STAY-GREEN and chlorophyll catabolic enzymes interact at light-harvesting complex II for chlorophyll detoxification during leaf senescence in Arabidopsis

Abstract

During leaf senescence, plants degrade chlorophyll to colorless linear tetrapyrroles that are stored in the vacuole of senescing cells. The early steps of chlorophyll breakdown occur in plastids. To date, five chlorophyll catabolic enzymes (CCEs), NONYELLOW COLORING1 (NYC1), NYC1-LIKE, pheophytinase, pheophorbide a oxygenase (PAO), and red chlorophyll catabolite reductase, have been identified; these enzymes catalyze the stepwise degradation of chlorophyll to a fluorescent intermediate, pFCC, which is then exported from the plastid. In addition, STAY-GREEN (SGR), Mendel's green cotyledon gene encoding a chloroplast protein, is required for the initiation of chlorophyll breakdown in plastids. Senescence-induced SGR binds to light-harvesting complex II (LHCII), but its exact role remains elusive. Here, we show that all five CCEs also specifically interact with LHCII. In addition, SGR and CCEs interact directly or indirectly with each other at LHCII, and SGR is essential for recruiting CCEs in senescing chloroplasts. PAO, which had been attributed to the inner envelope, is found to localize in the thylakoid membrane. These data indicate a predominant role for the SGR-CCE-LHCII protein interaction in the breakdown of LHCII-located chlorophyll, likely to allow metabolic channeling of phototoxic chlorophyll breakdown intermediates upstream of nontoxic pFCC.

Figure 1.

Accelerated Leaf Yellowing of Arabidopsis Plants Constitutively Expressing GFP-Tagged SGR or CCEs during Dark-Induced Senescence. Three-week-old plants grown under long-day conditions were used in this study. Photographs were taken from whole plants (A) or detached leaves (B) before (0 DDI; [B]) or after incubation in darkness for 4 d (4 DDI; [A] and [B]). WT, wild type. Bar = 5 cm.

Figure 2.

CCEs Interact with LHCII in Vivo. In vivo interactions of tagged SGR (A), RCCR (B), NYC1 (C), NOL (D), PPH (E), or PAO (F) with photosystem proteins were examined with α-Lhcb1, α-Lhca1, and α-CP43. Membrane-enriched fractions of 3-week-old GFP- or GST-tagged plants at 0 DDI were used for pull-down experiments with α-GFP– (GFP-IP) or α-GST–conjugated beads (GST-IP). Total protein extracts from nonsenescent rosette leaves of 35S:SGR-GFP plants were used as a positive control (Park et al., 2007), and 35S:GFP and 35S:CLH-GST plants were used as negative controls. Input levels of tagged proteins and of Lhcb1, Lhca1, and CP43 (all detected with respective antibodies) are shown as loading controls. Note that CLH, whose participation in chlorophyll breakdown has been questioned recently (Schenk et al., 2007), was unable to pull down LHCII (C).

Figure 3.

Coimmunoprecipitation of All CCEs and SGR in Senescing Chloroplasts. 35S:GFP and 35S:PPH-GFP transgenic plants grown for 3 weeks under long-day conditions were transferred to darkness and sampled at 3 DDI. Membrane-enriched fractions were used for in vivo pull-down assays. For this, GFP was immunoprecipitated (GFP-IP) with α-GFP–conjugated beads. Native SGR, RCCR, PAO, NYC1, and NOL in the input samples (left panel) and the pulled fractions (right panel) were detected using respective antibodies. The expression of GFP (negative control) and PPH-GFP were confirmed by α-GFP.

Figure 4.

Interactions among SGR and Five CCEs in Yeast Two-Hybrid Assays. β-Galactosidase (β-Gal) activities in yeast two-hybrid assays were measured by a liquid assay using chlorophenol red-β-d-galactoside (CPRG) as substrate according to the Yeast Protocols Handbook (Clontech). Empty bait or prey plasmids (−) were used as negative controls. Values are the average of relative activity from four colonies, and error bars represent sd.

Figure 5.

Interactions between SGR and CCEs by in Vitro Pull-Down Assays. Equal fresh weight of rosette leaves of two 3-week-old Arabidopsis plants expressing GFP- or GST-tagged SGR and CCEs were cohomogenized. Membrane-enriched fractions were used for pull-down assays with α-GFP–conjugated beads (GFP-IP), followed by immunoblot analysis using α-GST. Five combinations, including SGR-RCCR (A), SGR-NOL (B), SGR-NYC1 (C), SGR-PAO (D), and SGR-PPH (E), were examined. 35S:GFP and 35S:CLH-GST plants were used as negative controls (nc). Input levels of tagged proteins detected with respective antibodies are shown. Note that none of the GFP-tagged CCEs were able to pull down GST-tagged CLH, indicating that CLH is not part of a chlorophyll breakdown complex. This is in agreement with recent data questioning the involvement of CLH in chlorophyll breakdown (Schenk et al., 2007).

Figure 6.

In Vivo Interactions among SGR and CCEs Analyzed by BiFC. (A) For BiFC assays, construct pairs expressing fusions between SGR, PAO, NYC1, or NOL, and the N- or C-terminal half of YFP (YFPn or YFPc, respectively), were coexpressed in Arabidopsis mesophyll protoplasts isolated from 0 DDI green or 4 DDI senescing leaves. Confocal microscopy analysis was performed after 24 h. Note that YFP fluorescence was not detected at 0 DDI. Further positive BiFC interactions among SGR and CCEs are shown in Supplemental Figure 6 online. Auto, chlorophyll autofluorescence. Bars = 10 μm. (B) As a positive control for chloroplast-located BiFC interaction, PRK and CP12 fusions were used. Note that positive interaction was detected in both 0 DDI and 4 DDI protoplasts. Bars = 10 μm. (C) BiFC interaction between PAO and RCCR was positive in the wild-type protoplasts at 4 DDI but negative in the Arabidopsis sgr mutant nye1-1. As a control, positive interaction of PRK and CP12 was demonstrated in nye1-1. Bars = 10 μm.

Figure 7.

PAO Localizes to the Thylakoid Membrane. (A) Transient expression of PAO-GFP in the Arabidopsis mesophyll protoplasts. Note that GFP fluorescence largely superimposed chlorophyll autofluorescence (Auto), while an envelope control, TIC110-GFP, specifically labeled the surrounding of chloroplasts. (B) Immunoblot analysis of chloroplast membranes isolated from the 2 DDI senescent leaves after Suc density gradient centrifugation. PAO was visualized in gradient fractions using α-PAO, and its migration was compared with marker proteins from thylakoids (CAB), envelope (TOC75), and plastoglobules (PGL35). PAO largely comigrated with CAB, but not with TOC75 or PGL35, indicating thylakoid localization.

Figure 8.

Tentative Model of Chlorophyll Breakdown in a Senescing Mesophyll Cell. A model depicting the current knowledge about the topology of the PAO pathway, including the results from this work. Breakdown of chlorophyll to pFCC tentatively occurs in an enzyme complex located at LHCII in the thylakoid membrane. The exact composition of the complex remains unknown, but the results presented here indicate that SGR and all five CCEs known so far participate. After its formation, pFCC is exported from the chloroplast and modified in the cytosol. Modified FCCs are then imported into the vacuole and nonenzymatically converted to respective NCCs. Note that the location and/or participation of HMCR and metal chelating substance (MCS) in the multicomplex is unclear. For simplicity, LHCI and core complex proteins of the photosystems have been omitted. Likewise, the recently discovered hypermodified FCCs formed from pFCC in a branched pathway are not shown (Hörtensteiner and Kräutler, 2011). ABC, ABC transporter. [See online article for color version of this figure.]