Accumulation of Carotenoids and Expression of Carotenoid ... · Accumulation of Carotenoids and Expression of Carotenoid Biosynthetic Genes during Maturation in ... Method A was optimized

Accumulation of Carotenoids and Expression ofCarotenoid Biosynthetic Genes during Maturation inCitrus Fruit1

Masaya Kato*, Yoshinori Ikoma, Hikaru Matsumoto, Minoru Sugiura, Hiroshi Hyodo, andMasamichi Yano

Department of Citrus Research, National Institute of Fruit Tree Science, Shimizu-okitsunakacho, Shizuoka424–0292, Japan (M.K., Y.I., H.M., M.S., M.Y.); and Department of Biological Sciences, Faculty of Agriculture,Shizuoka University, 836 Ohya, Shizuoka 422–8529, Japan (H.H.)

The relationship between carotenoid accumulation and the expression of carotenoid biosynthetic genes during fruitmaturation was investigated in three citrus varieties, Satsuma mandarin (Citrus unshiu Marc.), Valencia orange (Citrussinensis Osbeck), and Lisbon lemon (Citrus limon Burm.f.). We cloned the cDNAs for phytoene synthase (CitPSY), phytoenedesaturase (CitPDS), �-carotene (car) desaturase (CitZDS), carotenoid isomerase (CitCRTISO), lycopene �-cyclase (CitLCYb),�-ring hydroxylase (CitHYb), zeaxanthin (zea) epoxidase (CitZEP), and lycopene �-cyclase (CitLCYe) from Satsuma manda-rin, which shared high identities in nucleotide sequences with Valencia orange, Lisbon lemon, and other plant species. Withthe transition of peel color from green to orange, the change from �,�-carotenoid (�-car and lutein) accumulation to�,�-carotenoid (�-car, �-cryptoxanthin, zea, and violaxanthin) accumulation was observed in the flavedos of Satsumamandarin and Valencia orange, accompanying the disappearance of CitLCYe transcripts and the increase in CitLCYbtranscripts. Even in green fruit, high levels of �,�-carotenoids and CitLCYe transcripts were not observed in the juice sacs.As fruit maturation progressed in Satsuma mandarin and Valencia orange, a simultaneous increase in the expression ofgenes (CitPSY, CitPDS, CitZDS, CitLCYb, CitHYb, and CitZEP) led to massive �,�-xanthophyll (�-cryptoxanthin, zea, andviolaxanthin) accumulation in both the flavedo and juice sacs. The gene expression of CitCRTISO was kept low or decreasedin the flavedo during massive �,�-xanthophyll accumulation. In the flavedo of Lisbon lemon and Satsuma mandarin,massive accumulation of phytoene was observed with a decrease in the transcript level for CitPDS. Thus, the carotenoidaccumulation during citrus fruit maturation was highly regulated by the coordination of the expression among carotenoidbiosynthetic genes. In this paper, the mechanism leading to diversity in �,�-xanthophyll compositions between Satsumamandarin and Valencia orange was also discussed on the basis of the substrate specificity of �-ring hydroxylase and thebalance of expression between upstream synthesis genes (CitPSY, CitPDS, CitZDS, and CitLCYb) and downstream synthesisgenes (CitHYb and CitZEP).

Carotenoids are essential components of the pho-tosynthetic apparatus in plants, algae, and cyanobac-teria, in which they protect against photooxidativedamage and contribute to light harvesting for photo-synthesis (Goodwin, 1980). In higher plants, thebright yellow, orange, and red colors provided bycarotenoids accumulate in the chromoplasts of flow-ers and fruits. In these tissues, plants exploit carote-noids as colorants to attract pollinators and agents ofseed dispersal. In addition, epoxy-carotenoids, vio-laxanthin, and neoxanthin are precursors for planthormone abscisic acid (Rock and Zeevaart, 1991).Some carotenoids serve as precursors for vitamin A,which is essential to human and animal diets, and as

antioxidants, which play a role in reducing the risk ofcertain forms of cancer (Olson, 1989).

Citrus is a complex source of carotenoids, with thelargest number of carotenoids found in any fruit(Gross, 1987). Carotenoid concentration and compo-sition vary greatly among citrus varieties and dependon the growing conditions (Gross, 1987). During cit-rus fruit development, massive accumulation of caro-tenoids occurred concomitantly with the degradationof chlorophyll. Mandarin varieties, such as Satsumamandarin (Citrus unshiu Marc.), accumulated�-cryptoxanthin (�-cry) predominantly in the fla-vedo and juice sacs in mature fruit (Goodner et al.,2001; Ikoma et al., 2001). In contrast, mature sweetorange (Citrus sinensis Osbeck) accumulated violax-anthin isomers predominantly in fruit (Molnar andSzabolcs, 1980; Lee and Castle, 2001), in which 9-cis-violaxanthin was found to be the principal carot-enoid (Molnar and Szabolcs, 1980). Because �-crywas detected to a minor extent in sweet orange va-rieties, the difference in �-cry concentration can beused as a discriminating factor among mandarin,

1 This work was supported by the Research and DevelopmentProgram for New Bio-industry Initiative of the Bio-oriented Tech-nology Research Advancement Institution.

* Corresponding author; e-mail [email protected]; fax81–543– 69 –2115.

Article, publication date, and citation information can be foundat http://www.plantphysiol.org/cgi/doi/10.1104/pp.103.031104.

824 Plant Physiology, February 2004, Vol. 134, pp. 824–837, www.plantphysiol.org © 2004 American Society of Plant Biologists www.plantphysiol.orgon July 8, 2018 - Published by Downloaded from Copyright © 2004 American Society of Plant Biologists. All rights reserved.

http://www.plantphysiol.org

orange, and their hybrids (Goodner et al., 2001).Mature lemon (Citrus limon Burm.f.) showed lightyellow color in the flavedo and juice sacs. This lightcoloration was primarily because of a small concen-tration of total carotenoids, which was much lowerthan that in navel orange (Citrus sinensis Osbeck;Yokoyama and Vandercook, 1967). Thus, Satsumamandarin (variety accumulating �-cry), Valencia or-ange (variety accumulating violaxanthin), and Lis-bon lemon (variety accumulating small amounts ofcarotenoids) are useful experimental materials toinvestigate the molecular mechanism regulating ca-rotenoid concentration and composition in fruit be-cause the carotenoid profiles in mature fruit weremuch more diverse among these three citruses.These materials are especially useful for undertand-ing the mechanism of xanthophyll accumulation infruit because massive xanthophyll accumulationdoes not occur in common experimental materials,such as tomato (Lycopersicon esculentum) and Arabi-dopsis.

The pathway of carotenoid biosynthesis in plants isillustrated in Figure 1 (Cunningham and Gantt, 1998;Ronen et al., 1999; Isaacson et al., 2002; Park et al.,2002). The first committed step in carotenoid biosyn-thesis is a head-to-head condensation of two mole-cules of geranylgeranyl pyrophosphate (C20) to formcolorless phytoene (phy; C40) catalyzed by phytoenesynthase (PSY). Phytoene desaturase (PDS) and

�-carotene (car) desaturase (ZDS) introduce four dou-ble bonds into phy to yield lycopene (lyc) via phyt-ofluene, and neurosporene. Recently, Park et al.(2002) and Isaacson et al. (2002) isolated the geneencoding carotenoid isomerase (CRTISO) from Ara-bidopsis and tomato, respectively, by map-basedcloning. CRTISO functions as the isomerization ofpoly-cis-carotenoids to all-trans-carotenoids. The cy-clization of lyc is a crucial branching point in thispathway, yielding �-car with one �-ring and one�-ring and �-car with two �-rings, in which twocyclases, namely, lycopene �-cyclase (LCYb) and ly-copene �-cyclase (LCYe), are responsible for thesereactions (Cunningham et al., 1996). �-Car is con-verted into lutein (lut) by sequential hydroxylations,which are catalyzed by �-ring hydroxylase and �-ringhydroxylase (HYb). �-Car is converted to zeaxanthin(zea) via �-cry by two-step hydroxylation, which iscatalyzed by HYb. Furthermore, zea is converted toviolaxanthin via antheraxanthin by zea epoxidase(ZEP).

Carotenoid biosynthesis and its regulation havebeen studied in various plant species, such as Arabi-dopsis (Pogson et al., 1996; Park et al., 2002), tomato(Giuliano et al., 1993; Fraser et al., 1994; Ronen et al.,1999; Isaacson et al., 2002), pepper (Capsicum annuum;Bouvier et al., 1996, 1998), tobacco (Nicotiana tabacum;Busch et al., 2002), and alga (Steinbrenner and Lin-den, 2001). Bramley (2002) reviewed carotenoid

Figure 1. Carotenoid biosynthetic pathway inplants. Eight cDNAs, CitPSY, CitPDS, CitZDS,CitCRTISO, CitLCYb, CitHYb, CitZEP, and CitL-CYe, were cloned from the Satsuma mandarinflavedo and used for RNA probes for northern-blot analyses in this study. GGPP, Geranylgera-nyl pyrophosphate.

Carotenoid Biosynthesis in Citrus Fruit

Plant Physiol. Vol. 134, 2004 825 www.plantphysiol.orgon July 8, 2018 - Published by Downloaded from Copyright © 2004 American Society of Plant Biologists. All rights reserved.


biosynthesis and regulation during ripening anddevelopment in tomato fruit. During tomato fruitripening, the expression of PSY and PDS increased(Giuliano et al., 1993; Fraser et al., 1994; Ronen et al.,1999; Isaacson et al., 2002), whereas the expressionof both LCYb and LCYe disappeared (Pecker et al.,1996; Ronen et al., 1999), leading to massive accu-mulation of lyc. Genes involved in carotenoid bio-synthesis were induced by various environmentalstimuli. It has been reported that PSY was inducedby light in tomato seedlings (Giuliano et al., 1993).Moreover, oxidative stress induced chromoplast-specific carotenoid genes in pepper fruit (Bouvier etal., 1996, 1998). In green alga, high light and saltstress elicited the gene expression of PSY and carot-enoid hydroxylase, resulting in a rapid accumula-tion of astaxanthin (Steinbrenner and Linden, 2001).

Also in citrus fruit, the gene expression of somecarotenoid biosynthetic enzymes was investigatedpreviously. These investigations showed that, in Sat-suma mandarin, the gene expression of PSY in-creased in the peel and juice sacs with the onset ofcoloration (Ikoma et al., 2001; Kim et al., 2001),whereas the gene expression of PDS and HYb re-mained constant once fruit was fully developed (Kimet al., 2001; Kita et al., 2001). To elucidate the regu-lation steps for carotenoid synthesis in citrus fruit,these investigations were insufficient because at leastsix genes (PSY, PDS, ZDS, LCYb, HYb, and ZEP)participate in xanthophyll accumulation, which oc-curs massively in Satsuma mandarin and Valenciaorange. Therefore, in the present study, the expres-sion of genes, including the six genes mentionedabove, was investigated simultaneously.

In this study, during fruit maturation, the concen-tration and composition of carotenoid and the ex-pression of carotenoid biosynthetic genes were inves-tigated in the flavedos and juice sacs of threevarieties, Satsuma mandarin, Valencia orange, andLisbon lemon. The cDNAs related to linear car bio-synthesis (PSY, PDS, and ZDS), the cDNA related tothe isomerization of poly-cis-carotenoids (CRTISO),the cDNAs related to cyclization (LCYb and LCYe),and the cDNAs related to hydroxylation and epoxi-dation (HYb and ZEP) were cloned. The expressionof these genes was analyzed during fruit maturationin the three varieties. The results were the first oftheir kind, to our knowledge, to show the simulta-neous expression of genes participating in the syn-thesis of xanthophylls in citrus fruit. This study wasalso the first, to our knowledge, to provide informa-tion on differences among the three varieties in theprofiles of gene expression of carotenoid biosyntheticenzymes. On the basis of the comparison of theseprofiles, the mechanism causing the diversity of ca-rotenoid accumulation in citrus fruit was discussed.The mechanism of xanthophyll accumulation in fruit,especially �-cry and violaxanthin accumulation, wasalso discussed.

RESULTS

Identification of Carotenoids

To identify and quantify carotenoids, three differ-ent gradient elution schedules of HPLC, methods Ato C, were used. Method A was optimized for all-trans-violaxanthin (t-vio), cis-violaxanthin (c-vio),lut, phy, �-cry, and �-car analyses. Method B wasoptimized to separate �-car and �-car peaks becausethe peak of �-car overlapped with that of �-car bymethod A. Method C was optimized for zea analysisbecause the peak of zea overlapped with that of anunknown carotenoid by method A (Table I).

Peaks 1, 3, 5 to 7, and 9 were identified as t-vio, lut,�-cry, �-car, �-car, and zea, respectively, by compar-ing their absorption spectra and retention times withthose of purchased authentic standards (Table I). Al-though the corresponding authentic standard wasobtained, lyc was not detected in our experimentalsamples.

We prepared the standard for phy. The acetoneextract obtained from phy-producing E. coli cells wasseparated by HPLC (method A). The peak eluted at56 min was isolated. The mass spectrum of the eluentshowed the molecular ion at mass-to-charge ratio(m/z) 545 ([M � H]�). The eluent exhibited the typ-ical absorption spectrum of phy. Thus, we used theeluent as a phy standard. Peak 4 was identified asphy by comparing its absorption spectrum and re-tention time with those of the phy standard (Table I).

Also, we prepared the standard for �-car. The ace-tone extract obtained from �-car-producing E. colicells was separated by HPLC (method B). The peakeluted at 12 min was isolated. The mass spectrumof the eluent showed the molecular ion at m/z 541([M � H]�). The eluent exhibited the typical absorp-tion spectrum of �-car. Thus, we used the eluent as a�-car standard. Peak 8 was identified as �-car bycomparing its absorption spectrum and retentiontime with those of the �-car standard (Table I).

The standard for c-vio was prepared from the fla-vedo of Satsuma mandarin fruit. The crude extract ofcarotenoids from the tissue was separated by HPLC(method A). The peak eluted at 31 min was isolated.The absorption maxima of the eluent (414, 436, and464 nm) were close to those of c-vio reported previ-ously (Tai and Chen, 2000; Table I). The epoxide testindicated that the eluent was 5,6,5�,6�-diepoxide be-cause the absorption maxima shifted from 414, 436,and 464 nm to 380, 401, and 427 nm, respectively,after the addition of HCl. The mass spectrum of theeluent showed the molecular ion at m/z 601 ([M �H]�). These results indicated that the eluent wasc-vio. Therefore, the eluent was used as a c-vio stan-dard. Peak 2 was identified as c-vio by comparing itsabsorption spectrum and retention time with those ofthe c-vio standard (Table I).

Kato et al.

826 Plant Physiol. Vol. 134, 2004 www.plantphysiol.orgon July 8, 2018 - Published by Downloaded from Copyright © 2004 American Society of Plant Biologists. All rights reserved.


Isolation and Identification of the cDNA Fragments ofCarotenoid Biosynthetic Genes

On the basis of the conserved amino acid sequencesamong plant species in carotenoid biosyntheticgenes, eight sets of degenerated primers were de-signed for each of PSY, PDS, ZDS, CRTISO, LCYb,HYb, ZEP, and LCYe (Table II). Reversetranscriptase-PCRs were performed with the flave-dos of Satsuma mandarin, Valencia orange, and Lis-bon lemon. In the case of Satsuma mandarin, for eachcarotenoid biosynthetic gene, we sequenced at least12 cDNAs among which the nucleotide sequenceswere completely identical except for the primer re-gions. Thus, a cDNA was selected for each carotenoidbiosynthetic gene.

Eight cDNA fragments from Satsuma mandarin forcarotenoid biosynthetic genes were designated asCitPSY (accession no. AB114648), CitPDS (accessionno. AB114649), CitZDS (accession no. AB114650), Cit-CRTISO (accession no. AB114651), CitLCYb (acces-sion no. AB114652), CitHYb (accession no. AB114653),CitZEP (accession no. AB114654), and CitLCYe (ac-cession no. AB114655; Table II). The nucleotide se-quences of all cDNAs isolated from Satsuma manda-rin showed high identity (�97.4% at nucleotidesequence level) to those of corresponding cDNAsisolated from Valencia orange (accession nos.AB114656–AB114663) and Lisbon lemon (accessionnos. AB114664–AB114671; Table II). Because theidentity of cDNAs among the three citrus varieties

was very high, the RNA probes were synthesizedfrom Satsuma mandarin cDNA for northern-blotanalyses, which were used to compare gene expres-sion among the three varieties. These high identitiessuggested that the differences among varieties in thehybridizing intensity of the probe to mRNA werenegligible. The cDNAs isolated from Satsuma man-darin also shared high similarities with other citrusspecies, such as Citrus � paradisi and Citrus maxima(�97.6% at the nucleotide sequence level; data notshown) and non-citrus species (�73.4% at the nucle-otide sequence level; Table II).

Changes in the Concentration of Carotenoid andExpression of Carotenoid BiosyntheticGenes in the Flavedo

The color of the flavedo changed from green toorange during fruit maturation. The green stages inSatsuma mandarin, Valencia orange, and Lisbonlemon were from August to September, from Augustto October, and from August to October,respectively.

During the green stage, �-car (�8.4 �g g�1), t-vio(�4.3 �g g�1), �-car (�5.0 �g g�1), and lut (�14.8 �gg�1) were predominant, although the amounts ofthese carotenoids were low in the three varieties (Fig.2). Phy, �-car, �-cry, zea, and c-vio were barely de-tected in the three varieties. During this stage, highgene expression of CitLCYe and CitCRTISO and low

Table I. Identification of carotenoids found in citrus fruit

Peaks Retention Time Absorption MaximaWave Length forQuantification

Carotenoids

min nm

Method APeak 1 17.7 418, 440, 470 452 All-trans-violaxanthinPeak 2 30.8 414, 436, 464 452 Cis-violaxanthinPeak 3 43.6 (422)a, 446, 474 452 LuteinPeak 4 55.6 (276), 288, (300) 286 PhytoenePeak 5 57.6 (426), 452, 478 452 �-CryptoxanthinPeak 6 64.1 (424), 448, 476 452 �-CaroteneStandard Ib 17.4 418, 440, 470 452 All-trans-violaxanthinStandard IIc 30.9 414, 436, 464 452 Cis-violaxanthinStandard IIIb 43.8 (422), 446, 474 452 LuteinStandard IVd 55.8 (276), 288, (300) 286 PhytoeneStandard Vb 57.9 (426), 452, 478 452 �-CryptoxanthinStandard VIb 64.3 (424), 448, 476 452 �-Carotene

Method BPeak 7 12.2 (427), 454, 478 453 �-CarotenePeak 8 12.6 382, 402, 426 400 �-CaroteneStandard VIIb 12.3 (427), 454, 478 453 �-CaroteneStandard VIIId 12.8 382, 402, 426 400 �-Carotene

Method CPeak 9 24.7 (424), 452, 478 452 ZeaxanthinStandard IXb 24.7 (424), 452, 478 452 Zeaxanthin

a Values in parentheses represent shoulder. Methods A to C represent three different gradient elution schedules of HPLC. Details of theschedules were described in “Materials and Methods.” b Purchased standard. c Standard prepared from Satsuma mandarin fla-vedo. d Standard prepared from Escherichia coli cells.




gene expression of CitPSY, CitZDS, and CitHYb wereobserved in the three varieties (Fig. 3). The geneexpression increased noticeably in CitLCYb of Sat-suma mandarin and Valencia orange and in CitZEPof the three varieties. The gene expression of CitPDSincreased rapidly in Satsuma mandarin but slowly inValencia orange and Lisbon lemon.

After the green stage, the concentration of �-car,�-car, and lut decreased or remained constant at alow level with a concomitant decrease in the geneexpression of CitLCYe in the three varieties (Figs. 2and 3). In contrast, �,�-xanthophylls (�-cry, zea,t-vio, and c-vio) accumulated massively in Satsumamandarin and Valencia orange. In Satsuma manda-rin, �-cry and c-vio became abundant (in January,48.7 �g g�1 and 48.1 �g g�1, respectively). In Valen-cia orange, c-vio became abundant (in February, 50.7�g g�1). With the transition of the peel color fromgreen to orange, the gene expression of CitPSY, Cit-PDS, CitZDS, CitLCYb, CitHYb, and CitZEP, whichmake up a necessary set of genes to produce �,�-xanthophylls, increased to maximum levels or re-mained high in Satsuma mandarin and Valencia or-ange. In Lisbon lemon, the concentration of �,�-xanthophylls remained low, although the expressionof a gene set to produce �,�-xanthophylls increasedslightly or remained at near a maximum level withthe transition of the peel color. The increased levelsof gene expression in Satsuma mandarin were higherthan those in Valencia orange and Lisbon lemon.

After �,�-xanthophylls increased, massive accumu-lation of phy started in Satsuma mandarin (Decem-ber–January). The concentration of phy (110.7 �gg�1) was much higher than that of �-car (6.8 �g g�1)

in January, when phy became the most abundantcarotenoid (Fig. 2). In Lisbon lemon, instead of mas-sive accumulation of �,�-xanthophylls, accumulationof phy was observed (December–February). Phy be-came the most abundant carotenoid in February,when the phy concentration (77.2 �g g�1) was 36.6-fold of �-car. In Valencia orange, the concentration ofphy was much lower than those in Satsuma manda-rin and Lisbon lemon even in the latter maturationstage (January–February). The gene expression ofCitPSY remained at near maximum levels in Satsumamandarin and Lisbon lemon during the periods inwhich phy increased (Fig. 3). In contrast, the geneexpression of CitPDS clearly decreased in Satsumamandarin and Lisbon lemon during this period. Thegene expression of CitCRTISO was kept low ordecreased in the three varieties, whereas �,�-xanthophylls and phy were accumulating.

Changes in the Concentration of Carotenoid andExpression of Carotenoid Biosynthetic Genes in theJuice Sacs

During the green stage of the flavedo (August–September in Satsuma mandarin and August–October in Valencia orange and Lisbon lemon), theconcentration of total carotenoids was low in thejuice sacs in the three varieties (Fig. 4). However, inSatsuma mandarin, the concentration of �-cry in thejuice sacs increased (5.1 �g g�1 in September) andwas much higher than that in the flavedo (16.5-foldin September). In Valencia orange and Lisbonlemon, no noticeable accumulation of �-cry wasdetected during this stage (undetectable level in

Table II. Primers for reverse transcriptase-PCR, lengths of amplified products, and comparison of nucleotide sequences of carotenoid biosyn-thetic genes from Satsuma mandarin with those from Valencia orange, Lisbon lemon, and non-citrus plant species

cDNA (Accession No.)Sense Primers (Upper) andAntisense Primers (Lower)

LengthIdentity on the Level of Nucleotide Sequencea (Species, Accession No.)

Citrus species Non-citrus species

bp %

CitPSY (AB114648) ATTCAGCCATTYMGAGAYATG 503 100.0 (Valencia orange, AB114656) 91.0 (Cucumis melo, Z37543)RAARTTGTTGTARTCATTSGC 98.5 (Lisbon lemon, AB114664) 90.3 (Arabidopsis, AY056287)

CitPDS (AB114649) CAAGRGATGTTCTWGGTGGA 545 99.2 (Valencia orange, AB114657) 93.5 (Tageles erecta, AF251014)RTTRCCATCYAARAAKGCCAT 99.6 (Lisbon lemon, AB114665) 92.9 (Narcissus pseudonarcissus,

X78815)CitZDS (AB114650) RGGHCAYGAGGTGGATATA 612 97.4 (Valencia orange, AB114658) 93.7 (Pepper, X89897)

ATDGGMCCACTCAARTAAAC 99.5 (Lisbon lemon, AB114666) 91.1 (N. pseudonarcissus, AJ224683)CitCRTISO (AB114651) AATGCTACMMGMTGGGATAC 568 98.9 (Valencia orange, AB114659) 84.6 (Tomato, AF416727)

WACTATCSCCMACRCARTATA 99.8 (Lisbon lemon, AB114667) 83.4 (Arabidopsis, AC011001)CitLCYb (AB114652) TATGGTGTTTGGGTGGATGA 747 99.4 (Valencia orange, AB114660) 89.8 (Pepper, X86221)

DGTCCTTGCYACCATATARC 100.0 (Lisbon lemon, AB114668) 89.4 (A. palaestina, AF321534)CitHYb (AB114653) ARRTCVGARMGRTYYACTTAT 568 99.8 (Valencia orange, AB114661) 92.0 (Pepper, Y09722)

CTTCTTCHAVYTCCTTDGGT 98.7 (Lisbon lemon, AB114669) 88.6 (T. erecta, AF251018)CitZEP (AB114654) ATWCARATACARAGYAAYGCWT 984 99.5 (Valencia orange, AB114662) 81.8 (Prunus armeniaca, AF159948)

RCCHARRTAWGCYTTGTAAGT 99.4 (Lisbon lemon, AB114670) 81.5 (Tomato, Z83835)CitLCYe (AB114655) YACRAAYAAYTAYGGYGTDTG 651 99.8 (Valencia orange, AB114663) 77.2 (A. palaestina, AF321535)

TAHGACCAYTCCTCYTCRTA 99.5 (Lisbon lemon, AB114671) 73.4 (Solanum tuberosum,AF321537)

a Identity between cDNAs was estimated with the exception of nucleotide sequences for primer regions.

Kato et al.



Valencia orange and 0.2 �g g�1 in Lisbon lemon inSeptember). No gene expression of CitLCYe wasdetected in the three varieties (Fig. 5). In Satsumamandarin, accompanying the accumulation of �-cry,the gene expression of CitPSY, CitPDS, CitZDS,CitLCYb, CitHYb, and CitZEP, which make up a setof genes to produce �,�-xanthophylls, increasedsimultaneously. In Valencia orange, the gene ex-pression of CitPDS, CitHYb, and CitZEP increasedclearly, whereas no noticeable increase in thegene expression of CitPSY, CitZDS, and CitLCYbwas observed. No noticeable increase in the gene

expression of CitPSY, CitZDS, and CitLCYb wasobserved in Lisbon lemon either. The geneexpression of CitCRTISO increased in the threevarieties.

After the green stage, massive accumulation ofcarotenoids, especially the accumulation of �,�-xanthophylls, occurred in Satsuma mandarin and Va-lencia orange (Fig. 4). In January, Satsuma mandarinaccumulated predominantly a �,�-xanthophyll, �-cry(15.9 �g g�1), which accounted for 59.6% of the totalidentified carotenoids. Other �,�-xanthophylls, t-vio(4.7%) and c-vio (4.4%), were at low concentrations

Figure 2. Carotenoid concentration in the flave-dos of three citrus varieties, Satsuma mandarin,Valencia orange, and Lisbon lemon, during fruitmaturation. Columns and bars represent themeans and SE (n � 3), respectively. Total car,Total carotenoids. The value for total carote-noids was the sum of identified carotenoids(phy, �-car, �-car, �-cry, zea, t-vio, c-vio, �-car,and lut).




during this month. In February, Valencia orange ac-cumulated predominantly t-vio and c-vio (2.4 and 9.6�g g�1, respectively), which accounted for 65.4% ofthe total identified carotenoids. The concentration ofthe other �,�-xanthophyll, �-cry (15.1%), was low inFebruary. In Lisbon lemon, the concentration of caro-tenoids was extremely low, although phy, �-car, and�-cry accumulated slightly (0.2, 0.1, and 0.4 �g g�1 inFebruary, respectively). Clearly, the gene expressionof a set of genes to produce �,�-xanthophylls

(CitPSY, CitPDS, CitZDS, CitLCYb, CitHYb, andCitZEP) increased, reaching a maximum after thegreen stage in Satsuma mandarin and Valencia or-ange (Fig. 5). In Lisbon lemon, the gene expression ofa set of genes to produce �,�-xanthophylls also in-creased. However, the increased levels of gene ex-pression were lower than those in Satsuma mandarinand Valencia orange. The gene expression of Cit-CRTISO increased to a maximum and subsequentlydecreased in the three varieties after the green stages.

Figure 3. Expressions of carotenoid biosyn-thetic genes in the flavedos of three citrus vari-eties, Satsuma mandarin, Valencia orange, andLisbon lemon, during fruit maturation. A,Chemiluminescent images of northern-blotanalyses. B, Quantification of mRNA abun-dance from the images. Total RNAs (10 �g perlane) from Satsuma mandarin, Valencia orange,and Lisbon lemon were separated on a gel. TheRNA on the gel was stained with ethidium bro-mide. The gel was blotted onto a nylon mem-brane. Eight blots were prepared and hybridizedwith the RNA probes for CitPSY, CitPDS,CitZDS, CitCRTISO, CitLCYb, CitHYb, CitZEP,and CitLCYe, respectively. Hybridization signalsare shown as chemiluminescent images ac-quired using a CCD camera (A). The bottompanel of A is a representative from eightethidium bromide-stained gels showing rRNA(28S). The abundance of mRNA was quantifiedfrom the chemiluminescent images by densi-tometry, normalized to rRNA (28S) in each lane,and expressed as arbitrary units for pixel inten-sities (B). A, August; S, September; O, October;N, November; D, December; J, January; F,February.

Kato et al.



Xanthophyll Composition and Gene ExpressionRelated to car and �,�-Xanthophyll Syntheses inSatsuma Mandarin and Valencia Orange Fruits

During 2 months of sampling just after the greenstage (October and November for Satsuma mandarinand November and December for Valencia orange),�,�-xanthophylls accumulated massively in both theflavedo and juice sacs, which showed a considerabledifference in xanthophyll composition. In the fla-vedo, the ratios of the predominant xanthophylls,�-cry to violaxanthin (t-vio � c-vio), were 0.43 inSatsuma mandarin (average of October and Novem-

ber) and 0.11 in Valencia orange (average of Novem-ber and December). In the juice sacs, the ratios were5.17 in Satsuma mandarin (average of October andNovember) and 0.10 in Valencia orange (average ofNovember and December).

To compare gene expression related to the synthesesof car and �,�-xanthophyll between Satsuma manda-rin and Valencia orange during a period of massive�,�-xanthophyll accumulation (October and Novem-ber for Satsuma mandarin and November and Decem-ber for Valencia orange), the transcript levels forCitPSY, CitPDS, CitZDS, CitLCYb, CitHYb, and CitZEP

Figure 4. Carotenoid concentration in the juicesacs of three citrus varieties, Satsuma mandarin,Valencia orange, and Lisbon lemon, during fruitmaturation. Columns and bars represent themeans and SE (n � 3), respectively. Total car,Total carotenoids. The value for total carote-noids was the sum of identified carotenoids(phy, �-car, �-car, �-cry, zea, t-vio, c-vio, �-car,and lut).




during these 2 months were averaged from the resultsof Figures 3B and 5B (Fig. 6). In the flavedo, the levelsfor transcripts involved in car synthesis (CitPSY, Cit-PDS, CitZDS, and CitLCYb) and �,�-xanthophyll syn-thesis (CitHYb and CitZEP) were higher in Satsumamandarin than in Valencia orange. In the juice sacs,the levels for CitPSY, CitPDS, CitZDS, and CitLCYbtranscripts involved in the biosynthesis from phy to�-car were higher in Satsuma mandarin. In contrast, thelevels for CitHYb and CitZEP transcripts involved in the

biosynthesis from �-cry to violaxanthin were muchhigher in Valencia orange than in Satsuma mandarin.

DISCUSSION

Carotenoid Accumulation and GeneExpression during the Green Stage

In the flavedo of the green stage, the low geneexpression of CitPSY and CitZDS, which produce

Figure 5. Expressions of carotenoid biosyn-thetic genes in the juice sacs of three citrusvarieties, Satsuma mandarin, Valencia orange,and Lisbon lemon, during fruit maturation. A,Chemiluminescent images of northern-blotanalyses. B, Quantification of mRNA abun-dance from the images. Total RNAs (10 �g perlane) from Satsuma mandarin, Valencia orange,and Lisbon lemon were separated on a gel. TheRNA on the gel was stained with ethidium bro-mide. The gel was blotted onto a nylon mem-brane. Eight blots were prepared and hybridizedwith the RNA probes for CitPSY, CitPDS,CitZDS, CitCRTISO, CitLCYb, CitHYb, CitZEP,and CitLCYe, respectively. Hybridization signalsare shown as chemiluminescent images ac-quired using a CCD camera (A). The bottompanel of A is a representative from eightethidium bromide-stained gels showing rRNA(28S). The abundance of mRNA was quantifiedfrom the chemiluminescent images by densi-tometry, normalized to rRNA (28S) in each lane,and expressed as arbitrary units for pixel inten-sities (B). A, August; S, September; O, October;N, November; D, December; J, January; F,February.

Kato et al.



linear cars, was responsible for the low concentrationof carotenoids in Satsuma mandarin, Valencia or-ange, and Lisbon lemon. It seems that a rapid in-crease in the gene expression of CitPDS of Satsumamandarin did not lead to an increase in the concen-tration of carotenoids in this stage because the ex-pression of the upstream gene CitPSY remained lowin this variety (Fig. 3). In contrast, in the juice sacs ofSatsuma mandarin, simultaneous increases in the ex-pression of the CitPSY, CitPDS, CitZDS, CitLCYb,CitHYb, and CitZEP genes, which make up a set ofgenes to produce �,�-xanthophylls, led to the accu-mulation of �-cry in this stage (Fig. 5). In juice sacs ofValencia orange and Lisbon lemon, no noticeableaccumulation of �,�-xanthophylls was detected dur-ing this stage because the gene expression of CitPSY,CitZDS, and CitLCYb did not increase noticeably.

In the flavedos of the three varieties, the geneexpression of CitLCYe in the green stage was higherthan that in the orange stage (Fig. 3). The cyclizationof lyc is a key branch point in the pathway of carot-enoid biosynthesis in plants and algae, in which thelycopene cyclases, LCYe and LCYb, are key enzymes(Cunningham et al., 1996). LCYe adds only one �-ringto form the monocyclic �-car, leading to the synthesisof �,�-carotenoids (�-car and lut) after LCYb intro-duces one �-ring to �-car (Ronen et al., 1999; Fig. 1),whereas LCYb introduces two �-rings, leading to thesynthesis of �,�-carotenoids (�-car, �-cry, zea, andviolaxanthin). In the flavedo, the high expression ofthe CitLCYe gene suggested that cyclization to the�-ring was more active than that in the orange stage,resulting in the predominant accumulation of �,�-carotenoids in the green stage. Thus, it was thoughtthat the pathway changing from �,�-carotenoid syn-thesis to �,�-carotenoid synthesis did not occur in theflavedo in the green stage. In the juice sacs, no tran-scripts for CitLCYe in the green and orange stageswere detected in the three varieties (Fig. 5). Ikoma etal. (2001) reported that the chlorophyll content in thejuice sacs of Satsuma mandarin was high in June,subsequently decreasing to undetectable levels inAugust. These results suggested that, in the juicesacs, �,�-carotenoid synthesis occurred before Au-gust because the LCYe gene was expressed exclu-sively in chloroplast-containing photosynthetic tis-sues (Ronen et al., 1999). Thus, it was thought that thepathway changing from �,�-carotenoid synthesis to�,�-carotenoid synthesis occurred in the juice sacsearlier than in the flavedo.

Xanthophyll Accumulation and GeneExpression during the Orange Stage

After the green stage of the flavedos in the threevarieties, the increase in the gene expression of CitL-CYb and decrease in the gene expression of CitLCYesuggested that pathway changing from �,�-carotenoid synthesis to �,�-carotenoid synthesis oc-

curred in the flavedo with the transition from thegreen stage to the orange stage (Fig. 3).

During the orange stage, simultaneous increases inthe expression of genes to participate in �,�-xanthophyll synthesis (CitPSY, CitPDS, CitZDS, CitL-CYb, CitHYb, and CitZEP) led to the massive accu-mulation of �,�-xanthophylls in the flavedos andjuice sacs of Satsuma mandarin and Valencia orange(Figs. 3 and 5). In Lisbon lemon, the gene expressionof a set of genes to produce �,�-xanthophylls alsoincreased. However, the levels of gene expression inLisbon lemon were much lower than those in Sat-suma mandarin and Valencia orange. Low gene ex-pression may lead to an extremely low concentrationof �,�-carotenoids in the flavedo and juice sacs ofLisbon lemon.

These gene expression profiles in Satsuma manda-rin and Valencia orange were different from those intomato observed previously during fruit ripening.Previous studies showed that, in tomato fruit accu-mulating lyc, the gene expression of PSY and PDSincreased and that of LCYe decreased at the breakerstage of ripening (Pecker et al., 1996; Giuliano et al.,1993; Ronen et al., 1999). These gene expression pro-files of tomato fruit were consistent with those ofcitrus fruit observed in the present study. In contrast,the expression profiles of the LCYb gene were differ-ent between tomato and citrus. The transcripts for

Figure 6. Comparisons of mRNA abundance for carotenoid biosyn-thetic genes between Satsuma mandarin and Valencia orange in theflavedos and juice sacs during massive xanthophyll accumulations(October and November for Satsuma mandarin and November andDecember for Valencia orange). The mRNA abundance was ex-pressed as arbitrary units for pixel intensities. The values for Satsumamandarin were averaged from those of October and November inFigures 3B and 5B. The values for Valencia orange were averagedfrom those of November and December in Figures 3B and 5B.




LCYb in the tomato fruit disappeared at this stage(Pecker et al., 1996; Ronen et al., 1999), althoughthose for LCYb obviously increased in Satsuma man-darin and Valencia orange (Figs. 3 and 5). This dif-ference in the gene expression profile of LCYb seemsto be the primary determinant of the difference in thecarotenoid profile between tomato and citrus fruit. Intomato fruit, the disappearances in the gene expres-sion of LCYe and LCYb led to massive accumulationof a final upstream product, lyc. In contrast, in Sat-suma mandarin and Valencia orange fruit, disap-pearances in the gene expression of LCYe and in-creases in the gene expression of LCYb led to themassive accumulation of �,�-xanthophylls. �,�-Xanthophyll accumulation, as observed in citrus, alsowas reported in tomato petal, in which the expres-sion of the LCYb gene was up-regulated, whereas theLCYe gene was not expressed, resulting in the accu-mulations of �,�-xanthophylls, such as violaxanthinand neoxanthin (Ronen et al., 1999). Recently, Ronenet al. (2000) identified another type of LCYb gene,Beta, by map-based cloning from the tomato mu-tants. The gene was expressed exclusively in thechromoplast-containing tissues of flowers and fruits.In citrus fruit, whether the gene plays an importantrole in �-car formation is unknown.

Isaacson et al. (2002) found a gene, CRTISO, whichencoded an authentic CRTISO that is required forcarotenoid desaturation. In the report, the gene wasexpressed in all green tissues, but it was up-regulatedduring fruit ripening and in the flowers of tomato. Incitrus, the gene expression of CitCRTISO remainedlow or decreased in the flavedo during the massive�,�-xanthophyll accumulation (Fig. 3). Thus, it is notclear whether the gene expression of CitCRTISO con-tributed to massive �,�-xanthophyll accumulation incitrus.

phy Accumulation and Gene Expression during theOrange Stage

The accumulation of upstream carotenoids, such asphytofluene and �-car, was reported previously inlemon (Yokoyama and Vandercook, 1967; Gross,1987). Recently, a mutant Pinalate orange, which hasa distinct yellow fruit instead of the typical orange(Citrus sinensis Osbeck), was characterized (Rodrigoet al., 2003). In this mutant, linear cars (phy, phyt-ofluene, and �-car) had accumulated. In lemon fruitof the present study, we observed a high level of phyrather than phytofluene, which was high in previousreports (Yokoyama and Vandercook, 1967; Gross,1987; Fig. 2). The result of carotenoid analysis sug-gested that lack of action in carotenoid desaturasesled to the massive accumulation of phy and a limitedaccumulation of �,�-xanthophylls in Lisbon lemon.Actually, the transcript levels for CitPDS decreasedremarkably in February in Lisbon lemon (Fig. 3).Massive accumulation of phy was also observed in

the flavedo of Satsuma mandarin after the accumu-lation of �,�-xanthophyll with a decrease in tran-script levels for CitPDS (in December and January;Figs. 2 and 3). In these varieties, the accumulation ofphy could be explained primarily by the levels ofgene expression of CitPDS. phy accumulation wasalso reported in genetically transformed tobaccoplants. Busch et al. (2002) reported that the accumu-lation of 15-cis-phy was detected in tobacco plantsexpressing antisense RNA to PDS. This result sup-ports the present speculation that decreases in thegene expression of PDS lead to accumulation of phy.

Previously, Al-Babili et al. (1996) showed that PDSwas regulated posttranscriptionally in the chromo-plasts of N. pseudonarcissus. In the present study, thesame mechanism that is related to the posttranscrip-tional regulation of PDS also may be responsible forthe massive accumulation of phy.

Mechanism Leading to Diversity in XanthophyllCompositions between Satsuma Mandarin andValencia Orange

A clear difference in the �,�-xanthophyll composi-tion between Satsuma mandarin and Valencia orangefruits was observed in the juice sacs. The juice sacs ofSatsuma mandarin accumulated �-cry as a major ca-rotenoid in mature fruit, whereas those of Valenciaorange mainly accumulated violaxanthin isomers(Molnar and Szabolcs, 1980; Goodner et al., 2001;Ikoma et al., 2001; Lee and Castle, 2001; Fig. 4).

We isolated cDNAs encoding complete coding re-gions for HYb from Satsuma mandarin and Valenciaorange (M. Kato and Y. Ikoma, unpublished data).The deduced amino acid sequences between thesecDNAs were identical except for one amino acidresidue, which was located in the transit peptide.This result suggested that the amino acid sequencesbetween these cDNAs were completely identical inthe regions related to enzyme activity. Thus, wethought that the difference in the amino acid se-quences between these cDNAs was not responsiblefor the difference in the �,�-xanthophyll compositionbetween Satsuma mandarin and Valencia orange.

A previous study demonstrated that �-cry insteadof zea was mainly accumulated in E. coli cells carry-ing the truncated Arabidopsis HYb gene (Sun et al.,1996). The report speculated that HYb hydroxylatedthe �-rings of �-car with greater efficiency than thenot-yet hydroxylated �-ring of �-cry. Because of thehigh substrate specificity of HYb to �-car, HYbwould prefer the first step conversion from �-car to�-cry rather than the second step conversion from�-cry to zea under excessive �-car supply and/or lowHYb activity.

In citrus fruits, the substrate specificity of HYb andexpression balance between upstream synthesisgenes (CitPSY, CitPDS, CitZDS, and CitLCYb) anddownstream synthesis genes (CitHYb and CitZEP)

Kato et al.



seem important to determine the ratios of �-cry/violaxanthin. The higher expression of upstream syn-thesis genes and lower expression of the HYb genesuggested a higher supply of �-car and lower activityof CitHYb in the juice sacs of Satsuma mandarin thanin those of Valencia orange (Fig. 6). The increasedlevel of �-car in the juice sacs of Satsuma mandarinalso suggested a higher supply of �-car and loweractivity of CitHYb than in those of Valencia orange(Fig. 4). Therefore, it is thought that, under a balanceof high gene expression of upstream synthesis andlow gene expression of CitHYb (high supply of �-carand low activity of CitHYb), CitHYb catalyzes pre-dominantly the first step conversion by the high sub-strate specificity of HYb to �-car, leading to massiveaccumulation of �-cry. Moreover, because �-car wasrarely converted to zea via �-cry, and the gene ex-pression of CitZEP was lower in the juice sacs ofSatsuma mandarin than in those of Valencia orange,the accumulation of violaxanthin may be retarded inthis tissue (Fig. 6). In contrast, in the juice sacs ofValencia orange, CitHYb is likely to sufficiently cat-alyze the reaction to zea via �-cry because lower geneexpression of upstream synthesis and higher geneexpression of CitHYb (low supply of �-car and highactivity of CitHYb) were observed in the juice sacs ofValencia orange than in those of Satsuma mandarin(Fig. 6). Moreover, the intensity of gene expressionrelated to epoxidation, CitZEP, were much higher inthe juice sacs of Valencia orange than in those ofSatsuma mandarin (Fig. 6). Thus, zea would be rap-idly converted into violaxanthin in the juice sacs ofValencia orange.

In the flavedo, the gene expression involved incarotenoid biosynthesis (both upstream and down-stream syntheses) was much higher in Satsuma man-darin than in Valencia orange (Fig. 6). The varietaldifference in expression balance between upstreamsynthesis genes and downstream synthesis geneswas much smaller in the flavedo than in the juicesacs. Thus, a small difference between Satsuma man-darin and Valencia orange in the expression balanceof the genes resulted in a small difference in the�,�-xanthophyll composition in the flavedo.

CONCLUSION

In this paper, we investigated the relationship be-tween carotenoid accumulation and the expression ofcarotenoid biosynthetic genes during fruit matura-tion in three citrus varieties. Our results clearlyshowed that pathway changing from �,�-carotenoidsto �,�-carotenoid synthesis was caused in the flave-dos of Satsuma mandarin and Valencia orange by thedisappearance of transcripts for CitLCYe and the in-crease in transcripts for CitLCYb with the transitionof peel color from green to orange. In the juice sacs,pathway changing seems earlier than in flavedo be-cause transcripts for CitLCYe were not detected even

in green fruit. As fruit maturation progressed, inSatsuma mandarin and Valencia orange, a simulta-neous increase in the expression of genes (CitPSY,CitPDS, CitZDS, CitLCYb, CitHYb, and CitZEP) led tomassive �,�-xanthophyll accumulation in both theflavedo and juice sacs. In the flavedo of Lisbon lemonand Satsuma mandarin, massive accumulation ofphy was observed with a decrease in the transcriptlevel for CitPDS. The mechanism leading to diversityin �,�-xanthophyll compositions between Satsumamandarin and Valencia orange also was discussed.The substrate specificity of HYb and expression bal-ance between upstream synthesis genes (CitPSY, Cit-PDS, CitZDS, and CitLCYb) and downstream synthe-sis genes (CitHYb and CitZEP) seem responsible forthe considerable difference in the �,�-xanthophyllcompositions of the juice sacs between Satsuma man-darin and Valencia orange. Thus, the carotenoid ac-cumulation during citrus fruit maturation was highlyregulated by the coordination among the expressionof the carotenoid biosynthetic genes.

MATERIALS AND METHODS

Plant Materials

Satsuma mandarin (Citrus unshiu Marc.), Valencia orange (Citrus sinensisOsbeck), and Lisbon lemon (Citrus limon Burm.f.) cultivated at the NationalInstitute of Fruit Tree Science, Department of Citrus Research, Okitsu (Shi-zuoka, Japan) were used as materials. Fruit samples were collected period-ically from August to January for Satsuma mandarin and from August toFebruary for Valencia orange and Lisbon lemon. The flavedos and juice sacswere separated from sampled fruits, immediately frozen in liquid nitrogen,and kept at �80°C until use.

Carotenoid Identification in Citrus Fruit

�-car, lut, and t-vio were obtained from DHI Water and Environment(Horsholm, Denmark). �-car and zea were obtained from Extrasynthese(Genay, France). �-cry was obtained from Sokenkagaku (Tokyo). Thesecarotenoids were used for authentic standards.

A standard for phy was prepared from phy-producing Escherichia coli BL21 (DE3) cells carrying plasmid pACCRT-EB (Misawa et al., 1995). A stan-dard for �-car was also prepared from �-car-producing E. coli BL 21 (DE3)cells carrying plasmids pACCRT-EB and pET 30 (Novagen, Darmstadt,Germany), which harbors coding region of PDS cDNA from Satsuma man-darin (accession no. AB046992). These cells were grown in a Luria-Bertanimedium at 27°C to OD600 � 0.6 and induced with 1 mm isopropyl-�-d-thiogalactoside for 24 h at 27°C. For the preparation of phy and �-car, cellswere pelleted, frozen in liquid nitrogen, and suspended in acetone. After theacetone extract was partitioned into a diethyl ether phase, the phase wasrecovered and evaporated to dryness. Subsequently, the residue was dis-solved in a methyl tert-butyl ether (MTBE):methanol (1:1 [v/v]) solution.Phy and �-car were separated by HPLC using methods A and B, respec-tively. Details of methods A and B are presented in the next section. For theidentification of eluents, the absorption spectrum (220–550 nm) was moni-tored by a photodiode array detector (MD-910, Jasco, Tokyo) on an HPLCsystem. Moreover, fast atom bombardment mass spectrometry analysis wasperformed with a JMS-700 mass spectrometer (JEOL, Tokyo).

The standard of c-vio was prepared from the flavedo of Satsuma man-darin. The details of the carotenoid extraction method are presented in thenext section. Extract was separated by HPLC (method A). For the identifi-cation of the eluent, the absorption spectrum and the mass spectrum wereanalyzed as mentioned above. Moreover, an epoxide test was performed todetect the epoxy groups of carotenoids. The spectrum was recorded beforeand after the addition of 50 �L of 0.1 n HCl into a 100-�L carotenoidsolution dissolved in ethanol. A shift of about 20 nm to the shorter wave-




lengths indicated the presence of a monoepoxy group, whereas a shift ofapproximately 40 nm was indicative of a diepoxy (Davies, 1976).

Carotenoid Quantification in Citrus Fruit

For the analysis of carotenoid contents, samples were homogenized in40% (v/v) methanol containing 10% (w/v) magnesium carbonate basic.Pigments were extracted from the residues using an acetone:methanol (7:3[v/v]) solution containing 0.1% (w/v) 2,6-di-tert-butyl-4-methylphenol andpartitioned into diethyl ether. The extracts containing carotenoids esterifiedto fatty acids were saponified with 20% (w/v) methanolic KOH. After thesaponification, water-soluble extracts were removed from the extract byadding NaCl-saturated water. The pigments repartitioned into the diethyl-ether phase were recovered and evaporated to dryness. Subsequently, theresidue was redissolved in 5 mL of an MTBE:methanol (1:1 [v/v]) solution.

An aliquot (20 �L) was separated by a reverse-phase HPLC system(Jasco) fitted with a YMC Carotenoid S-5 column of 250- � 4.6-mm-i.d.(Waters, Milford, MA) at a flow rate of 1 mL min�1. The eluent wasmonitored by a photodiode array detector (MD-910, Jasco). The sample wasanalyzed by three different gradient elution schedules. To assay t-vio, c-vio,lut, �-cry, and �-car, the gradient elution schedule consisted of an initial 30min of 95% (v/v) methanol, 1% (v/v) MTBE, and 4% (v/v) water followedby a linear gradient of 6% (v/v) methanol, 90% (v/v) MTBE, and 4% (v/v)water for 60 min (method A). To assay �-car and �-car, the initial solventcomposition consisted of 50% (v/v) methanol, 46% (v/v) MTBE, and 4%(v/v) water followed by a linear gradient of 6% (v/v) methanol, 90% (v/v)MTBE, and 4% (v/v) water for 60 min (method B). zea was assayed by thegradient elution schedule of Rouseff and Raley (1996). The initial composi-tion consisted of 90% (v/v) methanol, 5% (v/v) MTBE, and 5% (v/v) waterfollowed by a linear gradient of 95% (v/v) methanol and 5% (v/v) MTBE for12 min, 86% (v/v) methanol, and 14% (v/v) MTBE for 8 min, 75% (v/v)methanol and 25% (v/v) MTBE for 10 min, and 50% (v/v) methanol and50% (v/v) MTBE for 20 min (method C).

The peaks were identified by comparing their specific retention times andabsorption spectra with the authentic standards. Standard curves for thecarotenoid quantification were prepared with those of the authentic stan-dards at 286 nm for phy, 400 nm for �-car, 452 nm for t-vio, c-vio, lut, �-cry,�-car, and zea, and 453 nm for �-car. The carotenoid concentration wasestimated by the standard curves and expressed as milligrams per gramfresh weight. Carotenoid quantification was performed in three replicates.

Isolation and Sequence Analysis of Citrus CarotenoidBiosynthetic Genes

Total RNA was extracted from the green and orange flavedos accordingto the method described by Ikoma et al. (1996). The first strand cDNA wassynthesized from 5 �g of the total RNA with the Ready-To-Go T-PrimedFirst Strand Kit (Amersham Bioscience, Little Chalfont, UK). CitPSY, Cit-PDS, CitZDS, CitCRTISO, CitLCYb, CitHYb, CitZEP, and CitLCYe were am-plified by PCR with the primers for these cDNAs designed by commonsequences that have been reported (Table II). The amplified cDNAs werecloned with a TA Cloning Kit (Invitrogen, San Diego). The sequences weredetermined using the Taq Dye Primer Cycle Sequencing Kit (Perkin ElmerApplied Biosystems, Foster City, CA).

Total RNA Extraction and Northern-Blot Hybridization

Total RNA was extracted from the flavedos and juice sacs of Satsumamandarin, Valencia orange, and Lisbon lemon fruits at different growingstages according to the method described by Ikoma et al. (1996). Aliquots oftotal RNA (10 �g) were separated on 1.2% (w/v) agarose-denaturing form-aldehyde gels containing 20 mm MOPS (pH 7.0), 0.5 mm Na-acetate, and 1mm EDTA. After electrophoresis for 2 h, the RNA was visualized withethidium bromide under UV light to ensure equal loading of RNA in eachlane. The RNA was transferred to nylon membranes (Roche DiagnosticsGmbH, Mannheim, Germany) with 20� SSC, and the blots were baked for3 h at 80°C.

CitPSY, CitPDS, CitZDS, CitCRTISO, CitLCYb, CitHYb, CitZEP, and CitL-CYe were labeled with a DIG RNA labeling kit (Roche Diagnostics GmbH)using T7 or SP6 RNA polymerase to synthesize the RNA probes. The blotswere prehybridized at 68°C in a solution containing 5� SSC, 50% (v/v)

formamide, 0.1% (w/v) N-lauroylsarcosine, 0.02% (w/v) SDS, and 2% (w/v)blocking regent (Roche Diagnostics GmbH) for 3 h. Hybridization wasperformed at 68°C in the same solution overnight. After the hybridization,the blots were washed twice at room temperature in 2� SSC and 0.1% (w/v)SDS for 5 min and then washed twice at 68°C in 0.1� SSC and 0.1% (w/v)SDS for 15 min. After the equilibration in buffer A (0.1 m maleic acid and0.15 mm NaCl [pH 7.5]) for 5 min at room temperature, the blots wereblocked with a 2% (w/v) blocking regent (Roche Diagnostics GmbH) inbuffer A for 30 min. Subsequently, the blots were incubated with Anti-Digoxygenin-AP Fab fragments (Roche Diagnostics GmbH) in the blockingbuffer for 30 min. The blots were washed four times for 8 min each in bufferA containing 0.3% (v/v) Tween 20 and equilibrated with a 0.1 m Tris-HClbuffer (pH 9.5) containing 0.1 m NaCl for 5 min. Chemiluminescence reac-tions were carried out with CDP-Star (Roche Diagnostics GmbH), a chemi-luminescent substrate. Chemiluminescent images of the blots were acquiredusing a CCD camera (Night Owl LB 981, Berthold Technologies, Bad Wild-bad, Germany). The pixel intensities of the bands on the chemiluminescentimages were estimated by subtracting the background signals with thesoftware analyzing the band intensity (Phoretix 1D Gel Analysis Software,Nonlinear Dynamics, Newcastle upon Tyne, UK).

ACKNOWLEDGMENT

We are grateful to Dr. Norihiko Misawa (Marine Biotechnology Institute,Iwate, Japan) for the gift of plasmids.

Received July 3, 2003; returned for revision September 7, 2003; acceptedOctober 28, 2003.

LITERATURE CITED

Al-Babili S, Lintig J, Haubruck H, Beyer P (1996) A novel, soluble form ofphytoene desaturase from Narcissus pseudonarcissus chromoplasts isHsp70-complexed and competent for flavinylation, membrane associa-tion, and enzymatic activation. Plant J 9: 601–612

Bouvier F, Backhaus RA, Camara B (1998) Induction and control ofchromoplast-specific carotenoid genes by oxidative stress. J Biol Chem273: 30651–30659

Bouvier F, d’Harlingue A, Hugueney P, Marin E, Marion-Poll A, CamaraB (1996) Xanthophyll biosynthesis: cloning, expression, functional recon-stitution, and regulation of �-cyclohexenyl carotenoid epoxidase frompepper (Capsicum annuum). J Biol Chem 271: 28861–28867

Bramley PM (2002) Regulation of carotenoid formation during tomato fruitripening and development. J Exp Bot 53: 2107–2113

Busch M, Seuter A, Hain R (2002) Functional analysis of the early steps ofcarotenoid biosynthesis in tobacco. Plant Physiol 128: 439–453

Cunningham FJ, Gantt E (1998) Genes and enzymes of carotenoid biosyn-thesis in plants. Annu Rev Plant Physiol Plant Mol Biol 49: 557–583

Cunningham FX, Pogson B, Sun Z, McDonald KA, DellaPenna D, Gantt E(1996) Functional analysis of the � and � lycopene cyclase enzymes ofArabidopsis reveals a mechanism for control of cyclic carotenoid forma-tion. Plant Cell 8: 1613–1626

Davies BH (1976) Carotenoids. In Goodwin TW, ed, Chemistry and Bio-chemistry of Plant Pigments, Vol 2. Academic Press, London, pp 38–165

Fraser PD, Truesdale MR, Bird CR, Schuch W, Bramley PM (1994) Carot-enoid biosynthesis during tomato fruit development. Plant Physiol 105:405–413

Giuliano G, Bartley GE, Scolnik PA (1993) Regulation of carotenoid bio-synthesis during tomato development. Plant Cell 5: 379–387

Goodner KL, Rouseff RL, Hofsommer HJ (2001) Orange, mandarin, andhybrid classification using multivariate statistics based on carotenoidprofiles. J Agric Food Chem 49: 1146–1150

Goodwin TW (1980) The Biochemistry of the Carotenoids, Ed 2, Vol 1.Chapman and Hall, London

Gross J (1987) Carotenoids: Pigments in Fruits. Academic Press, LondonIkoma Y, Komatsu A, Kita M, Ogawa K, Omura M, Yano M, Moriguchi T

(2001) Expression of a phytoene synthase gene and characteristic carot-enoid accumulation during citrus fruit development. Physiol Plant 111:232–238

Ikoma Y, Yano M, Ogawa K, Yoshioka T, Xu ZC, Hisada S, Omura M,

Kato et al.



Moriguchi T (1996) Isolation and evaluation of RNA frompolysaccharide-rich tissues in fruit for quality by cDNA library construc-tion and RT-PCR. J Jpn Soc Hortic Sci 64: 809–814

Isaacson T, Ronen G, Zamir D, Hirschberg J (2002) Cloning of tangerinefrom tomato reveals a carotenoid isomerase essential for the productionof �-carotene and xanthophylls in plants. Plant Cell 14: 333–342

Kim IJ, Ko KC, Kim CS, Chung WI (2001) Isolation and characterization ofcDNAs encoding �-carotene hydroxylase in Citrus. Plant Sci 161:1005–1010

Kita M, Komatsu A, Omura M, Yano M, Ikoma Y, Moriguchi T (2001)Cloning and expression of CitPDS1, a gene encoding phytoene desatu-rase in Citrus. Biosci Biotechnol Biochem 65: 1424–1428

Lee HS, Castle WS (2001) Seasonal changes of carotenoid pigments andcolor in Hamlin, Earlygold, and Budd Blood orange juices. J Agric FoodChem 49: 877–882

Misawa N, Satomi Y, Kondo K, Yokoyama A, Kajiwara S, Saito T, OhtaniT, Miki W (1995) Structure and functional analysis of a marine bacterialcarotenoid biosynthesis gene cluster and astaxanthin biosynthetic path-way proposed at the gene level. J Bacteriol 177: 6575–6584

Molnar P, Szabolcs J (1980) �-Citraurin epoxide, a new carotenoid fromValencia orange peel. Phytochemistry 19: 633–637

Olson JA (1989) Provitamin-A function of carotenoids: the conversion of�-carotene into vitamin-A. J Nutr 119: 105–108

Park H, Kreunen S, Cuttriss AJ, DellaPenna D, Pogson BJ (2002) Identifi-cation of the carotenoid isomerase provides insight into carotenoid bio-synthesis, prolamellar body formation, and photomorphogenesis. PlantCell 14: 321–332

Pecker I, Gabbay R, Cunningham FX, Hirschberg J (1996) Cloning andcharacterization of the cDNA for lycopene �-cyclase from tomato revealsdecrease in its expression during fruit ripening. Plant Mol Biol 30:807–819

Pogson B, McDonald KA, Truong M, Britton G, DellaPenna D (1996)

Arabidopsis carotenoid mutants demonstrate that lutein is not essentialfor photosynthesis in higher plants. Plant Cell 8: 1627–1639

Rock CD, Zeevaart JAD (1991) The aba mutant of Arabidopsis thaliana isimpaired in epoxy-carotenoid biosynthesis. Proc Natl Acad Sci USA 88:7496–7499

Rodrigo MJ, Marcos JF, Alferez F, Mallent MD, Zacarıas L (2003) Charac-terization of Pinalate, a novel Citrus sinensis mutant with a fruit-specificalteration that results in yellow pigmentation and decreased ABA con-tent. J Exp Bot 54: 727–738

Ronen G, Carmel-Goren L, Zamir D, Hirschberg J (2000) An alternativepathway to �-carotene formation in plant chromoplasts discovered bymap-based cloning of Beta and old-gold color mutations in tomato. ProcNatl Acad Sci USA 97: 11102–11107

Ronen G, Cohen M, Zamir D, Hirschberg J (1999) Regulation of carotenoidbiosynthesis during tomato fruit development: expression of the gene forlycopene epsilon-cyclase is down-regulated during ripening and is ele-vated in the mutant Delta. Plant J 17: 341–351

Rouseff R, Raley L (1996) Application of diode array detection with a C-30reversed phase column for the separation and identification of saponifiedorange juice carotenoids. J Agric Food Chem 44: 2176–2181

Steinbrenner J, Linden H (2001) Regulation of two carotenoid biosynthesisgenes coding for phytoene synthase and carotenoid hydroxylase duringstress-induced astaxanthin formation in the green alga Haematococcuspluvialis. Plant Physiol 125: 810–817

Sun Z, Gantt E, Cunningham FX (1996) Cloning and functional analysis ofthe �-carotene hydroxylase of Arabidopsis thaliana. J Biol Chem 271:24349–24352

Tai CY, Chen BH (2000) Analysis and stability of carotenoids in the flowersof daylily (Hemerocallis disticha) as affected by various treatments. J AgricFood Chem 48: 5962–5968

Yokoyama H, Vandercook CE (1967) Citrus carotenoids: I. Comparison ofcarotenoids of mature-green and yellow lemons. J Food Sci 32: 42–48




Documents

Accumulation of Carotenoids and Expression of Carotenoid ... · Accumulation of Carotenoids and Expression of Carotenoid Biosynthetic Genes during Maturation in ... Method A was optimized