•Overexpression of AtPAP2, a purple acid phosphatase (PAP) with a unique C-terminal hydrophobic motif in Arabidopsis, resulted in earlier bolting and a higher seed yield. Metabolite analysis showed that the shoots of AtPAP2 overexpression lines contained higher levels of sugars and tricarboxylic acid (TCA) metabolites. Enzyme assays showed that sucrose phosphate synthase (SPS) activity was significantly upregulated in the overexpression lines. The higher SPS activity arose from a higher level of SPS protein, and was independent of SnRK1.
•AtPAP2 was found to be targeted to both plastids and mitochondria via its C-terminal hydrophobic motif. Ectopic expression of a truncated AtPAP2 without this C-terminal motif in Arabidopsis indicated that the subcellular localization of AtPAP2 is essential for its biological actions.
•Plant PAPs are generally considered to mediate phosphorus acquisition and redistribution. AtPAP2 is the first PAP shown to modulate carbon metabolism and the first shown to be dual-targeted to both plastids and mitochondria by a C-terminal targeting signal.
•One PAP-like sequence carrying a hydrophobic C-terminal motif could be identified in the genome of the smallest free-living photosynthetic eukaryote, Ostreococcus tauri. This might reflect a common ancestral function of AtPAP2-like sequences in the regulation of carbon metabolism.
In this study, we show that AtPAP2 is dual-targeted to mitochondria and plastids via a novel and unique C-terminal transmembrane targeting signal. Overexpression of AtPAP2 significantly enhanced the growth rate and seed yield of Arabidopsis, through increased SPS activity.
Materials and Methods
Plant materials and growth conditions
Wild-type (WT) Arabidopsis thaliana (L.) Heynh. ecotype Columbia (Col-0) was used in this study. The T-DNA mutant of AT1G13900 (Salk_013567, ecotype Col-0) and AT2G03450 (Salk_129905, ecotype Col-0) were obtained from The Arabidopsis Information Resource (TAIR, http://www.arabidopsis.org). Chilled Arabidopsis seeds were surface-sterilized with 20% (v/v) bleach for 15 min, washed, and plated on Murashige and Skoog (MS) medium supplemented with 2% (w/v) sucrose for 10 d. Seedlings with the same size were transferred to soil under a 16-h light (22°C) : 8-h dark (18°C) regime (long day, LD) or an 8-h light (22°C) : 16-h dark (18°C) regime (short day, SD) under a light intensity of 120–150 μmol m−2 s−1. The pots were placed in the growth chamber in a randomized design. Bolting time was measured when the primary inflorescence reached 1 cm above the rosette leaves. This observation of phenotype was repeated at least three times (n =10–15).
A fragment of AtPAP2 cDNA corresponding to the N-terminus of AtPAP2 (amino acids (a.a.) 25–144) was amplified by primers P2AbF and P2AbR. This region was selected to avoid cross-reactivity toward other AtPAPs. The PCR product was digested with SacI and KpnI and fused with the N-terminal His-tag of vector pRSET-A (Invitrogen, Hong Kong, China) and the resulting plasmids were transformed into Escherichia coli strain BL21 (DE3). The overexpressed fusion protein in the inclusion bodies was solubilized in extraction buffer A (20 mM Tris–HCl, pH 7.5, 150 mM NaCl, 8 M urea) and purified by a HisTrap FF column (GE Healthcare, Hong Kong, China) using buffer A containing 100 mM imidazole. The eluted protein was dialyzed in 20 mM PBS, pH 7.2 to remove urea. The antigen was further gel-purified before being used for rabbit immunization. The specificity of this antiserum was confirmed by western blotting analysis on the proteins isolated from WT and the T-DNA insertion line (Fig. S3d).
Expression of AtPAP2 under phosphate starvation and sugar treatments
Arabidopsis plants were grown in phosphate (Pi) starvation MS medium (0 mM Pi) for 5 d and then transferred to complete MS medium for 0, 2, 4, 6 or 8 d. Alternatively, plants were grown on MS agar for 7 d before transfer to Pi starvation medium for 0, 2, 4, 7 or 10 d (del Pozo et al., 1999). For sugar treatments, 5-d-old seedlings on MS agar plates were transferred to solid MS medium containing a sucrose, glucose, sorbitol, or mannitol gradient (% w/v) for 24 h. Total protein isolated from seedlings was analyzed by western blot analysis using AtPAP2 antiserum.
Generation of AtPAP2 overexpression lines in Arabidopsis
The full-length coding region of the AtPAP2 cDNA (AT1G13900) was amplified by Pfx DNA polymerase (Invitrogen, Hong Kong) using primers P2YF and P2NR. The resulting product (1971 bp) was initially cloned into vector pGEM-T Easy (Promega, Hong Kong, China) and then subcloned into the binary vector pBA002 downstream of the cauliflower mosaic virus (CaMV) 35S promoter (pBA002-CaMV35: AtPAP2). The vector was then introduced into Agrobacterium tumefaciens strain GV3101 and then transformed by the floral dip method (Clough & Bent, 1998) into WT plants. Homozygous CaMV35S: AtPAP2 overexpression (OE) lines were selected on MS plates containing 5 mg l−1 glufosinate ammonium (Sigma-Aldrich, St. Louis, USA). Resistant lines were transferred to soil to grow to maturity, and their transgenic status was confirmed by genomic PCR and western blot analyses. Homozygous T3 seeds of the OE plants were used for further analysis. A construct of a truncated AtPAP2 fragment without its C-terminus (P2NC) was generated by PCR using primers P2YF and P2NCR and cloned in the pBA002 vector. AtPAP2 with a deleted signal peptide (P2NS) was amplified with primers P2NSF and P2NR and cloned into the pCXSN vector (Chen et al., 2009). All primers used are listed in Table S3.
Characterization of homozygous AtPAP2 T-DNA line
A T-DNA insertion line of AtPAP2 gene (Salk_013567) in the Col ecotype was obtained from TAIR. The homozygous T-DNA line was verified by genomic PCR screening using the T-DNA left border primer LBa1 and AtPAP2 gene-specific primer LP2 and reverse primer RP2 (SIGnAL database). The T-DNA insertion site was confirmed by DNA sequencing of the PCR product.
Subcellular localization of AtPAP2
The putative signal peptide (SP) and the transmembrane domain and C-terminal tail (TMD/CT) were PCR-amplified from full-length AtPAP2 cDNA and cloned into a GFP vector in pBI221 to generate SP-GFP, GFP-TMD/CT and SP-GFP-TMD/CT using full-length AtPAP2 cDNA as template and the primer sets listed in Table S3. Organelle markers were used to elucidate the subcellular localization of AtPAP2-GFP constructs. Transient expression in WT Arabidopsis PSB-D cell culture was performed as described previously (Miao & Jiang, 2007). Generally, PSB-D cells at day 5 after subculture were collected for protoplast preparation. Plasmids were electroporated into protoplasts and incubated at 26°C for 13 h before being observed under a confocal laser scanning microscope. Protoplasts were incubated for 5 min at room temperature with MitoTracker Orange (CMTMRos; Invitrogen, Hong Kong, China) before analysis. Confocal images were collected with a Fluoview FV1000 microscope (Olympus, Tokyo, Japan) with a ×60 objective water lens. The settings for collecting confocal images within the linear range were described previously (Tse et al., 2004). Images were processed using Adobe Photoshop software.
Organelle isolation and western blot analysis
Mitochondria and chloroplasts were isolated from 10-d-old Arabidopsis seedlings as previously described (Lister et al., 2007; Kubis et al., 2008). Proteins (25 μg) were resolved by SDS-PAGE and transferred to Hybond-C nitrocellulose membranes, then immunodetected as previously described (Carrie et al., 2008). The antibodies used, Tom40 (Carrie et al., 2009a,b) and Risp (Carrie et al., 2010), have been described previously. Antibodies to CoxII, Rubisco large subunit and Rubisco small subunit (SSU) were purchased from Agrisera (Vännäs, Sweden).
Measurement of soluble sugars and starch in Arabidopsis
Soluble sugars were extracted from Arabidopsis using the chloroform/methanol method (Antonio et al., 2007). Liquid chromatography-MS/MS spectrophotometry was employed for sugar measurement. Filtered samples (10 μl) were injected into an HP 1100 HPLC system (Agilent Technologies, Palo Alto, CA, USA) connected to a zwitterionic ZIC-HILIC column (3.5 μm, 150 mm × 2.1 mm i.d.; SeQuant, Umea, Sweden) (Antonio et al., 2008). Separation was performed using a solvent system of 0.1% (v/v) formic acid in water (A) and 0.1% (v/v) formic acid in methanol (B) with a linear gradient of 20–60% B over 25 min. Flow rate was maintained at 0.2 ml min−1 and the elution was monitored by a diode-array detector (200–600 nm). Multiple reaction monitoring (MRM) experiments were conducted at the electrospray ionization interface of an API 2000 QTRAP system (Applied Biosystems, Hong Kong, China) operating in negative mode. Starch measurement was carried out according to the method of Focks & Benning (1998) using a starch assay kit (Sigma-Aldrich, St. Louis, USA).
Pull-down assays using recombinant AtSnRK1.1, AtSnRK1.2, AtGRIK1, AtGRIK2 and 14-3-3 protein
Vectors for expressing GST-GRIK1, GST-GRIK2, His-AtSnRK1.1 (a.a. 1–341) and His-AtSnRK1.2 (a.a. 1–342) were gifts from Dr Shen Wei (Shen et al., 2009). The expression vector GST-BMH1 (containing a yeast 14-3-3 homolog) was a gift from Prof. Carol MacKintosh (University of Dundee, UK). The fusion protein His-AtSnRK1.1/His-AtSnRK1.2 was purified by Ni-NTA beads and the fusion protein GST-GRIK1/GST-GRIK2 by a GST-Trap column (Amersham Biosciences) (Moorhead et al., 1999). For His-AtSnRK1.1/His-AtSnRK1.2 pull-down assays, 20 μg of protein in 200 μl His binding buffer was coupled to 200 μl of Ni-NTA beads and incubated with 1 mg total soluble Arabidopsis protein for 2 h at 4°C. The bound proteins were washed and eluted in binding buffer containing 100 mM imidazole. For the GST-GRIK1/GST-GRIK2 and GST-14-3-3 pull-down assays, 1 mg total plant protein extract was mixed with 5 μg purified GST-fusion protein on 50 μl GST beads with gentle agitation for 2 h at 4°C. The beads were washed with washing buffer (50 mM HEPES-KOH, pH 7.5, 500 mM NaCl, 1 mM dithiothreitol) five times. Bound proteins were eluted in 50 mM Tris–HCl, pH 8.0, 10 mM reduced glutathione, 500 mM NaCl and subjected to immunoblot analysis.
Analysis of proteins in AtPAP2 transgenic Arabidopsis plants
Total soluble proteins were extracted according to the trichloroacetic acid method (Zhu et al., 2006). The samples were run in the pH 4–7 range and stained with SYPRO Ruby for total protein staining and Pro-Q Diamond for phosphoprotein staining. Images were analyzed using ImageMaster 2-D Platinum software (GE Healthcare, Hong Kong, China). The target spots were run on a MALDI-TOF/TOF mass spectrometer and identified by peptide mass fingerprinting with data searches carried out using MASCOT search tools (http://www.matrixscience.com) in the NCBI non-redundant public protein database. All of the peptide masses were assumed to be monoisotopic and corresponding to [M+H+]. Carboxymethyl was considered to be a fixed modification, while Oxidation (M), Phospho (ST) and Phospho (Y) were considered variable modifications. The peptide mass tolerance was set at ± 0.2 Da and the maximum number of missed cleavages was set at one. The identified proteins had more than five matched peptides, and the percentage of sequence coverage was > 20%. All of the positive protein identification scores were significant (MASCOT score > 60, P <0.05).
Metabolomic analysis by 1H-NMR and GC-MS
Shoots of 20-d-old Arabidopsis were freeze-dried immediately after harvest. Three biological replicates were used in this study. For 1H-NMR analysis, dried plant tissues (15 mg) were extracted at 50°C with D2O : CD3OD (80 : 20, 1 ml) containing 0.05% (w/v) d4-TSP (Ward et al., 2003; Beale et al., 2010). NMR data was collected using 128 scans on a Bruker 600 MHz Avance NMR spectrometer (Bruker BioSpin, Coventry, UK) with a water suppression pulse sequence. Data was scaled to an internal standard (d4-TSP) and binned to regions of equal width (0.01 ppm). For GC-MS analysis, lyophilized plant material (1.00 mg) was mixed with methoxyamine hydrochloride in anhydrous pyridine (50 μl, 16 mg ml−1, containing 40 μg ml−1 ribitol) and the samples were heated at 30°C for 90 min with agitation at 550 rpm. MSTFA + 1% TMCS (70 μl) was then added and the sample was heated for 30 min at 37°C with agitation at 550 rpm. The derivatized samples were equilibrated at room temperature for 2 h before analysis. GC-MS analysis was carried out on a Pegasus III time of flight mass spectrometer (Leco Corp., St. Joseph, MI, USA) coupled to an Agilent 6890N gas chromatograph (Agilent Technologies) system fitted with a Microseal septum (Merlin Instrument Co., Newark, DE, USA), a FocusLiner port liner (SGE Analytical Science, Milton Keynes, UK) packed with deactivated quartz wool and a DB-5ms capillary column (15 m × 0.18 mm I.D. × 0.18 μm d.f. with 5 m integrated guard column). Data were exported from ChromaTOF software as netCDF files and converted to MassLynx format using the DataBridge module of MassLynx v4.0 software (Waters, Elstree, UK) for semi-quantification in the QuanLynx module of MassLynx. A unique mass was assigned to each analyte observed in the dataset and its contribution (response factor) to the total ion current of its corresponding mass spectrum determined. For each analyte, peak areas for the unique mass were determined by integration of its corresponding extracted ion chromatogram, and the extracted ion chromatogram areas and response factors were used to calculate a reconstituted TIC peak area. Principal component analysis was carried out using Simca-P v.11 software (Umetrics, Umea, Sweden) with unit variance scaling. Significance was determined using ANOVA.
Identification of a conserved class of PAP, with a C-terminal hydrophobic motif, in photosynthetic eukaryotes
AtPAP2 (AT1G13900) was classified as a group IIa PAP, among the 29 AtPAPs (Li et al., 2002). The open reading frame of AtPAP2 encodes a polypeptide of 656 amino acids with a predicted molecular weight of 73.72 kDa and a pI of 6.11. Database searches revealed that the protein contains a metallophosphoesterase domain within amino acids 256–487. In the genomic DNA, AtPAP2 contains two exons. The program TMHMM predicts that AtPAP2 possesses a transmembrane motif at its C-terminus, suggesting that AtPAP2 is a membrane protein.
Phylogenetic analysis showed that PAPs with this C-terminal motif could only be found in plants, not in animals (Supporting Information Fig. S1). There are many more PAP-like sequences in the genomes of higher plants than in the genomes of primitive photosynthetic eukaryotes. For instance, there are 29 putative PAP genes in the genome of Arabidopsis (Li et al., 2002) and two in Ostreococcus tauri (Fig. S1). Unlike Arabidopsis, where two PAPs contain a C-terminal hydrophobic region (AtPAP2 and AtPAP9), only a single PAP-like sequence from other plant genomes was found to carry this additional hydrophobic motif at the C-terminus (Figs S1, S2, Table S1). Interestingly, in the genome of the smallest free-living photosynthetic eukaryote, O. tauri (Derelle et al., 2006), only two PAP-like sequences could be found among its 8166 ORFs, of which one (OtPAP2) carries a C-terminal hydrophobic motif (Table S1). In addition to this hydrophobic motif, all the sequences homologous to AtPAP2 carry an acidic amino acid doublet (e.g. EE, DD, ED) in the last 2–8 residues of the polypeptides (Table S1). These findings indicate that PAPs with a C-terminal hydrophobic motif are only found in photosynthetic eukaryotes, and have been conserved during the evolution of green plants.
Expression of AtPAP2 is not affected by phosphorus and sugar treatments
To examine the expression pattern of the AtPAP2 protein in various plant organs and at various developmental stages, a polyclonal antiserum was raised against recombinant AtPAP2 protein. A 74 kDa protein was detected in WT Arabidopsis but was absent in the T-DNA insertion line (Fig. S3). This indicates that the antiserum recognizes AtPAP2 specifically. AtPAP2 is highly expressed in siliques, stems, flowers, roots and senescing leaves, but its expression is relatively lower in leaves and mature seeds (Fig. S4). These results corroborate the results from RT-PCR experiments (Zhu et al., 2005). Since AtPAP2 was predicted to be a constitutively expressed protein by spot history analysis (data not shown), the level of AtPAP2 protein expression under various stimuli was examined. As shown in Fig. S5, inorganic phosphate starvation, sucrose, glucose, sorbitol and mannitol treatments did not affect the level of AtPAP2, confirming that its expression is constitutive.
AtPAP2 overexpression accelerates plant growth and produces higher seed yield
A homozygous T-DNA line of AtPAP2 that carries a T-DNA insertion site in the second exon (1864 bp/1971 bp) was identified (Fig. S3a). The insertion resulted in the loss of AtPAP2 expression, as indicated by the absence of detectable transcript and protein (Fig. S3). Under normal germination conditions on MS agar, the length of hypocotyls of the AtPAP2 T-DNA insertion line was shorter (c. 20–30%) than WT. To confirm whether the disruption of AtPAP2 was responsible for the phenotype, the T-DNA insertion line was complemented by introducing the full-length coding sequence of AtPAP2 into Arabidopsis under the control of the CaMV 35S promoter. Four homozygous lines were verified by western blot analysis and their hypocotyls were similar to WT seedlings (data not shown). Therefore, AtPAP2 expression indeed rescued the short hypocotyl phenotype of the T-DNA line.
Overexpression studies were used to dissect the roles of AtPAP2 in vivo. T3 homozygous AtPAP2 OE transgenic lines under the control of CaMV 35S promoter were generated. Multiple independent lines with increased levels of AtPAP2 protein, as confirmed by western blot analysis, were obtained. Two homozygous lines, OE7 and OE21, were randomly selected for subsequent studies. Compared to WT, both AtPAP2 OE lines displayed earlier bolting (Fig. 1) and produced more lateral branches and inflorescences. Typically, AtPAP2 OE plants started bolting earlier than both WT and the T-DNA line, c. 6 d earlier under LD growth conditions and c. 14 d earlier under SD conditions (Table 1a). At day 27 (LD) and day 41 (SD), when WT first bolted, AtPAP2 OE lines had produced significantly more cauline leaves and inflorescences than WT (Fig. S6). Conversely, the number of rosette leaves of AtPAP2 OE lines was fewer than WT (Fig. 1b, Table 1a), although the leaf area and total biomass of OE lines were similar to WT at 20 d old, at day 40, the total leaf area of AtPAP2 OE lines was only 65% that of WT, but the total biomass was 2-fold higher than WT (Fig. 1c). AtPAP2 OE plants not only matured earlier but also produced higher seed yields. In two independent experiments, the total seed yield of AtPAP2 OE7 and OE21 plants increased remarkably by 40–57% over WT, which was due to higher numbers of auxiliary branches and a higher silique density (Table 1b). No significant differences in plant growth or seed yield relative to WT were observed in the AtPAP2 T-DNA line, possibly due to redundancy of gene function.
Table 1. Growth phenotypes of transgenic Arabidopsis
(a) Bolting time
Long day (LD, 16 h : 8 h)
Short day (SD, 8 h : 16 h)
26.9 ± 1.2a
13.0 ± 0.8a
41.0 ± 4.7a
18.0 ± 3.0a
25.7 ± 0.7a
11.6 ± 1.1a
40.7 ± 4.9a
15.0 ± 3.0a
20.0 ± 1.1b
6.4 ± 0.5b
25.6 ± 1.3b
5.3 ± 0.5b
20.8 ± 0.6b
6.5 ± 0.7b
26.0 ± 1.1b
5.4 ± 0.5b
(b) Seed yield at maturity
No. of siliques per plant
Seed yield (g per plant)
Seed weight (mg per 100 seeds)
396 ± 89a
0.225 ± 0.058a
1.96 ± 0.14a
386 ± 70a
0.240 ± 0.049a
2.02 ± 0.16a
621 ± 76b
0.351 ± 0.050b
1.91 ± 0.10a
625 ± 94b
0.355 ± 0.066b
1.98 ± 0.03a
(c) Carbohydrate contents in 20-d-old Arabidopsis shoots
Sucrose (μg g−1 FW)
Hexose (μg g−1 FW)
Starch (mg g−1 FW)
AEI, average date of emergence of inflorescence.
NRL, no. of rosette leaves at the first appearance of inflorescence.
All seeds and siliques from a single plant were harvested after the plant was completely dried. The plants were grown under LD regime.
Sugars were extracted 8 h after illumination during the LD regime from 20-d-old Arabidopsis.
Seeds of wild-type (WT), T-DNA line and two AtPAP2 overexpression (OE) lines (OE7 and OE21) were germinated in the MS agar for 10 d, seedlings with the same size were transferred to a growth chamber. Statistical differences (P <0.05) in the same column for each line were based on one-way ANOVA analysis followed by Tukey’s Honestly Significant Differences (HSD) test using statistical program SPSS 11.5. Within each column, the values marked by different letters (a, b, c) are significantly different (P < 0.05). There were 8–15 replicates for each line. The data were reproducible in at least two independent experiments.
110.1 ± 33.3a
16.7 ± 7.1a
3.3 ± 0.4a
116.1 ± 15.6a
24.9 ± 7.3a
4.4 ± 1.1a
207.1 ± 22.5b
43.9 ± 9.2b
3.8 ± 1.0a
147.7 ± 17.3c
31.7 ± 4.8ac
3.5 ± 0.8a
AtPAP2 is dual-targeted to chloroplasts and mitochondria
The protein sequence of AtPAP2 carries a hydrophobic C-terminal motif (TMD/CT) predicted to carry a trans-membrane motif and it has already been identified as a chloroplast protein by proteomics (Kleffmann et al., 2004). By contrast, the SignalP server predicts that AtPAP2 has a 24 amino acid SP at its N-terminus to direct the protein to the secretory pathway. To investigate the targeting of AtPAP2, 30 amino acid residues from the N-terminus and 69 amino acid residues from the C-terminus were fused to green fluorescent protein (GFP) under the control of the CaMV 35S promoter (Fig. 2a). Red fluorescence protein (RFP) was fused to either the mitochondrial presequence of the ATP synthase F1-γ subunit from Nicotiana plumbaginifolia (F1–RFP) (Duby et al., 2001) as a mitochondrial targeting marker or to the targeting sequence (a.a. 1–79) of the small subunit of tobacco Rubisco (plastid-mCherry) (Dabney-Smith et al., 1999) as a chloroplast targeting marker. Transient expression of the GFP fusion constructs was performed in WT Arabidopsis PSB-D cell culture (Fig. 2a).
Confocal microscopy analysis showed that both GFP (Fig. 2b,c, green) and GFP fused to the N-terminal SP (SP-GFP) (Fig. 2d,e, green) localized to the cytosol, and did not co-localize with the mitochondrial F1-RFP (Fig. 2b,d, red) or plastid-mCherry (Fig. 2c,e, red) markers. However, GFP constructs containing the TMD/CT region of AtPAP2, including SP-GFP-TMD/CT (Fig. 2f,g) and GFP-TMD/CT (Fig. 2h,i), were targeted to both plastids and to mitochondria, which were also labeled by the MitoTracker Orange (Molecular Probes) (Fig. S7). In addition, the patterns of expression of all GFP constructs were distinct from that of the Golgi marker (RFP-ManI) (Cai et al., 2011), ER marker (RFP-HDEL) (De Caroli et al., 2011) (Fig. S8) and peroxisome marker (peroxisome-mCherry) (Nelson et al., 2007) (Fig. S9). These results confirmed the importance of the TMD/CT in dual targeting of the reporter GFP protein to mitochondria and plastids, and the TMD/CT alone is sufficient to direct the GFP protein to both organelles.
To confirm the subcellular localization of AtPAP2, protein extracts from isolated chloroplasts and mitochondria were probed with the antibody against AtPAP2. A protein band with an apparent molecular mass of 74 kD was detected in both mitochondria and chloroplasts. The purity of the mitochondria and chloroplasts was verified using antibodies raised against a variety of mitochondrial and chloroplast specific proteins (Fig. 2j).
C-terminal targeting is crucial for AtPAP2 function
To further study the function of the C-terminal motif and the putative signal peptide of AtPAP2, OE lines with AtPAP2 truncation of the C-terminal motif (a.a. 588–656) or its predicted SP (a.a. 1–24) were generated in Arabidopsis by Agrobacterium-mediated plant transformation (Fig. 3). Western blot analysis was employed to verify its overexpression in homozygous transgenic lines.
As shown in Fig. 3(a), transgenic Arabidopsis lines expressing SP-deleted AtPAP2 (P2NS, a.a. 25–656) exhibited phenotypes similar to the full-length AtPAP2 OE lines, such as earlier bolting and higher biomass. By contrast, when the C-terminal motif was deleted, the OE lines (P2NC, a.a. 1–588) grew like the WT Arabidopsis (Fig. 3a). These results suggest that the C-terminal motif of AtPAP2 is critical for its function in Arabidopsis. By contrast, the putative N-terminal signal peptide of AtPAP2 does not seem to be essential.
In AtPAP2 OE lines (P2OE and P2NS), two protein bands could be seen in membrane fractions but only the low molecular weight band appeared in the soluble fraction. Furthermore, the high molecular weight protein band was not seen in the membrane fraction from the P2NC lines (with no C-terminal motif). Therefore, it is likely that the high molecular weight protein band is a membrane-associated full-length AtPAP2 and the cleavage of its C-terminal motif releases it to the soluble fraction. Correlated with the growth phenotype, the presence of the membrane-associated high molecular weight form is essential to the robust growth phenotypes of the OE lines (Fig. 3b).
AtPAP2 overexpression changes the C metabolism of Arabidopsis
Sugars supplied by the shoot are the driving force for plant growth and anabolism. We analyzed the carbohydrate contents of leaves from 20-d-old plants halfway through the light period. The size and morphology of leaves from various lines were still similar at day 20 and therefore this age was chosen for further analysis. Both sucrose and hexose sugar contents were greatly elevated in the shoots of AtPAP2 OE plants (Table 1c), in which sucrose was 1.3- to 1.8-fold and hexose sugars were 1.9- to 2.6-fold that of WT (P <0.01). However, the starch level did not differ significantly between the various lines (Table 1c).
The increased sucrose and hexose sugars (Table 1c) in the AtPAP2 OE plants may be due to altered activity of enzymes involved in sugar metabolism. SPS activity was significantly enhanced in the 20-d-old AtPAP2 OE lines under both optimal Vmax and limiting Vlimit conditions, and was not affected by light/dark transitions, as in WT Arabidopsis (Huber et al., 1989). In addition, the activities of enzymes involved in sucrose catabolism, including sucrose synthase, acid and alkaline soluble invertase, insoluble cell wall invertases, were unaffected (Table 2). The enzyme activity of acid phosphatase was elevated in both OE lines but the total NR activity was only significantly elevated in one OE line (OE21, Table 2). The higher SPS activity could result from either a higher level of protein expression or a higher activity per enzyme molecule. Western blotting indicated that the SPS protein levels were significantly elevated in the OE lines during both day and night. The levels of the other enzymes, such as fructose-1,6-bisphosphatase (FBPase) and FBP aldolase in the sucrose synthetic pathway did not change significantly (Fig. 4).
Table 2. Enzyme assays: samples were collected 8 h after the light period (Day) and 4 h after the dark period (Night)
Values marked with a different letter were determined by the t test to be significantly different from wild-type (WT) and were determined by the ANOVA–Tukey test to be significantly different (P <0.05) from each other. There are three biological replicates in this study.
Sucrose phosphate synthase (μmol sucrose mg−1 total soluble protein h−1)
5.44 ± 0.90a
5.80 ± 0.59a
7.89 ± 1.08b
7.97 ± 0.95b
3.18 ± 0.32a
3.47 ± 0.12a
4.54 ± 0.92b
3.99 ± 0.68a
5.91 ± 0.57a
5.24 ± 0.72a
7.54 ± 0.97b
6.82 ± 0.75a
3.74 ± 0.31a
2.83 ± 0.16b
4.67 ± 0.18c
4.86 ± 0.54c
Sucrose synthase (μmol glucose mg−1 total soluble protein h−1)
12.46 ± 0.23a
12.35 ± 1.20a
12.43 ± 0.22a
12.76 ± 0.27a
12.47 ± 0.36a
12.63 ± 0.44a
12.54 ± 0.43a
12.94 ± 0.27a
Soluble invertase (μmol glucose mg−1 total soluble protein h−1)
0.71 ± 0.11a
0.60 ± 0.27a
0.91 ± 0.45a
0.90 ± 0.04a
0.70 ± 0.34a
1.30 ± 0.5a
0.88 ± 0.21a
0.69 ± 0.04a
8.48 ± 0.49a
8.06 ± 1.61a
8.04 ± 1.39a
8.94 ± 0.84a
6.50 ± 0.41a
5.25 ± 0.63a
6.80 ± 0.68a
6.83 ± 0.07a
Insoluble cell wall invertase (μmol sucrose mg−1 total soluble protein h−1)
1.11 ± 0.19a
0.90 ± 0.24a
1.17 ± 0.31a
1.20 ± 0.33a
1.08 ± 0.50a
0.85 ± 0.21a
1.41 ± 0.21a
1.11 ± 0.16a
6.08 ± 0.14a
6.35 ± 0.11a
5.09 ± 0.32a
5.27 ± 0.10a
6.94 ± 0.31a
7.57 ± 0.43a
6.17 ± 0.28a
6.77 ± 0.73a
Nitrate reductase (μmol nitrite g−1 FW h−1)
Middle of the day
0.14 ± 0.00a
0.15 ± 0.01a
0.17 ± 0.00a
0.24 ± 0.02b
Middle of the night
0.13 ± 0.01a
0.13 ± 0.05a
0.15 ± 0.04ab
0.20 ± 0.01c
Acid phosphatase activity (μmol Pi mg−1 total soluble protein h−1)
Middle of the day
8.33 ± 1.38a
8.05 ± 1.14a
18.46 ± 1.91b
16.59 ± 1.04b
Middle of the night
6.04 ± 0.11a
6.00 ± 0.81a
10.71 ± 0.12b
13.71 ± 0.56c
To investigate whether enhanced SPS activity was due to a change in the relative amounts of phosphorylated SPS vs total SPS, a 14-3-3 capture experiment was carried out. The recombinant 14-3-3 captured similar amounts of phosphorylated SPS in all four lines (Fig. 4). Hence, the increase in SPS activity in the OE lines was due to an increase of the unphosphorylated active form of enzyme. Theoretically, a suppression of SnRK1 kinase activity could enhance SPS activity. Hence, the expression levels of SnRK1 protein subunits and the phosphorylation status of the Thr175 residue on its activation loop were examined by western blotting (Fig. S10); the levels of the AtSnRK1.1, AtSnRK β1 and AtSnRK β2 subunits were unaltered. Blots using antiserum against phosphorylated-AMPK (T172), which cross-reacts with phosphorylated Thr175 on AtSnRK1 in Arabidopsis (Sugden et al., 1999), indicated that the phosphorylation status of T175 was unaltered. This suggested that the higher SPS activity in the OE lines was not mediated by the SnRK1 pathway.
Overexpression of AtPAP2 results in alteration of soluble plastid proteins
Two-dimensional gel analysis of total leaf soluble proteins revealed that two proteins involved in photosynthesis exhibited changes in the OE lines; first, PSBP-1, an oxygen-evolving enhancer protein (AT1G06680), was induced, and second, Rubisco activase (RCA, AT2G39730) showed a mobility shift (Fig. 5). PSBP-1 is required for high levels of oxygen evolution (Mayfield et al., 1987). RCA enhances photosynthetic rate by freeing Rubisco for the next round of C fixation (Portis, 2003; Portis et al., 2008). In Arabidopsis, there are two RCA genes (At1g73110 and At2g39730), and alternative splicing of At2g39730 produced three isoforms. Deletion of the short thermolabile isoform of At2g39730 (RCA1) caused decreased seed yield while transgenic Arabidopsis overexpressing thermostable RCA1 exhibited higher biomass and photosynthetic rate (Kurek et al., 2007). According to the PhosphAT database (phosphat.mpimp-golm.mpg.de), there is at least one phosphorylation site in both gene products (GLAYDpTSDDQ on At2G39730.1; AGGMEpTLGKV and VPLIVpTGNDF on AT1G73110.1). A change in phosphorylation status of RCA conceivably would affect pI, but whether this occurred was not determined.
Metabolomic analysis was carried out using a combination of NMR-MS and GC-MS. Shoot samples were collected from 20-d-old plants before bolting. Principal component analysis of 1H-NMR and GS-MS datasets (Fig. S11) was carried out to discern any global changes in the metabolomes of these lines. There was a very clear separation of the OE lines from WT and the T-DNA insertion line by each analytical technique. 1H-NMR was also able to discriminate the T-DNA line from WT samples, although this separation was much smaller. The main changes that could be discerned by 1H-NMR between the T-DNA samples and WT included a reduction in fumarate and citrate levels and several amino acids.
Combining the results of metabolomics analysis and LC-MS-MS measurements, metabolites which were increased in the OE lines included alanine, proline, several tricarboxylic acid (TCA) metabolites (citrate, fumarate, malate), sugars (glucose and sucrose) and a metabolite putatively identified as indole acetonitrile, a breakdown product of indole glucosinolate. Reduced metabolites in the OE lines included certain amino acids (asparagine, aspartate, glutamine, lysine, methionine, phenylalanine, threonine, tyrosine), GABA, ethanolamine, 4-hydroxybenzoic acid, ribose, phytol, linoleic acid, and linolenic acid (Table S2). In addition, OE lines contained significantly higher levels of putrescine and significantly lower levels of spermidine. The metabolites detected in the glycolysis, TCA and amino acid synthesis pathways are presented in Fig. 6.
The genomes of plant species generally contain multiple PAP-like sequences, among which, except Arabidopsis, usually only a single PAP gene carries a C-terminal hydrophobic motif. Amino acid sequence alignment of the 29 PAP-like sequences from the Arabidopsis genome showed that only two AtPAPs (AtPAP2, AT1G13900 and AtPAP9, AT2G03450) carry an additional hydrophobic motif (a.a. 614–636 of AtPAP2 and a.a. 604–626 of AtPAP9) at their C-termini (Table S1), which was predicted as a trans-membrane motif by the TMHMM program. Although AtPAP9 shares up to 72% sequence identity with AtPAP2 at protein level, AtPAP9 did not exhibit the biological function of AtPAP2. A homozygous T-DNA line of AtPAP9 (Salk_129905) and several homozygous AtPAP9 OE lines were identified by genomic PCR and RT-PCR analysis. All of them exhibited normal growth phenotypes, which were very different from that of AtPAP2 OE lines (data not shown). A double T-DNA insertion line of AtPAP2 and AtPAP9 was also generated through crossing the pollens of AtPAP2 insertion line and the pistil of AtPAP9 insertion line and the interruption of both genes were confirmed by genomic PCR analysis (Fig. S3). The double insertion line exhibited similar phenotypes with that of the AtPAP2 T-DNA line (data not shown). In addition, two PAP-like sequences could be identified in the genomes of the small free-living photosynthetic eukaryotes O. tauri (Derelle et al., 2006) and Micromonas pusilla (Worden et al., 2009), of which only one carries a hydrophobic C-terminal motif. O. tauri, belonging to the Prasinophyceae, is the smallest (size < 1 μm) free-living photosynthetic eukaryote known. This unicellular organism lacks a cell wall and each cell carries one chloroplast, one mitochondrion, one Golgi body and a large nucleus. The Prasinophyceae are a primitive group of marine green algae, which suggests an ancient ancestral function of AtPAP2-like proteins in the regulation of C partitioning in photosynthetic organisms.
In our study, AtPAP2 was shown to be dual-targeted to both chloroplasts and mitochondria, key organelles for two important energy processes in plant cells: photosynthesis and respiration (Nunes-Nesi et al., 2007, 2011). An increasing number of proteins in mitochondria and plastids have been found to be encoded by single nuclear genes, and are referred to as dual-targeted proteins (Peeters & Small, 2001). To date, as many as 47 different proteins from seven plant species have been reported to be dual-targeted (Carrie et al., 2009a). In Arabidopsis, 11 PKs and 10 PPs have been shown to be targeted to the chloroplast and all are nuclear encoded (Schliebner et al., 2008; Shapiguzov et al., 2010). In mitochondria, 10 PKs and one PP were identified in Arabidopsis (Heazlewood et al., 2004). All of these PKs and PPs are targeted by N-terminal targeting signals. The identification of a C-terminal dual-targeting signal in AtPAP2 is novel and AtPAP2 is the only PP that is dual-targeted to both organelles.
Overexpression of AtPAP2 in Arabidopsis leads to a faster growth rate. The higher biomass but lower total leaf area of 40-d-old transgenic plants provided indirect evidence of a higher photoassimilation rate per leaf area in the OE lines. The changes in RCA and PSBP-1 in the OE lines might also affect the photosynthetic activity of the OE lines. Whether the observed modification of RCA affects the activity of Rubisco and thus results in the enhanced growth rate of AtPAP2 OE lines is an interesting topic for further study.
Sucrose is the fuel and building block for sink tissues. AtPAP2 overexpression drastically enhanced the sucrose and hexose levels in rosette leaves while the starch level did not change significantly. SPS activity shows a high correlation with sucrose production in leaves (Baxter et al., 2003; Haigler et al., 2007) and overexpression of SPS promotes earlier flowering, fruit development, and increased biomass in tomato (Micallef et al., 1995) and tobacco (Park et al., 2008). In the two AtPAP2 OE lines, SPS was upregulated at the protein level, which resulted in a significant enhancement of its activity. SnRK1 can phosphorylate and inhibit SPS activity (Halford et al., 2003) under high sucrose levels. Since overexpression of AtSnRK1.1 resulted in phenotypes opposite to that of the AtPAP2 OE plants, such as delayed flowering and retarded growth (Baena-Gonzalez et al., 2007), it may be worth asking whether AtPAP2 is an antagonistic phosphatase of SnRK’s kinase activity. Under energy deficiency, activated SnRK1 phosphorylates SPS hence suppresses activity posttranslationally (Toroser et al., 1998; Moorhead et al., 1999). By contrast, overexpression of AtPAP2 resulted in higher SPS activity. Theoretically, a phosphatase (Sugden et al., 1999) that specifically dephosphorylates SnRK1 or its upstream kinases (e.g. GRIK) in vivo might suppress their biological activity and thus boost the activity of SPS. However, it is unlikely that AtPAP2 mediated its actions through GRIK or SnRK1 because the phosphorylation status of the Tl75 residue was unaltered in the OE lines; also, pull-down experiments only found an interaction between GRIK and SnRK (data not shown), but not between AtPAP2 and GRIK or SnRK1 (Fig. S10). Unlike SnRK1, which inactivates SPS by posttranslational modification, AtPAP2 overexpression enhances sucrose synthesis by increasing the expression of SPS.
The mechanism of how AtPAP2 stimulates plant growth remains unclear. As the acid phosphatase activity was significantly increased in the AtPAP2 OE lines (Table 2), AtPAP2 may act as a PP or as a phosphatase for inorganic P compounds. In the former scenario, AtPAP2 might exhibit its effects as a PP like PPH1, a chloroplast phosphatase that dephosphorylates light-harvesting complex II during state transition (Finazzi et al., 2002; Shapiguzov et al., 2010). In the latter scenario, AtPAP2 might increase the acid phosphatase activity in the chloroplasts and mitochondria, thus releasing more free inorganic phosphate (Pi). In chloroplasts, as Pi is an activator of Rubisco activity, production of triose phosphates might be enhanced by a higher Pi supply (Giersch & Robinson, 1987), and in turn provide more triose phosphate for sucrose synthesis. Similarly, the possibly higher Pi supply in the mitochondrial matrix could also modulate TCA flux and activate oxidative phosphorylation and ATP synthesis (Wu et al., 2007). The higher concentrations of fumarate and malate in OE lines (Fig. 6) might reflect a higher flux from succinyl-CoA to succinate (Johnson et al., 1998), another step in the TCA cycle that produces ATP (Meyer et al., 2010). ATP generated from the TCA cycle could be transported out of the mitochondria to sustain the high rate of sucrose synthesis (Krömer et al., 1993). The higher levels of organic acids of the TCA cycle in the OE lines did not result in a higher amino acid level of the aspartate family, indicating that the regulation and the flux of the TCA cycle is a complicated process. Alteration of key enzymes and their activities in the TCA cycle affect photosynthesis: downregulation of aconitase, the iron–sulfur subunit of succinate dehydrogenase and the mitochondrial malate dehydrogenase enhance the rate of photosynthesis (Carrari et al., 2003; Nunes-Nesi et al., 2005; Araujo et al., 2011), whereas inhibition of citrate synthase, succinyl CoA ligase, or isocitrate dehydrogenase show little effect on the rate (Studart-Guimaraes et al., 2007; Sienkiewicz-Porzucek et al., 2008, 2010; Sulpice et al., 2010). All these lines displayed changes in expression levels of organic acids and amino acids. Thus, the mechanism by which AtPAP2 influences both photosynthesis and the TCA cycle is likely to be complex and may be involved in a number of pathways or factors.
In summary, we have shown that ectopic overexpression of AtPAP2 in Arabidopsis causes dramatic effects on plant growth and C metabolism. Its biological activity is dependent on its dual targeting to both plastids and mitochondria. Whether overexpression of AtPAP can be used as a tool to boost the yield of important crop plants will be an interesting subject of future studies.
This project was supported by the General Research Fund (HKU772710M) and the Innovation and Technology Fund (ITS158/09) of the HKSAR, China. We thank Dr Yip’s lab at HKU for kindly providing plant vectors and technical support for plant cultures. We thank Aimee Llewellyn and Delia Corol at Rothamsted Research for preparation of the analytical extracts and for 1H-NMR data collection, Dr David Secco at the University of Western Australia for his comments on this manuscript.