Journal list menu

Volume 581, Issue 28 p. 5517-5525
Short communication
Free Access

The chlorophyllases AtCLH1 and AtCLH2 are not essential for senescence-related chlorophyll breakdown in Arabidopsis thaliana

Nicole Schenk

Nicole Schenk

Institute of Plant Sciences, University of Bern, Altenbergrain 21, CH-3013 Bern, Switzerland

These authors contributed equally to this work. Search for more papers by this author
Silvia Schelbert

Silvia Schelbert

Institute of Plant Biology, University of Zurich, Zollikerstrasse 107, CH-8008 Zurich, Switzerland

These authors contributed equally to this work. Search for more papers by this author
Marion Kanwischer

Marion Kanwischer

Max-Planck-Institute of Molecular Plant Physiology, Am Mühlenberg 1, D-14476 Golm, Germany

Search for more papers by this author
Eliezer E. Goldschmidt

Eliezer E. Goldschmidt

The Robert H. Smith Institute of Plant Sciences and Genetics in Agriculture, The Hebrew University of Jerusalem, Rehovot 76100, Israel

Search for more papers by this author
Peter Dörmann

Peter Dörmann

Max-Planck-Institute of Molecular Plant Physiology, Am Mühlenberg 1, D-14476 Golm, Germany

Search for more papers by this author
Stefan Hörtensteiner

Corresponding Author

Stefan Hörtensteiner

Institute of Plant Sciences, University of Bern, Altenbergrain 21, CH-3013 Bern, Switzerland

Institute of Plant Biology, University of Zurich, Zollikerstrasse 107, CH-8008 Zurich, Switzerland

Corresponding author. Address: Institute of Plant Biology, University of Zurich, Zollikerstrasse 107, CH-8008 Zurich, Switzerland.Search for more papers by this author
First published: 08 November 2007
Citations: 107

Abstract

One important reaction of chlorophyll (chl) breakdown during plant senescence is the removal of the lipophilic phytol moiety by chlorophyllase. AtCLH1 and AtCLH2 were considered to be required for this reaction in Arabidopsis thaliana. Here we present evidence against this assumption. Using green fluorescent protein fusions, neither AtCLH isoform localizes to chloroplasts, the predicted site of chlorophyll breakdown. Furthermore, clh1 and clh2 single and double knockout lines are still able to degrade chlorophyll during senescence. From our data we conclude that AtCLHs are not required for senescence-related chlorophyll breakdown in vivo and propose that genuine chlorophyllase has not yet been molecularly identified.

1 Introduction

Chlorophyllase catalyzes the hydrolysis of chlorophyll (chl) to chlorophyllide (Chlide) and phytol. It is considered as the first enzyme of chl breakdown during leaf senescence and fruit ripening [1, 2]. The enzyme is found in higher plants and algae, and chlorophyllases have been purified from different sources, including Citrus sinensis, Chenopodium album and Pheaodactylum tricurnutum [3-5]. In 1999, two research groups independently succeeded in cloning chlorophyllase genes (tentatively termed CLH) from C. sinensis (CsCLH1), C. album (CaCLH) and Arabidopsis thaliana (AtCLH1 and AtCLH2) [6, 7] and further genes have been described since then [2]. Surprisingly, not all molecularly identified CLHs contained a predicted chloroplast transit peptide. Thus, e.g. CaCLH was suggested to be imported into the vacuole via the ER [1, 7]. Consequently, in addition to the pheophorbide a oxygenase (PAO)- and red chlorophyll catabolite reductase (RCCR)-dependent pathway that operates in senescing chloroplasts [2], extraplastidial CLHs together with so far unknown oxidases were considered to degrade chl inside the vacuole [1].

Several lines of evidence question the existence of and need for such an extraplastidial chl degradation machinery and raise doubts on the participation of CLHs in senescence-related chl breakdown in vivo. (i) Alteration of AtCLH1 levels by RNA interference did not cause a senescence-related phenotype [8]. (ii) Heterologously expressed CLH from wheat exhibited an equally high specificity for chl and unrelated hydrophobic p-nitrophenyl esters [9]. (iii) Except for PAO, chl oxidizing activities have so far not been molecularly identified nor have breakdown products of respective activities been identified [2].

In this work we aimed at elucidating whether AtCLH1 and AtCLH2 participate in senescence-related chl breakdown in A. thaliana in vivo. By using green fluorescent protein (GFP) fusions, we show that neither AtCLH isoform is located in chloroplasts. Furthermore, single and double knockout mutants of AtCLHs are only marginally affected in chl breakdown during senescence. Collectively, our data strongly support the hypothesis that in A. thaliana CLHs are not essential for senescence-related chl breakdown.

2 Materials and methods

2.1 Plant material

A. thaliana T-DNA insertion lines SALK_124978 (designated clh1-1; ecotype Columbia) and SAIL_646_E09 (clh2-2; Columbia) were obtained from the European Arabidopsis Stock Center, Nottingham, UK. FLAG_076H05 (designated clh2-1; ecotype Wassilewskija) was from INRA, Versailles, France. Homozygous plants were identified by PCR using T-DNA and gene-specific primers as listed in Supplementary Table 1 (see Fig. 2A). The T-DNA positions were confirmed by sequencing. Double mutants were obtained by crossing clh1-1 with either clh2-1 or clh2-2. Double homozygous F 2-plants were identified by PCR as above. Plants were grown on soil in long day (16 h light) growth rooms under fluorescent light of 60–120 μmol photons m−2 s−1 at 22 °C. Because chl degradation and chl degradation-related parameters had been shown to be identical in Wassilewskija and Columbia (Col-0) [10], the latter was used as representative wild type only. Senescence was induced by placing 3- to 4-week-old detached leaves on wet filter paper and incubating in permanent darkness for up to 7 d.

figure image
Identification of A. thaliana clh mutants. (A) Gene structures of CLH1 (At1g19670) and CLH2 (At5g43860) showing the sites of T-DNA insertion in clh1-1, clh2-1 and clh2-2. Arrowheads indicate the location of primers used for PCR analyses. (B) RT-PCR analysis of CLH1 and CLH2 gene expression in clh single and double mutants. Actin2 (ACT2) was used as control. Numbers of PCR cycles performed are indicated on the right. (C) Northern blot analysis of CLH1 expression in the absence (−MeJA) or after exposure to 1 mM methyljasmonate (+MeJA) for 7 h. (D) Northern blot analysis of CLH2 expression. An additional 3.3 kb band cross-reacting with the CLH2 probe in lines containing the clh2-1 allele is labeled with an asterisk. See Section 3.2. for further details. The bottom parts of panels C and D show the RNA gel as loading control. Sizes of RNA marker bands are in kb.

2.2 Generation and confocal microscopic analysis of GFP constructs

cDNAs for AtCLH1 (pda09091) and AtCLH2 (pdz786270) were obtained from the RIKEN resource. The insert of pdz786270 was confirmed by sequencing. For C-terminal fusions (AtCLH-GFP), cDNAs were PCR-amplified using Pfu polymerase (Promega) with primers as listed in Supplementary Table 1. After restriction digest, fragments were cloned into XmaI/SpeI-digested pUC18-GFP5T-sp [11]. Likewise, for N-terminal fusions (GFP-AtCLH), PCR fragments were cloned into pUC18-spGFP6 [11]. For a fusion between the N-terminal region of AtCLH2 and GFP, the AtCLH2-GFP construct was used as template to amplify by inverse PCR the relevant plasmid part. After DpnI digest to remove the template, amplification products were restricted with SpeI and religated. Constructs were confirmed by sequencing. Free GFP expressed from pUC18-spGFP6 and TIC110-GFP expressed from pCL60-FLTIC110-GFP (F. Kessler, personal communication) were used as controls. A. thaliana mesophyll protoplast isolation and confocal microscopic analyses were performed essentially as described [10, 11].

2.3 Gene expression analysis

For RT-PCR analysis, total RNA was isolated from green leaves and reverse transcribed as described [12]. Semiquantitative PCR was performed using primers as listed in Supplementary Table 1. For Northern blot analysis of CLH1, RNA was isolated from 3-week-old plants grown on plates, which were incubated for 7 h in 20 mM MES, 10 μM Na2HPO4, pH 6.5 containing 1 mM methyl jasmonate or DMSO. For Northern blot analysis of CLH2, leaf material from plants grown for 6 weeks in long day conditions was used for RNA extraction. RNA electrophoresis (20 μg) and hybridization was according to standard procedures. A cDNA fragment derived from CLH1 and a PCR product amplified with PD415 and PD414 (Supplementary Table 1) served as a probe for CLH1 and CLH2, respectively.

2.4 Extraction and analysis of pigments

Fluorescent chlorophyll catabolites (FCCs) and non-fluorescent chlorophyll catabolites (NCCs) were extracted, quantified, and analyzed photometrically (chl) or by HPLC (NCCs, FCCs) according to published procedures [10, 12]. For the determination of in vivo contents of Chlide, a fast isolation method was used in order to minimize artificial production of Chlide by chlorophyllases that become rapidly activated upon leaf tissue rupture during extraction. Leaves were ground in liquid nitrogen and suspended in 10% (v/v) 0.2 M Tris–HCl pH 8 in acetone, cooled to −20 °C (5 ml g−1 fresh weight). After centrifugation (18 000 × g, 2 min, 4 °C) the supernatant was analyzed by HPLC as described [12]. Chlide/chl ratios were calculated from peak areas recorded at 665 nm.

2.5 Protein extraction and immunoblot analysis

After grinding of leaf disks (1 cm diameter), total proteins were extracted into 20 mM sodium phosphate pH 7.5, 1% (w/v) polyvinylpyrrolidone, 0.1% (v/v) β-mercaptoethanol (25 μl disk−1). Extracts were filtered through two layers of miracloth. Immunoblot analysis was performed as described [10] using anti-LHCII antibodies (1:2000) [13].

2.6 Chlorophyllase activity

Two different assays were employed to assess chlorophyllase activity. For an acetone-based assay, material from green leaves was extracted into 0.2 M Tris–HCl pH 8:acetone, 1:1 (v/v) (5 ml g−1 fresh weight) and incubated at 25 °C in the dark. Reactions were stopped by the addition of 1 volume of acetone. For separation of Chlide from chl, 0.5 volume of hexane were added. Chlide in the lower aqueous/acetone phase was quantified photometrically. The Triton-X100-based assay was performed essentially as described [14]. Chloroplasts were isolated from green or senescent leaves as described [15].

3 Results

3.1 AtCLH1 and AtCLH2 are located outside the chloroplast

Only for AtCLH2 (At5g43860), a majority of available servers that predict subcellular localization of plant proteins [16], indicate a plastidial localization (Table 1 ). Yet, the scores found with e.g. TargetP, PProwler or Predotar were rather low, compared to PAO and RCCR, which have been experimentally shown to localize to plastids [2]. In the case of AtCLH1 (At1g19670) none of the used prediction servers favored localization within the chloroplast. In order to investigate the subcellular localization of AtCLHs in A. thaliana protoplasts, C-terminal (AtCLH-GFP) and N-terminal (GFP-AtCLH) GFP fusions were analyzed. Neither of the two AtCLHs targeted GFP to the plastid. Instead, green fluorescence was detected in the cytosol, like free GFP (Fig. 1 A). Identical results were obtained with N-terminal fusions (data not shown). Cytosolic localization was also obtained when fusing GFP directly to the N-terminal 47 amino acids of AtCLH2, weakly predicted by ChloroP (Table 1) to contain a chloroplast transit peptide (data not shown). To investigate, whether targeting of AtCLHs to the chloroplast could be senescence-specifically regulated, C-terminal GFP fusions were expressed in protoplasts isolated from senescent leaves (Fig. 1B). Again, AtCLHs were not targeted to the plastids, but sometimes AtCLH2 was enriched in granular structures within the cytosol. The nature of these structures remains unknown.

figure image
In vivo targeting of GFP constructs to A. thaliana mesophyll protoplasts. After transfer of CLH1-GFP, CLH2-GFP, TIC110-GFP and free GFP in protoplasts isolated from green (A) or 3 d-dark incubated leaves (B), GFP fusions were monitored by confocal laser scanning microscopy. Lane 1, autofluorescence; lane 2, GFP fluorescence; lane 3, merge of GFP and autofluorescence; lane 4, bright field images. In panel B, arrows point to GFP-stained vesicles. Bar length, 10 μm.
Table Table 1. Prediction of subcellular localization of enzymes involved in chl degradation
Program a Prediction range b AtCLH1 At1g19670 Score AtCLH2 At5g43860 Score AtPAO At3g44880 Score AtRCCR At4g37000 Score
ChloroP 0–1 (>0.5) Not Chlpl 0.44 Chlpl 0.51 Chlpl 0.58 Chlpl 0.56
TargetP 0–1 (>0.5) mTP 0.02 (3) c mTP 0.08 (4) c mTP 0.04 (1) c mTP 0.21 (3) c
cTP 0.16 cTP 0.71 cTP 0.97 cTP 0.73
SP 0.26 SP 0.02 SP 0.01 SP 0.03
Other 0.73 Other 0.31 Other 0.1 Other 0.07
iPSort 0–1 d (<0.083) Not Mito or Chlpl 0.13 Not Mito or Chlpl 0.1 Mito or Chlpl 0 Mito or Chlpl 0
Chlpl Yes Chlpl Yes
PCLR 0–1 (>0.42) NonC 0.15 Chlpl 0.85 Chlpl 0.99 Chlpl 0.6
PProwler 0–1 mTP 0.09 mTP 0.14 mTP 0 mTP 0.01
cTP 0.01 cTP 0.61 cTP 0.99 cTP 0.98
SP 0.03 SP 0.17 SP 0 SP 0
Other 0.86 Other 0.07 Other 0 Other 0
Predotar 0–1 mTP 0.01 mTP 0.01 mTP 0.07 mTP 0.07
cTP 0.01 cTP 0 cTP 0.86 cTP 0.96
SP 0.01 SP 0 SP 0.01 SP 0
Other 0.97 Other 0.99 Other 0.13 Other 0.03
WoLF-PSORT 0–1 e Cyto 0.93 Chlpl 0.32 Chlpl 0.29 Chlpl 0.61
Mito or Chlpl 0.29 Mito 0.21 Mito or Chlpl 0.42
Cyto 0.21 Plas 0.21 Nucl 0.21
Mito 0.18 Golg 0.14 Mito 0.14
  • a For references, see [16] and http://aramemnon.botanik.uni-koeln.de.
  • b Threshold values are given in parentheses.
  • c Reliability classes 1 (highest) to 5 (lowest) are given in parantheses.
  • d Scoring according to average negative charge.
  • e Comparison done with 14 best neighbour proteins.

3.2 Isolation of AtCLH T-DNA insertion lines

A T-DNA insertion line for AtCLH1 (clh1-1) and two lines for AtCLH2 (clh2-1 and clh2-2) were obtained as shown in Fig. 2 A. Double mutants were produced by crossing clh1-1 with either clh2-1 (clh1-1 2-1) or clh2-2 (clh1-1 2-2). In each case, homozygous progeny was identified by PCR (data not shown). Expression of AtCLH1 and/or AtCLH2 transcripts was analyzed by Northern blots and/or by semiquantitative RT-PCR (Fig. 2B–D). Absence of CLH1 transcripts could be confirmed in all lines containing the homozygous clh1-1 allele. In clh2-2 and clh1-1 2-2, CLH2 transcripts were absent as well, but using RT-PCR, the clh2-1 allele gave strong signals at the correct size (Fig. 2B). Northern blot analysis using a CLH2 probe showed a strong 3.3 kb band, compared to the wild type CLH2 signal at 1.6 kb (Fig. 2D). This indicated that transcription from a strong promoter within the FLAG T-DNA had occurred. The search for translational start sites in the putative T-DNA–CLH2 transcript revealed that translation could not be initiated from this chimeric mRNA (data not shown), hence the occurrence of active CLH2 protein in clh2-1 and clh1-1 2-1 can be ruled out. Collectively, the homozygous genotype and the expression analysis of the different mutants strongly suggested that the respective mutants represent null alleles. None of the clh mutants exhibited any obvious phenotype during entire plant development, including natural senescence (data not shown).

3.3 Absence of AtCLHs has a minor effect on dark-induced leaf senescence

Under detached leaf senescence, all clh mutants showed leaf yellowing after 5–7 d and were hardly distinguishable from wild type (Fig. 3 A). Chl content decreased linearly during the senescence period in all lines. After 7 d some retention of chl was observed in the double knockout lines and in clh2-1, but yet more than 70% of the chl amount originally present had then been degraded (Fig. 3B). Increase of chl a/b ratios, a typical feature of senescence-related chl breakdown in A. thaliana [12], occurred in all investigated lines (results not shown). Loss of chl was accompanied by the loss of chl-binding proteins in all clh mutants and wild type as judged from immunoblots using anti-LHCII antibodies (Fig. 3C). Thereby loss of LHCII abundance was slightly delayed in CLH2-deficient lines.

figure image
Characterization of clh1 and clh2 single and double knockout lines during dark-induced senescence. (A) Phenotype of leaves. (B) Determination of total chl content. Values are means ± S.D. of at least two independent experiments with each three replicates. (C) Immunodetection of LHCII.

To determine the contribution of the two CLHs to overall chlorophyllase activity, in vitro activities were assessed in two different well-established assays, using either acetone or Triton X100 for solubilization (Fig. 4 A). Absence of CLH2 did only marginally affect chlorophyllase activity, but activities were drastically reduced in the absence of CLH1. The two analyzed double mutants retained significant activity (about 25–45% in the Triton X100-based assay), which was similar to the activity residing within isolated chloroplasts of the wild type (Fig. 4B). During dark-induced senescence, in vitro chlorophyllase activity slightly decreased in wild type and no significant increase in overall activity was evident in the two double knockout lines or in isolated wild type chloroplasts (Fig. 4B). Thus, neither the absence of cytosolic AtCLH1 and AtCLH2 in isolated wild type chloroplasts, nor clh1 clh2 double knockouts uncovered alternative senescence-regulated chlorophyllases. Possibly, genuine chlorophyllases are not senescence-regulated, or, more probably, their activity cannot be assessed under the applied assay conditions.

figure image
Determination of chlorophyllase activity in clh1 and clh2 single and double knockout lines. (A) Relative in vitro chlorophyllase activity was assessed after solubilization with either Triton X100 (gray) or acetone (black). See Section 2.6. for further details. Values are means ± S.D. of at least four independent experiments with each three replicates. (B) Chlorophyllase activity during dark-induced senescence. Assays were performed after Triton X100-solubilization of total leaf extracts of Col-0 (white), clh1-1 2-1 (black) and clh1-1 2-2 (gray), or of isolated Col-0 plastids (hatched). Values are means ± S.D. of two independent experiments with each three replicates. (C) Quantification of the chlorophyllide fraction (Chlide) of total green pigments in non-senescent leaves. Chlide and chl were separated by HPLC as described in Section 2.4. Values are means ± S.D. of two independent experiments with each three replicates. n.d., not determined.

To analyze in vivo chlorophyllase activities, Chlide contents in the different mutants were measured. In all lines, the Chlide fraction of total chl was below 0.5%. Chlide amounts largely correlated to the in vitro measured chlorophyllase activities (Fig. 4A) and, hence, Chlide accumulation in Col-0 and clh2 single knockout lines mainly resulted from some residual activity of AtCLH1 during extraction (Fig. 4C).

3.4 AtCLH mutants degrade chl via PAO/RCCR-catalyzed formation of FCCs and NCCs

The fact that clh mutants degraded chl at similar rates like wild type raised the question whether in the mutants breakdown of chl followed the PAO/RCCR pathway, which leads to the accumulation of FCCs and NCCs [2, 12]. Alternatively, oxidative or peroxidative pathways have been postulated [17], which could be active in the mutants, but this would not lead to the formation of FCCs and NCCs. To distinguish between these possibilities, FCC and NCC formation was analyzed (Fig. 5 ). In wild type, three known NCCs (At-NCC-1, 2 and 5) and three FCCs (At-FCC-1, 2 and 3 (pFCC-1)) were identified in respective HPLC chromatograms. The same catabolites were formed in single and double clh mutants (data shown for clh1-1, clh2-1 and clh1-1 2-1).

figure image
HPLC analysis of colorless chl breakdown products in clh mutants. Chl catabolites after 5 d of dark-induced senescence were separated by HPLC, and absorbance (A) and fluorescence (B) were recorded for the identification of A. thaliana NCCs and FCCs, respectively [12]. N1, At-NCC-1; N2, At-NCC-2; N5, At-NCC-5; F1, At-FCC-1; F2, At-FCC-2, F3, At-FCC-3 (= pFCC-1).

4 Discussion

Structure elucidation of NCCs from different plant species revealed the existence of dephytylated linear tetrapyrroles derived from chl a [2, 18]. Likewise, mutations in genes of different chl catabolic enzymes, such as A. thaliana pao1 and acd2-2, cause the accumulation of dephytylated intermediates [10, 12, 13] indicating that removal of the phytol moiety by chlorophyllase is an early step in senescence-related chl breakdown. Two independent groups reported the molecular nature of chlorophyllases (termed CLHs) in 1999 [6, 7], but the in vivo involvement of CLHs in chl breakdown during senescence has so far not been verified.

4.1 AtCLHs are located outside the plastid

Except for Citrus CLHs [9, 19] and (partially) for AtCLH2 (Table 1) [1, 7], all other CLHs were predicted to localize outside the chloroplast [1, 9]. Experimental confirmation of such prediction has so far only been obtained in two cases, i.e. CsCLH1 and Ginkgo biloba [19, 20]. Our finding of cytosolic localization of AtCLHs is in line with the fact that AtCLHs have so far not been identified in chloroplast proteome approaches [21, 22]. When senescent mesophyll protoplasts were used for transient expression, green fluorescence was sometimes found in cytoplasmic vesicles (Fig. 1B). The alternative extraplastidial chl breakdown pathway proposed by Takamiya et al. [1] was supposed to involve transfer of chl via chloroplast-derived lipophilic globules. These globules occur in a senescence-specific manner, and as demonstrated by confocal imaging contain chl [23]. It cannot be ruled out that in senescent cells, some AtCLH2 locates to such globules, where it might act as chlorophyllase in vivo. Indeed, clh2 mutants showed a slight delay in chl degradation, which was accompanied with some retention of LHCII protein, a feature known from different types of true stay green mutants [24]. Despite, the overall contribution of such a mechanism appears to be marginal.

4.2 Does a chl degradation pathway exist outside plastids?

PAO and RCCR have been shown to locate to chloroplasts [22, 25], suggesting that their respective substrates, pheophorbide a and red chl catabolite, are formed inside the plastid. Following this concept, it is highly likely to assume that also the preceding dephytylation and dechelation steps take place within the chloroplast. If an alternative (extraplastidial) pathway exists (see above), its contribution to overall chl breakdown during senescence can be estimated from mutants that are blocked in the PAO/RCCR pathway. Yet, chl is largely retained in pao1 and lls1 [12, 13] indicating that no significant chl degradation occurred through a bypass via such alternative degradation machinery. Furthermore, in several plant species including A. thaliana, the amounts of NCCs accumulating in senescent leaves correlate to the amount of chl present in green leaves [12], implying that the PAO/RCCR pathway is (almost) exclusively responsible to account for senescence-related chl degradation. Finally, neither AtCLH1 nor AtCLH2 expression is positively regulated with leaf senescence [26, 27], nor does overall activity increase during senescence (Fig. 4B). Our data indicate slight retention of chl degradation in the absence of AtCLH2, but not AtCLH1. Collectively, we like to conclude that AtCLH2 might marginally contribute to overall chl breakdown in A. thaliana leaf senescence, but rule out the requirement of AtCLH1. Nevertheless, overall in vitro wild type activity seems to be mainly dominated by CLH1, raising the question what function CLH1 might have? CLH1 expression is highly up-regulated by necrotrophic pathogens [8] and, thus, it might function in chl degradation in pathogen-infected tissue, i.e. after cell disintegration, thereby minimizing oxidative effects to the surrounding non-infected tissue.

4.3 Are other enzymes acting as chlorophyllases in vivo?

Expressed sequence tags of CLHs are available for different higher and lower plants, and algae. In contrast, within the genome of the diatom Pheaodactylum tricornutum, which contains a highly active chlorophyllase [5], sequences with significant homology to higher plant CLHs are not readily identifiable (S. Hörtensteiner, unpublished). This indicates, that chlorophyllase activity is not unequivocally linked to proteins with structural similarity to CLH, but implies that alternative enzymes could act as chlorophyllases in vivo. Our data support this assumption and provokes the notion that the genuine chlorophyllases might have escaped molecular identification so far. What is the molecular nature of alternative chlorophyllases? It is reasonable to assume that, like CLHs [28], alternative chlorophyllases would also contain motifs common to lipases or lipolytic enzymes. The A. thaliana genome encodes more than 70 potential lipase-like proteins, several of which are predicted to locate to the chloroplast. Future work analyzing respective candidate proteins might shed light on new chlorophyllases that possibly are required for senescence-related chl breakdown in vivo.

Acknowledgements

We thank F. Kessler for pCL60-FLTIC110-GFP and D. Rentsch for pUC18-spGFP6 and pUC18-GFP5T-sp. We thank S. Meier and C. Brinkmann for their help with confocal microscopy and I. Anders for technical support. This project was funded by the Swiss National Science Foundation (3100A0-105389) and the National Center of Competence in Research Plant Survival, research program of the Swiss National Science Foundation.

    Appendix A A

    Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.febslet.2007.10.060.