Maize Y9 Encodes a Product Essential for 15-cis-ζ-Carotene Isomerization

Carotenoids are a diverse group of pigments found in plants, fungi, and bacteria. They serve essential functions in plants and provide health benefits for humans and animals. In plants, it was thought that conversion of the C40 carotenoid backbone, 15-cis-phytoene, to all-trans-lycopene, the geometrical isomer required by downstream enzymes, required two desaturases (phytoene desaturase and ζ-carotene desaturase [ZDS]) plus a carotene isomerase (CRTISO), in addition to light-mediated photoisomerization of the 15-cis-double bond; bacteria employ only a single enzyme, CRTI. Characterization of the maize (Zea mays) pale yellow9 (y9) locus has brought to light a new isomerase required in plant carotenoid biosynthesis. We report that maize Y9 encodes a factor required for isomerase activity upstream of CRTISO, which we term Z-ISO, an activity that catalyzes the cis- to trans-conversion of the 15-cis-bond in 9,15,9′-tri-cis-ζ-carotene, the product of phytoene desaturase, to form 9,9′-di-cis-ζ-carotene, the substrate of ZDS. We show that recessive y9 alleles condition accumulation of 9,15,9′-tri-cis-ζ-carotene in dark tissues, such as roots and etiolated leaves, in contrast to accumulation of 9,9′-di-cis-ζ-carotene in a ZDS mutant, viviparous9. We also identify a locus in Euglena gracilis, which is similarly required for Z-ISO activity. These data, taken together with the geometrical isomer substrate requirement of ZDS in evolutionarily distant plants, suggest that Z-ISO activity is not unique to maize, but will be found in all higher plants. Further analysis of this new gene-controlled step is critical to understanding regulation of this essential biosynthetic pathway.


Introduction
Carotenoids are a diverse group of more than 750 naturally occurring pigments found in plants, fungi and bacteria (Britton et al., 2004). In higher plants, carotenoids serve as accessory pigments in photosynthesis and as photoprotectors at high light intensities. Apocarotenoids, carotenoid degradative products, are signals in plant development and in stress responses; their roles include extracellular rhizosphere signals that attract beneficial fungi and damaging parasitic plants that have opposite effects on plant yield (Milborrow, 2001;Bouvier et al., 2003;Booker et al., 2004;Schwartz et al., 2004;Simkin et al., 2004;Castillo et al., 2005;Matusova et al., 2005;Moise et al., 2005;Nambara and Marion-Poll, 2005). Increasing interest has also been placed on carotenoid content and composition of food crops, because of the manifold importance of carotenoids in human health (Fraser and Bramley, 2004;Wurtzel, 2004;Quinlan et al., 2007) .
In plant cells, carotenoids are synthesized in plastids from isoprenoid precursors through reactions catalyzed by nuclear encoded enzymes (DellaPenna and Pogson, 2006). The first committed step, mediated by phytoene synthase, is condensation of two molecules of geranylgeranyl pyrophosphate to form 15-cis phytoene, containing a central 15-15' cis double bond (Beyer et al., 1985;Dogbo et al., 1988;Misawa et al., 1994). A four-step enzymatic desaturation of 15-cis phytoene to all-trans-lycopene also requires an electron transport chain (Mayer et al., 1990). Lycopene cyclases then catalyze ring formation at both ends of all-trans lycopene to form carotenes which can be further hydroxylated to produce xanthophylls (Kim and DellaPenna, 2006;Quinlan et al., 2007).
In bacteria, four desaturation steps from 15-cis phytoene to all-trans lycopene are mediated by a single enzyme, CrtI (Linden et al., 1991). In contrast, plants employ two desaturases, phytoene desaturase (PDS), which forms trans double bonds at 11 and 11', concomitant with cis bond formation at existing 9 and 9' double bonds, while ζ-carotene desaturase (ZDS) forms cis double bonds at 7 and 7' (Breitenbach and Sandmann, 2005) (Fig. 1). In addition, plant desaturation steps require supplementary isomerization reactions to produce acceptable geometrical isomer substrates for the desaturases and for the following lycopene cyclization steps (Beyer et al., 1989). Such differences in isomerization capacities of the plant and bacterial desaturases are important considerations for metabolic engineering of carotenoid content in food crops and in influencing biological activities of plant-derived geometrical isomers, including Li, Murillo, and Wurtzel intestinal absorption and localization (Krinsky et al., 1990;Osterlie et al., 1999;Bjerkeng and Berge, 2000;Holloway et al., 2000;Patrick, 2000).
There has been confusion in the literature regarding the required number of carotene isomerases needed in plant carotenoid biosynthesis, especially since reports of cloning of "the carotene isomerase gene" encoding "CRTISO", from cyanobacteria and plants (Breitenbach et al., 2001;Masamoto et al., 2001;Isaacson et al., 2002;Park et al., 2002). Early in vitro studies using daffodil chromoplasts (Beyer et al., 1989) suggested that further progression in the plant carotenoid biosynthetic pathway beyond ζ-carotene (e.g. ζ-carotene desaturation and onwards), required an isomerase activity and only the 15-cis position was recognized as an isomerase target. Expression of Arabidopsis PDS and ZDS in E. coli also revealed a missing plant factor required for ζ-carotene desaturation that could be replaced by photoisomerization; ζ-carotene accumulated instead of the predicted prolycopene, unless cells were exposed to light (Bartley et al., 1999). In study of the coupled maize desaturases, it was proposed that there was a need for either two companion isomerases or a single isomerase having multiple substrates (Matthews et al., 2003). In an effort to identify the "missing isomerase gene," research groups looked at a number of pathway mutants that exhibited accumulation of atypical geometrical isomers. The tomato tangerine locus, for which recessive alleles condition accumulation of prolycopene (poly cis lycopene), led to cloning of the corresponding gene that encoded what was thought to be the long sought-after "Carotene Isomerase," and it was so named as CRTISO (Isaacson et al., 2002).
When CRTISO (Isaacson et al., 2004) was assayed, it was found that isomerase activity followed, rather than preceded, ζ-carotene desaturation which was inconsistent with earlier biochemical data indicating isomerization was needed prior to ζ-carotene desaturation; CRTISO was shown to be specific for adjacent double bonds at 7, 9 and 7 ',9' positions needed to convert prolycopene to all-trans lycopene, but it did not isomerize the single 15-15' cis double bond that ZDS substrate (Breitenbach and Sandmann, 2005). The in vitro assays of CRTISO (Isaacson et al., 2004), no longer supported the possibility of a multi-substrate enzyme, but suggested that indeed a second isomerase might exist. While light might compensate for this requirement in light-exposed tissue, there was still the question of whether any genetic locus was responsible for such activity in the dark or in dark-grown tissue. In considering the phenotype associated with a lesion affecting such a second isomerase, which we will call Z-ISO (15-cis-ζ-carotene isomerase), it seemed that mutants would be predicted to accumulate the PDS product (and Z-ISO substrate), 9,15,9'-tri-cis-ζ-carotene, which could be shown to photoisomerize to 9,9'-di-cisζ-carotene. In comparison, ZDS mutants should accumulate 9,9'-di-cis-ζ-carotene in a genetic background conditioning functional Z-ISO. Unlike ZDS mutants that are lethal because they accumulate ζ-carotene under both light and dark conditions, Z-ISO mutants should be nonlethal because the mutant phenotype only manifests in the dark and not in light-grown tissues.
Therefore, ZDS and Z-ISO mutants should both condition accumulation of ζ-carotene in the dark but only ZDS mutants will show accumulation in light grown tissues and exhibit an albino phenotype. We therefore searched for ζ-carotene accumulating mutants in maize (Wurtzel, 2004) to test for presence of the predicted Z-ISO geometrical isomer substrate that is distinguishable from the ZDS substrate. We identified the maize y9 locus as a candidate and demonstrated that recessive alleles condition accumulation of 9,15,9'-tri-cis-ζ-carotene in "dark" tissues. This finding supports the role of an additional genetic locus in plants that controls an upstream carotene isomerase activity, Z-ISO, which is independent of CRTISO. Comparison of the y9 phenotype with that of ζ-carotene accumulating mutants in other species shows that Z-ISO activity is not limited to maize.

Identification of a maize candidate Z-ISO mutant
Numerous mutations affecting maize carotenoid biosynthesis, including phytoene desaturation, have been reported (Wurtzel, 2004). Endosperm phenotypes include white endosperm (compared to the yellow) and vivipary caused by an absence of abscisic acid, a carotenoid cleavage product promoting seed dormancy. Absence of leaf carotenoids cause lethality and manifest as albino tissue due to pleiotropic effects on chloroplast biogenesis.
Pathway blocks are also associated with accumulation of pathway intermediates. Accumulation of ζ-carotene was predicted for mutations in genes controlling either ZDS or a putative Z-ISO; such accumulation was found for maize vp9 and y9 mutant endosperms (Robertson, 1975a).
These mutants had otherwise dissimilar phenotypes; y9 homozygous mutants were nonlethal recessives affecting only endosperm and leaves remained green, while vp9 mutants were lethal recessives affecting both endosperm and leaf tissues and plants were albino. When the maize ZDS gene was isolated (Matthews et al., 2003), it was mapped to vp9 on chromosome 7 and not to y9 on chromosome 10 (Robertson, 1975b); also, y9 had no affect on ZDS transcript accumulation, indicating that y9 did not encode ZDS or control ZDS transcript levels. Given that a block in isomerase activity would be light-complemented, we predicted that a carotenoid isomerase mutant should have an almost normal leaf phenotype as seen for y9 and not an albino phenotype as in vp9. Therefore, it was conceivable that y9 might encode CRTISO, although it did not accumulate prolycopene as did other CRTISO mutants (Isaacson et al., 2002).
Furthermore, mapping of CRTISO to chromosomes 2 and 4 (Wurtzel, 2004) (Conrad, Brutnell, and Wurtzel, unpublished) indicated that y9 did not encode CRTISO, because y9 mapped to chromosome 10. While y9 was not associated with ZDS or CRTISO, it remained a good candidate to consider for a possible genetic locus that might encode a factor necessary for putative Z-ISO activity. To validate that y9 was necessary for Z-ISO activity (and to further prove that light was insufficient and that a genetic locus was involved), it was necessary to demonstrate that y9 conditioned accumulation specifically of 9,15,9'-tri-cis-ζ-carotene in contrast to accumulation of 9,9'-di-cis-ζ-carotene in a ZDS mutant, such as vp9. Furthermore, while prior reports suggested that y9 had no leaf phenotype (Janick-Buckner et al., 2001), we predicted that light was photoisomerizing the accumulated 9,15,9'-tri-cis-ζ-carotene intermediate to relieve the pathway block such that plants appeared "green." Therefore, it was necessary to demonstrate that Li, Murillo, and Wurtzel in etiolated leaf tissue ζ-carotene accumulated because of the absence of photoisomerization. In contrast, photoisomerization would have no effect on the phenotype of a ZDS mutant.

ζ-carotene isomers in etiolated leaf and root tissues of y9 and vp9 mutants
If the y9 locus indeed controls expression of a new isomerase, then a lesion in the gene should manifest as 9,15,9'-tri-cis-ζ-carotene accumulation in dark grown plants, when light is not available to compensate for absence of the cis-trans conversion; y9 leaves of light-grown plants have been reported not to accumulate ζ-carotene (Janick-Buckner et al., 2001). Therefore, we germinated plants in the dark and collected etiolated leaves and roots, extracted carotenoids and analyzed by HPLC (Fig. 3, Table 1). In both etiolated leaves and roots of homozygous y9 plants, Li, Murillo, and Wurtzel 9,15,9'-tri-cis-ζ-carotene accumulated; no 9,9'-di-cis-ζ-carotene was detected nor were there any other downstream xanthophylls. This suggests that the carotenoid biosynthetic pathway in darkgrown y9 is completely blocked at this step, thereby preventing downstream synthesis leading to xanthophylls typically found in either light-grown y9 or in dark-grown maize seedlings carrying the dominant Y9 allele (data not shown). In comparison, homozygous vp9 roots and etiolated leaves accumulated mainly 9,9'-di-cis-ζ-carotene. No xanthophylls were observed in carotenoid extracts from etiolated leaves or roots of either mutant as compared to the profile in y9 endosperm. The absence of downstream carotenoids in dark-grown tissues further supports the hypothesis that the xanthophylls detected in y9 endosperm were due to some photoconversion in the field-grown plants. As predicted for a Z-ISO lesion, 9,15,9'-tri-cis-ζ-carotene was found to accumulate in dark-grown tissues of y9 plants. These results establish that y9 does interfere with carotenoid biosynthesis in the absence of light in leaves and roots; the only reason that y9 plants are normally green is that light compensates for the lesion.

Identification of other putative genetic loci needed for Z-ISO activity
Specific details on how Table 2 was developed are provided (see Methods). Nine mutants were examined, all of which exhibit significant accumulation of ζ-carotene (see % ζ-carotene isomers), and some of which have already been associated with a structural locus for ZDS or CRTISO. To distinguish between ZDS and isomerase mutants, we grouped the mutants into two classes, light non-responsive (ZDS mutants), where light has no effect on the phenotype, or light-responsive (isomerase mutants), where light can reverse the phenotype by compensation for a genetic lesion.
The first class included maize vp9 and sunflower nondormant-1 (nd-1); mutants in this class have been shown to be light-nonresponders and therefore albino, which is typical for ZDS lesions based on phenotype and in these cases, supported by gene analysis (Matthews et al., 2003;Conti et al., 2004). Members of this group show a low ratio of tri-cis /di-cis ζ-carotene isomers and accumulate 9,9'-di-cis-ζ-carotene, the ZDS substrate.
The second class of mutants have been shown to be light-responders; photoisomerization releases a pathway block and hence normal carotenoid accumulation in the light can be observed (e.g. in the case of plants, they are green and viable). This class could alternatively represent either CRTISO or Z-ISO isomerase mutants. To distinguish between the two isomerase mutant types, we predicted that dark-grown CRTISO mutants should accumulate prolycopene or proneurosporene because there is no block in tri-cis to di-cis isomerization, as evidenced by a lower tri-cis to di-cis ζ-carotene isomer ratio. On the other hand, dark-grown Z-ISO mutants will not accumulate prolycopene (or proneurosporene), but instead accumulate a high ratio of tri-cis to di-cis ζ-carotene isomers due to the block in the cis to trans conversion which is otherwise released by photoisomerization.
PG1 is included in this class because it has a close to normal pigment composition in light-grown cells, indicating it carries a mutation in an isomerase because it can be light complemented; were it a lesion in ZDS, it would exhibit ζ-carotene accumulation even in light-grown cells. The low tri-cis to di-cis ζ-carotene isomer ratio found in dark-grown PG1 cells also suggests that the mutation blocks CRTISO and not Z-ISO activity, which would otherwise have given a high ratio.
The Z-ISO class of light responders included the Euglena gracillis mutant W 3 BUL (Cunningham and Schiff, 1985) and maize y9. As expected for a block in Z-ISO activity, neither mutant accumulated prolycopene nor proneurosporene; both mutants exhibited a high ratio of tricis to di-cis ζ-carotene isomers, for which the tri-cis isomer was found to be photoconvertible to the di-cis isomer. This biochemical evidence suggests that both W 3 BUL and y9 loci affect Z-ISO and not CRTISO activity, and that Z-ISO activity is not limited to plants but is also present in a photosynthetic protist.

Discussion
We demonstrated that isomerization of 9,15,9'-tri-cis-ζ-carotene to 9,9'-di-cis-ζ-carotene is not simply light-mediated, but that it requires the product of a nuclear-encoded gene. We showed that in the absence of light, the 9,15,9'-tri-cis-ζ-carotene isomer accumulates in etiolated leaves and roots of maize plants carrying the recessive y9 allele, in comparison to a normal carotenoid composition in light-exposed y9 leaves (Janick-Buckner et al., 2001). Therefore, an activity, which we termed Z-ISO, is essential in dark-exposed tissues. However, even in lightexposed y9 tissues, there is evidence that Z-ISO activity is needed. Some striping seen in light-not entirely efficient in overcoming the pathway lesion associated with the recessive y9 allele and that Z-ISO activity may be required in photosynthetic tissues when plants are subjected to abiotic stress.
The Z-ISO activity, which we demonstrated to be genetically controlled in maize, is not unique to this species, as we deduced by examining a collection of ζ-carotene accumulating mutants in multiple species. By applying a standard convention for naming the ζ-carotene isomers, we showed that the mutants fell into three classes of ζ-carotene accumulating mutants: ZDS, CRTISO and Z-ISO. The ZDS class contains two plant genes, the sunflower nd-1 and maize vp9 loci; CRTISO includes mutants from plants, a cyanobacterium and a green alga; the Z-ISO mutants include those in maize (y9) and Euglena. Therefore, unlike bacteria that encode one four-step desaturase, plants and other photosynthetic organisms possess two desaturases (PDS and ZDS) and two isomerases (Z-ISO and CRTISO) which have alternating functions in the biosynthetic pathway.
Z-ISO functions upstream of CRTISO and we ruled out y9 as encoding CRTISO, although we don't as yet know whether y9 encodes Z-ISO or regulates its expression. It is unlikely that y9 encodes a factor that alters CRTISO activity, because if it did, then CRTISO mutants would not accumulate prolycopene, but instead would accumulate 9,15,9'-tri-cis-ζcarotene.
We named the isomerase activity that is upstream of CRTISO, Z-ISO, to reflect isomerization of the 15-cis double bond in ζ-carotene. We don't know if Z-ISO will act on 15cis bonds only in ζ-carotene, or also on 15-cis phytoene, or the intermediate 9,15-cis phytofluene.
It is unlikely that 15-cis phytoene is a substrate, as the primary phytoene isomer that was detected in plants is the 15-cis isomer and not trans (Jungalwalab and Porter, 1965). The y9 mutant phenotype also does not support 15-cis phytoene as being a substrate as it is the tri-cis-ζcarotene isomer that predominates in accumulation and not 15-cis phytoene that would accumulate were it the preferred substrate. Therefore, perhaps Z-ISO activity requires the unique conformation of the 15-cis double bond in ζ-carotene which also has the 9,9' cis double bonds, uniquely produced during PDS desaturation. If one looks at the structure of 9,15,9'-tri-cis ζcarotene, its conformation is quite distinct from that of 15-cis phytoene, which may be significant in substrate binding and recognition of the 15-cis double bond. Li, Murillo, and Wurtzel The naming of CRTISO, was optimistic at the time of its discovery and suggested that it catalyzed all carotene isomerizations, which we know now not to be the case. If one looks further at the isomerase substrates it becomes clear why CRTISO was not the sole carotene isomerase.
CRTISO isomerizes adjacent cis double bonds (Isaacson et al., 2004), whereas the Z-ISO substrate is a single cis double bond, making for a very different substrate structure. Therefore it is likely that Z-ISO is different from CRTISO in terms of protein structure which may explain why there has been no success in showing Z-ISO activity for any of the plant CRTISO homologs.
Characterization of the maize y9 locus has brought to light a new factor required in plant carotenoid biosynthesis. Future characterization of this locus will lead to isolation of all of the components needed for plant desaturation/isomerization. Biogenesis and regulation of the complex plant desaturase/isomerase metabolon is an important and future problem to address; elucidation will have direct impact on metabolic engineering of this pathway in plants.

Plant and bacterial materials
Maize ζ-carotene desaturase mutant vp9-Bot100 (Maize Co-op 706B) and mutant y9 (Maize Co-op X07C) were used in this study. For etiolated leaves and roots, seeds were geminated and grown in the dark for two weeks at 25 0 C. Endosperms were collected at 20 days after pollination from field grown plants as described (Gallagher et al., 2004). The plasmid pACCRT-EBP, carrying bacterial crtE, crtB and maize Pds genes confers accumulation of ζcarotene isomers in E. coli and was used to produce ζ-carotene standards for HPLC (Matthews et al., 2003); transformed cells were incubated for 48 hours at 37 0 C under darkness or in the light (50 μmol m -2 s -1 photon flux) prior to pigment extraction as described below.

Carotenoid Extraction
Carotenoids were extracted from approximately 100 mL bacterial culture or 1 g endosperm, etiolated leaf or root tissues taken from dark-grown plants as described (Kurilich and Juvik, 1999), dissolved in 1 mL methanol, and 50 μl injected for HPLC analysis.
Values shown are averages and standard deviation for 3 samples.