Introduction

Photosynthetic organisms have evolved regulatory mechanisms that allow them to adapt their photosynthetic performance to changes in light conditions (Rochaix 2007). In the short term, this involves the phosphorylation of several core subunits of photosystem II (PSII), as well as some of its light-harvesting proteins (LHCII; Vener 2007). In flowering plants, the reversible phosphorylation of thylakoid proteins is regulated by light, temperature and redox state (Allen 1992; Bergantino et al. 1995; Vener et al. 1998; Aro and Ohad 2003). Changes in thylakoid protein phosphorylation have been associated with a number of regulatory and adaptive responses, including state transitions (Wollman 2001) and long-term photosynthetic acclimation (Bonardi et al. 2005). Technological improvements in the fields of chromatography and mass spectrometry have enhanced our ability to identify phosphoproteins within complex mixtures of polypeptides, and a number of novel thylakoid phosphoproteins have been described in the last few years (Vener 2007). The PSI-D1 subunit was the first phosphoprotein to be isolated from light-adapted PSI complexes (Hansson and Vener 2003). In the same work, the TMP14 protein was also shown to be phosphorylated. The latter was recently found to be a subunit of PSI and renamed PSI-P (Khrouchtchova et al. 2005).

Several homologous protein kinases involved in the phosphorylation of thylakoid proteins have been identified recently, namely, the thylakoid-associated Ser/Thr protein kinases STT7 in Chlamydomonas reinhardtii (Depege et al. 2003), and STN7 and STN8 in Arabidopsis thaliana (Bellafiore et al. 2005; Bonardi et al. 2005; Vainonen et al. 2005). STT7 and STN7 are required for the phosphorylation of LHCII proteins, whereas STN8 is necessary for modification of PSII subunits. However, the possibility that STN7 and STN8 do not directly phosphorylate antenna and PSII core proteins, respectively, but are components of partially overlapping phosphorylation cascades cannot be excluded (Bonardi et al. 2005).

The aim of our study was to identify novel thylakoid phosphoproteins, and to this purpose the photosynthetic mutants psad1-1 and psae1-3 (Ihnatowicz et al. 2004, 2007) with a general and drastic increase in the phosphorylation of thylakoid proteins were analysed. In these mutants, the mature Lhca4 protein is phosphorylated near its N-terminus, at Thr16, and the analysis of psad1-1 stn7-1 and psae1-3 stn7-1 double mutants implies that the phosphorylation depends on STN7.

Materials and methods

Plant materials

The A. thaliana mutants psad1-1, psae1-3 and stn7-1 have already been described (Ihnatowicz et al. 2004; Bonardi et al. 2005; Ihnatowicz et al. 2007). The lhca4-1 mutant, corresponding to the line Salk_138555 (ecotype Col-0), was identified by screening the insertion flanking database SIGnAL (http://signal.salk.edu/cgi-bin/tdnaexpress). Details of the T-DNA insertion and the primers used for segregation analysis are given in Supplementary Fig. 1. The double mutants psad1-1 stn7-1, psae1-3 stn7-1 and psae1-3 lhca4-1 were generated by crossing the corresponding single mutants and identifying homozygous F2 plants by PCR. Wild-type (WT) Arabidopsis thaliana (L.) Heynh., ecotype Columbia 0 (Col-0), and mutant plants were grown as described (Pesaresi et al. 2002) and, unless otherwise indicated, all analyses were performed on light-adapted plants at the eight-leaf rosette stage.

Protein complex isolation and immunoblot analysis

Leaves were harvested in the middle of the light period, and thylakoids were prepared as described previously (Bassi et al. 1985). For the isolation of PSI complexes, the membranes were washed twice with 5 mM EDTA (pH 7.8), centrifuged (5 min, 10000g) and resuspended in ddH2O to reach 2 mg/ml of total chlorophyll concentration. After solubilization with 2% (w/v) β-dodecyl maltoside (10 min, 4°C), PSI isolation and fractionation by 16–23% gradient SDS-PAGE was carried out as described before (Jensen et al. 2000) and proteins were visualized by Coomassie Blue staining. Western analysis of thylakoid proteins (for phosphothreonine) or PSI complexes (for phospothreonine and Lhcb2) was performed as reported (Pesaresi et al. 2002) and signals were detected using the Enhanced Chemiluminescence Western Blotting Kit (Amersham Biosciences).

Phosphopeptide analysis

PSI complexes were subjected to proteolysis with trypsin and chymotrypsin (overnight at 37°C) and undigested proteins were removed by ultrafiltration (Microcon YM-3, Millipore). The phosphorylated and non-phosphorylated peptides at the Thr residue were obtained from PSL GmbH (Heidelberg, Germany).

Purified samples were analysed by Multidimensional Protein Identification Technology (MudPIT) (for details see Supplementary Methods). MS/MS data were analysed with the SEQUEST algorithm (Eng et al. 1994). To ensure reliable protein identification, only peptides with a ΔCn score > 0.1 were considered. In addition, a peptide had to be partially tryptic or chymotryptic to be accepted and the cross-correlation scores of single, double and triple charged peptides had to be >1.8, >2.5 and >3.5, respectively. Peptides were manually evaluated as described elsewhere (Washburn et al. 2001).

Results

A novel 22-kDA phosphoprotein is present in psad1-1 and psae1-3 thylakoids

Because phosphorylation of thylakoid proteins is increased under conditions, which augment the reduction of the plastoquinol pool (Aro and Ohad 2003), the PSI mutants psad1-1 and psae1-3, both of them showing a reduction by about 60% of the subunits of the stromal ridge of PSI (Varotto et al. 2000; Ihnatowicz et al. 2004, 2007), were examined for the presence of novel thylakoid phosphoproteins. As expected, phosphorylation of PSII and LHCII proteins was markedly increased in light-adapted psad1-1 and psae1-3 thylakoids, as revealed by Western analysis with an antibody specific for phosphothreonine (pThr) residues (Fig. 1). In addition to thylakoid phosphoproteins with apparent molecular weights between 8 and 20 kDa, which include PSI-P (Hansson and Vener 2003; Khrouchtchova et al. 2005) and TSP9 (Carlberg et al. 2003), a novel species of about 22 kDa appears in the mutant thylakoids (see asterisk in Fig. 1).

Fig. 1
figure 1

Thylakoid protein phosphorylation. a Identical amounts of thylakoid proteins corresponding to 5 μg of total chlorophyll from WT, psad1-1 and psae1-3 leaves were fractionated by SDS-PAGE and visualized by Coomassie Blue staining. Note that only the portion of the Coomassie-stained gel containing phosphoproteins is shown. b Immunolabelling was performed with a pThr-specific antibody. The results shown are representative of those obtained in three independent experiments

Full size image

Phosphorylation of the novel PSI phosphoprotein depends on STN7

Because the migration behaviour of the 22-kDa phosphoprotein resembles that of PSI antenna proteins (LHCI), PSI complexes were isolated from WT, psad1-1 and psae1-3 leaves, fractionated by SDS-PAGE (Fig. 2a), and subjected to Western analyses with pThr- (Fig. 2b) and Lhcb2-specific antibodies (Fig. 2c). In addition to the highly phosphorylated LHCII fraction (pLHCII) described before (Pesaresi et al. 2002), the 22-kDa phosphoprotein, which comigrates with Lhca1 and Lhca4, was found to be associated with PSI in psad1-1 and psae1-3, but not in WT leaves (Fig. 2).

Fig. 2
figure 2

PSI phosphorylation. a Identical amounts of PSI complexes isolated from WT and mutant (psae1-3 stn7-1, psad1-1 stn7-1, psae1-3, psad1-1) leaves were fractionated by SDS-PAGE and visualized by Coomassie Blue staining. Positions of subunits previously identified by immunodetection (Jensen et al. 2000) are indicated. b Western analyses for pThr residues were performed on identical amounts of PSI complexes derived from WT and mutant leaves. c A similar filter to that shown in (b), but probed with an Lhcb2-specific antibody

Full size image

To test whether phosphorylation of the novel PSI protein depends on the kinase STN7, the double mutants psad1-1 stn7-1 and psae1-3 stn7-1 were generated and their PSI phosphorylation patterns were analysed as above. The novel PSI phosphoprotein was not detectable in the double mutants (Fig. 2), implying that its phosphorylation requires STN7 activity. The pLHCII-PSI complex is also absent in the double mutants, which supports the notion that phosphorylation of LHCII is essential for its interaction with PSI (Pesaresi et al. 2002).

The N-terminal portion of Lhca4 is phosphorylated at Thr16

To identify the novel PSI phosphoprotein, PSI complexes isolated from psad1-1 and psae1-3 thylakoid membranes were subjected to proteolysis with trypsin and chymotrypsin, and analysed by automated multidimensional protein identification technology (MudPIT), which combines biphasic liquid chromatography with electrospray ionization tandem mass spectrometry (MS/MS). In both psae1-3 (Fig. 3a) and psad1-1 (Supplementary Fig. 2), but not in WT samples, the phosphorylated peptide sequence TGSLAGDNGFDPLGLAEDPENLK, which originates from Lhca4, was identified with a cross correlation value of 4.15 for the doubly charged fragment ion, where the b- and y-ion series showed close-to-complete coverage of this peptide. The m/z ratio of the precursor ion selected for MS/MS analysis was 1,206.8, which was 0.63 Da higher than expected. However, for database search algorithms like SEQUEST, an error of ±1 to 2 Da is tolerable for reliable characterization of peptides.

Fig. 3
figure 3

Sequencing of a phosphopeptide released, by proteolysis with trypsin and chymotrypsin, from PSI complexes isolated from psae1-3 leaves. a Fragmentation spectrum (MS2) of the doubly protonated and phosphorylated peptide from psae1-3. b MS2 spectrum of the synthesized phosphorylated peptide tGSLAGDNGFDPLGLAEDPENLK. c MS2 spectrum of the synthesized non-phosphorylated peptide TGSLAGDNGFDPLGLAEDPENLK. Note that no neutral loss of H3PO4 (NL; m/z = −49) is observed. The b- and y-ions are indicated in the spectra and in the corresponding peptide sequence. The peptide fragment that underwent neutral loss of H3PO4 is indicated as NL. The lowercase t designates pThr

Full size image

In addition, the typical neutral loss (NL) of a phosphorylated peptide was obvious in the spectrum (see peak indicated as “NL”). These data, together with the fact that the novel phosphoprotein could be detected with a pThr-specific antibody (see Figs. 1 and 2) but not with an antibody specific for phosphoserine (data not shown), imply that Lhca4 is indeed phosphorylated at Thr16 in psad1-1 and psae1-3 plants. To further investigate this aspect, the Lhca4 peptide was synthesized in its threonine-phosphorylated (Fig. 3b) and threonine-non-phosphorylated forms (Fig. 3c). The MS/MS spectrum of the phosphorylated peptide is very similar to the spectrum shown in Fig. 3a and the predominant NL of the phosphate group (−49) is characteristic for the phosphorylated peptide. In contrast, the non-phosphorylated peptide shows no NL. Since the b-ion series is not as predominant as the y-ion series, only very little but still some NL loss is observed from fragment ions (e.g. predominant b18  m/z = 1812 shows −98 loss at m/z = 1714; Fig. 3a and Supplementary Fig. 2).

To confirm that the novel phosphoprotein detected corresponds to pLhca4, an insertion mutant for LHCA4 was isolated and crossed to psae1-3 plants to generate the double mutant psae1-3 lhca4-1. Western analysis of thylakoid protein preparations from WT, psae1-3 and psae1-3 lhca4-1 plants with a pThr-specific antibody showed that thylakoid membranes isolated from the double mutant resembled those of psae1-3, except for the specific absence of the novel phosphoprotein (Fig. 4). This strongly suggests that pLhca4 is the novel phosphoprotein that accumulates in an STN7-dependent manner in the mutants psad1-1 and psae1-3. The possibility that Lhca1 is also phosphorylated cannot be excluded, because depletion of Lhca4 results in a concomitant decrease in the level of Lhca1 (Fig. 4a; Klimmek et al. 2005). However, only non-phosphorylated Lhca1 peptides were detected by MudPIT analysis (data not shown).

Fig. 4
figure 4

Thylakoid protein phosphorylation in WT, psae1-3 and psae1-3 lhca4-1 leaves. a Identical amounts of thylakoid proteins (corresponding to 5 μg of total chlorophyll) from WT, psae1-3 and psae1-3 lhca4-1 leaves were fractionated by SDS-PAGE and visualized by Coomassie Blue staining. Note that only the portion of the Coomassie-stained gel containing phosphoproteins is shown. b Western analysis was performed with a pThr-specific antibody. The results shown are representative of those obtained in three independent experiments

Full size image

As in the case of the majority of thylakoid phosphoproteins, the phosphorylatable Thr in Lhca4 is found in the N-terminal region, which protrudes into the stroma. However, the target amino acid residue is not strictly conserved in most flowering plants (Fig. 5), thus a conserved physiological function of N-terminal Lhca4 phosphorylation appears rather unlikely.

Fig. 5
figure 5

Comparison of N-terminal sequences of mature Lhca4 proteins. The N-terminal amino acid sequence of the mature A. thaliana Lhca4 protein (At3g47470; gi: 23197770) was compared with Lhca4 sequences from S. tuberosum (gi: 100801744), V. vinifera (gi: 147843386), P. sativum (gi: 6470348), B. vulgaris (gi: 33438474), P. sylvestris (gi: 22753) and O. sativa (gi: 125561654). Black boxes indicate strictly conserved amino acids and grey boxes indicate closely related ones. The region from which the phosphopeptide (pp) was derived is indicated (see also Fig. 3); the asterisk refers to the phosphorylated threonine

Full size image

Discussion

Taken together, our data clearly show that the over-reduction of the plastoquinone pool associated with impaired PSI oxidation in psae1-3 and psad1-1 thylakoids leads to the accumulation of a novel thylakoid phosphoprotein, pLhca4, phosphorylated at Thr16 of the mature protein in an STN7-dependent manner. This can be accounted for by hyperactivation of the kinase STN7 and/or by a change in the accessibility of Lhca4. Because the phosphorylatable Thr residue in Lhca4 is not strictly conserved in most flowering plants and the pLhca4 protein could not be detected in WT plants under standard conditions (unpublished results of our lab; Vener 2007), the identified phosphopeptide is most probably a peculiarity of the extremely high reduction of the PQ pool in mutant thylakoids.

Nevertheless, our findings suggest that, under the extreme redox conditions prevailing in psad1 and psae1 mutants, the cytochrome b 6 /f complex, STN7 and PSI must be located in close proximity to each other. The reduced number of disks per granum observed in psae1 mutant thylakoids (Pesaresi et al. 2002) might induce a general reorganization of thylakoid complex distribution, and thus make PSI complexes accessible to thylakoid kinases. A similar process, although more subtle, might take place in WT plants, thus making possible the phosphorylation of the PSI-D1 and -P proteins (Hansson and Vener 2003; Khrouchtchova et al. 2005). Additionally, the general reorganization of thylakoid protein complex distribution upon over-reduction of the thylakoid electron transport chain is associated with important adaptative processes, such as state transitions (Wollman 2001). Certainly, the hyperactivation of thylakoid kinases in psad1 and psae1 might also contribute to the phosphorylation of Lhca4 and to the hyperphosphorylation of other thylakoid proteins.