In vitro characterization of a carotenoid cleavage dioxygenase from Nostoc sp. PCC 7120 reveals a novel cleavage pattern, cytosolic localization and induction by highlight


*E-mail; Tel. (+49) 761203 8454; Fax (+49) 761203 2675.


Carotenoid oxygenases catalyse the cleavage of C-C double bonds forming apocarotenoids, a diverse group of compounds, including retinoids and the precursors of some phytohormones. Some apocarotenoids, like β-ionone (C13), are ecologically important volatiles released by plants and cyanobacteria. In this work, we elucidated the activity of the Nostoccarotenoid cleavage dioxygenase (NosCCD, previously named NSC1) using synthetic and cyanobacterial substrates. NosCCD converted bicyclic and monocyclic xanthophylls, including myxoxanthophylls, glycosylated carotenoids that are essential for thylakoid and cell wall structure. The products identified revealed two different cleavage patterns. The first is observed with bicyclic xanthophylls and is identical with that of plant orthologues, while the second is novel and occurs upon cleavage of monocyclic substrates at the C9-C10 and C7′-C8′ double bonds. These properties enable the enzyme to produce a plenitude of different C10 and C13 apocarotenoids. Expression analyses indicated a role of NosCCD in response to highlight stress. Western blot analyses of Nostoc cells revealed NosCCD as a soluble enzyme in the cytosol, which also accomodates NosCCD substrates. Incubation of the corresponding fraction with synthetic substrates revealed the activity of the native enzyme and confirmed its induction by highlight.


Carotenoids are isoprenoid pigments synthesized by all photosynthetic organisms and some non-photosynthetic bacteria and fungi. Animals do not synthesize carotenoids de novo and rely on their diet to meet their needs of these essential compounds, which function as provitamins A, antioxidants and colorants. In photosynthetic organisms, carotenoids exert a vital role in protecting the photosynthetic apparatus from photooxidation and represent essential components of the light-harvesting and of reaction centre complexes (For review see Cunningham and Gantt, 1998; Hirschberg, 2001; Fraser and Bramley, 2004; DellaPenna and Pogson, 2006).

Environmental factors such as low temperature, CO2 limitation and high irradiance can perturb the balance between excitation energy absorption by the chlorophyll antenna and its use (Horton et al., 2001). The accumulation of excitation energy favours the production of triplet excited chlorophyll molecules (3Chl) that can interact with O2 to generate the highly toxic singlet oxygen (1O2). Carotenoids exert their protective function by quenching 3Chl and 1O2. In addition, the xanthophyll zeaxanthin produced through the violaxanthin cycle upon thylakoid luminal acidification, is known to mediate energy dissipation by quenching 1Chl, a phenomenon designated as non-photochemical quenching (npq; Müller et al., 2001). Zeaxanthin is also supposed to play a role in protecting highly unsaturated glycerolipids against peroxidation (Havaux and Niyogi, 1999). In addition, zeaxanthin and other xanthophylls may stabilize the thylakoid membranes, like cholesterol and other membrane-spanning lipids. Under elevated temperature and/or high irradiance, zeaxanthin is thought to decrease membrane fluidity and to increase its thermostability (Havaux, 1998).

Cyanobacteria, the evolutionary oldest oxygenic photosynthetic organisms, are regarded as the origin of plant chloroplasts. The major cyanobacterial carotenoids are β-carotene, its dihydroxylated derivative zeaxanthin, ketocarotenoids, such as echinenone and canthaxanthin, and carotenoid glycosides, such as myxoxanthophylls (Takaichi and Mochimaru, 2007). Cyanobacterial xanthophylls protect from highlight stress (Miœkiewicz et al., 2000; Schäfer et al., 2006; Schagerl and Müller, 2006) and/or UV-B irradiation (Ehling-Schulz et al., 1997). Low temperatures represent a further environmental factor that involves the protective function of xanthophylls. The filamentous cyanobacterium Plectonema boryanum UTEX 485, for instance, exhibited reduced chlorophyll a contents and concomitantly higher levels of myxoxanthophyll both when grown at low temperature or under highlight conditions (Miœkiewicz et al., 2000). Similarly, decreased growth temperatures led to marked increases in the amounts of polyunsaturated glycerolipids and myxoxanthophylls in the thylakoid membranes of the filamentous cyanobacterium Cylindrospermopsis raciborskii (Várkonyi et al. 2002). Myxoxanthophylls are also required to maintain normal cell wall structure and thylakoid organization in Synechocystis sp. PCC 6803 (Mohamed et al., 2005).

Apart from these functions, carotenoids serve as precursors of several physiologically important compounds synthesized through oxidative cleavage and generally known as apocarotenoids. Representative examples are the ubiquitous chromophore retinal, the chordate morphogen retinoic acid and the phytohormone abscisic acid (ABA; for review see Moise et al., 2005; Bouvier et al., 2005; Auldridge et al., 2006a). Moreover, a yet unidentified apocarotenoid is thought to act as an essential long-distance signal molecule involved in establishing normal plant architecture (Beveridge, 2006; Mouchel and Leyser, 2007). Apocarotenoids, such as saffron and bixin, play roles as plant pigments of economical value (for review see Bouvier et al., 2005). In addition, a group of C15-apocarotenoids, the strigolactones, are essential signalling molecules attracting both symbiotic arbuscular mycorrhizal fungi and parasitic plants (for review see Bouwmeester et al., 2007).

In general, the biosynthesis of apocarotenoids is catalysed by carotenoid oxygenases, non-haem iron enzymes common in all taxa (for review see Moise et al., 2005; Bouvier et al., 2005; Auldridge et al., 2006a; Kloer and Schulz, 2006). The question whether these enzymes utilize a monooxygenase or dioxygenase mechanism is still a subject of controversy (Leuenberger et al., 2001; Schmidt et al., 2006). The identification of carotenoid oxygenases has become possible through their sequence homology to VP14 (viviparous14) from maize, the first carotenoid cleavage enzyme described. VP14 catalyses the oxidative cleavage of 9-cis-violaxanthin and 9′-cis-neoxanthin leading to xanthoxin, the precursor of the phytohormone abscisic acid (Schwartz et al., 1997). This reaction seems to be a key regulatory step in the biosynthesis of ABA in several plants (Qin and Zeevaart, 1999; Iuchi et al., 2001). Based on their substrate specificity, VP14 and its orthologues have been classified as nine-cis-epoxy-carotenoid dioxygenases (NCEDs). As indicated by the plenitude of diverse apocarotenoids synthesized, plants possess a second group of carotenoid oxygenases, the so-called carotenoid cleavage dioxygenases (CCDs), which act on different substrates. The CCDs of higher plants contribute to diverse physiological processes, including the regulation of the outgrowth of lateral shoot buds (Schwartz et al., 2004; Booker et al., 2004), plastid development (Næsted et al., 2004) and the formation of volatile compounds such as β-ionone (Schwartz et al., 2001).

C13-apocarotenoids, also known as norterpenoids/norisoprenoids such as β-ionones and damascones, constitute an essential aroma note in tea, grapes, roses, tobacco and wine (Rodríguez-Bustamante and Sánchez, 2007). The first identified enzyme mediating the formation of C13-apocarotenoids was the CCD1 from Arabidopsis thaliana (AtCCD1; Schwartz et al. 2001). Orthologues have been characterized from several plant species, such as crocus, tomato, grape, melon and petunia (Bouvier et al., 2003; Simkin et al., 2004a,b; Mathieu et al., 2005; Ibdah et al., 2006). Plant CCD1s cleave numerous cyclic and linear all-trans-carotenoids at the C9-C10 and C9′-C10′ double bonds into a C14-dialdehyde and two C13-ketones. The role of CCD1s in the formation of carotenoid-derived volatiles in planta was suggested by loss-of-function experiments in tomato fruits and petunia flowers, which led to significant decreases in the emission of C13-apocarotenoids such as β-ionone (Simkin et al., 2004a,b). CCD1 activity can dominate carotenoid turnover in planta as shown with Arabidopsis ccd1 mutants exhibiting elevated carotenoid contents in their seeds (Auldridge et al., 2006b).

Algae and cyanobacteria release volatile organic compounds, including apocarotenoids, which debase freshwater, especially upon eutrophication (Jüttner, 1984). Secreted apocarotenoid volatiles, such as geranylacetone and β-ionone, may play ecological roles in providing competitive advantages, e.g. by inhibiting the growth of surrounding phytoplankton (Jüttner, 1979; Ikawa et al., 2001). Carotenoid-derived volatiles are produced by divergent microorganisms such as the cyanobacteria Synechocystis and Microcystis, the red alga Cyanidium caldarium and the green alga Trentepihlia iolithus (for review, see Bouvier et al., 2005). However, little is known about the conditions that induce their formation and the enzymes involved.

The carotenoid oxygenase family is represented in microorganisms like fungi (Prado-Cabrero et al., 2007) and eubacteria, including cyanobacteria. In fact, the lignostilbene-α,β-dioxygenase from Pseudomonas paucimobilis, which mediates lignin degradation instead of carotenoid cleavage, was the first molecularly identified member of the family (Kamoda and Saburi, 1993). Recently, we have characterized the membrane-bound apocarotenoid-cleaving and retinal-forming enzymes SynACO and NosACO from Synechocystis sp. PCC 6803 and Nostoc sp. PCC 7120 respectively (Ruch et al., 2005; Scherzinger et al., 2006).

The Nostoc sp. PCC 7120 carotenoid oxygenase, encoded by the ORF all1106 and previously named as NSC1, was shown to cleave the C9-C10 double bond of β-apo-8′-carotenal (C30) leading to β-ionone (C13) and the corresponding dialdehyde (C17; Marasco et al., 2006). Due to its homology to plant CCD1s, we refer here to this enzyme as NosCCD and report on it as the first myxoxanthophyll converting oxygenase. NosCCD cleaves the C7′-C8′/C9-C10 and the C9′-C10′/C9-C10 double bonds of monocyclic and bicyclic carotenoids respectively. These activities lead to the formation of both C13 and C10 apocarotenoids depending on the structure of the carotenoid substrates. The abundance of NosCCD is induced by highlight and, surprisingly, the enzyme is found in the soluble fraction of Nostoc cells, which also accommodates the suitable substrates.


In vitro assays with purified NosCCD

To characterize NosCCD in vitro, the corresponding gene was cloned and the enzyme expressed in Escherichia coli cells as GST fusion. The protein was purified using glutathione sepharose and released by the protease Factor Xa (Fig. S1). To investigate whether the enzyme cleaves apocarotenals other than β-apo-8′-carotenal, purified NosCCD was incubated for 30 min with β-apo-8′-carotenal (C30) as positive control, β-apo-10′-carotenal (C27) and their corresponding 3-OH-derivatives (for structure see Fig. S2C). Subsequent HPLC analyses (Fig. S2A) of the incubations with the hydroxylated substrates revealed the formation of 3-OH-β-ionone (C13) as a common product (Fig. S2A, peak a; structure is depicted in Fig. S2C, a) beside the compounds apo-8′,10-apocarotene-dial (C17) and apo-10′,10-apocarotene-dial (C14) from 3-OH-β-apo-8′-carotenal (C30) and 3-OH-β-apo-10′-carotenal (C27) respectively (Fig. S2A, peaks b and c). The identities of these products were confirmed by LC-MS analyses showing the expected [M + H]+ molecular ion of 209.18, 257.17 and 217.17 for 3-OH-β-ionone, the C17- and the C14-dialdehyde respectively. Moreover, the MS2-spectra of the two dialdehydes were consistent with their suggested structures (Fig. S2B and C, compounds b and c). Analogously, β-ionone was formed with the non-hydroxylated β-apo-8′-carotenal and β-apo-10′-carotenal as the substrate (data not shown). These data demonstrate that NosCCD cleaves the C9-C10 double bond of cyclic apocarotenals independent of their chain length and the occurrence of the 3-hydroxy functional group.

Nostoc sp. PCC 7120 cells accumulate several monocyclic and bicyclic carotenoids (Takaichi et al., 2005) representing putative substrates for carotenoid oxygenases. To explore the cleaving activities of NosCCD with these compounds, we incubated the purified enzyme with the myxoxanthophyll myxol 2′-fucoside (Fig. 2, compound 7), one of the major pigments accumulated in Nostoc sp. PCC 7120 (Takaichi et al., 2005). The in vitro assays were performed for 2 h, and lipophilic components were then extracted and analysed by HPLC (Fig. 1-I). The enzyme converted the myxoxanthophyll into three products, two of which (b and c; Fig. 1-I) exhibited elution patterns, UV-Vis- and MS-spectra (the latter not shown) identical to those obtained from the incubation with 3-OH-β-apo-8′-carotenal (C30). Thus, the compounds b and c represented 3-OH-β-ionone (C13) and apo-8′,10-apocarotene-dial (C17) respectively. The formation of these two compounds implies two cleavage sites at the C9-C10 and C7′-C8′ double bonds (Fig. 2, compound 7). The first site is localized on the cyclic side of the molecule, while the second is at its linear fucosylated end. Accordingly, the cleavage of the myxoxanthophyll should lead to a third product with a C10 isoprenoid chain. This product is represented by the third compound (Fig. 1-I, peak a). It eluted earlier than 3-OH-β-ionone, indicating its higher polarity that is in accordance with the presumed structure. To prove its identity, the compound was collected from HPLC and subjected to LC-MS analysis. As shown in Fig. 1-II, the predominant ion coeluting with compound a showed a mass of 389.04 that corresponds to the acetate-adduct [M + CH3COO]- of the assumed fucosylated C10-aldehyde. The formation of such adducts during LC-MS analyses of carotenoid glycosides was frequently observed (data not shown). To elucidate the NosCCD activities with other monocyclic substrates, two further Nostoc myxoxanthophylls, 4-ketomyxol 2′-fucoside and 4-hydroxymyxol 2′-fucoside (Takaichi et al., 2005), and the intermediates myxol and γ-carotene (Takaichi et al., 2006; for structures see Fig. 2, compounds 5 and 6) were employed. The assays were incubated for 2 h and analysed by HPLC (data not shown). All four substrates were cleaved at the C9-C10 and C7′-C8′ double bonds, according to the pattern determined for the monocyclic myxoxanthophyll myxol 2′-fucoside.

Figure 2.

The structure of the carotenoid substrates and the products formed in vitro. Bicyclic substrates like zeaxanthin (1), echinenone (2), canthaxanthin (3) and astaxanthin (4) were cleaved at the C9-C10 and C9′-C10′ double bonds, resulting in two C13 products and apo-10′,10-apocarotene-dial (C14). The cleavage of monocyclic substrates, i.e. γ-carotene (5), myxol (6) and myxoxanthophylls (e.g. myxol 2′-fucoside; 7) occurred at the C9-C10 and C7′-C8′ double bonds, yielding C10- and C13-compounds and apo-8′,10-apocarotene-dial (C17).

Figure 1.

HPLC analyses of the incubations of NosCCD with myxoxanthophylls and zeaxanthin. The in vitro assays were performed using 50 μg of purified NosCCD in a total volume of 200 μl.
I. Incubation with the myxoxanthophyll myxol 2′-fucoside (M) and zeaxanthin (Z) led to an obvious change in colour (shown in the insets) indicating cleaving activity. HPLC analyses showed the conversion of zeaxanthin (Z) into the products (b) and (d) representing 3-OH-β-ionone (C13) and apo-10′,10-apocarotene-dial (C14). The peak (e) does not represent a carotenoid. The myxoxanthophyll (M) was converted into 3-OH-β-ionone (product b), apo-8′,10-apocarotene-dial (C17; product c) and product a (fucosylated C10-aldehyde). The products b, c, d were identified based on their UV-Vis spectra (insets) and elution patterns (see also Fig. S2). The nature of the products imply a cleaving of zeaxanthin at the C9-C10 and C9′-C10′ and of myxoxanthophyll at the C9-C10 and C7′-C8′ double bonds (for structures Fig. 2).
II. The LC-MS analysis of product a confirmed its identity as a fucosylated C10-aldehyde showing the expected molecular ion [M + CH3COO]- of 389.04.

To investigate the activity of NosCCD with bicyclic cyanobacterial carotenoids, 2 h incubations were carried out using purified enzyme and zeaxanthin (Fig. 2, compound 1) as substrate. As suggested by HPLC analyses (Fig. 1-I), these incubations led to the formation of products identical to those obtained from the conversion of 3-OH-β-apo-10′-carotenal (C27, Fig. S2A), i.e. 3-OH-β-ionone (C13) and apo-10′,10-apocarotene-dial (C14; Fig. S2C, compounds a and c). These data prove that the bicyclic substrate was cleaved at the C9-C10 and C9′-C10′ double bonds and are consistent with the previously described pattern of the cleaving of β-carotene (Marasco et al., 2006). This represents a marked difference in the regional specificity of cleavage as compared with open-chain substrates (see above). As deduced from the formation of the common C14-dialdehyde (apo-10′,10-apocarotene-dial) from all of them (data not shown), the bicyclic echinenone, canthaxanthin and astaxanthin (Fig. 2, compounds 2, 3, 4 respectively) were cleaved, like zeaxanthin, at the C9-C10 and C9′-C10′ double bonds yielding the respective cyclic products, namely β-ionone, 4-oxo-β-ionone and 3-OH-4-oxo-β-ionone.

The relative amounts of the dialdehydes produced were also used to estimate the cleaving efficiency with respect to the different substrates (Table 1). In standard 2 h assays, we observed about 75% conversions of the monocyclic myxoxanthophylls and myxol, while only about 15% of the bicyclic xanthophylls zeaxanthin, echinenone, canthaxanthin and astaxanthin were cleaved. The enzyme exhibited weak activities when incubated with carotenes and converted the monocyclic γ-carotene only at a low level of about 5%. Similarly, only barely detectable activity was observed upon the incubation with β-carotene. The structures of the substrates and the products formed are depicted in Fig. 2. However, the preference of NosCCD for monocyclic xanthophylls may be due to their higher solubility compared with the non-oxygenated γ-carotene or β-carotene. Therefore, our data do not allow conclusions on the enzyme's preferences in cellula.

Table 1.  Cleaving efficiency of cyanobacterial carotenoids in standard 2 h in vitro assays.
SubstrateConversion rate
  1. Assays were performed using 50 μg purified NosCCD in a total volume of 200 μl and analysed by HPLC. Conversion rates were calculated from peak areas of the reaction products relative to those of the substrates. Values represent the average of four independent incubations ± SD.

myxoxanthophylls, myxol74 ± 7%
Zeaxanthin, canthaxanthin, echinenone, astaxanthin15 ± 4%

Expression analyses of NosCCD

Several cyanobacterial species increase their xanthophyll content in response to lowered temperatures. Such changes can be achieved by upregulating the respective biosynthetic capacities, or by decreasing degradation processes. The carotenoid degrading activities of NosCCD and, especially its assumed preference for xanthophylls signals a possible role in the latter process.

To investigate this possibility, we determined the carotenoid levels and compositions in 24 h intervals in cultures pre-grown at 30°C and subsequently transferred to 20°C. In parallel, the expression of NosCCD was analysed as well as of CrtL(diox), a key enzyme in the synthesis of myxoxynthophylls (Mohamed and Vermaas, 2006). The transcript levels of Crhc, a known cold-inducible RNA helicase (Chamot and Owttrim, 2000), were determined as a control. As shown in Fig. 3A, lowering the temperature led to a decrease in both chlorophyll a and total carotenoids. The impact on chlorophyll a (−27%) was more pronounced than on total carotenoids (−19%), leading to reduced Chla/Car ratios. However, the carotenoid composition changed in favour of the xanthophylls, while β-carotene decreased (Fig. 3A).

Figure 3.

Carotenoid content and expression of NosCCD during long-term acclimation to lowered temperature.
A. Relative contents of β-carotene and xanthopylls were monitored over a period of 5 days after transfer to 20°C. The values shown in the table below were determined at the starting and end-points of the treatment. Total carotenoid and chlorophyll a contents were measured spectrophotometrically. The relative amount of each carotenoid was calculated from peak areas of a MaxPlot 350–550 nm.
B. Transcript levels of NosCCD, CrtL(diox) and CrhC during the long-term cold acclimation. The analyses were performed using real-time RT-PCR. The transcript levels are expressed as multiples of the T0 controls, which were normalized to 1. Values are the average of three independent experiments ± SD.

The enhanced xanthophyll levels may reflect decreased levels of NosCCD. However, the data presented in Fig. 3B show that low temperature did not lead to the assumed downregulation of NosCCD transcript levels which, similar to those of the carotenoid biosynthetic gene CrtL(diox), were rather slightly enhanced (about twofold). The control gene Crhc was upregulated, as expected. Thus, in essence, the increased xanthophyll contents did not correlate with the NosCCD expression levels.

To account for possible short-term responses, the cold treatment was applied for 48 h and transcript levels were then determined at the time points depicted in Fig. 4A. Here again, we did not observe a reduction of NosCCD transcript levels.

Figure 4.

Real-time RT-PCR analyses of the transcript levels of NosCCD and CrtL(diox) under cold or highlight stress.
A. Expression of NosCCD (black, top) and CrtL(diox) (grey, bottom) in cultures exposed to cold stress for 0.5–48 h. Cultures were pre-grown at 30°C and then shifted to 20°C.
B. Expression of both genes in response to exposure to highlight for 0.5–6 h. Cultures were grown at a light intensity of 20 μmol m−2s−1 and exposed then to 280 μmol m−2s−1. Transcript levels were determined at the time points shown and are expressed as multiples of the To controls, which were normalized to 1. Values represent the average of three independent experiments ± SD.

Highlight treatment led to a steady induction of NosCCD (Fig. 4B), reaching levels that were about three times higher than those observed under lowered temperature. The induction of the carotenoid biosynthetic CrtL(diox) gene was similarly strong. This simultaneous upregulation of carotenoid biosynthesis and degradation translated into an about 8% increase of the relative xanthophyll amounts, and to an about 20% decrease of the Chla/Car ratios representing highlight specific steady-state conditions (data not shown).

Localization of NosCCD

Carotenoid cleavage in vivo requires accessibility by NosCCD which is therefore expected to be membrane-bound. However, Western blot analyses of soluble and membrane fractions of Nostoc cells showed the localization of NosCCD in the soluble fraction (Fig. 5).

Figure 5.

Western blot analysis of NosCCD in membrane and soluble fractions of Nostoc cells.
A. Coomassie-stained SDS-PAGE gel of the different cell fractions. An amount of 25 μg total protein was loaded on each lane. The lanes refer to the following fractions: total cell lysate (1), soluble fraction (2) and membrane fraction (3).
B. Western blot analysis of the different fractions using anti-NosCCD-antibodies. NosCCD was detected in the total lysate and the soluble fraction in which the enzyme is enriched due to the fractionation of the cells.

To confirm, we analysed the capability of the soluble Nostoc cell fraction to cleave the model substrates 3-OH-β-apo-10′-carotenal and 3-OH-β-apo-8′-carotenal. Soluble fractions from cells grown under standard and highlight conditions were investigated expecting higher levels of the enzyme in the latter. In a control experiment, we used soluble fractions isolated from the Synechocystis sp. PCC 6803 ΔSynDiox2 knock-out-strain which contains the retinal-forming and membrane-bound SynACO as the sole carotenoid oxygenase (Scherzinger et al., 2006). As shown in Fig. 6A, 3-OH-β-apo-10′-carotenal was converted by the soluble fraction into apo-10′,10-apocarotene-dial, as with the purified protein. Highlight conditions resulted in enhanced cleavage activity leading to increased product amounts, as determined by product peak integration (Fig. 6B). The highlight induction increased with the time of exposure, up to 8 h. Similar results were obtained from the incubations of 3-OH-β-apo-8′-carotenal (data not shown). In contrast, no cleavage activity was detected with soluble fractions isolated from the Synechocystis sp. PCC 6803 ΔSynDiox2 knock-out strain.

Figure 6.

HPLC analyses after incubation of 3-OH-β-apo10′-carotenal with soluble fractions.
A. 3-OH-β-apo10′-carotenal was incubated with soluble fractions obtained from Synechocystis ΔSynDiox2 (1), untreated Nostoc cultures (2) and Nostoc cultures exposed to highlight for 4 (3) and 8 h (4). The HPLC analyses show the formation of apo10′,10 apocarotene-dial (a; inset) after four hours of incubation with the Nostoc fractions (2–4), which was not detectable in the Synechocystis control (1). The highlight treatment led to an obvious increase in the cleaving activity, which correlated with the exposure time.
B. Relative amounts of the dialdehyde product formed by untreated and highlight-treated soluble Nostoc fractions, calculated from peak areas of a MaxPlot 300–550 nm. The values show the increase of the activity due to highlight stress.

The Nostoc cytoplasm contains NosCCD substrates

The cytosolic localization of NosCCD implies the occurrence of cytosolic carotenoids. To clarify, lipophilic compounds were extracted from the cytosolic fractions of Nostoc cells and analysed by HPLC. In addition, we determined the ‘soluble’ carotenoids of Synechocystis that lacks CCD1 activity. As shown in Fig. 7, the cytoplasm of both cyanobacteria harbour significant amounts of carotenoids. According to chromatographic behaviour and UV-Vis spectra, the soluble fraction of Synechocystis contained mainly 3′-OH-echinenone beside traces of echinenone and zeaxanthin, while the carotenoid pattern of the Nostoc-cytoplasm was more complex consisting of the NosCCD-substrates myxoxanthophylls, canthaxanthin, echinenone and traces of β-carotene. However, the amount of carotenoids in Nostoc-cytoplasm was much lower than in Synechocystis (Fig. 7C). The compositions of the carotenoids in the soluble fractions differed markedly from those of the corresponding total extracts (Fig. 7A and B). Therefore, it can be excluded that the carotenoids detected in the cytoplasm were due to membrane contaminations.

Figure 7.

HPLC analyses of cytosolic carotenoids in Nostoc (A) and Synechocystis cultures (B).
A and B. The lipophilic pigments were isolated from cytosolic fractions (I) and whole cells (II). Pigments were identified, based on their UV-Vis spectra and chromatographic properties, as myxoxanthophylls (1), canthaxanthin (2), echinenone (3), β-carotene (4), zeaxanthin (5), 3′-OH-echinenone (6) and chlorophyll a (Ch).
C. Carotenoid content and pattern of the Nostoc (N) and Synechocystis (S) cytosolic fractions. The two cyanobacteria species exhibited marked differences in the amount and the pattern of cytosolic carotenoids. Total carotenoid content was measured spectrophotometrically. The relative amounts of each carotenoid was calculated from peak areas of a MaxPlot 350–550 nm. Data represent the average of three independent measurements ± SD.


The filamentous freshwater cyanobacterium Nostoc sp. PCC 7120 contributes to cyanobacterial blooms that represent a serious problem for the management of drinking water (Kim et al., 2006). The excessive growth of algae and cyanobacteria is associated with an intensive release of volatile compounds like the C13-ketone β-ionone (Jüttner, 1984). β-ionone and related compounds are also produced by higher plants through breakdown of carotenoids, catalysed by CCD1 enzymes (Bouvier et al., 2005; Auldridge et al., 2006a). The Nostoc sp. PCC 7120 enzyme NosCCD, previously named NSC1, was shown to form β-ionone from β-apo-8′-carotenal (Marasco et al., 2006). NosCCD exhibits 44% identity to AtCCD1 and is a member of a group of cyanobacterial carotenoid oxygenases that may represent the plant CCD1 precursor. The characterization of NosCCD revealed a novel cleavage pattern, unexpected solubility and induction by stress conditions.

Using in vitro assays with the purified enzyme, we demonstrated that NosCCD is a carotenoid oxygenase that cleaves bicyclic carotenoids at the C9-C10 and C9′-C10′ double bonds leading to the C14-compound apo-10′,10-apocarotene-dial and to a variety of C13-apocarotenoids, depending on the ring modifications of the substrates. This cleavage pattern is consistent with the previous report on this enzyme (Marasco et al., 2006) and identical to the one of plant CCD1s (Bouvier et al., 2005; Auldrige et al., 2006a). In contrast, the simultaneous cleavage of the C7′-C8′ and C9-C10 double bonds in monocyclic substrates is novel and has not been shown for any carotenoid cleaving enzyme so far. It is very surprising to note that NosCCD (and plant CCD1s; data not shown) can cleave carotenoid monoglycosides (myxoxanthophylls) that are very hydrophilic and bulky compared with ‘standard’ carotenoids. A very high degree of sterical flexibility of the NosCCD substrate binding site must be expected that also readily accomodates much smaller and less polar compounds like 3-OH-β-apo-10′-carotenal (C27) and 3-OH-β-apo-8′-carotenal (C30). The cleavage of the latter substrates led – irrespective of their chain lengths – to 3-OH-β-ionone formation. This strict regional specificity of cleavage with monocyclic apocarotenals is paralleled by previously published findings with the cyanobacterial oxygenases SynACO and NosACO, which yield retinal and its oxygenated homologues from a variety of monocyclic apocarotenals irrespective of their chain lengths (Ruch et al., 2005; Scherzinger et al., 2006). Based on the structural elucidation of SynACO (Kloer et al., 2005), it appears reasonable to assume that these substrates enter the NosCCD binding tunnel with their linear ends and that the regional specificity of cleavage is attained by arresting the β-ionone ring at the entrance. The distance between the β-ionone ring and the non-haem iron reaction centre, which is identical in the C30 and C27 substrates, then determines the common C9-C10 cleavage site. However, this carotenoid oxygenase structure alone and the only available so far does not allow to explain the large diversity of substrates converted and the two different cleavage patterns observed.

During the preparation of this manuscript, a publication by Vogel et al. (2008) appeared reporting on a re-investigation of the plant CCD1s cleaving activities in carotenoid-accumulating E. coli cells. Based on GC-MS analyses, the authors showed that plant CCD1s converted the non-cyclic lycopene into 6-methyl-5-hepten-2-one, a widely occurring plant volatile. These data showed that the plant CCD1s cleaved the C5′-C6′/C5-C6 double bonds of linear carotenoids, in addition to their known C9′-C10′/C9-C10 target sites, and demonstrate their capability to distinguish between linear and cyclic substrate end-groups. This capability is shared by the cyanobacterial enzyme in principle, however, NosCCD recognizes exclusively the C7′-C8′ double bond of linear carotenoid ends.

The previous report on the enzymatic activity of NosCCD (designated NSC1; Marasco et al., 2006) described in vitro cleavage of β-apo-8′-carotenal at the C9-C10 double bond leading to β-apo-8′,10-carotene-dial. The authors also observed the formation of the two additional dialdehydes β-apo-8′,8-carotene-dial and β-apo-10′,10-carotene-dial. The former could be the result of a single enzymatic cleavage of the C7-C8 double bond localized at the cyclic end, while the latter may be formed through the enzymatic cleavage of the C9′-C10′/C9-C10 double bonds. However, the very low amounts of these two additional products and the described instability of the major product β-apo-8′,10-carotenedial in the in vitro system used there indicate that these additional dialdehydes may be formed by a non-specific activity. In contrast, we did not observe any cleavage of the C7-C8 double bond or any significant formation of β-apo-10′,10-carotene-dial or β-apo-8′,8-carotene-dial upon the incubation with β-apo-8′-carotenal or its 3-hydroxy derivative. The novel C7′-C8′ cleavage site determined here is restricted to the acyclic end of monocyclic C40-carotenoids, such as myxoxanthophyll, which were not investigated by Marasco et al. (2006). Upon NosCCD expression, Marasco et al. (2006) observed decreased substrate contents in several E. coli strains accumulating different carotenoids and concluded that the enzyme was capable to cleave a variety of carotenoid substrates; however, no cleavage products were detected. Indeed, carotenoid-accumulating E. coli strains represent a fast and simple system frequently used to elucidate the activities of cleaving enzymes (Cunningham et al., 1993; Redmond et al., 2001). However, decreased carotenoids alone do not necessarily imply a cleavage activity in this model system but can be the result of interference with carotenoid biosynthesis through pleiotropic effects caused by the overexpression. We occasionally observe bleaching of such cells even when the expressed oxygenase gene is inactivated by mutation.

Confirming the solubility of NosCCD, we did not detect any cleaving activity in carotenoid-accumulating E. coli strains that accomodate the substrates in membranes (data not shown). Therefore, we employed in vitro studies that allowed to investigate a wide set of substrates, many of which can not be synthesized in E. coli so far. The cleavage pattern and the wide substrate specificity with respect to end-group oxygenation even including glycosylation predestine NosCCD for the production of a plenitude of C10- and C13-compounds, including the volatiles β-ionone (C13) and geranial (trans-citral; C10).

As suggested by cell fractionation, NosCCD occurs in the soluble fraction of the Nostoc cells and does not associate with membranes. This is corroborated by the cleavage activity detected in the cytosol of normally grown and highlight-treated Nostoc cells. It is reasonable to assume that these activities are due to NosCCD, despite the presence of the two additional carotenoid oxygenases NosACO (Accession No: BAB75983; Scherzinger et al., 2006) and NosDiox2 (Accession No: BAB76594, named also as NSC3; Marasco et al., 2006). This is because first, the cleavage patterns observed with the purified enzyme and the soluble cell extract were identical. Second, NosACO (Scherzinger et al., 2006) and NosDiox2 (D. Scherzinger et al., unpubl. data) have an entirely different cleavage pattern, and third, NosACO is membrane-bound and relies on membrane-localized apocarotenoids for activity (Scherzinger et al., 2006). Finally, the absence of detectable activities in the soluble fraction of the Synechocystis sp. PCC 6803 ΔSynDiox2 knock-out strain (Scherzinger et al., 2006) points to the specificity of the Nostoc soluble activities. In the light of the endosymbiont origin, it is worth mentioning that the plant CCD1 orthologues maintain an unexpected localization; namely, they are not localized in plastids where carotenoids are synthesized and where they usually occur. Plant CCD1s do not possess transit peptides and represent cytosolic enzymes (Bouvier et al., 2005; Auldrige et al., 2006a). However, it was shown that tomato CCD1s may be associated with the outer plastid membrane (Simkin et al., 2004a).

The solubility of NosCCD and the cleaving of lipophilic carotenoids is contradictory at first glance. However, xanthophyll substrates are present in the cytoplasm of Nostoc cells with canthaxanthin and echinenone as the major components, while the Synechocystis cytoplasm contained mainly 3′-OH-echinenone. Beside these qualitative differences, the carotenoid content in Synechocystis cytosol was much higher than in Nostoc. This lower level might be due to the NosCCD degrading activity which is absent in Synechocystis.

In cyanobacteria, cytosolic carotenoids are known to occur non-covalently bound to proteins, including the orange carotenoid proteins (OCPs) and the red carotenoid proteins (RCPs; for review see Kerfeld, 2004). The nature of the carotenoid bound by the Synechocystis OCP, a 35 kDa protein, is still the subject of controversy (Kerfeld, 2004) but may be 3′-OH-echinenone (Wu and Krogmann, 1997). Highly conserved homologues of the Synechocystis OCP are found in all sequenced cyanobacterial genomes, with the exception of Prochlorococci (Kerfeld, 2004). The Nostoc OCP encoded by ORF all3149 exhibits 85% similarity to the Synechocystis OCP. Other Nostoc ORFs (all4940, all4941, all1123, all3221 and all4783) encode polypeptides sharing homology with either the C-terminal or the N-terminal domain of OCPs and may constitute further soluble carotenoid binding proteins (Kerfeld, 2004).

The colocalization of OCPs and NosCCD may indicate a functional interaction determining cytosolic carotenoid levels. It has been suggested that OCPs function as carotenoid transport proteins shuttling xanthophylls and myxoxanthophylls between thylakoid membranes and the outer membrane and the cell wall (Kerfeld, 2004), which are known to harbour carotenoids (Jürgens and Weckesser, 1985; Jürgens et al., 1989; Mohamed et al., 2005). Several lines of evidence also demonstrated a photoprotective role of OCPs. For instance, it was shown that these proteins are effective quenchers of singlet oxygen (Kerfeld, 2004) and recent analyses of a Synechocystis-ΔOCP strain revealed their role in the dissipation of excess light energy in the phycobilisome (Wilson et al., 2007). Supporting photoprotective functions, the transcript levels of OCP in Synechocystis increased markedly upon highlight stress (Hihara et al., 2001). Similar results were reported on soluble carotenoid proteins from Anacystis nidulans (Masamoto et al., 1987) and Synechococcus sp. strain PCC 7942 (Reddy et al., 1989). The photoprotective role of OCPs may be associated with a destruction of the pigments bound, caused by reactive oxygen species leading to apocarotenoid-formation. Considering the highlight induction of the NosCCD transcript levels and its apocarotenoid-cleaving activities, shown for C30 and C27 substrates, it may be speculated that this enzyme is responsible for scavenging cytosolic apocarotenoids. In this case, β-ionone and related cleavage products would represent by-products of the carotenoid photoprotective activities. This assumed function differs markedly from the homologous enzyme NosACO (Scherzinger et al., 2006) which provides retinal, the chromophore of the Nostoc (Anabaena) sensory rhodopsin (Jung et al., 2003), by cleaving membrane embedded apocarotenals. However, expression analyses revealed, like in the case of NosCCD, a marked induction of NosACO upon highlight stress (data not shown). This points to a possible second function of NosACO, namely the scavenging of membrane apocarotenoids which may arise from photodestructive processes.

Experimental procedures

Culture of cyanobacteria and stress treatments

Cultures of Nostoc sp. PCC 7120 were cultivated in liquid BG110 medium supplemented with 5 mM NaHCO3. Synechocystis sp. PCC 6803 and the Synechocystis sp. PCC 6803 ΔSynDiox2 knock-out strain (Scherzinger et al., 2006) were grown in liquid BG11 medium or BG11 medium containing 15 μg ml−1 chloramphenicol respectively (Rippka et al., 1979). Cells were cultured in Erlenmeyer flasks with constant illumination of 20 μmol m−2s−1 at 30°C.

For cold or highlight experiments the growth temperature was shifted to 20°C or cultures were exposed to a light intensity of 280 μmol m−2s−1 respectively. Cultures were harvested by filtration at the indicated time points.

Cloning of all1106/NosCCD

The gene all1106 which codes for NosCCD (NCBI Accession No. NP_485149) was amplified with the oligonucleotides NosCCD-for: 5′-ATGGTAAAAGATTCACTCACTTTC-3′ and NosCCD-back: 5′-TTAAATTCTCCTAAAATTCAACTGTTC-3′. The PCR reaction was carried out using approximately 100 ng of genomic DNA, 100 ng of each primer, 200 μM dNTPs and 1 μl of Advantage® cDNA Polymerase Mix (BD Biosciences, Franklin Lakes, NJ) in the buffer provided, as follows: 2 min initial denaturation at 94°C followed by 32 cycles (30 s 94°C, 30 s 60°C, 2 min 68°C) and 10 min final polymerization at 68°C. The resulting PCR product was purified using GFXTM PCR DNA and Gel Band Purification Kit (Amersham Biosciences, Piscataway, NJ), and cloned into the pCR2.1®-TOPO® vector (Invitrogen, Paisley, UK) to yield pNosCCD-CR2.1. The product was verified by sequencing.

Protein expression and purification

NosCCD was expressed as a glutathion-S-transferase (GST) fusion. For this purpose, the corresponding gene was excised from pNosCCD-CR2.1 with EcoRI, and ligated into EcoRI digested and calf intestine alkaline phosphatase-treated pGEX-5X-1 (Amersham Biosciences, Piscataway, NJ) to yield the plasmid pGEX-5X-NosCCD. BL21 E. coli cells were then transformed and grown at 37°C in 2xYT-medium. Induction was performed at an OD600 of 0.5 with 0.2 mM IPTG. After an incubation for additional 4 h at 28°C, cultures were harvested. The fusion protein was purified using glutathione-sepharose 4B (Amersham Biosciences, Piscataway, NJ), and NosCCD was released by overnight treatment with the protease factor Xa in PBS containing 0.1% Triton X-100 (v/v) at room temperature. Purification steps and protein expression were controlled by SDS-PAGE.

Enzymatic assays

Substrates were purified using thin-layer silica-gel plates (Merck, Darmstadt, Germany). Plates were developed with light petroleum/diethylether/acetone (40:10:10, v/v/v). Substrates were scraped off in dim daylight and eluted with acetone. Echinenone, myxoxanthophylls and β-carotene were purified from Nostoc sp. PCC 7120, zeaxanthin from Synechocystis sp. PCC 6803. Synthetic apocarotenals were kindly provided by BASF (Ludwigshafen, Germany). Canthaxanthin and γ-carotene were obtained from Carotenature (Lupsingen, Switzerland), and astaxanthin was from Sigma (Deisenhofen, Germany). Enzyme assays were performed in a total volume of 200 μl as described previously (Scherzinger et al., 2006) with some modifications. Fifty microlitres of ethanolic substrate solution (160 μM; 80 μM for β-carotene and astaxanthin) was mixed with 50 μl of ethanolic 4% octyl-β-glucoside solution, dried using a vacuum centrifuge and then re-suspended in 100 μl of 2× incubation buffer containing 2 mM TCEP, 0.6 mM FeSO4 and 2 mg ml−1 catalase (Sigma, Deisenhofen, Germany) in 200 mM HEPES-NaOH, pH 7.8. Purified NosCCD was then added to a final concentration of 100 ng μl−1 for apocarotenal assays or 250 ng μl−1 for incubations with C40-carotenoids, and assays were incubated for 30 min and 2 h at 28°C respectively. The incubations were stopped by adding one volume of acetone and partitioned twice against two volumes of light petroleum/diethylether (1:4, v/v). Lipophilic supernatants were combined, dried and resolved in chloroform.

Localization of NosCCD

Fractionation of cells was performed according to Murata and Omata (1988) with some modifications. About 400–500 ml of Nostoc cultures were harvested, washed in 5 mM HEPES-NaOH pH 7.0 and re-suspended in 10 mM HEPES-NaOH pH 7.0 containing 600 mM sucrose, 5 mM EDTA and 0.2% (w/v) lysozyme (Sigma, Deisenhofen, Germany). Cells were incubated under shaking for 1 h at room temperature and then washed twice with 10 mM HEPES-NaOH pH 7.0 containing 600 mM sucrose, centrifuged and finally re-suspended in PBS. The cells were broken by four passages through a French pressure cell at 18 000 psi, then PMSF was added to a final concentration of 1 mM. Cell debris were removed by centrifugation for 5 min at 4000 g. Soluble fractions and membranes were separated by ultracentrifugation for 4 h at 130 000 g and 4°C. Membranes were washed with PBS containing 1 mM PMSF, pelleted by ultracentrifugation at 130 000 g for 30 min and re-suspended in the same buffer. Equal amounts of protein were then subjected to SDS-PAGE. Western blot analyses were performed using mouse polyclonal anti-NosCCD antibodies and the ECLTM Western Blotting Analysis System (Amersham Biosciences, Piscataway, NJ).

In vitro assays with soluble fractions

About 500 ml of Nostoc or Synechocystis cultures were harvested on Supor 800 filters (Pall Life Sciences, Ann Arbor, MI), frozen in liquid nitrogen and homogenized using a Mixer Mill MM301 (Retsch, Haan, Germany) for 1 min at 30 Hz. The resulting powder was transferred to the incubation buffer, consisting of 100 mM HEPES-NaOH, pH 7.8, 2 mM TCEP, 0.5 mM FeSO4, 1 mg ml−1 catalase and a mixture of protease inhibitors (0.2 mM PMSF, 1 mM benzamidine, 5 μg ml−1 leupeptin, 3.5 μg ml−1 pepstatin, 20 μg ml−1 antipain). The suspension was centrifuged for 10 min at 3200 g and 4°C, and the supernatant was subjected to ultracentrifugation for 2 h at 100 000 g and 4°C. The protein concentration of the resulting supernatant was measured and adjusted to equal concentrations with incubation buffer. For in vitro assays, substrates were mixed with an ethanolic solution of octyl-β-glucoside, dried in a vacuum centrifuge and solubilized directly in 400 μl of the supernatants containing 2 μg μl−1 of cyanobacterial proteins. The final concentrations for the substrates were 10 μM and 1% (w/v) for octyl-β-glucoside. Incubation time ranged from 4 to 12 h at 28°C. The assays were extracted as given above.

Determination of carotenoid content and composition

Cultures were harvested on GF/C microfibre filters (Whatman, Maidstone, UK) and homogenized as given above. Broken cells were extracted twice with acetone/MeOH (7:3, v/v), 2 mM Tris-HCl, pH 7.8 and once with light petroleum. Carotenoid and chlorophyll a content was measured spectrophotometrically and determined using the equations of Chamovitz et al. (1993) and Porra (2002) respectively. For analyses of carotenoid composition the extracts were subjected to HPLC analyses.

Determination of cytoplasmic carotenoid content

About 500–800 ml of Nostoc and Synechocystis cultures were harvested on GF/C filters and homogenized frozen as given above. The resulting powder was transferred to a buffer containing 100 mM Tris-HCl, pH 7.8, 1 mM DTT and 1 mM PMSF. Cell debris were removed by centrifugation at 3200 g for 10 min at 4°C. The resulting supernatant was ultracentrifuged for 1.5 h at 100 000 g at 4°C. The supernatant was ultracentrifuged again for 4 h at 130 000 g and 4°C. Membrane impurities were further precipitated with 20% ammonium sulphate (Wu and Krogmann, 1997) for 30 min on ice and subsequent centrifugation at 8000 g and 4°C for 20 min. The protein content of the cytoplasmic fractions was measured as mentioned below. The lysates were extracted twice with light petroleum/diethylether (1:4, v/v). Carotenoid content was measured spectrophotometrically and calculated using the E1% value of the main carotenoid.

Analytical methods

Substrates were quantified spectrophotometrically at their individual λmax using extinction coefficients as given by Barua and Olson (2000) or Davies (1976). Protein concentration was determined using the Bio-Rad protein assay kit (Bio-Rad, Hercules, CA). For HPLC analyses, a Waters system (Eschborn, Germany) equipped with a photodiode array detector (model 996) was used. A YMC-Pack C30-reversed phase column (250 × 4.6 mm i.d., 5 μm; YMC Europe, Schermbeck, Germany) was used with the solvent systems B: MeOH/tert-butylmethyl ether/water (120:4:40, v/v/v) and A: MeOH/tert-butylmethyl ether (500:500, v/v). The column was developed at a flow rate of 1 ml min−1 with a linear gradient from 100% B to 43% B within 45 min, then to 100% A and a flow rate of 2 ml min−1 within 1 min, maintaining the final conditions for another 14 min at a flow rate of 2 ml min−1. For analysis of in vitro assays with Nostoc cell lysates a more polar B solvent was used (MeOH/tert-butylmethyl ether/water (50:5:45, v/v/v) with the same program.

For analyses of cyanobacterial carotenoid patterns the following system consisting of the solvents A: MeOH/water (2/1, v/v), B: MeOH/water (9/1, v/v) and C: MeOH was chosen. Separation was carried out on a MN Nucleosil 100 C18-reversed phase column (250 × 4 mm i.d., 10 μm; Macherey-Nagel, Düren, Germany). The column was developed at a flow rate of 2 ml min−1 with a linear gradient from 100% A to 100% B within 5 min, maintaining these conditions for 5 min, then to 87.5% B and 12.5% C in 2 min and to 100% C in 1 min maintaining these conditions for 25 min.

LC-MS analyses

The cleavage products were collected from HPLC. LC-MS analyses were performed using a Thermo Finnigan LTQ mass spectrometer coupled to a Surveyor HPLC system (Thermo Electron, Waltham, MA). Separations were carried out using a Thermo Hypersil GOLD C18-reversed phase column (150 × 4.6 mm i.d., 3 μm) with the solvent system A: MeOH/water/tert-butylmethyl ether (50/45/5, v/v/v) and B: MeOH/water/tert-butylmethyl ether (27/3/70, v/v/v) with the water containing 0.1 g l−1 ammonium acetate. The column was developed at a flow rate of 450 μl min−1 with 90% A and 10% B for 5 min, followed by a linear gradient to 5% A and 95% B in 10 min, then increasing the flow to 900 μl min−1 within 4 min and maintaining these final conditions for 5 min. The identification of 3-OH-β-ionone and carotenoid dialdehydes were carried out using APCI in positive mode. Nitrogen was used as sheath and auxiliary gas which were set to 20 and 5 units respectively. The vaporizer temperature was 225°C and capillary temperature 175°C. The source current was set to 5 μA, the capillary voltage and tube lens settings were 49 and 125 V respectively. For analysis of the fucosylated C10 cleavage product, APCI in negative mode was used. The column was developed at a flow rate of 450 μl min−1 with 100% A for 5 min, followed by a linear gradient to 5% A and 95% B within 10 min, then increasing the flow to 900 μl min−1 within 4 min and maintaining these final conditions for 5 min. Sheath and auxiliary gas were set to 30 and 10 units respectively. The vaporizer and capillary temperatures were 225°C and 150°C respectively. The source current was set to 5 μA, the capillary voltage and tube lens settings were −60 and −125 V respectively.

TaqMan real-time RT-PCR

About 100 ml Nostoc cultures were harvested and homogenized as given above. Total RNA was isolated with Trizol reagent (Invitrogen, Paisley, UK) according to the manufacturer's instructions. The RNA was further treated with DNaseI and cleaned up with the RNeasy Mini Kit (Qiagen, Hilden, Germany). First-strand cDNA synthesis was performed using the TaqMan Reverse Transcription Reagents (Applied Biosystems, Warrington, UK) according to the manufacturer's protocol. Primers and TaqMan MGB probes were designed from cDNA sequences of Nostoc sp. PCC 7120, using Primer Express software (Applied Biosystems). The following primers and probes were used: NosCCD for 5′-AACACTGGTAATACTGCCCTCATCT-3′, NosCCD rev 5′-GCCGCCTTCCCATAATGC-3′, NosCCD probe 5′-CACGCAGGACAACTA-3′; Nos16S for 5′-AGCAGCCGCGGTAATACG-3′, Nos16S rev 5′-CGCTTTACGCCCAATCATTC-3′, Nos16S probe 5′-ATGCAAGCGTTATCC-3′, Nos crhC for 5′-CGCCGCCGGTCTACTG-3′, Nos crhC rev 5′-AACGTTACGGCGTGCTACTTC-3′, Nos crhC probe 5′-CGCCCGCAGATGA-3′, Nos crtL for 5′-TTTCGCCCACGGCTTCTAT-3′, Nos crtL rev 5′-TCACCCCAACTCGGTGGTA-3′, Nos crtL probe 5′-TGCCAGCTAGCAATT-3′.

Specific mRNA levels were quantified by real-time RT-PCR (ABIPrism 7000, Applied Biosystems) using 16S rRNA levels for normalization. Reporter (5′ end) dyes for the TaqMan MGB probes were 6FAM, except for 16S rRNA, where VIC was used. The relative quantity of transcripts was calculated using the comparative threshold cycle (CT) method (Livak, 1997). Data were normalized first to the corresponding 16S rRNA levels and then expressed as relative to untreated control transcript levels.


This work was supported by the Deutsche Forschungsgemeinschaft (DFG) Grant 892/1–3 and by the HarvestPlus programme ( We are indebted to Dr Peter Beyer and to Dr Jorge Mayer for valuable discussions. We thank Dr Hansgeorg Ernst for providing the apo-carotenoids and David Scheuring for his help in the in vitro assays.