Cell. Mol. Life Sci. 63 (2006) 22912303 1420-682X/06/202291-13 DOI 10.

1007/s00018-006-6176-6 Birkhuser Verlag, Basel, 2006

Cellular and Molecular Life Sciences

Review Structural and biological aspects of carotenoid cleavage

D. P. Kloer and G. E. Schulz* Institut fr Organische Chemie und Biochemie, Albert-Ludwigs-Universitt, Albertstr. 21, 79104 Freiburg im Breisgau (Germany), Fax: +49 761 203 6161, e-mail: [email protected] Received 19 April 2006; received after revision 26 May 2006; accepted 15 June 2006 Online First 11 August 2006 Abstract. Apo-carotenoid compounds such as retinol (vitamin A) are involved in a variety of cellular processes and are found in all kingdoms of life. Instead of being synthesized from small precursors, they are commonly produced by oxidative cleavage and subsequent modification of larger carotenoid compounds. The cleavage reaction is catalyzed by a family of related enzymes, which convert specific substrate double bonds to the corresponding aldehydes or ketones. The individual family members differ in their substrate preference and the position of the cleaved double bond, giving rise to a remarkable number of products starting from a limited number of carotenoid substrate molecules. The recent determination of the structure of a member of this family has provided insight into the reaction mechanism, showing how substrate specificity is achieved. This review will focus on the biochemistry of carotenoid oxygenases and the structural determinants of the cleavage reaction.

Keywords. Abscisic acid, apo-carotenoid, BCO (beta-carotene cleavage oxygenase), carotenoid cleavage, dioxygenases, monooxygenases, non-heme iron proteins, retinal, RPE65 (retinal pigment epithelium 65 kDa protein). Introduction Carotenoids are isoprenoid compounds present in the membranes of all phototrophic as well as many heterotrophic organisms. They are derived from the C40 compound phytoene by cyclization, isomerization and oxidation [1]. Carotenoids have an extended conjugated double-bond system and absorb light in the UV-vis range. They can be oxidatively cleaved to smaller compounds called apo-carotenoids (Fig. 1). Like their parent molecules, apo-carotenoids are non-polar isoprenoids that possess an extended conjugated double-bond system, and have therefore been intensely studied for their properties as antioxidants and free radical scavengers [2]. Apo-carotenoids with altered spectroscopic and physicochemical properties can be obtained by adding functional groups through oxygenation, isomerization and double-bond cleavage. From a chemical point of view, their distinct and rigid shape combined with a tailorable amphipathic behavior makes these com* Corresponding author.

pounds ideally suited as chromophores and signaling molecules, in agreement with their roles observed in all kingdoms of life. Most apo-carotenoids are not synthesized from smaller molecules, but rather result from the oxidative cleavage of larger isoprenoids and subsequent modification in a process termed oxidative remodeling [35]. It is now known that the initial step of this remodeling is achieved through members of the carotenoid cleavage oxygenase (CCO) family, which show a high degree of regio- and stereo-specificity giving rise to defined products with distinct biological roles. These enzymes use dioxygen but it has not yet been clarified whether both oxygen atoms end up in the produced apo-carotenoids [68], that is, whether the frequently used name dioxygenases is correct. As a measure of caution, we therefore use the more general term oxygenases. The recent determination of the structure of a member of the CCO family has provided for the first time structural insight into this enzyme family [9]. Here, we review current knowledge about the biochemistry of CCO enzymes.

2292

D. P. Kloer and G. E. Schulz

Carotenoid-cleaving enzymes

Figure 1. Educts, products and cleavage positions of the substrates of CCO enzymes. (a). Safranal is obtained through cleavage of zeaxanthin at the 7,8 (or 7,8) positions while -ionone and retinal are directly produced by cleavage of -carotene at the 9,10 (or 9,10) and 15,15 positions, respectively. In contrast to the symmetric -carotene with its 5,6 and 5,6 double bonds, the asymmetric substrates -carotene and lutein have a 4,5 double bond rendering position 6 chiral. The same double bond is found in -carotene, which is only cyclic at one end. (b) Cleavage of a number of 9-cis-epoxycarotenoids at the 11,12 position results in the C15 apo-carotenoid xanthoxin, which is further processed to abscisic acid. (c) The linear C24 core of bixin is synthesized by cleavage of lycopene at both the 5,6 and 5,6 positions. Phytoene can also be cleaved at the 9,10 (9,10) positions. (d) Apo-carotenoids are themselves substrates for further cleavage.

Cell. Mol. Life Sci. Vol. 63, 2006

Review Article

2293

Biological role of apo-carotenoids In animals, the most important apo-carotenoid is the C20 compound retinal (15-apo--carotenal) and its derivatives (Fig. 1). Retinal plays a crucial role as the chromophore of rhodopsin in the vertebrate visual cycle [10]. Four other retinoids are also used as visual chromophores in the animal kingdom: (3,4)-didehydroretinal, (3R)-3-hydroxyretinal, (3S)-3-hydroxyretinal, and (4R)-4-hydroxyretinal [11]. All-trans-retinal can be oxidized to all-trans-retinoic acid and converted to the 9-cis isomer, which play an important role as signaling molecules in the immune system [12], in development [13] and in cancer [14]. They bind to nuclear retinoic acid (RAR) and retinoid X (RXR) receptors. Therefore, the use of natural and synthetic retinoids in cancer therapy has received appreciable attention [15]. Retinal also serves as the chromophore in bacteriorhodopsin as well as in bacteriorhodopsin-like proton pumps in halophilic archaea [16], eubacterial species [17] and in the fungus Leptosphaeriamaculans [18], which is the first eukaryotic example for such a pump. The same chromophore is also found in sensory rhodopsins of halophilic archaea [19] and eubacteria [20]. In plants and cyanobacteria, apo-carotenoids are found in large amounts in the thylakoid membrane, where they act as accessory and photoprotective pigments [21]. Depending on the size of the chromophore, apo-carotenoids can absorb visible light and are useful as color pigments. Well-known examples are the bright orange-red color of saffron, mainly due to glycosides derived from the C20 apo-carotenoid crocetin (Fig. 1), and the orange-red color of annatto, arising from the C24 apo-carotenoid bixin [22]. When used to color plant tissues, apo-carotenoids are found in large amounts in specialized plastids called chromoplasts. Small apo-carotenoid volatiles are aroma compounds: the C13 volatile -ionone, for example, is a characteristic component of rose scent. Several examples of apo-carotenoids acting as chemoattractants, repellants, growth stimulators and inhibitors have been described [23]. It has long been known that certain apo-carotenoids act as hormones in the regulation of plant growth and architecture. The best characterized example is abscisic acid (ABA), which plays an important role in the regulation of drought tolerance, seed development and sugar sensing [24, 25]. It is derived from the C15 apo-carotenoid xanthoxin (Fig. 1b). The biosynthesis of ABA is mainly regulated at the initial cleavage step leading to xanthoxin [24]. Evidence has been put forward that apo-carotenoids derived from all-trans carotenoid precursors play a role as signaling molecules in auxin-mediated apical dominance [26]. Analysis of Arabidopsisthaliana (mouse-ear cress) mutants led to the discovery of the two novel carotenoid cleavage oxygenases CCD7 and CCD8, which act in concert to produce an apo-carotenoid that promotes

shoot branching [27]. The exact chemical nature of the produced apo-carotenoid and its mode of action are still unknown. However, the active compound is likely derived from 13-apo--carotenone, the C18 cleavage product of CCD7 (Table 1), because loss of either CCD7 or CCD8 led to the same bushy phenotype [28]. The various roles of apo-carotenoids in plants have been recently reviewed [5, 23]. Biosynthesis of apo-carotenoids The recurrent structural motifs found in the large variety of known apo-carotenoids indicate that they are derived from full-length carotenoids, presumably along common biosynthetic pathways. The first established conversion of a carotenoid to an apo-carotenoid was that of ,carotene (provitamin A, C40) to retinal (vitamin A, C20) in animals. The existence of a specific enzyme catalyzing this reaction was postulated 40 years ago [29]. The isolation and subsequent characterization [30] of the enzyme showed that it cleaves the 15,15-double bond of -carotene to yield two molecules of retinal. In animals, the symmetric or asymmetric cleavage of C40 carotenoid compounds is the only known biosynthetic route to retinoids. All enzymes of the CCO family cleave an internal C-C double bond using dioxygen [35]. The reaction gives rise to the corresponding aldehydes or ketones. It is frequently regulated and followed by further biosynthetic steps to form the functional apo-carotenoid [3133]. However, not all apo-carotenoids in all organisms are synthesized by cleavage of a larger precursor molecule. In the case of ABA, a direct biosynthetic pathway has recently been described in fungi [34]. Here, the C15 isoprenoid farnesyldiphosphate is cyclized and oxidized in several steps to form ABA, eliminating the need for a carotenoid precursor. Specificity of CCO enzymes Of the more than 200 putative CCO enzymes to be found in sequence databases, 26 have been characterized and tested for in vitro cleavage activity. They are listed in Table 1, where they are grouped according to the type of cleaved substrate and the position of the scissile bond. All CCO enzymes cleave either one double bond or two such bonds at structurally equivalent positions, and most show a high degree of specificity with respect to the accepted substrate. The initially characterized CCO enzymes constitute the BCO (beta-carotene cleavage oxygenase) subfamily, all members of which process all-trans-,-carotene. BCOI catalyzes the cleavage of the 15,15 double bond result-

2294

D. P. Kloer and G. E. Schulz

Carotenoid-cleaving enzymes

Table 1. Functionally characterized carotenoid cleavage oxygenases. Name 1 BCOI BCOI BCOI BCOI BCOII CCD7 Source Drosophilamelanogaster Gallusgallus Musmusculus Homosapiens Musmusculus Arabidopsisthaliana Established substrate 2
-caro -caro, -caro -caro -caro -caro -caro

Cleaved double bond 15,15 15,15 15,15 15,15 9,10 9,10

Product size C20 C20 C20 C20 C27, C13 C27, C13 C15, C25 C15, C25 C15, C25 C15, C25 C15, C25 C15, C25 C15, C25 C15, C25 C15, C25 C15, C25

Functional product retinal retinal retinal retinal

Reference 35 36 37 38

possibly retinal 40 unknown 28

VP14 NCED1 NCED1 NCED1 NCED3 NCED2 NCED3 NCED5 NCED6 NCED9

Zeamays Phaseolusvulgaris Citrussinensis Perseaamericana Perseaamericana Arabidopsisthaliana Arabidopsisthaliana Arabidopsisthaliana Arabidopsisthaliana Arabidopsisthaliana

9-cis-vio, -neo, -zea, -anth, -luxR, -luxS 9-cis-vio, 9-cis-neo 9-cis-vio, 9-cis-neo 9-cis-vio, 9-cis-neo 9-cis-vio, 9-cis-neo 9-cis-vio, 9-cis-neo 9-cis-vio, 9-cis-neo 9-cis-vio, 9-cis-neo 9-cis-vio, 9-cis-neo 9-cis-vio, 9-cis-neo

11,12 11,12 11,12 11,12 11,12 11,12 11,12 11,12 11,12 11,12

ABA ABA ABA ABA ABA ABA ABA ABA ABA ABA

67 41 32 42 42 43 43 43 43 43

CCD1 CCD1 CCD1 CCD CCD1 CCD1

Arabidopsisthaliana Lycopersiconesculentum Cucumismelo Crocussativus Vitisvinifera Petuniahybrida

-caro, lut, zea, vio, 8-apo-al, 9-cis-vio, -neo -caro, lut, zea, 9-cis-vio, 9-cis-neo -caro, -caro, lyco, phyt

9,10 (9,10) 9,10 (9,10) 9,10 (9,10) 9,10 (9,10) 9,10 (9,10) 9,10 (9,10)

C13, C14 3 C13, C14 3 C13, C14 3 C13, C14 3 C13, C14 3 C13, C14 3 C10, C20 3 C8, C24 3 C20, C5-C15 3 C18, C9
3

C13 volatile C13 volatile C13 volatile C13 volatile C13 volatile C13 volatile safranal + crocetin bixin

45 48 46 3 47 33

zea zea
-caro

ZCD

Crocussativus

zea

7,8 (7,8)

LCD

Bixaorellana

lyco apo-al and -ol, apo-lycopenal and -ol 10-apo-al

5,6 (5,6)

22

ACO CCD8
1

Synechocystis PCC 6803 Arabidopsisthaliana

15,15 13,14

retinal unknown

49 28

Abbreviations: BCOI, beta-carotene cleavage oxygenase I; CCD1, carotenoid cleavage dioxygenase 1; VP14, viviparous maize protein 14; NCED1, nine-cis-epoxycarotenoid dioxygenase 1; ZCD, zeaxanthin cleavage dioxygenase; LCD, lycopene cleavage dioxygenase; ACO, apo-carotene cleavage oxygenase. 2 All substrates are in all-trans configuration unless otherwise indicated. All sizes are C except for ACO and CCD8, which process sub40 strates of up to C35 and C27, respectively. Abbreviations: caro, carotene; vio, violaxanthin; neo, neoxanthin; zea, zeaxanthin; anth, antheraxanthin; luxR, luteoxanthin 8R; luxS; luteoxanthin 8S; lut, lutein; lyco, lycopene; apo-al, apo--carotenal; phyt, phytoene. The molecule structures are shown in Figure 1. 3 Ring-free dialdehyde product.

Cell. Mol. Life Sci. Vol. 63, 2006

Review Article

2295

ing in two retinal molecules (Fig. 1a). Substrates with cis-double bonds are generally not accepted. BCOI enzymes from fruit fly [35], chicken [36], mouse [37] and human [38] sources have been cloned and biochemically characterized. For human BCOI, small variations in the substrate such as an -ionone ring on one side are tolerated, but at least one unsubstituted -ionone ring must be present for efficient cleavage [39]. In contrast to the symmetric cleavage of -carotene by BCOI, the enzyme BCOII catalyzes the asymmetric cleavage at the 9,10 position (Fig. 1a), producing 10-apo--carotenal (C27) and -ionone (C13). BCOII enzymes have been cloned from mouse, zebrafish and human sources [40]. In addition to -carotene, murine BCOII also cleaves lycopene at the 9,10 position. Another enzyme catalyzing the asymmetric cleavage of carotene at the 9,10 position is CCD7 from Arabidopsis [28]. However, CCD7 has not been classified as a BCO because it is only distantly related to the animal BCOI enzymes and because its preferred natural substrate has yet to be determined. The characterization of a viviparous maize mutant defective in ABA biosynthesis led to the discovery of VP14, an enzyme catalyzing the cleavage of 9-cis-epoxycarotenoids at the 11,12 position (Fig. 1b). Several orthologs of VP14 have been cloned from bean [41], orange [32], avocado [42] and Arabidopsis [43]. Collectively, these enzymes are grouped into the NCED (nine-cis-epoxycarotenoid dioxygenase) subfamily, of which maize VP14 is the best characterized member. It cleaves the 9-cis isomers of zeaxanthin, violaxanthin, luteoxanthin (both 8S and 8R forms), antheraxanthin and neoxanthin [44]. Except for 9-cis-zeaxanthin, all of these substrates have a 5,6-epoxy-3-hydroxy--ionone ring close to the cleavage site. Moreover, all of them yield xanthoxin as the C15 cleavage product, which in turn is the precursor of ABA. Since the most common 9-cis-epoxycarotenoids in plants are violaxanthin and neoxanthin, cleavage assays with members of the NCED family were mostly performed using only these two substrates. The range of 9-cis-epoxycarotenoid substrates that are accepted and cleaved is probably larger. As a common feature, none of the NCED family members accept an all-trans substrate. A third group of CCO enzymes cleaves substrates at the 9,10 as well as the 9,10 position and shows a broad substrate specificity. These enzymes constitute the CCD (carotenoid cleavage dioxygenase) subfamily named after CCD1 from Arabidopsis [45]. In contrast to murine BCOII, which cleaves at the same position but only accepts all-trans--carotene, CCD1 processes a wide range of all-trans- and 9-cis-carotenoids as well as epoxycarotenoids, among them the C30 apo-carotenoid all-trans-8apo--carotenal [8, 45]. When 9-cis substrates are used, CCD1 generally cleaves at the 9,10 site yielding C27 apo-carotenoids. Several other enzymes have been as-

signed to the CCD subfamily based on about 80% sequence identity with CCD1 from Arabidopsis. The recently characterized enzyme CCD1 from melon is able to cleave the all-trans substrates phytoene and lycopene, as well as - and -carotene at the 9,10 (9,10) positions to the respective C13 products [46]. However, CCD from crocus [3] and CCD1 from Vitisvinifera (grapevine) [47] were only assayed for cleavage activity towards zeaxanthin, in which case 3-hydroxy--ionone is obtained as the C13 product. Similarly, CCD1 from petunia was shown to cleave -carotene yielding -ionone [33]. Because of the close sequence relationship within the subfamily (Table 1), all of these enzymes are expected to accept a range of substrates similar to that of CCD1 from Arabidopsis andtomato [48]. Two enzymes cleaving at positions closer to the ionone rings have been characterized. ZCD (zeaxanthin cleavage dioxygenase) from crocus cleaves at the 7,8 (7,8) positions [3] giving rise to one C20 and two C10 cleavage products that are used for the biosynthesis of safranal and crocetin glycosides, respectively. LCD (lycopene cleavage dioxygenase) from Bixaorellana (lipstick tree) cleaves all-trans-lycopene at the 5,6 (5,6) positions, which is the first committed step in the biosynthesis of bixin [22]. The resulting C24 dialdehyde is oxidized to the corresponding diacid and esterified at one end (Fig. 1c). LCD is the only CCO enzyme that exclusively cleaves an acyclic substrate. Neither -carotene nor zeaxanthin are accepted. Apart from the numerous enzymes acting on C40 carotenoids, some enzymes accept only apo-carotenoids. For example, ACO (apo-carotenoid cleavage oxygenase) from the cyanobacterium Synechocystis sp. PCC 6803 cleaves the 15,15 double bond of various all-trans-apocarotenoids (Fig. 1d). It accepts apo-carotenals and apocarotenols as well as the corresponding 3-hydroxy derivatives, ranging from 12-apo (C25) to 4-apo (C35). The reaction results in all-trans-retinal or -3-hydroxy-retinal (C20) and a dialdehyde product of variable size. Larger substrates are neither bound nor cleaved [49]. Notably, ACO cleaves acyclic apo-8-lycopenal at positions other than 15,15 [50]. The Arabidopsis enzyme CCD8 cleaves 10-apo--carotenal, the cleavage product of CCD7, at the 13,14 position to produce 13-apo--carotenone (C18) and a C9 dialdehyde [28]. The CCD7 and CCD8 enzymes are thought to act in concert to produce a compound involved in the regulation of shoot branching. Interestingly, RPE65 (retinal pigment epithelium 65-kDa protein) shows a significant degree of sequence homology to the CCO family. However, its members do not cleave, but rather isomerize a specific double bond in the visual cycle. After photoisomerization of the retinal chromophore in rhodopsin from 11-cis to all-trans, the retinal is removed from the apo-protein opsin and esterified. RPE65 then binds the retinyl ester and re-isomerizes the

2296

D. P. Kloer and G. E. Schulz

Carotenoid-cleaving enzymes

11,12-trans double bond to the cis configuration in order for the visual cycle to recommence [5153]. Carotenoid cleavage by other enzymes Beside members of the CCO family, a number of rather unspecific carotenoid-degrading enzymes are known to be active during the uptake of foodstuffs. Fatty acid hydroperoxides produced by lipoxygenases, for example, have been shown to mediate -carotene cleavage in gastric mucosal homogenates [54]. The substrate was cleaved at several positions, resulting in apo-carotenals of sizes from C18 (13-apo--carotenone) to C30 (8-apo-carotenal). Similarly, lycopene was cleaved at several positions using rat intestinal homogenates [55]. Cleavage of -carotene to produce -ionone was also shown to occur in the presence of radicals generated by xanthine oxidase [56]. Some microorganisms possess membranebound or extracellular enzymes for the degradation of environmental carotenoids [57]; for example, a peroxidase from the fungus Lepistairina cleaves -carotene mainly at the 9,10 double bond to form -ionone [58]. Further characterization of these enzymes is to be expected in the future. Subcellular location of CCO enzymes The substrates of CCO enzymes are generally large, lipophilic molecules that are sparingly water soluble and therefore bound to carrier proteins or dissolved in membranes. In plants, carotenoids are synthesized and stored in plastids. In the thylakoid membrane, they are used as accessory and photoprotective pigments. They are also found in chromoplast membranes and stromata, where they accumulate in large quantities giving rise to brightly colored structures. CCO enzymes involved in the remodeling of carotenoids are therefore often targeted to plastids. The respective signal peptides are present in the sequences of some, but not all plant members. In the case of VP14 from the NCED subfamily (Table 1), the import of the soluble cytosolic protein into chloroplasts led to two versions of the enzyme, a soluble stromal version and a second version bound to the thylakoid membrane [59]. The enzyme NCED1 from beanis also active in thylakoids [41]. CCD7 and CCD8 as well as their homologues from Arabidopsis are targeted to plastids, except for CCD1, which is cytosolic like CCD1 from tomato [48] and CCD from crocus [3]. Although most carotenoids are located in the thylakoid membrane, significant amounts of -carotene have been found in the outer envelope of pea chloroplasts [21] which likely supply the cytosolic CCD1 enzymes. Up to now, little is known about how the cytoplasmic carotenoid oxy-

genases access their substrates. However, the biochemical and structural data of the CCO family member ACO from Synechocystis point to a monotopic membrane association mediated by a large non-polar surface patch close to the active center [9]. It remains to be seen whether this also applies to the other CCO enyzmes. Structure of CCO enzymes The only established structure of a CCO family member is that for ACO from Synechocystis PCC 6803 [9]. The chainfold of ACO is a seven-bladed propeller (Fig. 2). Five propeller blades consist of four antiparallel strands each, while two blades are five-stranded. The bottom side of the propeller is relatively flat and contains short loops connecting the strands to each other (Fig. 3a). In contrast, the top side of the propeller is crowded where extended loops with short helices form a large dome. As observed in numerous propeller structures, the active center of ACO is located near the propeller axis on the top side. The most prominent feature is a large tunnel that enters the protein from one side in a direction perpendicular to the propeller axis (Fig. 4a), passes the active center defined by the iron, and then proceeds to an exit approximately parallel to the propeller axis. The shape of the tunnel is mostly defined by the extended loops of the dome and only slightly by the rigid propel-

Figure 2. Sketch of the chainfold of ACO showing the seven-bladed propeller as the basic motif. The positions of the four histidines coordinating the iron at the center of the propeller are marked by dots. The four loops and the N terminus forming the dome above the active center are dark gray, and the N-terminal extension of the propeller is hatched (1 to 2).

Cell. Mol. Life Sci. Vol. 63, 2006

Review Article

2297

Figure 3. The structure of ACO from SynechocystisPCC 6803. (a) Ribbon representation of the enzyme showing the strands of the propeller in blue, the residues of the dome above the active center in green, and the N-terminal extension in yellow. The bound substrate (3R)-3-hydroxy-8--apo-carotenol, the catalytic iron with its four coordinating histidines as well as two water molecules are shown in a ball-and-stick representation. The view is along the entrance of the active-center tunnel below the dome. The tunnel exit is indicated by a red line. (b) Active-center and substrate-binding geometry. The apo-carotenoid substrate is shown in orange with the scissile 15,15 double bond in yellow. The vacant iron coordination site is indicated by a transparent gray sphere and the weakly defined second water molecule is shown as a transparent red sphere. Some distances are given.

Figure 4. Substrate binding in ACO shown in a surface representation. Positively charged, negatively charged and non-polar residues are given in blue, red and yellow, respectively. The catalytic iron atom is indicated by a red circle, the bound substrate is black/red. (a) Cut through the molecule approximately along the propeller axis, revealing a pocket penetrating from the bottom along the propeller axis to the center, where it almost reaches the catalytic iron. Moreover, the cut shows the substrate-binding tunnel that enters the protein from the right side, changes its direction near the center and ends in the tunnel exit at top left. (b) Top view of the substrate (3R)-3-hydroxy-8--apocarotenol in the cut enzyme. The 15,15-double bond of the substrate is at the catalytic iron. On both sides of the cleavage site, the 13,14 and 13,14 double bonds assume a cis-cis configuration on binding, giving rise to a doubly kinked chain. Further entry of the substrate into the tunnel is prevented by the narrow entrance. The view is approximately along the propeller axis.

ler scaffold. This allows the individual CCO family members to modulate substrate specificity by changing length and sequence of the connecting loops, while retaining the rigid propeller scaffold for stability. Indeed, sequence alignments within the CCO family clearly show that the most highly conserved regions are within the strands forming the propeller as demonstrated in Figure 5. This strongly suggests that all members of the CCO family

have the same propeller chainfold and may therefore be modeled along the lines of the ACO structure. In ACO, a second cavity enters the protein from the bottom side, along the propeller axis (Fig. 4a). This is an intrinsic feature of the propeller architecture and is therefore not expected to vary greatly within the CCO family. The pocket ends shortly before the active center. Its functional significance is not clear. It may serve as

2298

D. P. Kloer and G. E. Schulz

Carotenoid-cleaving enzymes

an auxiliary tunnel for dioxygen, which, however, would have to pass through a short dense area of the protein to reach the iron. ACO contains a mononuclear iron center consisting of four histidine side chains holding the catalytic iron with their N atoms. The iron-binding site is rigid as it does not change on removal of the metal ion [9]. In the crystal structure of ACO with a ferrous iron and a substrate molecule, the coordination geometry is best described as octahedral with two missing ligands (Fig. 3b). The fifth coordination position is occupied by a water molecule, while the sixth position is empty. The active center location was confirmed by mutations of the iron-ligating histidines, which in mouse BCOI [60] and in human RPE65 [61] led to a total loss of enzymatic activity. Within the large group of non-heme-iron-containing oxygenases, the iron is usually coordinated by a combination of histidine, aspartate or glutamate and water ligands [62]. Mononuclear iron centers with four histidine ligands are rare, ACO being the fifth example of such a coordination environment. Three of the four histidines are fixed by hydrogen bonds to glutamates. In the entire CCO family as well as in the isomerizing RPE65 group, the histidines and glutamates are strictly and well conserved, respec-

tively (Fig. 5). The nitrogen ligands certainly stabilize the ferrous state of the iron. Unfortunately, no redox potential of any CCO enzyme has yet been determined. Structural determinants of substrate specificity The substrate molecule (3R)-3-hydroxy-8-apo--carotenol bound to ACO in the crystal provides some clues regarding the recognition of the carotenoids by the CCO enzymes. ACO accepts apo--carotenals (and -ols) with lengths from C25 (12-apo) to C35 (4-apo) and cleaves at the 15,15 position [49], always yielding the same C20 compound retinal (Fig. 1d). Larger substrates such as carotene (C40) are not cleaved at all, whereas acyclic apolycopenals are cleaved at multiple positions [50]. In the crystal structure, the -ionone ring of the substrate is invisible, presumably due to its mobility. However, the visible part of the apo-carotenoid indicates clearly that the ionone ring is located on the enzyme surface because it cannot pass through the narrow entrance of the substrate tunnel (Fig. 4). As sketched in Figure 6, the tunnel and its entrance work as a ruler, determining the cleavage position relative to the ionone ring. In addition, the entrance

Figure 5. Sequence alignment of six members of the carotenoid cleavage oxygenase family showing the chain segments forming the loops of the dome above the active center (boxes). In pairwise comparisons these enzymes have at most 21% identical residues. Residue conservation is indicated by shading. The four iron-ligating histidines (*) and the glutamates fixing them () are labeled. The dots at the bottom mark every tenth residue of ACO. The enzyme codes are explained in Table 1; CCD1 is from Arabidopsis, BCOI is from mouse. The numbers denote the position of the last printed residue of the line. Some sequences are truncated: the N-terminal marks (#) stand for 113, 38, 52, 0, 69 and 11 missing residues (top to bottom), and the C-terminal marks (#) denote 8, 9, 0, 46, 1 and 0 residues.

Cell. Mol. Life Sci. Vol. 63, 2006

Review Article

2299

Figure 6. Schematic sketch of apo-carotenoid cleavage in ACO showing the tunnel with entrance, kink and exit. The dome and the non-polar patch near the tunnel entrance are indicated. Top: Apocarotenoid substrates with only one -ionone group can enter the enzyme and are held in position by the narrow tunnel entrance. The distance between the catalytic iron and the entrance determines the cleavage position. Bottom: Substrates with -ionone rings on both sides cannot enter the tunnel and are therefore not cleaved.

prevents -carotene cleavage because neither of the two substrate ends can enter the tunnel (Fig. 6). This explains the selectivity towards apo-carotenoids. The upper size limit for the apo-carotenoids is dictated by the available linear space between the tunnel entrance and the end of the central cavity. The ruler model is consistent with the observation that substrate binding does not cause an appreciable induced-fit displacement. It should be noted that besides retinal, ACO produces dialdehydes of different lengths that possess higher solubility. ACO has a large non-polar patch near the entrance (Fig. 6), which is probably used to enter into the interior of a membrane from where it picks up the nonpolar educt. After cleavage, the non-polar product retinal is again released to the membrane, whereas the more soluble second product dialdehyde probably dissociates into the cytosol. This is quite possible if the dialdehyde leaves ACO through the exit instead of through the entrance. Accordingly, the observed non-polar entrance-polar exit construction around the active center seems to be explained by the physicochemical properties of educts and products. An unexpected feature of the ACO crystal structure was the presence of two cis double bonds in the bound substrate that were located on either side of the cleaved 15,15 double bond (Figs. 3b, 6), even though the all-trans form had been used in the cocrystallization experiment. Obviously, the doubly kinked conformation is enforced by the shape of the active center. However, the isomerization

has to be supported catalytically, because the barrier of methyl-substituted double bonds in an extended conjugated double-bond system is about 100 kJ/mol. Numerous examples of proteins that achieve isomerization of double bonds are known. For example, the activation barrier of the retinal chromophore in rhodopsin is lowered by the surrounding protein to such an extent that thermal isomerization of the 11-cis to the all-trans form occurs at ambient temperatures [63, 64]. In ACO, the isomerization is most likely catalyzed by the iron atom in conjunction with a ligated oxygen species. Incidently, the reported isomerization of the ACO ligand is similar to the isomerization activity of RPE65, converting all-trans to 11-cis retinal [65]. Although the structure of ACO prohibits cleavage of carotene (Fig. 6), the homology between ACO and, for example, the carotene-cleaving BCOI (Table 1) indicates that the structural difference should be comparatively small. Some changes of the dome structure above the iron would suffice to convert the tunnel to a cleft, which may then close in an induced fit on a bound -carotene. The peptide segments forming the dome are boxed in Figure 5. They show large variations, rendering any attempt to model the dome of another CCO family member highly hazardous. Several CCO enzymes cleave two double bonds at substrate positions related to each other by symmetry or pseudo-symmetry [45, 48]. Examples are the 5,6 and 5,6 bonds in lycopene (symmetric, Fig. 1), the 7,8 and 7,8 bonds in zeaxanthin (symmetric), or the 9,10 and 9,10 bonds in -carotene (symmetric) or -carotene (pseudosymmetric). Because of the symmetry of substrate and cleavage, CCO enzymes have been proposed to act as dimers [45]. Dimerization is not necessary, however, because an asymmetric polypeptide may well form an almost symmetric binding site. Moreover, if the strongest binding occurs at the symmetric central part of the substrate molecule, pseudo-symmetric substrates with different end structures like lutein or -carotene (Fig. 1a) will also be accepted in both orientations. A similar case of binding the pseudo-symmetric molecule NAD+ in both orientations is known [66]. Presumably, these enzymes cleave their C40 substrate once, release the products, bind the primary apo-carotenoid product again in reverse orientation, and cleave it a second time on the other side. Therefore, this subgroup of the CCO family should be able to cleave apo-carotenoids. This conclusion was confirmed for CCD1 from Arabidopsis, which accepted an apo-carotenoid invitro[8]. Reaction mechanism In general, enzymatic carotenoid cleavage reactions in vitro are performed in the presence of a reductant such

2300

D. P. Kloer and G. E. Schulz

Carotenoid-cleaving enzymes

as ascorbate [67], DTT [68], TCEP [39] or excess Fe2+ [35] in order to keep the iron in its ferrous state. However, it is unlikely and has never been demonstrated that the reductant participates stoichiometrically in the reaction itself. All electrons in the products can be derived exclusively from the carotenoid and dioxygen. Nevertheless, it is clear that all enzymes of the CCO family are iron dependent. BCOI enzymes isolated from various animal sources have long been known to be inhibited by ironchelating agents such as 1,10-phenantroline and 2,2-dipyridine [30, 69]. Moreover, the cleavage activity of VP14 was lost on the removal of iron with EDTA and restored by reconstitution with Fe2+, but not with Fe3+ [67]. Also, BCOI activity in the small intestine of rats was shown to increase with rising iron levels [70]. Dependence on ferrous iron was also shown for RPE65, which was inactivated by iron chelation, but recovered on addition of Fe2+, though not on addition of Cu2+, Zn2+, Mg2+ or Fe3+ [65]. An inductively coupled plasma (ICP) emission analysis revealed one bound iron atom per protomer in chicken BCOI [7] and bovine RPE65 [65]. This is consistent with the crystal structure of ACO [9], and we conclude that the reaction mechanism requires one ferrous iron atom near the scissile double bond. While the requirement of Fe2+ and O2 for the reaction is clear, the nature of the involved oxygen species has not yet been clarified. Due to their conjugated double bonds, carotenoids are highly reactive, so that unspecific oxygenation and cleavage can occur in a number of chemical, thermal or enzymatic processes. In contrast, the reactions catalyzed by the CCO enzymes are well defined. The mechanism of these reactions, however, remains controversial. It is clear that the C-C double bond is cleaved to the corresponding aldehydes or ketones using dioxygen as reactant and ferrous iron as a cofactor, but whether the reaction follows a monooxygenase or a dioxygenase pathway has not been established. In the first case, the scissile double bond is first epoxidized by activated dioxygen and the epoxide is then opened by a water molecule (Fig. 7). Subsequently, the resulting vicinal diol is cleaved, leading to the incorporation of one oxygen atom from dioxygen and another one from a water molecule into the products. According to the alternative dioxygenase mechanism, the double bond accepts dioxygen to form a dioxetane intermediate that decays to two aldehydes or ketones (Fig. 7). To discriminate between these two possibilities, several isotope-labeling experiments have been performed invitro and invivo. Experiments using leaves from a waterstressed Xanthiumstrumarium (cocklebur) in an 18O2 atmosphere showed an 18O incorporation into the carboxyl group of ABA [71]. As the 18O label was probably introduced by the oxidative cleavage of the 11,12-double bond of a 9-cis-epoxycarotenoid to form xanthoxin (Fig. 1b), the presumed full substitution with 18O was interpreted in terms of a dioxygenase mechanism. This may have been

Figure 7. Discussed reaction mechanisms of CCO enzymes. The oxygen species at the iron is putative. The essentials of the monooxygenase mechanism passing through an epoxide and of the dioxygen mechanism with a dioxetane intermediate are given as alternatives. The putative oxygen derived from water is shaded.

a premature conclusion, because the second product of the cleavage reaction had not been analyzed. By itself, the label at ABA is not conclusive because it may arise if the epoxide were opened regio-specifically in a mono-oxygenase mechanism, and regio-specificity is a common feature of enzymes. For an avian -carotene-15,15-oxygenase BCOI (Table 1), the reaction mechanism was analyzed using 18O-labeled water and 17O-labeled dioxygen during invitro cleavage assays with recombinant protein [6]. The asymmetric carotene (Fig. 1a) instead of the symmetric -carotene was used, in order to obtain two separable reaction products. In this experiment, 17O and 18O were found in equal amounts in both products, clearly supporting a monooxygenase mechanism, because a dioxygenase mechanism would have led to the incorporation of 17O alone [7]. The observed labeling symmetry can be explained by -carotene binding in both orientations to the enzyme, which is consistent with the structure of ACO, where the ring of the substrate is only loosely bound [9]. Alternatively, the opening of the epoxide by water may not occur regio-specifically which, however, seems unlikely for an enzyme.

Cell. Mol. Life Sci. Vol. 63, 2006

Review Article

2301

The experiment was criticized as being inconclusive, because the reductase used for converting the produced aldehydes to alcohols may have scrambled the oxygens [8]. The reductase was required to fix the oxygen in a stable alcohol in order to avoid an oxygen exchange at the aldehyde stage. This criticism awaits experimental testing. Recently, a further labeling experiment was performed with recombinant CCD1 from Arabidopsis [8] using 8-apo--carotenal as substrate in an 18O2 atmosphere. Cleavage at the 9,10 position resulted in -ionone (C13) and a C17 dialdehyde. Almost all of the produced -ionone and some dialdehyde carried the 18O label, which was suggested to prove the dioxygenase mechanism [8]. While the -ionone ketone is stable, the C17 dialdehyde is known to exchange rapidly the oxygen in water. Unfortunately, the required 100% labeling of one of the two aldehyde groups could not be demonstrated with certainty, so this experiment cannot be considered conclusive. In summary, the BCOI experiment [6] would be conclusive for a mono-oxygenase mechanism if the applied reductase were proven to keep the aldehyde oxygen during the reaction with NADH, while the CCD1 [8] experiment requires a more direct proof of the full label at the dialdehyde. Conclusions Since the discovery of carotenoid cleavage by specific oxygenases, much insight has been gained into this fascinating family of enzymes. Although the structure determination of a member of the CCO family in complex with a bound substrate has revealed the architecture of the active center, the details of the reaction and the nature of the involved oxygen species have yet to be clarified. It seems likely that the question whether these enzymes are monooxygenases or dioxygenases, i.e. whether one or both atoms of dioxygen are incorporated into the products, will soon be answered by conclusive isotope labeling experiments. In the ACO crystal structure, the isomerization of two specific substrate C-C double bonds upon binding has been observed. At present, it is unclear whether this is a speciality of ACO, or a general feature of substrate binding in CCO enzymes, or is even required for the reaction mechanism. The RPE65 group functioning as non-cleaving double-bond isomerases points to the latter possibility. Understanding substrate specificity in the CCO family remains a fascinating endeavor. The structure of ACO shows how the correct substrate double bond is selected for a series of apo-carotenoids. However, a different selection process should exist in enzymes cleaving alltrans- or 9-cis-C40 carotenoids, or in enzymes cleaving a

broad range of both all-trans and 9-cis substrates such as CCD1 from Arabidopsis. The structure of ACO showed that all these enzymes have a different dome structure on top of a stable propeller scaffold and, furthermore, that the loops forming the dome can easily vary during evolution. Consequently, novel substrate specificities should arise frequently during evolution. Unfortunately, the structure of ACO does not permit the modeling of the dome of homologous enzymes at a reasonable confidence level, because the variations are too extensive. Further structures of the CCO family members are needed to explain the great variety of substrate specificities.
Acknowledgements. We thank L. Bhm for reading the manuscript. 1 2 Rmer, S. and Fraser, P. D. (2005) Recent advances in carotenoid biosynthesis, regulation and manipulation. Planta 221, 15331535. Escribano, J., Alonso, G. L., Coca-Prados, M. and Fernandez, J. A. (1996) Crocin, safranal and picrocrocin from saffron (Crocussativus L.) inhibit the growth of human cancer cells invitro. Cancer Lett. 100, 2330. Bouvier, F., Suire, C., Mutterer, J. and Camara, B. (2003) Oxidative remodeling of chromoplast carotenoids: identification of the carotenoid dioxygenase CsCCD and CsZCD genes involved in Crocus secondary metabolite biogenesis. Plant Cell 15, 4762. Giuliano, G., Al-Babili, S. and von Lintig, J. (2003) Carotenoid oxygenases: cleave it or leave it. Trends Plant Sci. 8, 145149. Camara, B. and Bouvier, F. (2004) Oxidative remodeling of plastid carotenoids. Arch. Biochem. Biophys. 430, 1621. Leuenberger, M. G., Engeloch-Jarret, C. and Woggon, W. D. (2001) The reaction mechanism of the enzyme-catalyzed central cleavage of beta-carotene to retinal.Angew. Chem. Int. Ed. Engl. 40, 26132617. Woggon, W. D. (2002) Oxidative cleavage of carotenoids catalyzed by enzyme models and beta-carotene 15,15-monooxygenase. Pure Appl. Chem. 74, 13971408. Schmidt, H., Kurtzer, R., Eisenreich, W. and Schwab, W. (2006) The carotenase AtCCD1 from Arabidopsisthaliana is a dioxygenase. J. Biol. Chem. 281, 98459851. Kloer, D. P., Ruch, S., Al-Babili, S., Beyer, P. and Schulz, G. E. (2005) The structure of a retinal-forming carotenoid oxygenase. Science 308, 267269. Spudich, J. L., Yang, C. S., Jung, K. H. and Spudich, E. N. (2000) Retinylidene proteins: structures and functions from archaea to humans. Annu. Rev. Cell Dev. Biol. 16, 365392. Seki, T., Isono, K., Ozaki, K., Tsukahara, Y., Shibata-Katsuta, Y., Ito, M., Irie, T. and Katagiri, M. (1998) The metabolic pathway of visual pigment chromophore formation in Drosophila melanogasterall-trans (3S)-3-hydroxyretinal is formed from all-trans retinal via (3R)-3-hydroxyretinal in the dark. Eur. J. Biochem. 257, 522527. Spilianakis, C. G., Lee, G. R. and Flavell, R. A. (2005) Twisting the Th1/Th2 immune response via the retinoid X receptor: lessons from a genetic approach. Eur. J. Immunol. 35, 3400 3004. Lampert, J. M., Holzschuh, J., Hessel, S., Driever, W., Vogt, K. and von Lintig, J. (2003) Provitamin A conversion to retinal via the beta,beta-carotene-15,15-oxygenase (bcox) is essential for pattern formation and differentiation during zebrafish embryogenesis. Development 130, 21732186. Abu, J., Batuwangala, M., Herbert, K. and Symonds, P. (2005) Retinoic acid and retinoid receptors: potential chemopreventive and therapeutic role in cervical cancer. Lancet Oncol. 6, 712720.

4 5 6

7 8 9 10 11

12

13

14

2302

D. P. Kloer and G. E. Schulz

Carotenoid-cleaving enzymes 34 Inomata, M., Hirai, N., Yoshida, R. and Ohigashi, H. (2004) The biosynthetic pathway to abscisic acid via ionylideneethane in the fungus Botrytiscinerea. Phytochemistry 65, 26672678. 35 von Lintig, J. and Vogt, K. (2000) Filling the gap in vitamin A research: molecular identification of an enzyme cleaving betacarotene to retinal. J. Biol. Chem. 275, 1191511920. 36 Wyss, A., Wirtz, G., Woggon, W., Brugger, R., Wyss, M., Friedlein, A., Bachmann, H. and Hunziker, W. (2000) Cloning and expression of beta,beta-carotene 15,15-dioxygenase. Biochem. Biophys. Res. Commun. 271, 334336. 37 Redmond, T. M., Gentleman, S., Duncan, T., Yu, S., Wiggert, B., Gantt, E. and Cunningham, F. X. Jr (2001) Identification, expression, and substrate specificity of a mammalian betacarotene 15,15-dioxygenase. J. Biol. Chem. 276, 65606565. 38 Yan, W., Jang, G. F., Haeseleer, F., Esumi, N., Chang, J., Kerrigan, M., Campochiaro, M., Campochiaro, P., Palczewski, K. and Zack, D. J. (2001) Cloning and characterization of a human beta,beta-carotene-15,15-dioxygenase that is highly expressed in the retinal pigment epithelium. Genomics 72, 193202. 39 Lindqvist, A. and Andersson, S. (2002) Biochemical properties of purified recombinant human beta-carotene 15,15-monooxygenase. J. Biol. Chem. 277, 2394223948. 40 Kiefer, C., Hessel, S., Lampert, J. M., Vogt, K., Lederer, M. O., Breithaupt, D. E. and von Lintig, J. (2001) Identification and characterization of a mammalian enzyme catalyzing the asymmetric oxidative cleavage of provitamin, A. J. Biol. Chem. 276, 1411014116. 41 Qin, X. and Zeevaart, J. A. (1999) The 9-cis-epoxycarotenoid cleavage reaction is the key regulatory step of abscisic acid biosynthesis in water-stressed bean. Proc. Natl. Acad. Sci. USA 96, 1535415361. 42 Chernys, J. T. and Zeevaart, J. A. (2000) Characterization of the 9-cis-epoxycarotenoid dioxygenase gene family and the regulation of abscisic acid biosynthesis in avocado. Plant Physiol. 124, 343353. 43 Tan, B. C., Joseph, L. M., Deng, W. T., Liu, L., Li, Q. B., Cline, K. and McCarthy, D. R. (2003) Molecular characterization of the arabidopsis 9-cis epoxycarotenoid dioxygenase gene family. Plant, J. 35, 4456. 44 Schwartz, S. H., Tan, B. C., McCarty, D. R., Welch, W. and Zeevaart, J. A. (2003) Substrate specificity and kinetics for VP14, a carotenoid cleavage dioxygenase in the ABA biosynthetic pathway. Biochim. Biophys. Acta 1619, 914. 45 Schwartz, S. H., Qin, X. and Zeevaart, J. A. (2001) Characterization of a novel carotenoid cleavage dioxygenase from plants. J. Biol. Chem. 276, 2520825211. 46 Ibdah, M., Azulay, Y., Portnoy, V., Wasserman, B., Bar, E., Meir, A., Burger, Y., Hirschberg, J., Schaffer, A. A., Katzir, N., Tadmor, Y. and Lewinsohn, E. (in press) Functional characterization of CmCCD1, a carotenoid cleavage dioxygenase from melon. Phytochemistry. 47 Mathieu, S., Terrier, N., Procureur, J., Bigey, F. and Gunata, Z. (2005) A carotenoid cleavage dioxygenase from Vitisvinifera L.: functional characterization and expression during grape berry development in relation to C13-norisoprenoid accumulation. J. Exp. Bot. 56, 27212731. 48 Simkin, A. J., Schwartz, S. H., Auldridge, M., Taylor, M. G. and Klee, H. J. (2004) The tomato carotenoid cleavage dioxygenase 1 genes contribute to the formation of the flavor volatiles beta-ionone, pseudoionone, and geranylacetone. Plant J. 40, 882892. 49 Ruch, S., Beyer, P., Ernst, H. and Al-Babili, S. (2005) Retinal biosynthesis in Eubacteria: invitro characterization of a novel carotenoid oxygenase from Synechocystis sp. PCC 6803. Mol. Microbiol. 55, 10151024. 50 Ruch, S. (2005) Biochemische und strukturelle Charakterisierung einer Retinal-bildenden Carotinoid-Oxygenase aus Synechocystis sp. PCC 6803. PhD thesis, Albert-Ludwigs-Universitt, Freiburg.

15 Simeone, A. M. and Tari, A. M. (2004) How retinoids regulate breast cancer cell proliferation and apoptosis. Cell. Mol. Life Sci. 61, 14751484. 16 Oesterhelt, D., Meentzen, M. and Schuhmann, L. (1973) Reversible dissociation of the purple complex in bacteriorhodopsin and identification of 13-cis and all-trans-retinal as its chromophores. Eur. J. Biochem. 40, 453463. 17 Beja, O., Spudich, E. N., Spudich, J. L., Leclerc, M. and DeLong, E. F. (2001) Proteorhodopsin phototrophy in the ocean. Nature 411, 786789. 18 Waschuk, S. A., Bezerra, A. G. Jr, Shi, L. and Brown, L. S. (2005) Leptosphaeria rhodopsin: bacteriorhodopsin-like proton pump from a eukaryote. Proc. Natl. Acad. Sci. USA 102: 68796883. 19 Spudich, J. L. and Luecke, H. (2002) Sensory rhodopsin II: functional insights from structure. Curr. Opin. Struct. Biol. 12, 540546. 20 Vogeley, L., Sineshchekov, O. A., Trivedi, V. D., Sasaki, J., Spudich, J. L. and Luecke, H. (2004) Anabaena sensory rhodopsin: a photochromic color sensor at 2.0 . Science 306, 13901393. 21 Markwell, J., Bruce, B. D. and Keegstra, K. (1992) Isolation of a carotenoid-containing sub-membrane particle from the chloroplastic envelope outer membrane of pea (Pisum sativum). J. Biol. Chem. 267, 1393313937. 22 Bouvier, F., Dogbo, O. and Camara, B. (2003) Biosynthesis of the food and cosmetic plant pigment bixin (annatto). Science 300, 20892091. 23 Bouvier, F., Isner, J. C., Dogbo, O. and Camara, B. (2005) Oxidative tailoring of carotenoids: a prospect towards novel functions in plants. Trends Plant. Sci. 10, 187194. 24 Schwartz, S. H., Qin, X. and Zeevaart, J. A. (2003) Elucidation of the indirect pathway of abscisic acid biosynthesis by mutants, genes, and enzymes. Plant Physiol. 131, 15911601. 25 Taylor, I. B., Sonneveld, T., Bugg, T. D. H. and Thompson, A. J. (2005) Regulation and manipulation of the biosynthesis of abscisic acid, including the supply of xanthophyll precursors. J. Plant Growth Regul. 24, 253273. 26 Xiong, L. and Zhu, J.-K. (2003) Regulation of abscisic acid biosynthesis. Plant Physiol. 133, 2936. 27 Booker, J., Auldridge, M., Wills, S., McCarty, D., Klee, H. and Leyser, O. (2004) MAX3/CCD7 is a carotenoid cleavage dioxygenase required for the synthesis of a novel plant signaling molecule. Curr. Biol. 14, 12321238. 28 Schwartz, S. H., Qin, X. and Loewen, M. C. (2004) The biochemical characterization of two carotenoid cleavage enzymes from Arabidopsis indicates that a carotenoid-derived compound inhibits lateral branching. J. Biol. Chem. 279, 4694046945. 29 Olson, J. A. and Hayaishi, O. (1965) The enzymatic cleavage of beta-carotene into vitamin A by soluble enzymes of rat liver and intestine. Proc. Natl. Acad. Sci. USA 54, 13641370. 30 Lakshman, M. R., Chansang, H. and Olson, J. A. (1972) Purification and properties of carotene 15,15-dioxygenase of rabbit intestine. J. Lipid Res. 13, 477482. 31 Iuchi, S., Kobayashi, M., Yamaguchi-Shinozaki, K. and Shinozaki, K. (2000) A stress-inducible gene for 9-cis-epoxycarotenoid dioxygenase involved in abscisic acid biosynthesis under water stress in drought-tolerant cowpea. Plant Physiol. 123, 553562. 32 Rodrigo, M. J., Alquezar, B. and Zacarias, L. (2006) Cloning and characterization of two 9-cis-epoxycarotenoid dioxygenase genes, differentially regulated during fruit maturation and under stress conditions, from orange (Citrussinensis, L. Osbeck). J. Exp. Bot. 57, 633643. 33 Simkin, A. J., Underwood, B. A., Auldridge, M., Loucas, H. M., Shibuya, K., Schmelz, E., Clark, D. G. and Klee, H. J. (2004) Circadian regulation of the PhCCD1 carotenoid cleavage dioxygenase controls emission of beta-ionone, a fragrance volatile of petunia flowers. Plant Physiol. 136, 35043514.

Cell. Mol. Life Sci. Vol. 63, 2006 51 Kuksa, V Imanishi, Y., Batten, M., Palczewski, K. and Moise, ., A. R. (2003) Retinoid cycle in the vertebrate retina: experimental approaches and mechanisms of isomerization. Vision Res. 43, 29592981. 52 Saari, J. C. (2001) The sights along route 65. Nat. Genet. 29, 89. 53 Jin, M., Li, S., Moghrabi, W. N., Sun, H. and Travis, G. H. (2005) Rpe65 is the retinoid isomerase in bovine retinal pigment epithelium. Cell 122, 449459. 54 Yeum, K. J., Lee-Kim, Y. C., Yoon, S., Lee, K. Y., Park, I. S., Lee, K. S., Kim, B. S., Tang, G., Russell, R. M. and Krinsky, N. I. (1995) Similar metabolites formed from beta-carotene by human gastric mucosal homogenates, lipoxygenase, or linoleic acid hydroperoxide. Arch. Biochem. Biophys. 321, 167174. 55 dos Anjos-Ferreira, A. L., Yeum, K. J., Russell, R. M., Krinsky, N. I. and Tang, G. (2004) Enzymatic and oxidative metabolites of lycopene. J. Nutr. Biochem. 15, 493502. 56 Bosser, A. and Belin, J. M. (1994) Synthesis of -ionone in an aldehyde/xanthine oxidase/-carotene system involving free radical formation. Biotechnol. Prog. 10, 129133. 57 Zorn, H., Langhoff, S., Scheibner, M. and Berger, R. G. (2003) Cleavage of beta,beta-carotene to flavor compounds by fungi. Appl. Microbiol. Biotechnol. 62, 331336. 58 Zorn, H., Langhoff, S., Scheibner, M., Nimtz, M. and Berger, R. G. (2003) A peroxidase from Lepistairina cleaves beta,betacarotene to flavor compounds. Biol. Chem. 384, 10491056. 59 Tan, B. C., Cline, K. and McCarty, D. R. (2001) Localization and targeting of the VP14 epoxy-carotenoid dioxygenase to chloroplast membranes. Plant J. 27, 373382. 60 Poliakov, E., Gentleman, S., Cunningham, F. X. Jr, Miller-Ihli, N. J. and Redmond, T. M. (2005) Key role of conserved histidines in recombinant mouse beta-carotene 15,15-monooxygenase-1 activity. J. Biol. Chem. 280, 2921729223. 61 Takahashi, Y., Moiseyev, G., Chen, Y. and Ma, J. X. (2005) Identification of conserved histidines and glutamic acid as key residues for isomerohydrolase activity of RPE65, an enzyme of

Review Article

2303

62 63 64 65

66

67 68

69 70

71

the visual cycle in the retinal pigment epithelium. FEBS Lett. 579, 54145418. Que, L. Jr. and Ho, R.Y. (1996) Dioxygen activation by enzymes with mononuclear non-heme iron active sites. Chem. Rev. 96, 26072624. Ala-Laurila, P., Donner, K. and Koskelainen, A. (2004) Thermal activation and photoactivation of visual pigments. Biophys. J. 86, 36533662. Barlow, R. B., Birge, R. R., Kaplan, E. and Tallent, J. R. (1993) On the molecular origin of photoreceptor noise. Nature 366, 6466. Moiseyev, G., Takahashi, Y., Chen, Y., Gentleman, S., Redmond, T. M., Crouch, R. K. and Ma, J. X. (2006) RPE65 is an iron(II)-dependent isomerohydrolase in the retinoid visual cycle. J. Biol. Chem. 281, 28352840. Ruf, A., Rolli, V., de Murcia, G. and Schulz, G. E. (1998) The mechanism of the elongation and branching reaction of poly(ADP-ribose) polymerase as derived from crystal structures and mutagenesis. J. Mol. Biol. 278, 5765. Schwartz, S. H., Tan, B. C., Gage, D. A., Zeevaart, J. A. and McCarty, D. R. (1997) Specific oxidative cleavage of carotenoids by VP14 of maize. Science 276, 18721874. Ershov, I. V., Dmitrovskii, A. A. and Bykhovskii, V. (1993) Properties of beta-carotene-15,15-dioxygenase, stabilized during purification with lutein and dithiothreitol. Biokhimiia 58, 416423. Olson, J. A. (1989) Provitamin A function of carotenoids: the conversion of beta-carotene into vitamin, A. J. Nutr. 119, 105 108. During, A., Fields, M., Lewis, C. G. and Smith, J. C. (1999) Beta-carotene 15,15-dioxygenase activity is responsive to copper and iron concentrations in rat small intestine. J. Am. Coll. Nutr. 18, 309315. Creelman, R. A. and Zeevart, J. A. D. (1984) Incorporation of oxygen into abscisic acid and phaseic acid from molecular oxygen. Plant Physiol. 75, 166169.