5A) than SF191. In contrast to SF4, SF216 contained similar amounts of carotenoids in dark or light, indicating differences in their respective carS mutations.

Because SF21 was obtained from a carotenoid-over- producing strain, a mutagenesis experiment was used to obtain a T5-derived carS mutant, termed here SF191. This deregulated strain contained more carote- noids than SF4 (Fig. 5A), as expected from the pres- ence of two carB genes, one with the carB36 mutation

Conidia collected from SF191 were grown in the same media and screened for the generation of yellow

FEBS Journal 276 (2009) 4582–4597 ª 2009 The Authors Journal compilation ª 2009 FEBS

4586

A. Prado-Cabrero et al.

Alteration of Fusarium phytoene desaturase

Fig. 3. Alignment of predicted structures for phytoene desaturases from the fungi Fusarium fujikuroi (CarB Ff, accession number CAD19989.2), Neurospora crassa (AL-1 Nc, XP964713), Xhanthophyllomyces dendrorhous (CrtI Xd, AAO53257) and Phycomyces blakeslee- anus (CarB Pb, CAA55197.1), the bacteria Rhodobacter sphaeroides (CrtI Rs, YP353345), the archaea Sulfolobus solfataricus (CrtI Ss, NP344226), and the plant Arabidopsis thaliana (PDS3 At, Q07356). The comparison also includes the A. thaliana f-carotene desaturase (ZDS1 At, Q38893) and the human all-trans-retinol 13,14-reductase (RetSat Hs, Q6NUM9). Structures were deduced with the program 3D-PSSM. Broad rectangles represent predicted a helices, and thin rectangles represent predicted b sheets. The conserved b–a–b dinucleo- tide-binding domain is indicated in black. The a-helix-rich segment is shaded in grey. Polarity of the helices of this region and the presence of basic residues in the F. fujikuroi enzyme are indicated (•,hydrophilic; , moderately hydrophobic; s, highly hydrophobic; each short line below is either a lysine or an arginine residue). Vertical lines and arrows indicate mutations; SF21, Pro170 fi Leu; SF73, Trp449 fi Stop; SF98, Gly504 fi Asp (described in this work); A486, Glu426 fi Lys; C5, Ser444 fi Phe and Leu446 fi Phe; S442, Glu482 fi Lys, described by Sanz et al. [37]. The asterisk marks the mutation in the R. sphaeroides PDS enzyme that provides the ability to carry out a fourth desaturation.

indicating that the fifth desaturation is less efficiently achieved than the preceding four in the E. coli back- ground. Expression of carB36 resulted in lycopene amounts at least comparable with those formed by CarB, whereas the production of 3-4 didehydrolyco- pene was reduced approximately sixfold (Fig. 6A). Similar results were obtained through introducing the two cDNAs in a lycopene-producing E. coli strain, expressing the bacterial four-step desaturase gene crtI. As shown in Fig. 6B, the activities of CarB and CarB36 led to similar lycopene contents, whereas the amounts of 3,4-didehydrolycopene were approximately eightfold higher in the carB-expressing strain.

colonies. Two yellow colonies were identified in a screening of (cid:2) 5000; both strains, called SF214 and SF215, were sensitive to hygromycin, indicating loss of the integrated plasmid. As predicted, both mutants lacked the FokI restriction site, present in the wild-type carB allele, but not in the carB36 mutant allele (Fig. 4D), confirming the expected allele replacement. Like SF21, SF214 and SF215 contained low amounts of neurosporaxanthin. Moreover, HPLC analyses of their neutral carotenoid fractions showed a pattern very similar to that of SF21 either in dark- or in light- grown cultures (Fig. 5B). This result strongly indicated that the carB36 mutation is responsible for the yellow phenotype, i.e. the defective CarB capacity to carry out the fifth desaturation step in the neurosporaxan- thin biosynthetic pathway.

Heterologous expression of the carB36 allele

strains and the

To further confirm the effect of the carB36 mutation on enzyme activity, wild-type carB and carB36 cDNAs were cloned and expressed in different carotene-pro- ducing E. coli resulting carotene patterns were determined (Fig. 6). Expression of wild- type carB in a phytoene producing E. coli strain resulted in an efficient desaturation to lycopene accom- panied by a lower production of 3,4-didehydrolycopene,

The carotenoid pattern of F. fujikuroi indicates that the substrate for the fifth desaturation step is c-caro- tene rather than lycopene. Therefore, we expressed the in a c-carotene-accumulating E. coli two cDNAs strain, engineered by introduction of the bacterial desaturase CrtI and the N. crassa cyclase ⁄ phytoene synthase AL-2 [34]. Compared with CarB36, the activ- ity of CarB led to a sevenfold higher quantity of toru- lene (Fig. 6C). Similarly, the conversion of lycopene to 3,4-didehydrolycopene, achieved in parallel in the same cells, was much higher in CarB-expressing cells. Taken together, the usage of the E. coli system confirmed the the carB36 mutation on the fifth specific effect of desaturation reaction.

FEBS Journal 276 (2009) 4582–4597 ª 2009 The Authors Journal compilation ª 2009 FEBS

4587

A. Prado-Cabrero et al.

Alteration of Fusarium phytoene desaturase

A

C

B

D

Fig. 4. carB allele replacement. (A) Physical map of the pB21H integration at the homologous carB sequence by a single recombination event in the genome of strain SF1. The mutation in the carB36 allele is indicated by a star. The carB probe, relevant BamHI sites used for Southern blot analyses and expected fragment sizes are indicated. (B) Southern blot analyses of the recipient strain SF1 and 10 transfor- mants. Squares highlight transformants whose hybridization pattern indicates the incorporation of a single copy of the plasmid at the carB locus. The transformant T5 was used in further experimental steps. (C) Physical map of molecular events leading to loss of the plasmid pB21H by a single recombination at the homologous carB sequence in the genome of a carS strain derived from T5. The recombination shown occurs at the opposite side from the one that produced the plasmid integration, leaving the mutated carB allele in the genome. (D) Electrophoretic profiles of the PCR products obtained with primers flanking the mutation site at allele carB36 using DNA from wild-type, SF21, SF191, SF214 and SF215 strains and digested with FokI. Interpretation of expected bands is depicted on the right scheme. The SF21 mutation leads to the loss of a FokI restriction site.

Further alterations in carB activity

amino acids, which include the putative carotene-bind- ing domain [37], making the accumulation of minor amounts of carotenoids an unexpected result. The SF98 carB allele contains a G1657 fi A transition, leading to a Gly504 fi Asp replacement in the caro- tene-binding domain (Fig. 3). The phenotype of the SF73 and SF98 mutants could be also caused by a combined effect of these mutations with the carB36 Pro170 fi Leu substitution, also present in these strains.

To determine further residues essential for other desat- uration steps, mutagenesis experiments of the SF21 strain were performed, leading to the isolation of low- pigmented strains. Two of them were SF73 and SF98, which exhibited a pale greenish hue. HPLC analyses of these two mutants revealed the accumulation of phyto- ene and lower amounts of f-carotene (Fig. 7). SF98 also contained minor amounts of c-carotene and b-car- otene, whereas SF73 was hardly able to desaturate f-carotene. These SF73 and SF98 carotene patterns suggested the occurrence of further mutations in the carB gene, resulting in more impaired PDSs.

Relation between carotenoid biosynthesis and expression of the car genes in carS and carS ⁄ carB mutants

As shown in Figs 7 and 8, the striking difference in the carotenoid content between SF21 and its precursor strain SF4 was less pronounced in the SF21-derived

Sequence analysis of the corresponding carB genes showed a G1493 fi A transition in the SF73 carB allele, resulting in a Trp449 fi Stop mutation. The predicted truncated protein lacks the C-terminal 121

FEBS Journal 276 (2009) 4582–4597 ª 2009 The Authors Journal compilation ª 2009 FEBS

4588

A. Prado-Cabrero et al.

Alteration of Fusarium phytoene desaturase

A

B

Fig. 5. Effect of wild-type and carB36 alleles on Fusarium fujikuroi carotenogenesis. (A) Carotenoids accumulated by the mutants SF1, SF4, SF21, SF191 and SF216. (B) Neutral carotenoids accumulated by the mutants SF21, SF214 and SF215. The analyses were carried out on mycelial samples from dark or light-grown cultures (5 WÆm)2). The data show average of two independent experiments.

A

B

C

Fig. 6. CarB36 desaturation activity of CarB and CarB36 in Escherichia coli strains producing different carotene substrates. Based on HPLC analyses, the data show carotenoid compositions of three E. coli strains accumulating different carotenoid intermediates and expressing wild-type thioredoxin-carB, -carB36 or thioredoxin (control). The three E. coli strains were engineered by introducing the following enzymes: (A) phytoene synthase (phytoene accumulation in the control); (B) phytoene synthase and the bacterial desaturase CrtI (lycopene accumula- tion); (C) phytoene synthase, CrtI and the Neurospora phytoene synthase ⁄ lycopene cyclase AL-2 (lycopene and c-carotene accumulation). The data show average and standard deviation of three independent experiments. 3,4ddl = 3,4-didehydrolycopene.

carotenogenic enzymes, we carried out northern blot experiments.

mutants SF73 and SF98. Furthermore, the latter mutants regained the light induction of carotenogene- sis, which had disappeared in the parental strain SF21 (Fig. 8). To check whether the differences in carotene content are a result of altered mRNA levels for the

As expected, carRA or carB mRNAs were undetect- able in dark-grown mycelia of the wild-type and SF1 strains and highly induced after 1 h exposure to light

FEBS Journal 276 (2009) 4582–4597 ª 2009 The Authors Journal compilation ª 2009 FEBS

4589

A. Prado-Cabrero et al.

Alteration of Fusarium phytoene desaturase

Fig. 7. Scheme of the carotenoids accumulated by mutants SF21, SF73 and SF98 under continuous illumination. The pathway on the left includes only detected carotenoids. Phytoene, f-carotene, c-carotene, b-carotene, and neurosporaxanthin are abbreviated as P, f, c, b and Nx, respectively. Circle surfaces are proportional to carotenoid amounts, indicated in lgÆg)1 dry mass.

tion reaction, which is caused by a single amino acid exchange in the wild-type enzyme.

(Fig. 8). mRNA levels in dark-grown SF4 were similar to those of illuminated wild-type and SF1 strains and exhibited a significant increase because of light expo- sure, correlating with enhanced carotenoid production. Similar expression patterns were observed for SF21, SF73 and SF98, indicating that the SF21 increased carotenoid content is not caused by enhanced tran- script levels.

Discussion

CarB shares a similar structural organization with other PDS enzymes, as revealed by our secondary structure predictions using the 3d-pssm protein-fold recognition program [38]. The same overall structure, the b–a–b dinucleotide-binding domain including [35,36], is displayed by phylogenetically distant PDS- related enzymes like the f-carotene desaturase from Arabidopsis thaliana, which shows only 14% sequence identity to CarB. Several carB mutations formerly investigated in the zygomycete P. blakesleeanus are located in a region close to the carboxy-end of the pro- tein [37], tentatively associated with binding of the carotenoid substrate [39]. One of these mutants, S442, exhibits a defective PDS with partial activity for the first two desaturations, leading to the accumulation of significant amounts of f-carotene [40]. In this study, the SF21- we identified two pale greenish strains, derived mutants SF73 and SF98, exhibiting a pheno- type similar to that of the P. blakesleeanus S442 caused by mutations affecting the same protein domain.

the symmetric reaction is

The carotenoid biosynthetic pathway of the orange- pigmented F. fujikuroi includes a sequence of five desaturation steps, consisting of two pairs of equivalent reactions at symmetrical sites in the carotene skeleton, predictably interrupted by a cyclization of the interme- diate neurosporene and completed by a fifth reaction in an outer position to produce torulene (Fig. 1). We have identified a yellow-pigmented mutant, SF21, exhibiting a novel carotenoid pattern. Consistent with its deep yellow colour, SF21 accumulates large amounts of c- and b-carotene and minor amounts of the final product neurosporaxanthin, indicating a specific defect on the CarB ability to catalyse the fifth desaturation reaction. Previous studies indicated that CarB is responsible for all desaturation steps of the pathway. Lack of mutants whose end product is any of the partially desaturated intermediates was inter- preted as an indication of the achievement of the five reactions by a single desaturase [18]. The investigations of the carB36 allele and the encoded enzyme, reported here, provide solid support to this assumption. More- over, our results show that CarB36 desaturase is unique in the specific impairment of the fifth desatura-

The different carotenoid patterns of SF73 and SF98 reflect defective PDSs maintaining certain activities with respect to the first pair of reactions but different capacities to perform the second pair of desaturations. The leaky activity of the SF98 desaturase resulted in the accumulation of significant amounts of f-carotene and c-carotene. However, we could not detect their respective precursors in the pathway, i.e. phytofluene or b-zeacarotene, indicating that when a desaturation readily reaction occurs, achieved. A similar result was found with the mutant SF73, possessing a more severely impaired desaturase,

FEBS Journal 276 (2009) 4582–4597 ª 2009 The Authors Journal compilation ª 2009 FEBS

4590

A. Prado-Cabrero et al.

Alteration of Fusarium phytoene desaturase

the carboxy domain formerly interpreted as involved in carotene binding. A single mutation in the same a- helix-rich domain of the three-step PDS of Rhodobacter sphaeroides (Fig. 3) allows this enzyme to recognize neurosporene as a substrate [41], supporting a relevant role for this PDS segment in substrate recognition. This a-helix-rich domain is similar to the proposed membrane surface-binding domain of protoporphyri- nogen IX oxidase (PPO) [42], an enzyme structurally related to PDSs. The PPO domain is characterized by the presence of amphipathic a helices rich in basic amino acids that interact with the phospholipid head groups of the lipid bilayer, embedding partially into the membrane and constituting a pore, which enables entrance of the hydrophobic substrate. As in other organisms, fungal PDSs are membrane-bound proteins [43,44] that act on hydrophobic substrates occurring in the lipid bilayer. The a helices of the PDS domain mentioned above have different hydrophobicities, and five of them contain basic amino acids that could interact with phospholipid head groups (Fig. 3). PDS enzymes might employ a membrane-binding and sub- strate-uptake mechanism similar to that of PPO. The Fusarium carB36 mutation could alter the conforma- tion of a putative pore, preventing the recognition and ⁄ or entrance of c-carotene.

Fig. 8. Total carotenoid contents and mRNA levels for genes carRA and carB in the wild-type and the mutants SF1, SF4, SF21, SF73 and SF98. D: grown in the dark. L: grown under continuous illumi- nation. Northern blot analyses were performed with total RNA sam- ples. rRNA bands are shown below each panel as load controls. The bars below each northern blot show the ratios of signal intensi- ties to rRNA controls; the values are expressed relative to the maximum in each panel, taken as 1.

which lost the ability to carry out the second pair of desaturations but maintained a low, but significant, capacity to produce f-carotene. However, despite its low desaturating activity, no phytofluene was accumu- lated. Interestingly, the SF73 desaturase represents a truncated CarB lacking the C-terminus, which includes the presumed substrate-binding domain. Hence, partic- ipation of other protein segments in carotene binding must be concluded.

The proline residue replaced in the predicted F. fu- jikuroi CarB36 protein is found in the PDSs AL-1 from N. crassa and CrtI from X. dendrorhous, presum- ably able to carry out five desaturations [33,45,46]. Conversely, the PDSs from the b-carotene-producing zygomycetes M. circinelloides, P. blakesleeanus and B. trispora, which carry out only four desaturations, contain an aliphatic residue instead of proline at the same position. However, this rule seems not to be the ascomycete C. nicotiane, valid for the PDS of described as producing b-carotene [27], because this enzyme is highly similar to CarB from F. fujikuroi ((cid:2) 70% identical amino acids), including the con- served proline residue. Carotene analysis of the close relative C. cruenta shows different carotenoids, but none of them result from a fifth desaturation [47]. the Based on our observations, a side branch of carotenoid pathway in C. nicotiane involving a fifth desaturation cannot be discarded. This is actually the case in X. dendrorhous, where the four-desaturation pathway into b-carotene and astaxanthin coexists with a lateral production of torulene [45,46]. The preva- lence of astaxanthin biosynthesis implies a highly effi- cient cyclase activity, which competes with the fifth desaturation step.

The carB36 mutation is

located in a predicted a-helix-rich protein domain, apparently distant from

Former studies proposed a mechanism of action for fungal PDSs organized as oligomers. In P. blakeslee-

FEBS Journal 276 (2009) 4582–4597 ª 2009 The Authors Journal compilation ª 2009 FEBS

4591

A. Prado-Cabrero et al.

Alteration of Fusarium phytoene desaturase

results provide

a novel

X. dendrorhous could deviate the biosynthesis further toward b-carotene and, therefore, favour astaxanthin production, which is used by the aquaculture feed industry to provide colour to salmon and crustaceans [51]. As key enzymes in their biosynthesis, PDSs rep- resent obvious targets for improved biotechnological production of carotenoids. E. coli and yeast cells have been engineered to produce industrially relevant carotenoids [52-54], and special attention has been paid to PDSs, whose desaturation activities have been altered by molecular breeding and directed evo- lution [41,55]. Moreover, the bacterial PDS CrtI was employed to generate transgenic plants accumulating b-carotene, such as high b-carotene tomato [56,57], canola [58], Golden Rice [59,60] and Golden Potato [61]. Our example on how subtle changes in the sequences of these pro- teins may have drastic effects on their biosynthetic capacities.

Experimental procedures

Strains and growth conditions

anus, each PDS monomer carries out a single desatura- tion and transfers the desaturated product to the next monomer, the process being repeated up to the accom- plishment of four desaturation steps [32,37]. A similar complex was proposed for the AL-1 PDS of N. crassa, but in this case the complex would be able to incorpo- rate partially desaturated carotenoids, either natural intermediates or unusual hydoxylated substrates [33]. The incorporation of different carotenoids was ele- gantly demonstrated with a bacterial PDS, which proved able to introduce further desaturations in unnaturally long carotenoid substrates [48]. In the case of the F. fujikuroi CarB, the identification of b-zeacar- otene implies the release of neurosporene by an even- tual CarB complex and the latter incorporation of the cycled substrate. Some PDS-related enzymes work as homodimers, in which each monomer carries out all the desaturations. A well-known example is provided by PPO, which introduces six desaturations in proto- porphyrinogen [49]. This may be also the case of the Fusarium CarB enzyme. If b-zeacarotene and c-caro- tene are incorporated independently by the CarB enzyme, CarB36 could then be impaired in the ability to incorporate c-carotene rather than in the desaturase activity.

strain

precursor

FKMC1995 is a wild-type strain of F. fujikuroi (Gibber- ella fujikuroi mating population C) [62]. SF1 is a spontane- ous nitrate reductase mutant obtained from FKMC1995 upon growth selection on media supplemented with KClO3 [25]. SF4 is a carotenoid-overproducing mutant obtained from SF1 conidia exposed to N-methyl-N¢-nitro-N-nitroso- guanidine [63]. SF21 is a yellow-pigmented mutant obtained by the same procedure from SF4 (Table 1). SF21 was deposited in the Spanish Type Culture Collection (‘Cole- ccio´ n Espan˜ ola de Cultivos Tipo’, University of Valencia, Burjasot, Spain) under the name CECT 20527. For carotenoid analysis of F. fujikuroi,

strains were grown for 9 days at 22 (cid:2)C on DG minimal medium [63] with 3 gÆL)1 l-asparagine instead of NaNO3 (DGasn med- ium). Light-grown cultures were exposed to 5 WÆm)2 pro- duced by a battery of five white fluorescent lights (Sylvania Standard F36 ⁄ 154-t8).

The total amount of carotenoids accumulated by SF21 is at least three times higher than in its neuros- poraxanthin-overproducing SF4, although their carRA and carB mRNA levels remain unaltered. The amount of carotenoids in vivo is pre- sumably determined by the balance between biosyn- thetic activities and carotenoid degradation rates. Therefore, the higher carotenoid content in SF21 may be explained either by a higher activity or the CarB36 enzyme compared with the wild-type counterpart or by a higher stability of c-carotene and b-carotene com- pared with torulene and neurosporaxanthin. This may actually be the case for torulene, as indicated by the fast decoloration of a torulene-accumulating mutant upon aging (unpublished observation). The decreased accumulation of carotenoids in the carS mutant with two carB genes – wild-type and carB36 – compared with the SF21 mutant (Fig. 5) supports the hypothesis on differential carotenoid stability.

For northern blot analyses, 4 · 106 conidia were grown on 100 mL of DGasn broth in 15-cm Petri dishes for 3 days at 30 (cid:2)C in the dark. For light-induction, cultures were exposed to 25 WÆm)2 white light for 1 h at 30 (cid:2)C. Mycelia were then immediately collected, dried, frozen in liquid nitrogen and stored at )80 (cid:2)C.

in other

implications

species. For

Recombinant DNA procedures

Genomic DNA was extracted according to Giordano et al. [64] from mycelial samples harvested and washed by filtra- tion, frozen in liquid nitrogen, and ground to a fine powder the carB alleles were in a cold mortar. Sequences of

The c-carotene and b-carotene accumulation result- ing from the carB36 mutation may have biotechnologi- example, cal replacement of the carB wild-type allele by carB36 in Fusarium venenatum, a fungus used by the food indus- try as the source for mycoprotein [50], may result in the predominant synthesis of c-carotene and b-caro- tene and give added value to mycoprotein for human consumption. Furthermore, a similar mutation in

FEBS Journal 276 (2009) 4582–4597 ª 2009 The Authors Journal compilation ª 2009 FEBS

4592

A. Prado-Cabrero et al.

Alteration of Fusarium phytoene desaturase

RNA extraction and northern hybridization

primers

obtained from two overlapping PCR products generated with 5¢-TGGGCGAGCTCATGAGCGACATT AAGAAATCTG-3¢ and 5¢-CGCTCAGAACGACACCG TTTG-3¢, covering 11 bp upstream of the start codon and the carB coding sequence, and 5¢- the first 957 bp of CGTTGAGGCACTGGTTAACG-3¢ and 5¢-CGAGAAT CATGGACATAGAC-3¢, covering the last coding 1048 and 88 bp downstream of the stop codon. The sequences of each allele were determined from two clones obtained from independent PCR, and compared with that of the wild-type F. fujikuroi strain FKMC1995 (accession number AJ426418). DNA sequencing was accomplished by Newbio- technic (Seville, Spain) using an ABI Prism(cid:3) 3100 Genetic Analyser (Applied Biosystems, Foster City, CA, USA). (reverse)

For northern blot analysis, 15 lg of total RNA, extracted with the Perfect RNA eukaryotic mini kit (Eppendorf, Hamburg, Germany), was applied. Transferred RNA was stained for 5 min in 0.02% methylene blue ⁄ 0.3 m sodium acetate, pH 5.2, and rRNA bands were used as load con- trols. Filters were hybridized with probes labelled with the nonisotopic digoxigenin labelling kit (Roche). Nonisotop- ically labelled single-stranded antisense probes of genes car- RA and carB were generated by asymmetric PCR, as described previously [66]. Initial PCR was achieved with primers 5¢-TCCGGCGCATTTCCTATC-3¢ (forward) and for gene 5¢-ATCTATGAATCTATGACCTC-3¢ carRA and 5¢-GGTACTGGTGTTCCTGTCTG-3¢ (for- ward) and 5¢-CCGATCAGATAGTTGTCACG-3¢ (reverse) for gene carB. The resulting PCR products were used as substrates for subsequent asymmetric amplifications with their respective reverse primers. Signal intensities were esti- mated by densitometric analysis using the imagej 1.42q software (http://rsbweb.nih.gov/ij/).

PCR was performed with 50 ng genomic DNA, 0.2 mm dNTPs, 1 lm of each primer and 0.5 lL of the Expand PCR System (Boehringer-Mannheim, Mannheim, Germany). Reaction mixtures were heated at 94 (cid:2)C for 2 min followed by 35 cycles of denaturation (94 (cid:2)C for 30 s), annealing (55 (cid:2)C for 30 s) and polymerization (68 (cid:2)C for 1 min), and by a final polymerization at 68 (cid:2)C for 10 min. Amplified DNA fragments were purified using Wizard Minicolumns (Promega, Madison, WI, USA), and cloned into pGEM-T Easy vector (Promega).

Plasmid constructions for carB expression in E. coli

carB allelic replacement

the two-step following

Targeted mutagenesis with the SF21 carB36 allele was per- formed replacement method described by Ferna´ ndez-Martı´ n et al. [20]. SF1 protoplasts were prepared and transformed with the plasmid pB21H, carrying the carB36 allele, following the protocol described by Proctor et al. [65]. To construct pB21H, the carB allele was amplified with primers 5¢-CAGCTTGCTCACAAT CATCC-3¢ and 5¢-CTCATTCCAAGCCCGAGAAAC-3¢. The 4.5-kb product was purified with the GFX(cid:4) PCR DNA and Gel Band Purification Kits GFX (Amersham Biosciences, Piscataway, NJ, USA) and ligated to pGEM-T Easy (Pro- mega). The carB allele was then excised using NotI and ligated into NotI-treated pHJA2 [20], yielding plasmid pB21H.

Transformants and revertants were identified by restric- tion analyses of PCR fragments obtained with primers Car- BG-2F (5¢-TGGGCGAGCTCATGAGCGACATTAAGAA ATCTG-3¢) and CarBG-3R (5¢-CGCTCAGAACGACA CCGTTTG-3¢). The presence of the carB36 mutation was checked using a FokI (Takara Shuzo, Kyoto, Japan) restriction site occurring in the wild-type and absent in the mutant allele.

FEBS Journal 276 (2009) 4582–4597 ª 2009 The Authors Journal compilation ª 2009 FEBS

4593

For Southern blot analysis, 5 lg total DNA was digested, separated in 0.8% agarose electrophoresis, trans- ferred to a nylon filter and hybridized with a carB probe labelled with the nonisotopic digoxigenin labelling kit (Roche, Mannheim, Germany) by Klenow, according to the manufacturer. The probe was obtained from wild-type genomic DNA by PCR with the primers CarBG-2F and CarBG-3R. For the cloning of carB and carB36 cDNAs, 5 lg of total RNA, isolated from wild-type and SF21 mycelia grown for 1 h under white light, were subjected to cDNA synthesis using SuperScript(cid:4) RNaseH- reverse transcriptase (Invitro- gen, Paisley, UK) following the manufacturer’s instructions. Two microlitres of cDNA were used for the amplification of carB using the primers 5¢-ATGAGCGACATTAAGAA ATCTG-3¢ and 5¢-CTAATTCGCAGCAATGACAAG-3¢. The PCR was performed using 500 nm of each primer, 150 lm dNTPs and 1 unit of Phusion(cid:4) High-Fidelity DNA Polymerase (Finnzymes, Espoo, Finland) in the buffer provided by the manufacturer. The reactions consisted of 30 s of initial denaturation at 98 (cid:2)C, 32 cycles of 98 (cid:2)C for 15 s, 58 (cid:2)C for 30 s and 72 (cid:2)C for 90 s and 10 min of final polymer- ization at 72 (cid:2)C. The resulting products were purified, cloned into the pJET1.2 ⁄ blunt(cid:4) cloning vector (Fermentas, Vlinius, Lithuania) to yield pJET–CaRB and pJET–CaRB36, respec- tively, and confirmed by sequencing. The same amplified into pBAD ⁄ cDNAs products were cloned afterwards THIO-TOPO(cid:3) TA (Invitrogen), allowing the expression of a thio-fusion under the control of an arabinose-inducible pro- moter, to yield pThio–CarB and pThio–CarB36, respectively. For c-carotene production, the al-2 cDNA was obtained from total N. crassa RNA extracted from 2-day-old mycelial samples grown at 30 (cid:2)C in 25 mL liquid Vogel’s media. Five micrograms of total RNA were subjected to cDNA synthesis as described above. Two microlitres of the cDNA were then used for amplification of al-2 using the primers 5¢-TCC AAGCTTCTATATGACAATAGCGCC-3¢ and 5¢-CCAG GATCCGTCTACTGCTCATACAAC-3¢, deduced from the

A. Prado-Cabrero et al.

Alteration of Fusarium phytoene desaturase

public sequence database (accession no. XM_960632) and carrying a HindIII and a BamHI restriction site, respectively. The resulting PCR product was purified and cloned into the pCR2.1(cid:3)-TOPO(cid:3) vector (Invitrogen) to yield pCR–AL2. The product was verified by sequencing. A DNA fragment harbouring the lac promoter upstream of the al-2 cDNA was amplified by PCR from pCR–AL2 using the primers 5¢- AAGCGGCCGCGCGCCCAATACGC-3¢ and 5¢-CCGT TTCTAGAGGATCCGTCTACTGC-3¢, which contain a NotI and a XbaI restriction site, respectively. The obtained fragment was digested with NotI and XbaI and ligated into accordingly treated pFarbeR, which enables lycopene accu- mulation [34], to yield pGamma, allowing the synthesis of c- carotene upon induction with isopropyl b-d-thiogalactoside. Validity of the constructs was verified by sequencing.

E. coli in vivo tests for CarB activity

(model 996) and a YMC-Pack C30-reversed phase column (250 · 4.6 mm i.d., 5 lm; YMC Europe, Schermbeck, Germany). The column was developed at a flow-rate of 1 mLÆmin)1 with a linear gradient from 100% B (MeOH ⁄ TBME ⁄ H2O 5 : 1 : 1 v ⁄ v ⁄ v) to 50% A (MeOH ⁄ TBME 1 : 4 v ⁄ v) within 15 min, then to 100% A within 7 min, followed by another 10 min under the same conditions. The flow-rate was then enhanced to 2 mLÆmin)1, and the sepa- ration was continued for further 8 min. Using a Maxplot (400–500 nm), coloured carotenoid peaks were integrated at their individual kmax, whereas phytoene and phytofluene peaks were integrated at 286 and 348 nm, respectively. Normalization and quantification were performed using an internal a-tocopherol acetate standard according to Hoa et al. [69]. Carotenoid amounts were calculated according to a b-carotene standard curve. Peaks were then normalized to correct for their individual molar extinction coefficients [70] relative to that of b-carotene, using the correction factors 1.970 for phytoene, 1.820 for phytofluene, 1.650 for f-carotene, 0.806 for c-carotene, 0.720 for lycopene, 0.774 for torulene and 0.800 for 3,4-didehydrolycopene. cells were

Sequence and secondary structure analysis

of LB medium containing

Alignments were carried out with the multi-processor ver- sion 1.81 of the clustal w program using the server of the Centre for Molecular and Biomolecular Informatics (Nijmegen, The Netherlands). Secondary structure ana- lysis was carried out with the threading program 3d-pssm [38].

Acknowledgements

lycopene- and c-carotene-accumulating E. coli Phytoene-, strains were generated by transforming TOP 10 cells with the plasmids pPhytoene [34], pFarbeR and pGamma, then respectively. Carotenoid-accumulating transformed with pThio–CarB, pThio–CarB36 and the control plasmid pThio (pBAD ⁄ THIO-TOPO(cid:3) TA). Fifty millilitres kanamycin (50 lgÆmL)1), ampicillin (100 lgÆmL)1) and 0.2 mm isopro- pyl b-d-thiogalactoside, the latter to induce the expression of the genes for the synthesis of phytoene, lycopene and c-carotene, were grown overnight. Afterwards, the cultures were grown at 28 (cid:2)C to D600 = 0.5 and induced with 0.2% arabinose (w ⁄ v). Forty millilitres of the cultures were reaching D600 (cid:2) 1.2. Carotenoids were harvested after extracted from the cell pellets by sonication in 20 mL CHCl3 ⁄ MeOH (2 : 1, v ⁄ v). After centrifugation, epiphases were discarded and organic phases were isolated. Interpha- ses were then re-extracted with 20 mL CHCl3. The organic phases were combined, dried under vacuum and resus- pended in 100 lL CHCl3, of which 10 lL were subjected to HPLC.

Chemical analyses

We are indebted to Dr Peter Beyer for valuable discus- sions. We thank Erdmann Scheffer for skilful technical assistance. This work was supported by the Spanish Government (Ministerio de Ciencia y Tecnologı´ a, pro- jects BIO2003-01548 and BIO2006-01323), the Andalu- sian Government (project P07-CVI-02813), a grant to Dr Peter Beyer by the Bill & Melinda Gates Founda- tion as part of the Grand Challenges in Global Health Initiative, and the Deutsche Forschungsgemeinschaft (DFG) Grant AL892-1-4.

References

1 Cunningham FX & Gantt E (1998) Genes and enzymes of carotenoid biosynthesis in plants. Annu Rev Plant Physiol Plant Mol Biol 49, 557–583. 2 Britton G, Liaaen-Jensen S & Pfander H (2004) Carote- noids Handbook. Birkha¨ user, Basel. For carotenoid analysis, mycelial samples were separated from agar and lyophilized before extraction. Carotenoids were extracted and analysed as described by Arrach et al. [67]. Neutral carotenoids were separated from neurospora- xanthin by Al2O3 chromatography [68]. The composition of neutral carotenoid mixtures was determined by HPLC as described by Arrach et al. [67]. HPLC analyses for each strain were performed at least three times. TLC was carried out on Silica gel 60 (Sigma) run with toluene ⁄ light petro- leum 5 : 95 v ⁄ v.

FEBS Journal 276 (2009) 4582–4597 ª 2009 The Authors Journal compilation ª 2009 FEBS

4594

3 Fraser PD & Bramley PM (2004) The biosynthesis and nutritional uses of carotenoids. Prog Lipid Res 43, 228– 265. Carotenoid analyses of E. coli cells were performed by HPLC using a Waters system Alliance 2695 (Eschborn, Germany) equipped with a photodiode array detector

A. Prado-Cabrero et al.

Alteration of Fusarium phytoene desaturase

4 Telfer A (2005) Too much light? How beta-carotene 21 Linnemannsto¨ ns P, Prado MM, Ferna´ ndez-Martı´ n R,

protects the photosystem II reaction centre Photochem Photobiol Sci 4, 950–956. Tudzynski B & Avalos J (2002) A carotenoid biosynthe- sis gene cluster in Fusarium fujikuroi: the genes carB and carRA. Mol Genet Genomics 267, 593–602. 22 Prado-Cabrero A, Estrada AF, Al-Babili S & Avalos J

5 Moise AR, von Lintig J & Palczewski K (2005) Related enzymes solve evolutionarily recurrent problems in the metabolism of carotenoids. Trends Plant Sci 10, 178– 186. 6 Chambon P (1996) A decade of molecular biology of (2007) Identification and biochemical characterization of a novel carotenoid oxygenase: elucidation of the cleavage step in the Fusarium carotenoid pathway. Mol Microbiol 64, 448–460. retinoic acid receptors. FASEB J 10, 940–954. 7 Rao AV & Rao LG (2007) Carotenoids and human health. Pharmacol Res 55, 207–216. 8 Hirschberg J (2001) Carotenoid biosynthesis in flower- ing plants. Curr Opin Plant Biol 4, 210–218. 23 Thewes S, Prado-Cabrero A, Prado MM, Tudzynski B & Avalos J (2005) Characterization of a gene in the car cluster of Fusarium fujikuroi that codes for a protein of the carotenoid oxygenase family. Mol Genet Genomics 274, 217–228.

9 Sandmann G (2009) Evolution of carotene desaturation: the complication of a simple pathway. Arch Biochem Biophys 483, 169–174.

24 Prado-Cabrero A, Scherzinger D, Avalos J & Al-Babili S (2007) Retinal biosynthesis in fungi: characterization of the carotenoid oxygenase CarX from Fusarium fujikuroi. Eukaryot Cell 6, 650–657. 25 Prado MM, Prado-Cabrero A, Fernandez-Martin R &

10 Albrecht M, Ruther A & Sandmann G (1997) Purifica- tion and biochemical characterization of a hydroxyneu- rosporene desaturase involved in the biosynthetic pathway of the carotenoid spheroidene in Rhodobact- er sphaeroides. J Bacteriol 179, 7462–7467. Avalos J (2004) A gene of the opsin family in the carot- enoid gene cluster of Fusarium fujikuroi. Curr Genet 46, 47–58.

26 Schmidhauser TJ, Lauter FR, Russo VE & Yanofsky C (1990) Cloning, sequence, and photoregulation of al-1, a carotenoid biosynthetic gene of Neurospora crassa. Mol Cell Biol 10, 5064–5070. 11 Wieland B, Feil C, Gloria-Maercker E, Thumm G, Lechner M, Bravo JM, Poralla K & Gotz F (1994) Genetic and biochemical analyses of the biosynthesis of the yellow carotenoid 4,4¢-diaponeurosporene of Staph- ylococcus aureus. J Bacteriol 176, 7719–7726. 27 Ehrenshaft M & Daub ME (1994) Isolation, sequence,

and characterization of the Cercospora nicotianae phyto- ene dehydrogenase gene. Appl Environ Microbiol 60, 2766–2771. 12 Linden H, Misawa N, Saito T & Sandmann G (1994) A novel carotenoid biosynthesis gene coding for zeta-caro- tene desaturase: functional expression, sequence and phylogenetic origin. Plant Mol Biol 24, 369–379.

13 DellaPenna D & Pogson BJ (2006) Vitamin synthesis in plants: tocopherols and carotenoids. Annu Rev Plant Biol 57, 711–738. 28 Ruiz-Hidalgo MJ, Benito EP, Sandmann G & Eslava AP (1997) The phytoene dehydrogenase gene of Phyc- omyces: regulation of its expression by blue light and vitamin A. Mol Gen Genet 253, 734–744.

14 Giuliano G, Giliberto L & Rosati C (2002) Carotenoid isomerase: a tale of light and isomers. Trends Plant Sci 7, 427–429. 15 Isaacson T, Ohad I, Beyer P & Hirschberg J (2004) 29 Velayos A, Blasco JL, Alvarez MI, Iturriaga EA & Eslava AP (2000) Blue-light regulation of phytoene dehydrogenase (carB) gene expression in Mucor circi- nelloides. Planta 210, 938–946.

Analysis in vitro of the enzyme CRTISO establishes a poly-cis-carotenoid biosynthesis pathway in plants. Plant Physiol 136, 4246–4255. 16 Moise AR, Kuksa V, Imanishi Y & Palczewski K 30 Rodrı´ guez-Saiz M, Paz B, De La Fuente JL, Lo´ pez- Nieto MJ, Cabri W & Barredo JL (2004) Blakeslea trispora genes for carotene biosynthesis. Appl Environ Microbiol 70, 5589–5594. (2004) Identification of all-trans-retinol:all-trans-13,14- dihydroretinol saturase. J Biol Chem 279, 50230–50242.

17 Tudzynski B (2005) Gibberellin biosynthesis in fungi: genes, enzymes, evolution, and impact on biotechnol- ogy. Appl Microbiol Biotechnol 66, 597–611. 31 Verdoes JC, Misawa N & van Ooyen AJ (1999) Cloning and characterization of the astaxanthin biosynthetic gene encoding phytoene desaturase of Xanthophyllomyces dendrorhous. Biotechnol Bioeng 63, 750–755. 18 Avalos J & Cerda´ -Olmedo E (1987) Carotenoid mutants

of Gibberella fujikuroi. Curr Genet 11, 505–511. 19 Aasen AJ & Jensen SL (1965) Fungal carotenoids II.

32 Aragon CM, Murillo FJ, de la Guardia MD & Cerda- Olmedo E (1976) An enzyme complex for the dehydro- genation of phytoene in Phycomyces. Eur J Biochem 63, 71–75. The structure of the carotenoid acid neurosporaxanthin. Acta Chem Scand 19, 1843–1853. 20 Ferna´ ndez-Martı´ n R, Cerda´ -Olmedo E & Avalos J

FEBS Journal 276 (2009) 4582–4597 ª 2009 The Authors Journal compilation ª 2009 FEBS

4595

(2000) Homologous recombination and allele replace- ment in transformants of Fusarium fujikuroi. Mol Gen Genet 263, 838–845. 33 Hausmann A & Sandmann G (2000) A single five-step desaturase is involved in the carotenoid biosynthesis pathway to beta-carotene and torulene in Neuros- pora crassa. Fungal Genet Biol 30, 147–153.

A. Prado-Cabrero et al.

Alteration of Fusarium phytoene desaturase

34 Estrada AF, Maier D, Scherzinger D, Avalos J &

47 Inomata M, Hirai N, Yoshida R & Ohigashi H (2004) Biosynthesis of abscisic acid by the direct pathway via ionylideneethane in a fungus, Cercospora cruenta. Biosci Biotechnol Biochem 68, 2571–2580. Al-Babili S (2008) Novel apocarotenoid intermediates in Neurospora crassa mutants imply a new biosynthetic reaction sequence leading to neurosporaxanthin forma- tion. Fungal Genet Biol 45, 1497–1505. 35 Armstrong GA, Alberti M & Hearst JE (1990) Con-

48 Tobias AV & Arnold FH (2006) Biosynthesis of novel carotenoid families based on unnatural carbon back- bones: a model for diversification of natural product pathways. Biochim Biophys Acta 1761, 235–246. 49 Dailey TA & Dailey HA (1998) Identification of an served enzymes mediate the early reactions of caroten- oid biosynthesis in nonphotosynthetic and photosynthetic prokaryotes. Proc Natl Acad Sci USA 87, 9975–9979.

FAD superfamily containing protoporphyrinogen oxid- ases, monoamine oxidases, and phytoene desaturase. Expression and characterization of phytoene desaturase of Myxococcus xanthus. J Biol Chem 273, 13658–13662. 50 Wiebe MG (2002) Myco-protein from Fusarium venena- tum: a well-established product for human consump- tion. Appl Microbiol Biotechnol 58, 421–427. 36 Bartley GE, Schmidhauser TJ, Yanofsky C & Scolnik PA (1990) Carotenoid desaturases from Rhodobact- er capsulatus and Neurospora crassa are structurally and functionally conserved and contain domains homolo- gous to flavoprotein disulfide oxidoreductases. J Biol Chem 265, 16020–16024. 51 Johnson EA (2003) Phaffia rhodozyma: colorful odys- sey. Int Microbiol 6, 169–174.

52 Schmidt-Dannert C (2000) Engineering novel carotenoids in microorganisms. Curr Opin Biotechnol 11, 255–261. 53 Lee PC & Schmidt-Dannert C (2002) Metabolic engi- 37 Sanz C, Alvarez MI, Orejas M, Velayos A, Eslava AP & Benito EP (2002) Interallelic complementation pro- vides genetic evidence for the multimeric organization of the Phycomyces blakesleeanus phytoene dehydroge- nase. Eur J Biochem 269, 902–908. 38 Kelley LA, MacCallum RM & Sternberg MJ (2000)

neering towards biotechnological production of carote- noids in microorganisms. Appl Microbiol Biotechnol 60, 1–11. Enhanced genome annotation using structural profiles in the program 3D-PSSM. J Mol Biol 299, 499– 520. 39 Armstrong GA, Alberti M, Leach F & Hearst JE 54 Mijts BN, Lee PC & Schmidt-Dannert C (2004) Engi- neering carotenoid biosynthetic pathways. Methods Enzymol 388, 315–329.

(1989) Nucleotide sequence, organization, and nature of the protein products of the carotenoid biosynthesis gene cluster of Rhodobacter capsulatus. Mol Gen Genet 216, 254–268. 55 Schmidt-Dannert C, Umeno D & Arnold FH (2000) Molecular breeding of carotenoid biosynthetic path- ways. Nat Biotechnol 18, 750–753. 40 Bejarano ER, Govind NS & Cerda-Olmedo E (1987)

n-carotene and other carotenes in a Phycomyces mutant. Phytochemistry 26, 2251–2254. 41 Wang CW & Liao JC (2001) Alteration of product

specificity of Rhodobacter sphaeroides phytoene desatur- ase by directed evolution. J Biol Chem 276, 41161– 41164. 56 Rosati C, Aquilani R, Dharmapuri S, Pallara P, Maru- sic C, Tavazza R, Bouvier F, Camara B & Giuliano G (2000) Metabolic engineering of beta-carotene and lyco- pene content in tomato fruit. Plant J 24, 413–419. 57 Romer S, Fraser PD, Kiano JW, Shipton CA, Misawa N, Schuch W & Bramley PM (2000) Elevation of the provitamin A content of transgenic tomato plants. Nat Biotechnol 18, 666–669.

42 Arnould S & Camadro JM (1998) The domain structure of protoporphyrinogen oxidase, the molecular target of diphenyl ether-type herbicides. Proc Natl Acad Sci USA 95, 10553–10558. 58 Ravanello MP, Ke D, Alvarez J, Huang B & Shew- maker CK (2003) Coordinate expression of multiple bacterial carotenoid genes in canola leading to altered carotenoid production. Metab Eng 5, 255–263.

43 Fraser PD & Bramley PM (1994) The purification of phytoene dehydrogenase from Phycomyces blakeslee- anus. Biochim Biophys Acta 1212, 59–66. 44 Mitzka-Schnabel U (1985) Carotenogenic enzymes from 59 Ye X, Al-Babili S, Kloti A, Zhang J, Lucca P, Beyer P & Potrykus I (2000) Engineering the provitamin A (beta-carotene) biosynthetic pathway into (carotenoid- free) rice endosperm. Science 287, 303–305. Neurospora. Pure Appl Chem 57, 667–669.

60 Al-Babili S, Hoa TT & Schaub P (2006) Exploring the potential of the bacterial carotene desaturase CrtI to increase the beta-carotene content in Golden Rice. J Exp Bot 57, 1007–1014. 45 An GH, Cho MH & Johnson EA (1999) Monocyclic carotenoid biosynthetic pathway in the yeast Phaf- fia rhodozyma (Xanthophyllomyces dendrorhous). J Bio- sci Bioeng 88, 189–193. 61 Diretto G, Al-Babili S, Tavazza R, Papacchioli V, 46 Verdoes JC, Sandmann G, Visser H, Diaz M, van

FEBS Journal 276 (2009) 4582–4597 ª 2009 The Authors Journal compilation ª 2009 FEBS

4596

Mossel M & van Ooyen AJ (2003) Metabolic engineer- ing of the carotenoid biosynthetic pathway in the yeast Xanthophyllomyces dendrorhous (Phaffia rhodozyma). Appl Environ Microbiol 69, 3728–3738. Beyer P & Giuliano G (2007) Metabolic engineering of potato carotenoid content through tuber-specific overexpression of a bacterial mini-pathway. PLoS ONE 2, e350.

A. Prado-Cabrero et al.

Alteration of Fusarium phytoene desaturase

62 O’Donnell K, Cigelnik E & Niremberg HI (1998) Molecular systematics and phylogeography of the Gibberella fujikuroi species complex. Mycologia 90, 465–493. 63 Avalos J, Casadesu´ s J & Cerda´ -Olmedo E (1985) 66 Di Pietro A & Roncero MI (1998) Cloning, expression, and role in pathogenicity of pg1 encoding the major extracellular endopolygalacturonase of the vascular wilt pathogen Fusarium oxysporum. Mol Plant Microbe Interact 11, 91–98.

Gibberella fujikuroi mutants obtained with UV radiation and N-methyl-N’-nitro-N-nitrosoguanidine. Appl Envi- ron Microbiol 49, 187–191. 67 Arrach N, Schmidhauser TJ & Avalos J (2002) Mutants of the carotene cyclase domain of al-2 from Neuros- pora crassa. Mol Genet Genomics 266, 914–921. 64 Giordano W, Avalos J, Fernandez-Martin R, Cerda-

68 Avalos J & Cerda´ -Olmedo E (1986) Chemical modifica- tion of carotenogenesis in Gibberella fujikuroi. Phyto- chemistry 25, 1837–1841.

FEBS Journal 276 (2009) 4582–4597 ª 2009 The Authors Journal compilation ª 2009 FEBS

4597

Olmedo E & Domenech CE (1999) Lovastatin inhibits the production of gibberellins but not sterol or caroten- oid biosynthesis in Gibberella fujikuroi. Microbiology 145, 2997–3002. 69 Hoa TT, Al-Babili S, Schaub P, Potrykus I & Beyer P (2003) Golden Indica and Japonica rice lines amenable to deregulation. Plant Physiol 133, 161–169. 70 Britton G (1985) General carotenoids methods. Methods Enzymol 111, 113–149. 65 Proctor RH, Hohn TM & McCormick SP (1997) Resto- ration of wild-type virulence to Tri5 disruption mutants of Gibberella zeae via gene reversion and mutant com- plementation. Microbiology 143, 2583–2591.