This Article Abstract Full Text (PDF) Supplemental Data All Versions of this Article: 19/1/369    most recent tpc.106.044404v1 Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in ISI Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (4) Citing Articles via Google Scholar Google Scholar Articles by Pruzinská, A. Articles by Hörtensteiner, S. Search for Related Content PubMed PubMed Citation Articles by Pruzinská, A. Articles by Hörtensteiner, S. Agricola Articles by Pruzinská, A. Articles by Hörtensteiner, S.

In Vivo Participation of Red Chlorophyll Catabolite Reductase in Chlorophyll Breakdown^[W]

Adriana Pruinská^a,1,2, Iwona Anders^a,1, Sylvain Aubry^a, Nicole Schenk^a, Esther Tapernoux-Lüthi^b, Thomas Müller^c, Bernhard Kräutler^c and Stefan Hörtensteiner^a,3

^a Institute of Plant Sciences, University of Bern, CH-3013 Bern, Switzerland
^b Institute of Plant Biology, University of Zürich, CH-8008 Zürich, Switzerland
^c Institute of Organic Chemistry and Center of Molecular Biosciences, Leopold-Franzens-University Innsbruck, A-6020 Innsbruck, Austria

³ To whom correspondence should be addressed. E-mail shorten{at}ips.unibe.ch; fax 41-31-631-49-42.

	ABSTRACT

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION METHODS REFERENCES

 
A central reaction of chlorophyll breakdown, porphyrin ring opening of pheophorbide a to the primary fluorescent chlorophyll catabolite (pFCC), requires pheophorbide a oxygenase (PAO) and red chlorophyll catabolite reductase (RCCR), with red chlorophyll catabolite (RCC) as a presumably PAO-bound intermediate. In subsequent steps, pFCC is converted to different fluorescent chlorophyll catabolites (FCCs) and nonfluorescent chlorophyll catabolites (NCCs). Here, we show that RCCR-deficient Arabidopsis thaliana accumulates RCC and three RCC-like pigments during senescence, as well as FCCs and NCCs. We also show that the stereospecificity of Arabidopsis RCCR is defined by a small protein domain and can be reversed by a single Phe-to-Val exchange. Exploiting this feature, we prove the in vivo participation of RCCR in chlorophyll breakdown. After complementation of RCCR mutants with RCCRs exhibiting alternative specificities, patterns of chlorophyll catabolites followed the specificity of complementing RCCRs. Light-dependent leaf cell death observed in different RCCR-deficient lines strictly correlated with the accumulation of RCCs and the release of singlet oxygen, and PAO induction preceded lesion formation. These findings suggest that RCCR absence causes leaf cell death as a result of the accumulation of photodynamic RCC. We conclude that RCCR (together with PAO) is required for the detoxification of chlorophyll catabolites and discuss the biochemical role(s) for this enzyme.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION METHODS REFERENCES

 
Senescence is the final stage of leaf development, which, except for some cases in which the regreening of yellow, senescent leaves has been reported (Thomas et al., 2003), is followed by the death of the whole leaf. Among different biochemical and physiological processes, chlorophyll degradation is integrally associated with leaf senescence and, as the most obvious sign of senescence, is widely used for senescence quantification. Chlorophyll is degraded in a pathway that is probably active in all higher plant species (Kräutler and Matile, 1999; Pruinská et al., 2003; Gray et al., 2004) and that leads to the formation of linear tetrapyrrolic breakdown products (nonfluorescent chlorophyll catabolites [NCCs]) that are deposited in the vacuoles of senescing cells.

The chlorophyll breakdown pathway has well been established in recent years (for reviews, see Kräutler, 2002; Eckhardt et al., 2004; Hörtensteiner, 2006; Kräutler and Hörtensteiner, 2006). It starts with the successive removal of phytol and magnesium by chlorophyllase (Takamiya et al., 2000) and magnesium-chelating substance (Suzuki and Shioi, 2002), respectively, followed by the conversion of pheophorbide (pheide) a to primary fluorescent chlorophyll catabolite (pFCC) (Mühlecker et al., 1997). The latter conversion has been attributed to the action of two enzymes, pheophorbide a oxygenase (PAO) and red chlorophyll catabolite reductase (RCCR) (Hörtensteiner et al., 1995; Rodoni et al., 1997), with red chlorophyll catabolite (RCC) occurring as an intermediate, but recent data suggest the involvement of an additional protein RCC forming factor (RFF), which together with PAO is required for RCC formation (Pruinská et al., 2005). pFCC is modified at several peripheral side positions by common or species-specific reactions, thus producing a species-specific set of modified FCCs. Finally, a primary active transport system (Hinder et al., 1996) at the tonoplast imports these FCCs into the vacuole, where they spontaneously convert to the respective NCCs, catalyzed by the acidic vacuolar pH (Oberhuber et al., 2003). In contrast with many other species, in which FCC conversion is fast and, hence, FCCs are barely detectable in senescence, Arabidopsis thaliana accumulates substantial amounts of both NCCs and FCCs in senescent leaves (Pruinská et al., 2005).

RCCR exhibits an intriguing specificity toward reduction of the C20/C1 double bond of RCC (see Figure 5C below). Thus, depending on the plant species used as a source of RCCR in an in vitro assay, together with PAO and pheide a as substrate, one of two possible C1 epimers of pFCC, pFCC-1 (Mühlecker et al., 1997) or pFCC-2 (= 1-epi-pFCC) (Mühlecker et al., 2000), is formed (Hörtensteiner et al., 2000b). RCCR is a soluble protein identical to Arabidopsis Accelerated Cell Death2 (ACD2). RCCR/ACD2 localizes to chloroplasts, as judged by chloroplast import experiments (Wüthrich et al., 2000), but in young seedlings and in response to stress it partially locates to mitochondria as well (Mach et al., 2001; Yao et al., 2004; Yao and Greenberg, 2006). RCCR/ACD2 is a single-copy gene in Arabidopsis (At4g37000) that is constitutively expressed; accordingly, RCCR activity is rather constant during all phases of leaf development (Pruinská et al., 2005). By contrast, PAO gene expression, protein abundance, and activity are highly upregulated during senescence (Pruinská et al., 2005). Furthermore, PAO has been shown to be located at the plastid inner envelope membrane (Pruinská et al., 2003; Kleffmann et al., 2004). Despite these differences in localization and patterns of activity, biochemical evidence suggests a close interaction between PAO and RCCR during catalysis. Thus, the intermediate RCC formed by the activity of PAO (and RFF) does not accumulate in vivo, and in vitro, RCC occurs only in trace amounts in respective assays (Rodoni et al., 1997). RCC seems to be efficiently reduced to pFCC by RCCR without release from PAO, which has been interpreted as a kind of metabolic channeling (Hörtensteiner, 1999). Further evidence for the PAO–RCCR interaction has been derived from the finding that in vitro the stereospecificity of RCCR is lost in the absence of PAO (Rodoni et al., 1997; Wüthrich et al., 2000).

View larger version (25K): [in this window] [in a new window] [as a PowerPoint slide]   Figure 5. HPLC Analysis of Chlorophyll Catabolites in acd2-2. (A) and (B) Chlorophyll catabolites of leaves from acd2-2 and Col-0 after 5 d of dark-induced senescence were separated by reverse-phase HPLC. A320 (A) and fluorescence (B) were recorded. For the identification and characterization of RCCs (R1 to R4), FCCs (F2 to F4), and NCCs (N1 to N5), see Table 1. Arrows indicate stereoisomers of known Col-0 catabolites (N3', N5', and F2') that were detected in acd2-2. Note that the A320 traces of acd2-2 and Col-0 between 50 and 80 min are identical to the respective traces in Figure 3A. (C) Chemical constitutions of RCCs, FCCs, and NCCs. Sites of peripheral modification present in the different chlorophyll catabolites (see Table 1) are indicated (R2 and R3). Relevant carbon atoms are labeled in RCCs. (D) Accumulation of NCCs (gray bars), FCCs (black bars), and RCCs (white bars) during detached leaf senescence in Col-0 and acd2-2. Results of a single representative experiment are shown. Data are means of three replicates. Error bars indicate SD. (E) Sums of chlorophyll catabolites of Col-0 (gray bars) or acd2-2 (black bars) shown in (D).

The intentions of this work were to elucidate whether RCCR participates in chlorophyll breakdown in vivo and to analyze whether the accumulation of RCC triggers cell death in RCCR mutants. To this end, we characterized different RCCR-deficient lines in relation to chlorophyll breakdown. We show that during dark-induced leaf senescence, RCC and RCC-like compounds accumulate in these mutants. Leaf cell death only occurred in lines that also accumulate RCCs and correlated to the light-dependent release of singlet oxygen. Surprisingly, FCCs and NCCs (i.e., downstream products of the site of RCCR action) were found in the mutants as well, further questioning the role of RCCR in chlorophyll breakdown. By molecular engineering of RCCRs, we show that a single amino acid exchange alters the C1 stereospecificity of At-RCCR from pFCC-1 to the formation of pFCC-2 (1-epi-pFCC). Exploiting this feature, acd2-2 was functionally complemented with Arabidopsis RCCRs exhibiting alternative C1 stereospecificities. Thereby, the pattern of FCCs and NCCs accumulating during senescence followed the specificity of the complementing RCCR, thus demonstrating the in vivo participation of RCCR in chlorophyll breakdown. Finally, we further substantiate earlier evidence of a biochemical interaction between RCCR and PAO. We conclude that RCCR is truly involved in chlorophyll breakdown, although its proposed catalytic role in the RCC-to-pFCC reduction remains to be confirmed. Furthermore, we propose that RCC accumulation is a major inducer of cell death observed in RCCR-depleted leaves.

	RESULTS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION METHODS REFERENCES

 
Isolation of Arabidopsis T-DNA Insertional Mutants and Transgenic Lines Deficient in RCCR
In this study, Arabidopsis lines of different accessions that are defective in RCCR activity (Figure 1C ; see Supplemental Figure 1B online) were used. acd2-2, obtained by ethyl methanesulfonate mutagenesis of ecotype Columbia (Col-0), contains a G-to-A mutation at the 3' splice site of the single intron of At4g37000 (RCCR/ACD2 gene) (Figure 1A) (Mach et al., 2001). Attempts to overexpress RCCR in the C24 background produced several lines, including line 68-13-1, in which constitutive expression of At-RCCR from pSC4 (Figure 1B) caused silencing of the endogenous RCCR gene of Arabidopsis. Silencing of RCCR in 68-13-1 was confirmed by RCCR activity measurements (Figure 1C) and immunoblot analysis (Figure 1D) and was stable for several generations. In contrast with acd2-2, in which neither RCCR protein nor activity could be detected, traces of RCCR proteins were present in 68-13-1, and 2.1% of the respective wild type activity was detected.

View larger version (29K): [in this window] [in a new window] [as a PowerPoint slide]   Figure 1. Characterization of RCCR-Deficient Lines in Arabidopsis. (A) Gene structure of RCCR (At4g37000) showing the point mutation of the 3' splice site of the intron present in acd2-2. (B) At-RCCR gene construct used for the production of the cosuppressing line 68-13-1. TP, chloroplast transit peptide of Arabidopsis RCCR. (C) RCCR activity in RCCR-deficient lines and respective wild types. RCCR was extracted from mature green leaves, and activities were determined in coupled PAO/RFF/RCCR assays. Values are means of three independent determinations. Error bars indicate SD. (D) Immunoblot of RCCR extracts used to measure enzyme activities as shown in (C). For each line, two independent extracts were analyzed. Gel loadings are based on equal amounts of fresh weight. (E) Cell-death phenotype of RCCR-deficient lines. Plants were grown under long-day conditions (100 µmol·m–2·s–1) for 3 weeks.

In addition to these lines, a T-DNA insertion line, acd2-9, was isolated by screening the Arabidopsis Knockout Facility (Wassilewskija [Ws] background) at the University of Wisconsin (Krysan et al., 1999). T-DNA insertion before the third codon of At4g37000 was confirmed by PCR and by cloning of the left T-DNA border (see Supplemental Figure 1A online). Although RCCR activity was largely lost (see Supplemental Figure 1B online), acd2-9 did not show cell death symptoms, nor was the senescence behavior affected (see Supplemental Figures 1C and 1D online). Yet, constitutive silencing of RCCR expression in Ws using an RNA interference strategy resulted in RCCR-deficient lines that exhibited symptoms of cell death, but compared with the respective silencing in Col-0, lesions developed only at later stages of plant development (see Supplemental Figures 1E and 1F online). The differences between acd2-9 and silencing lines in the development of premature cell death were confirmed both at the seedling stage and in mature leaves (see Supplemental Figures 1G and 1H online). In addition, analysis of chlorophyll catabolites (see below for details on the analyses of chlorophyll catabolites) further corroborated these differences: patterns of nongreen chlorophyll catabolites were identical in acd2-9 and Ws but were altered significantly in the silencing lines (see Supplemental Figure 2 online). The observed differences in these lines were difficult to explain, so we decided to focus our further studies on acd2-2 and 68-13-1.

Dark-Induced Senescence of RCCR Mutants Does Not Result in a Stay-Green Phenotype
To analyze the senescence behavior of the different RCCR mutants, detached leaves were incubated in permanent darkness. Under these conditions, lesions did not occur, confirming earlier data showing that light is required for lesion formation in acd2 mutants (Greenberg et al., 1994; Mach et al., 2001; Yao et al., 2004). Senescence progression in the wild-type lines was evident by a progressive yellowing of leaves during dark treatment. acd2-2 and 68-13-1 differed from their respective wild types by a faint red-orange coloration of the leaves after 5 to 7 d. Extraction into methanol, which does not extract protein-bound chlorophyll, caused a red coloration, indicating that red pigment(s) accumulated during the course of dark-induced senescence in acd2-2 and 68-13-1 (Figure 2A , insets). Red coloration of extracts was not detectable during natural senescence, or within the lesions of acd2-2 or 68-13-1, but red pigments accumulated when individual attached leaves were covered with aluminum foil to induce senescence (data not shown).

View larger version (35K): [in this window] [in a new window] [as a PowerPoint slide]   Figure 2. Senescence of Detached, Dark-Incubated Leaves of RCCR-Deficient Lines. (A) Phenotypes of individual detached leaves of acd2-2 and 68-13-1 and corresponding wild types after incubation in the dark for 0, 3, 5 and 7 d. Note the orange-red coloration in acd2-2 and 68-13-1 leaves after 5 to 7 d. Insets, methanolic extracts of leaves after 7 d in the dark indicate the accumulation of red pigments in acd2-2 and 68-13-1. (B) Degradation of chlorophyll in Col-0 (gray bars) and acd2-2 (black bars) during the course of detached leaf senescence. Data are means of a single representative experiment with three replicates. Error bars indicate SD. (C) Degradation of total proteins in Col-0 (gray bars) and acd2-2 (black bars) during the course of detached leaf senescence. Data are means of a single representative experiment with three replicates. Error bars indicate SD.

RCCR Mutants Accumulate RCC and RCC-Like Pigments during Chlorophyll Breakdown
It seemed possible that the red color observed in senescent leaf extracts of acd2-2 and 68-13-1 (Figure 2) was caused by the accumulation of RCC, the substrate of RCCR. To analyze this in more detail, extracts from leaves after 5 d of dark-induced senescence were analyzed by reverse-phase HPLC. An identical set of previously characterized FCCs and NCCs (Pruinská et al., 2005; Müller et al., 2006) were identified in the different wild-type lines, but surprisingly, several FCCs and NCCs were also present in the RCCR-deficient lines (see Figure 5 below). In addition to the rather polar FCCs and NCCs, a total of four less polar fractions exhibiting a typical ultraviolet/visible (UV/Vis) spectrum of RCCs (Kräutler, 2003) were detected in acd2-2 and 68-13-1 but not in the wild-type lines (Figure 3A ). Following an established nomenclature for colorless chlorophyll catabolites (Ginsburg and Matile, 1993), these four peaks were tentatively named At-RCC-1 through At-RCC-4 (see Figure 5 below and Table 1 for tentative structures of RCCs). By comparison with authentic RCC standards (Figure 3A) (Kräutler et al., 1997) and using mass spectrometry (MS; see Methods for MS data), At-RCC-3 and At-RCC-4 were identified as the primary cleavage product of pheide a, RCC, and an isomer, respectively. As RCC has been shown to isomerize readily at C13² in aqueous solution (Kräutler et al., 1997), At-RCC-4 may be assumed to have the structure of the C13² epimer of At-RCC-3. At-RCC-1 comigrated with a major polar red pigment accumulating in the medium of degreened Chlorella protothecoides cultures (Figure 3A). Mass analysis indicated that it represents the C13²-demethylated form of RCC (Table 1). By the same means, At-RCC-2 was characterized as another modified RCC, as the decarboxylation product of At-RCC-1 (Engel et al., 1991). In analogy with the accumulation of modified FCCs and NCCs during senescence (Pruinská et al., 2005; Müller et al., 2006), At-RCC-1 would be the result of a hydrolyzing enzymatic activity and At-RCC-2 would represent its chemical decarboxylation product.

View larger version (11K): [in this window] [in a new window] [as a PowerPoint slide]   Figure 3. RCCR-Deficient Lines Accumulate RCC and RCC-Like Compounds. (A) Partial HPLC traces (A320) of senescent leaf extracts. Detached leaves of the indicated lines were dark-incubated for 5 d. RCC and RCC-like compounds (R1 to R4) were identified by their characteristic spectra (Hörtensteiner, 1999) and by comparison with a RCC standard (Kräutler et al., 1997) and RCC-containing Chlorella protothecoides extracts (Engel et al., 1996). For further identification and characterization of R1 to R4, see Table 1. In the C24 spectrum, the asterisk indicates an unknown peak with a retention time similar to that of R1 that does not have a RCC spectrum. (B) Subcellular localization of RCCs in acd2-2. Partial HPLC traces (A320) of extracts from senescent acd2-2 leaves and from protoplasts, evacuolated protoplasts, and isolated vacuoplasts. A RCC standard was analyzed as a control.

View larger version (36K): [in this window] [in a new window] [as a PowerPoint slide]   Figure 7. Complementation of acd2-2. (A) Schemes of the constructs used for complementation of acd2-2. pGr-At-RCCR, wild type Arabidopsis RCCR construct. In pGr-At-RCCR-X, the region identical to the engineered region of RCCR protein X (At-RCCR-X; see Figure 6D) had been replaced with the respective tomato RCCR sequence (black region). Arabidopsis lines acd2-2+At-RCCR and acd2-2+At-RCCR-X, obtained after transformation of acd2-2 with pGr-At-RCCR and pGr-At-RCCR-X, respectively, were further analyzed. TP, chloroplast transit peptide of Arabidopsis RCCR. (B) Complementation of the acd2-2 cell-death phenotype in seedlings of acd2-2+At-RCCR and acd2-2+At-RCCR-X. Col-0 was used as a control. Plants were grown on Murashige and Skoog medium for 1 week, incubated in the dark for another week, and then reexposed to light for 1 d. (C) RCCR activity of complemented lines. RCCR was extracted from green mature leaves, and activities were determined. Data from representative single point determinations are shown. Experiments were repeated twice with similar results. (D) and (E) HPLC traces of leaf extracts from complemented acd2-2 lines. Chlorophyll catabolites of leaves after 5 d of dark-induced senescence were separated by reverse-phase HPLC, and fluorescence (D) and A320 (E) were monitored. For the identification and characterization of FCCs (D) and NCCS and RCCs (E), see Table 1. For coinjection, equal quantities of extracts from acd2-2+At-RCCR and acd2-2+At-RCCR-X were mixed and analyzed by HPLC. As controls, extracts of acd2-2 and a pFCC1/2 standard, containing both pFCC-1 (F3) and pFCC-2 (F4) (Rodoni et al., 1997), were used.

Cell Death in RCCR Mutant Leaves Correlates with the Accumulation of RCCs and the Induction of PAO
As shown in Figures 2A and 3, RCC and RCC-like compounds accumulated during senescence in acd2-2 and 68-13-1. This finding indicated that the occurrence of RCCs correlated with the formation of leaf lesions in respective lines. To investigate this interconnection in more detail, either detached leaves of acd2-2 (data not shown) or attached leaves that were individually covered with aluminum foil were incubated in the dark before ion leakage from leaf discs was determined as a measure of cell death. RCC content increased during the incubation period (Figure 4A ) and positively correlated with a light-dependent cell death reaction (Figure 4C). Thus, cell death was faster after longer dark treatment (i.e., at higher concentrations of RCCs). By contrast, Col-0 exhibited only some weak tissue leakage toward the end of the senescence period (Figure 4B). Similar results were obtained for 68-13-1 (data not shown). Preliminary experiments indicate that lesion formation can also be induced by infiltration of leaf discs with RCC-containing extracts of C. protothecoides cultures. Under these conditions, fast cell death occurred in acd2-2 in a light- and concentration-dependent manner, but Col-0 leaf discs also exhibited a cell death reaction at high RCC concentrations (A. Pruinská and S. Hörtensteiner, unpublished data).

View larger version (18K): [in this window] [in a new window] [as a PowerPoint slide]   Figure 4. Cell Death of acd2-2 and Induction of PAO Correlate with the Accumulation of RCCs. (A) Amounts of different RCCs (gray bars, At-RCC-1; black bars, At-RCC-2; white bars, At-RCC-3; hatched bars, At-RCC-4) accumulating during dark-induced leaf senescence in acd2-2. RCCs did not accumulate in Col-0 (data not shown). Results of a single representative experiment are shown. Experiments were performed three times (each single determinations) with similar results. (B) and (C) Determination of ion leakage as a measure of cellular death in Col-0 (B) and acd2-2 (C). Before reexposure to light (150 µmol·m–2·s–1) for up to 8 h, leaves were incubated in the dark for 0 d (open squares), 2 d (open circles), 4 d (closed circles), 6 d (closed squares), or 8 d (closed triangles). Results of the experiment corresponding to (A) are shown. Data are means of three replicates. Error bars indicate SD. Experiments were performed three times with similar results. (D) Release of singlet oxygen in acd2-2 after dark incubation. Detached Col-0 and acd2-2 leaves were incubated in the dark for 0 d (gray bars), 2 d (black bars), or 3 d (white bars) before reexposure to light for 0, 3, and 10 min. Leaf discs were infiltrated with DanePy under green safelight, and subsequently, singlet oxygen trapping was measured as relative quenching of DanePy fluorescence (Hideg et al., 1998). Results of a single representative experiment are shown. Data are means of 12 replicates. Error bars indicate SE. (E) Representative immunoblots of PAO protein abundance in acd2-2 leaf areas showing different degrees of cell death as indicated. Left, leaf with a lesion at the tip; right, two independent acd2-2 leaves without visible symptoms. Gel loadings are based on equal areas of leaf tissue.

In contrast with pao1, in which pheide a accumulated in the necrotic tissue (Pruinská et al., 2005), RCCs could not be detected within acd2-2 lesions of light-grown plants, probably because of the RCC's instability in light (data not shown). The formation of RCC requires the activity of PAO; therefore, we investigated whether the amount of PAO correlates with lesion formation in acd2-2. Leaves of similar age but with different severity of cell death were collected, and PAO abundance was estimated by immunoblot analysis (Figure 4E). In all samples exhibiting lesions, PAO was more abundant than in phenotypically healthy leaves. In addition, in leaves that just started to develop small lesions at the tip, a gradient of PAO could be observed. In contrast with the leaf base (i.e., the region most distant from the lesions), PAO abundance was increased within the area of lesions but also in the phenotypically still healthy middle region of the leaves (Figure 4E). Thus, the occurrence of lesions correlated with PAO abundance. Furthermore, the results suggested that before the development of visible lesions, PAO levels are increased, likely causing increased amounts of RCC in respective areas. These results are in agreement with the observation that, unlike in pao1, pheide a did not accumulate in healthy or necrotic tissues of RCCR-depleted lines (data not shown), thus excluding the possibility of pheide a being a trigger of cell death in the mutants.

RCCR-Deficient Lines Accumulate NCCs and FCCs
Five NCCs (At-NCC-1 to At-NCC-5) and three FCCs (At-FCC-1 to At-FCC-3) have been identified during chlorophyll breakdown in Arabidopsis (Pruinská et al., 2005). To our surprise, At-NCC-1, At-NCC-3, At-NCC-5, At-FCC-2, and At-FCC-3 (pFCC-1) were detected in HPLC runs of senescent leaf extracts of acd2-2 (Figures 5A and 5B ). The identity of chlorophyll catabolites was confirmed by HPLC-MS analysis (see Methods for MS data) and/or chromatography with authentic standards (Table 1). Further analysis indicated the presence in acd2-2 of C1 stereoisomers of certain catabolites (Table 1). Thus, in addition to pFCC-1, its C1 stereoisomer, pFCC-2, was present as well (Figure 5B). Furthermore, a NCC with an identical mass but slightly different retention time accompanied At-NCC-5 (arrow in Figure 5A) and was tentatively named C1-epi-At-NCC-5 (N5') (see Supplemental Figure 2C online). The same observation was also made for At-NCC-3 and At-FCC-2, in which double peaks were observed (arrows in Figures 5A and 5B), which suggested the presence of respective NCC and FCC isomers (N3' and F2'), but to date their identity has not been unambiguously confirmed (Table 1; see Methods for MS data). Identical results were obtained with 68-13-1 (data not shown). These results suggest a reduction of RCC in the absence of RCCR. RCC-to-pFCC reduction in acd2-2 was accompanied by a loss in stereospecificity, indicating that a stereounselective reduction by unknown mechanisms had occurred in the mutant.

During the course of dark-induced senescence, NCCs and FCCs accumulated progressively in Col-0, whereas in acd2-2, FCCs remained almost constant (Figure 5D). RCCs accounted for nearly 50% of the catabolites found in acd2-2 after 7 d, but the total amount of tetrapyrrolic catabolites was similar to that in the wild type (Figure 5E).

Exchange of a Single Amino Acid Changes the Stereospecificity of At-RCCR
RCCR exhibits an intriguing stereospecificity toward the reduction of the C20/C1 double bond of RCC; thus, depending on the source of the enzyme, one of two pFCC isomers, pFCC-1 or pFCC-2, is formed, indicating the presence of two types of RCCRs (type-1 and type-2 RCCR, respectively) (Hörtensteiner et al., 2000b; Wüthrich et al., 2000). Screening the GenBank/EMBL data libraries for the presence of RCCR sequences revealed deduced proteins or protein fragments from >20 species of different taxa within angiosperms, gymnosperms, and mosses (see Supplemental Figure 4A online), but RCCR-like sequences were not readily identifiable from the available genomes of photosynthetic prokaryotes and Chlamydomonas reinhardtii (data not shown). Figure 6A shows an alignment of At-RCCR with selected RCCRs from different species. The sequences were obtained by resequencing of ESTs. With identities between 36% (Marchantia polymorpha) and 84% (canola [Brassica napus]) the sequences were homologous with At-RCCR, and all (near) full-length proteins contained N-terminal chloroplast transit peptides as predicted by ChloroP (Emanuelsson et al., 1999). Identity as RCCRs was confirmed by heterologous expression of predicted mature forms of RCCRs from Arabidopsis and tomato (Solanum lycopersicum) (see below), canola, Pinus taeda, and Citrullus lanatus (data not shown). Furthermore, in vitro analysis of the RCCR activities in protein extracts from 21 different species confirmed the presence of RCCR in each species. With the exception of the Brassicaceae and a few Gramineae species, which produced pFCC-1, pFCC-2 was formed in all other species. One exception was hexaploid wheat (Triticum aestivum), in which both pFCCs occurred simultaneously (see Supplemental Figure 4A online). Subsequent analysis of diploid and tetraploid Triticum and Aegilops ancestors (Isidore et al., 2005) showed that in modern wheat the B and D genomes contribute a type-2 and type-1 RCCR, respectively, whereas the A genome contributes either two enzymes with different specificities or a mixed-type RCCR (see Supplemental Figure 4B online).

View larger version (64K): [in this window] [in a new window] [as a PowerPoint slide]   Figure 6. Engineering of the C1 Stereospecificity of RCCRs. (A) Sequence alignment of RCCRs from different species. The sequences were aligned using the program DIALIGN. Black shading with white letters, gray shading with white letters, and gray shading with black letters reflect 80, 60, and 40% sequence conservation, respectively, with Blosum62 similarity groups enabled. GenBank/EMBL/Swissprot accession numbers are as follows: Q8LDU4 for Arabidopsis thaliana RCCR, CAJ80766.1 (fragment) for tomato (Solanum lycopersicum) RCCR, CAJ80767.1 (fragment) for canola (Brassica napus) RCCR, CAJ80768.1 (fragment) for watermelon (Citrullus lanatus) RCCR, CAJ87104.1 for meadow fescue (Festuca pratensis) RCCR, CAJ80770.1 (fragment) for loblolly pine (Pinus taeda) RCCR, and CAJ80769.1 (fragment) for liverwort (Marchantia polymorpha) RCCR. Dashes indicate gaps. Single, double, and triple overlined sequences indicate regions of Arabidopsis RCCR that were replaced with the respective tomato RCCR regions in engineered RCCR proteins M, X, and Y, respectively, as shown in (C) and (D). The asterisk indicates Phe-218 of Arabidopsis RCCR, whose replacement with Val-214 of tomato RCCR (F218V in [E]) changes the C1 stereospecificity of Arabidopsis RCCR. The arrowheads indicate cleavage sites of the chloroplast transit peptides of Arabidopsis and tomato RCCR as predicted by ChloroP. (B) to (E) The middle panels show schemes of engineered RCCR proteins consisting of parts of the predicted mature RCCRs from tomato (Le-RCCR) and Arabidopsis (At-RCCR). The left and right panels show RCCR-dependent production of the C1 stereoisomers, pFCC-1 and pFCC-2, respectively. aa, amino acids. (B) Engineered RCCRs were expressed as His6-tagged proteins in E. coli, and after extraction, RCCR activities were determined. Activities are based on identical volumes of RCCR-containing E. coli extracts and are related to the activities of the predicted mature Arabidopsis (At-RCCR) and tomato (Le-RCCR) enzymes. Data of single representative experiments are shown. (C) Of 17 engineered RCCRs, protein M showed a change in stereospecificity. (D) Dissection of the exchanged region of protein M. (E) A single Phe-to-Val exchange in At-RCCR is sufficient to change the specificity.

Formation of FCCs and NCCs Follows the Stereospecificity of Engineered RCCRs Used for acd2-2 Complementation
The presence in RCCR mutants of catabolites occurring downstream of RCCR (see above) raised doubts about the participation of RCCR in chlorophyll breakdown. To address this question, we complemented acd2-2 using At-RCCR and chimeric protein X (At-RCCR-X) (Figure 6D; see above) that exhibit alternative stereospecificities. Both constructs (Figure 7A ) complemented the phenotype conferred by acd2-2 (Figure 7B). Respective transgenic lines (acd2-2+At-RCCR and acd2-2+At-RCCR-X) exhibited normal plant development and leaf senescence and were indistinguishable from the wild type (data not shown). The specificity of the complementing RCCRs was confirmed in assays using soluble protein extracts of respective lines as a source of RCCR (Figure 7C). Finally, in vivo formation of chlorophyll catabolites from senescent leaves of complemented lines was analyzed by HPLC. Thereby, the different in vitro stereospecificities of At-RCCR and At-RCCR-X were reflected by differences in retention times of FCCs (Figure 7D) and NCCs (Figure 7E) accumulating in respective lines. In particular, pFCC-1 and pFCC-2 exclusively occurred in acd2-2+At-RCCR and acd2-2+At-RCCR-X, respectively. Furthermore, patterns of pFCCs, At-FCC-2, At-NCC-3, and At-NCC-5 peaks accumulating in acd2-2, were identical to the fingerprint of FCC/NCC peaks after coinjection of acd2-2+At-RCCR and acd2-2+At-RCCR-X extracts. Together, these data confirm that RCCR participates in chlorophyll breakdown inside the chloroplast.

Does RCCR Form a Protein Complex with PAO?
The formation of pFCC from pheide a requires the activity of three protein components, PAO, RFF, and RCCR (Rodoni et al., 1997; Pruinská et al., 2005), and different lines of evidence (see Introduction) indicate that PAO and RCCR (and probably RFF) act together, thereby channeling the intermediary RCC. This suggests that PAO and RCCR may interact, possibly through the formation of a stable protein complex. To analyze this notion in more detail, PAO activity and protein abundance during leaf senescence were investigated in the RCCR-deficient lines acd2-2 and 68-13-1. As shown previously (Pruinská et al., 2005), in the wild-type lines PAO activity transiently increased during the course of dark-induced senescence (Figure 8A ; data shown for Col-0) concomitant with an increase in PAO protein levels (Figure 8B). By contrast, PAO levels and activities remained low in the respective RCCR-deficient lines (Figures 8A and 8B; PAO activity data shown for acd2-2). This finding indicated that the increase of PAO level during senescence is affected by the absence of RCCR. However, it remains to be shown whether this is attributable to changes in senescence-related PAO expression (Pruinská et al., 2005) or to reduced PAO stability. Because typical senescence parameters such as chlorophyll and protein degradation are normal in RCCR-deficient lines (Figure 2), senescence induction and progression seem not to be affected in the mutants, and it can be assumed that most likely the absence of RCCR decreases PAO stability. To this end, a possible interaction of PAO and RCCR was investigated using a bacterial two-hybrid system. Protein interaction, as demonstrated by reconstitution of ß-galactosidase activity, was indeed demonstrated in cells coexpressing PAO and RCCR, but not when either of them was coexpressed with the respective control plasmid (Figure 8D). Subsequently, different methods, such as coimmunoprecipitation, binding of PAO to RCCR-immobilized columns, and chemical cross-linking, were used to investigate whether PAO and RCCR form a protein complex. However, none of these attempts was successful (data not shown), indicating that despite a likely interaction during catalysis, PAO and RCCR do not form a stable protein complex.

View larger version (37K): [in this window] [in a new window] [as a PowerPoint slide]   Figure 8. Analysis of PAO Activity, and Interaction of PAO and RCCR. (A) PAO activity during dark-induced senescence of Col-0 (bars) compared with acd2-2, presented as the acd2-2:Col-0 PAO activity ratios (solid line). Note that at all measured times, the acd2-2:Col-0 PAO activity ratio was <1 (dotted line). Data are means of three independent measurements. Error bars indicate SD. (B) Representative immunoblots of PAO extracts used for measuring the activities in (A) and in 68-13-1 and C24 (data not shown). Gel loadings are based on equal amounts of fresh weight. (C) BacterioMatch two-hybrid screen demonstrating interaction between PAO and RCCR. Reporter strains were cotransformed with the plasmids as indicated, and interaction was analyzed by reconstitution of ß-galactosidase activity. 1, interaction of PAO and RCCR; 2, positive control; 3 and 4, combinations of control plasmids with PAO or RCCR constructs.

	DISCUSSION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION METHODS REFERENCES

In RCCR Mutants, RCC Is Modified and Transported to the Vacuole
During dark-induced senescence, four different types of RCC were identified in RCCR-deficient lines. Noteworthy, in C. protothecoides, which is devoid of a reductase with the function of RCCR, the breakdown of chlorophyll has been shown to end in the production of several RCC-like compounds, which also differ from each other by peripheral modifications, particularly at the methoxycarbonyl group at C13 (Engel et al., 1991, 1996). Thus, one major chlorophyll breakdown product in C. protothecoides is C13²-demethyl RCC, which is identical to At-RCC-1. This pigment, together with its decarboxylation product (Engel et al., 1991), At-RCC-2, was (partially) localized to the vacuolar fraction. By contrast, RCC (At-RCC-3) or its isomer (At-RCC-4) was not found in the vacuole. This finding indicated that RCC was transported from the plastid, its site of production by PAO, to the vacuole and that modification of RCC occurred before vacuolar import (i.e., in the plastid or cytosol). It can be concluded that in respective mutants the absence of RCCR causes the accumulation of RCC, which is in part metabolized similar to pFCC, the first identifiable product of porphyrin cleavage in wild-type plants (Mühlecker et al., 1997, Pruinská et al., 2005). In wild-type plants, RCC is rapidly converted to pFCC (i.e., probably before release from PAO [see below] and export from the plastid could occur). Furthermore, chlorophyll catabolite-modifying enzymes and transporters seem to work in a rather unspecific manner. Pheophorbidase, an enzyme that catalyzes the C13 demethylation of pheides a and b, has been identified and cloned recently by Shioi and coworkers (Shioi et al., 1996; Suzuki et al., 2006), and preliminary results from our laboratory indicate that the same enzyme can use pFCC as a substrate (S. Aubry, A. Pruinská, and S. Hörtensteiner, unpublished data). Thus, it is possible that RCC (At-RCC-3) also is demethylated by this activity to give rise to At-RCC-1. Export of FCCs from the plastid has been shown to involve an active process (Matile et al., 1992), but the nature of the transporter has not been elucidated. Transport of chlorophyll catabolites across the tonoplast probably requires the activity of members of the multidrug resistance–associated protein subfamily of ATP binding cassette transporters, which exhibit a rather broad substrate specificity (Martinoia et al., 2000, Klein et al., 2006). Thus, At-MRP2 and At-MRP3 are able to transport NCCs in vitro (Lu et al., 1998; Tommasini et al., 1998), although FCCs have been indicated to be the natural substrates for vacuolar uptake (Kräutler, 2003; Oberhuber et al., 2003). Likely, these transporters are responsible for the transmembrane transport of RCC-like pigments in the mutants.

The Biochemical Function of RCCR
RCCR is widely distributed among higher plants, but RCCR sequences and activity are not found in green algae or photosynthesizing prokaryotes (Hörtensteiner et al., 2000b; Wüthrich et al., 2000). This is in agreement with its role in RCC reduction: when C. protothecoides is grown under heterotrophic conditions and low nitrogen in the dark, RCC and RCC-like chlorophyll catabolites are excreted into the medium (Engel et al., 1996). Likewise, a red bilin pigment has been described as the final degradation product of chlorophyll metabolism in C. reinhardtii (Doi et al., 1997). Higher plants have to cope with chlorophyll catabolites internally, and it has been argued that the invention of RCCR required for the removal of potentially phototoxic RCC was a prerequisite for land colonization by plants (Matile et al., 1999). It seems likely that RCCR of higher plants evolved from bilin reductases of cyanobacteria that catalyze the Fd-dependent reduction of biliverdin to phycobilins. RCCR is distantly related to these bilin reductases and to phytochromobilin reductase (HY2) of higher plants (Frankenberg et al., 2001). Therefore, the similarity is restricted to a few highly conserved amino acids, including Phe-218, which defines the C1 stereospecificity of At-RCCR. Interestingly, different bilin reductases specifically produce the 3Z isomers of their respective products, although the thermodynamically more stable 3E isomers are substrates for incorporation of the respective holoproteins. The nature of 3Z/3E isomerization remains elusive (Frankenberg et al., 2001). The stereospecificity of RCCR is irrelevant for functionality, as demonstrated by the complementation of acd2-2 with RCCRs exhibiting different stereospecificities; hence, it does not affect plant development or senescence. Nevertheless, the importance of RCCR is evident from the cell-death phenotype of RCCR mutants (see above).

The analysis of chlorophyll catabolites found in RCCR-deficient lines showed that in addition to the accumulation of RCCs, downstream catabolites of RCCR action (i.e., FCCs and NCCs) accumulate, although with an apparent loss of C1 stereospecificity. This finding, together with a cell-protective role, which is independent of chlorophyll (Yao and Greenberg, 2006), raised doubts regarding the catalytic activity of RCCR and its involvement in chlorophyll catabolism. By complementation of acd2-2 with type-1 and type-2 RCCRs, we now confirm the in vivo participation of RCCR in chlorophyll breakdown.

If RCCR mutants are able to reduce RCC, what might be the mechanism involved and what is the catalytic activity of RCCR in this reaction? Preliminary data indicate that HY2 cannot be used for RCC reduction in vitro (A. Pruinská and S. Hörtensteiner, unpublished data). This finding corroborates the known distinct substrate preferences of the members of the HY2 family of ferredoxin-dependent bilin reductases (Frankenberg et al., 2001). Furthermore, it excludes a functional substitution by a related enzyme but leaves open the possibility of an unrelated activity taking over RCCR function in the mutants. As shown for the bilin reductase, phycocyanobilin:ferredoxin oxidoreductase (PcyA) (Frankenberg and Lagarias, 2003), RCCR is devoid of cofactors that are required for electron transfer to RCC (J. Berghold, I. Primoi, S. Hörtensteiner, and B. Kräutler, unpublished data). A reduction mechanism by electron transfer from Fd might be active in RCC reduction by RCCR (Oberhuber and Kräutler, 2002). A similar radical mechanism involving direct electron transfer from Fd has been postulated for PcyA (Tu et al., 2004). Indeed, by purely nonenzymatic, electrochemical means, RCC has been converted to roughly equal ratios of both pFCC-1 and its C1 epimer (Oberhuber and Kräutler, 2002; M. Oberhuber, J. Berghold, and B. Kräutler, unpublished data). This raises the possibility that in the mutants the redox conditions inside the chloroplast enable a (slow) reduction of the accumulating RCC(s) without the involvement of proteins. Pheide a oxygenation by PAO is limited in vitro (Rodoni et al., 1997), probably through product (RCC) inhibition of the enzyme, and only after reduction to pFCC by RCCR is the catabolite released from PAO (Hörtensteiner et al., 1998; Kräutler, 2003). This view is supported by biochemical data suggesting a catalytic cooperation between PAO and RCCR (Rodoni et al., 1997; Wüthrich et al., 2000) and implies a physical interaction. Here, we provide further evidence for such a PAO–RCCR interaction. First, the absence of RCCR in RCCR mutants affects PAO activity and protein abundance, indicating that RCCR may be required for the stabilization of PAO. Nevertheless, PAO levels were sufficient to prevent the accumulation of pheide a in RCCR mutants, unlike pao1, in which cell death has been attributed to pheide a accumulation (Pruinská et al., 2005). Second, the PAO–RCCR interaction could be demonstrated in a bacterial two-hybrid system. It has been argued that RCCR might not be a reductase itself but rather could accelerate the RCC-to-pFCC conversion in the chlorophyll catabolic pathway by enhancing electron transfer reactions and defining a regioselective and stereoselective reduction of RCC at C1 (Kräutler, 2002, 2003; Oberhuber and Kräutler, 2002; Hörtensteiner, 2004, 2006), but preliminary coupled PAO/RFF/RCCR assays (Pruinská et al., 2005) using RCC as substrate did not produce pFCCs, questioning the possibility of a direct electron transfer via PAO (A. Pruinská and S. Hörtensteiner, unpublished data).

Thus, it can be concluded that RCCR is a protein factor that, besides a possible direct function in protection from cellular death (Yao and Greenberg, 2006), is truly involved in chlorophyll breakdown. Its importance is manifested by the cell-death phenotype of RCCR mutants and the accumulation of RCCs, which likely triggers cell death. The biochemical function of RCCR in the reduction of RCC to pFCC remains to be confirmed, but data obtained here and elsewhere (Rodoni et al., 1997; Wüthrich et al., 2000) indicate that RCCR is not a reductase itself but rather functions as a kind of chaperone (Oberhuber and Kräutler, 2002; Kräutler, 2003), promoting the catalytic conversion of RCC to pFCC.

	METHODS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION METHODS REFERENCES

 
Plant Material and Senescence Induction
As wild types of Arabidopsis thaliana, ecotypes Col-0, C24, and Ws were used. Seeds of acd2-2 (Greenberg et al., 1994) were kindly provided by J.T. Greenberg (University of Chicago). acd2-9 (Ws background) was isolated from the Wisconsin Knockout Facility (Krysan et al., 1999) using the RCCR/ACD2 (At4g37000)–specific primer AtKO3 (5'-TAGCGTGTACAACACGGGAGGAAACTTCT-3') and the T-DNA left border primer JL202 (5'-CATTTTATAATAACGCTGCGGACATCTAC-3'). PCR products were analyzed by DNA gel blot analysis (Hörtensteiner et al., 2000a) using a digoxygenin-labeled (Dig High Prime; Roche) RCCR probe. In two subsequent rounds, positive signals were obtained in pools of 2025 (#16) and pools of 225 (#142; CSH142) (Krysan et al., 1999). Seed stock CSH142, containing seeds from 25 pools of 9 were obtained through the ABRC, and a RCCR T-DNA insertion was identified in a pool of 9 (#26; CSJ3526). One of the individual lines from seed stock CSJ3526 was renamed acd2-9, and homozygous siblings were used for further analysis. The presence of the T-DNA insertion within the RCCR gene in acd2-9 was confirmed by sequencing the PCR product generated from genomic DNA with the gene-specific primer At-RCCR7 (5'-AGGAGTTGTGATTGGAGGCG-3') and JL202. For the relative locations of primers, see Supplemental Figure 1 online.

Plants were grown on soil either in short-day (8 h/16 h) or long-day (16 h/8 h) growth rooms under fluorescent light of 60 to 120 µmol photons·m^–2·s^–1 at 22°C. For senescence induction, leaves from 3- to 4-week-old (long-day) or 8-week-old (short-day) plants were excised and incubated in permanent darkness on wet filter paper for up to 7 d at ambient temperature. Alternatively, individual attached leaves were wrapped with aluminum foil (Pruinská et al., 2005). For the induction of cell death in seedlings, seeds were surface-sterilized and grown on Murashige and Skoog agar plates without sucrose for 7 d under long-day conditions. Plates were then wrapped in aluminum foil for 7 d and analyzed after reexposure to light for 1 d.

Canola (Brassica napus) was grown on soil, and senescence of cotyledons was induced by dark incubation of 10-d-old seedlings in the dark (Hörtensteiner et al., 1995). Based on publicly available RCCR sequences (see Supplemental Figure 4 online), seeds or leaf material from respective species were obtained from the Botanical Gardens of Bern and Zürich, Switzerland, or from local garden shops. Seeds from diploid and tetraploid Triticum and Aegilops species were a gift from B. Keller (University of Zürich). For a few species, plant material could not be obtained, and following the assumption that RCCR stereospecificity is conserved within genera (Hörtensteiner et al., 2000b), a different species of the same genus was used instead.

Biocomputational Methods and Data Sources
Homologs of At-RCCR were searched for and identified with the BLAST neighborhood search algorithm (Altschul et al., 1997) through database searches with the National Center for Biotechnology Information, the Chlamydomonas reinhardtii genomic database, Chlamy Center (http://www.chlamy.org/chlamydb.html), and the CyanoBase (http://www.kazusa.or.jp/cyano/cyano.html). Protein sequences were derived from conceptual translation of cDNA sequences or overlapping ESTs. RCCR homologs were aligned (Figure 6) using the programs DIALIGN (Morgenstern, 2004) and GENEDOC (http://www.psc.edu/biomed/genedoc). For the phylogenetic tree shown in Supplemental Figure 4 online, RCCR protein sequences were aligned using ClustalW and the tree was calculated using neighbor joining with the program MEGA3.1 (Kumar et al., 2004). Accession numbers for the sequences used are indicated in the figure legends.

Constructs and Production of Transgenic Plants
RCCR Cosuppression
Arabidopsis line 68-13-1, in which the expression of the endogenous RCCR is cosuppressed, was obtained during the attempt to overexpress RCCR in C24. For this, full-length Arabidopsis RCCR (At-RCCR) was cloned behind the cauliflower mosaic virus 35S promoter containing a double enhancer element (35SDE). The open reading frame of At-RCCR was amplified from pGEM-At-RCCR (Wüthrich et al., 2000) with primers At-RCCR6 (5'-GGATCCATGGCGATGATATTTTGCAACAC-3') and At-RCCR3 (5'-TGATCATAATCTAGAGAACACCGAAAGC-3') (restriction sites in italics) and recloned into pGEM-T Easy (Promega), yielding pSC1a. A SphI/BglII fragment of pER30 (E. van der Graaff, personal communication) containing 35SDE was ligated to SphI/BamHI-restricted pSC1a, yielding pSC2a. Subsequently, the uidA gene of the binary vector pGPTV-KAN (Becker et al., 1992) was replaced by the 35SDE-At-RCCR fragment of pSC2a to give pSC4. Arabidopsis plants (C24) were transformed using the flower dip method (Sidler et al., 1998). Thirty kanamycin-resistant plants were analyzed for levels of RCCR by immunoblot analysis. After isolation of homozygous siblings, line 68-13-1, which showed strong reduction of RCCR levels (Figure 1D), was analyzed further.

RCCR RNA Interference Plants
Primer combinations SIL4 (5'-CTCGAGCCATGGATAGGAAGTTGGATACATTGC-3')/SIL5 (5'-GGTACCATCTCTCAATATCTCCTCC-3') and SIL6 (5'-GGATCCATAGGAAGTTGGATACATTGC-3')/SIL7 (5'-ATCGATGATCTCTCAATATCTCC-3') were used to amplify by PCR two 420-bp fragments of At-RCCR from pGEM-RCCR (Wüthrich et al., 2000). Appropriate restriction sites linked to the primers enabled a two-step cloning of the PCR products in opposite directions into pHannibal (Wesley et al., 2001). After digestion with NotI, the RNA interference construct was introduced into NotI-restricted pGreen0029 (Hellens et al., 2000) to produce pCSIL. Arabidopsis plants (Col-0 or Ws) were transformed (Sidler et al., 1998), and homozygous kanamycin-resistant T2 plants (Col-0+pCSIL#2-2, Col-0+pCSIL#7-1, Ws+pCSIL#5-4, and Ws+pCSIL#7-1) were isolated for further analysis.

Complementation of acd2-2
Full-length At-RCCR was ampl