PAPP5 Is Involved in the Tetrapyrrole Mediated Plastid Signalling during Chloroplast Development

The initiation of chloroplast development in the light is dependent on nuclear encoded components. The nuclear genes encoding key components in the photosynthetic machinery are regulated by signals originating in the plastids. These plastid signals play an essential role in the regulation of photosynthesis associated nuclear genes (PhANGs) when proplastids develop into chloroplasts. One of the plastid signals is linked to the tetrapyrrole biosynthesis and accumulation of the intermediates the Mg-ProtoIX and its methyl ester Mg-ProtoIX-ME. Phytochrome-Associated Protein Phosphatase 5 (PAPP5) was isolated in a previous study as a putative Mg-ProtoIX interacting protein. In order to elucidate if there is a biological link between PAPP5 and the tetrapyrrole mediated signal we generated double mutants between the Arabidopsis papp5 and the crd mutants. The crd mutant over-accumulates Mg-ProtoIX and Mg-ProtoIX-ME and the tetrapyrrole accumulation triggers retrograde signalling. The crd mutant exhibits repression of PhANG expression, altered chloroplast morphology and a pale phenotype. However, in the papp5crd double mutant, the crd phenotype is restored and papp5crd accumulated wild type levels of chlorophyll, developed proper chloroplasts and showed normal induction of PhANG expression in response to light. Tetrapyrrole feeding experiments showed that PAPP5 is required to respond correctly to accumulation of tetrapyrroles in the cell and that PAPP5 is most likely a component in the plastid signalling pathway down stream of the tetrapyrrole Mg-ProtoIX/Mg-ProtoIX-ME. Inhibition of phosphatase activity phenocopied the papp5crd phenotype in the crd single mutant demonstrating that PAPP5 phosphatase activity is essential to mediate the retrograde signal and to suppress PhANG expression in the crd mutant. Thus, our results suggest that PAPP5 receives an inbalance in the tetrapyrrole biosynthesis through the accumulation of Mg-ProtoIX and acts as a negative regulator of PhANG expression during chloroplast biogenesis and development.


Introduction
The chloroplasts house the photosynthetic light reactions where sunlight is converted into chemical energy in the form of NADPH and to an electrochemical gradient over the thylakoid membrane that is subsequently used to synthesize ATP.Plastids are also the location of a number of vital metabolic pathways, including primary carbon metabolism and the biosynthesis of fatty acids, amino acids and tetrapyrroles.Plastids exhibit a very clear developmental program where all plastids are derived from proplastids present in meristematic cells either direct or via the dark-grown intermediate form known as etioplasts.The etioplasts contain a characteristic lattice-like membrane structure known as the prolamellar body.The prolamellar body also contains the precursor of chlorophyll, protochlorophyllide, bound to its reducing enzyme protochlorophyllide oxidoreductase (POR).Following exposure to light the prolamellar body forms the thylakoid membrane and POR is activated to convert protochlorophyllide into chlorophyllide a, which is subsequently converted into chlorophyll a and b [1].Plants can detect almost all wavelengths of light using three major classes of photoreceptors: the red/far-red light absorbing phytochromes, the blue/UV-A light absorbing cryptochromes and phototropins, and the UV-B sensing UV-B receptors [2].These photoreceptors perceive light signals and initiate intracellular signalling pathways involving proteolytic degradation of signalling components and large reorganization of the transcriptional program to modulate plant growth and development [2].When dark grown seedlings are exposed to light as much as one-third of the nuclear encoded genes show transcription changes [3] and among the genes dramatically up-regulated in the light are genes encoding chloroplast-targeted proteins.Thus, it is clear that the initiation of chloroplast development in the light is dependent on nuclear encoded components through so called anterograde mechanisms [4].Although the control of chloroplast development appears to be overwhelmingly under nuclear control there are several reports demonstrating that nuclear genes encoding key components in the photosynthetic machinery are regulated by signals originating in the plastids, so called retrograde mechanisms [5,6,7].
During chloroplast development a tight stoichiometric assembly of nuclear-encoded and plastid-encoded proteins together with chlorophylls and carotenoids is essential.To achieve this there must be a communication between the chloroplast and the nucleus through retrograde signalling.The first evidence of the existence of a ''plastid signal'' came from studies of mutants with morphologically aberrant plastids.These include mutants with defective plastid protein synthesis such as the plastid ribosome-deficient albostrians barley mutant and the Brassica napus al mutant [8,9,10].These mutants demonstrated reduced expression of nuclearencoded plastid components suggesting that a plastid signal was emitted to repress the nuclear encoded photosynthesis genes [10].We now know that several different plastid processes produce signals that regulate specific sets of genes or regulons and several molecular candidates for plastid signals have been described [11].Plastid signals are essential to the plant both during the initial developmental stages (biogenic control) and in adult stage to face changes in the environment (operational control) [12].
One of the plastids signals described to regulate the expression of photosynthesis associated nuclear genes (PhANGs) is linked to the tetrapyrrole biosynthesis and accumulation of the intermediates the Mg-ProtoIX and its methyl ester Mg-ProtoIX-ME [13,14,15,16,17,18].Mg-ProtoIX and Mg-ProtoIX-ME have been shown to accumulate when plants and algae are exposed to factors that give rise to oxidative stress such as exposure to low temperatures and inhibitors of photosynthetic electron transport [18,19,20,21].The reason for the accumulation of these particular intermediates could be explained by the fact that the aerobic cyclase reaction was shown, both in Arabidopsis and cucumber, to be extremely sensitive to oxidative stress [20,21].Thus, impaired flux through chlorophyll biosynthesis and the accumulation of Mg-ProtoIX/Mg-ProtoIX-ME is an indicator of changes in the environment and results in changes in PhANG expression.Whether accumulation of Mg-ProtoIX is itself an important part of the tetrapyrrole-mediated signal remains unclear [22,23].In order to investigate whether Mg-ProtoIX itself is an important part of the tetrapyrrole-mediated signal, cytosolic ligands of Mg-ProtoIX were isolated from a proteomic study [17] and the Phytochrome-Associated Protein Phosphatase 5 (PAPP5) was isolated as a one of the putative Mg-ProtoIX interacting proteins [17].PAPP5 is a type 5 serine/threonine protein phosphatase that dephosphorylates biologically active Pfr phytochromes and enhances the phytochrome-mediated photoresponses.Depending on the specific serine residues dephosphorylated by PAPP5, phytochrome stability and affinity for a downstream signal transducer, NDPK2, were enhanced [24].Chloroplast biogenesis and development is triggered by exposure to light [1,25] and possibly PAPP5 is the suggested link between plastid and light signalling networks [26,27].
In order to elucidate if there is a biological link between PAPP5 and the tetrapyrrole mediated signal we prepared double mutants between the papp5 mutant and the chl27/crd Arabidopsis mutant.CHL27/CRD is a nuclear gene encoding one of the potential subunits of the aerobic Mg-protoporphyrin monomethyl ester cyclase complex and the crd mutant over-accumulates Mg-ProtoIX and Mg-ProtoIX-ME and the tetrapyrrole accumulation affects the retrograde signalling in the crd mutant [28].Surprisingly, introducing the papp5 mutation into the crd background reverts the pale phenotype and restores normal chloroplast development and light stimulated induction of PhANG expression in crd.Our results suggest that PAPP5 is a component in the tetrapyrrole-mediated plastid signalling pathway and genetic, physiological and biochemical data supporting this model will be discussed.

Plant material and growth conditions
The different Arabidopsis thaliana lines used in this study were Col-0 wild type, crd/chl27 (Salk_009052) and papp5 (Salk_021153) (Figure S1).T-DNA insertion lines were obtained from NASC stock centre.For aseptic growth, seeds were surface-sterilized with 95% Ethanol and 0.5% Triton X100 (v/v), washed tree times with 95% ethanol, stratified at 4uC for 2 days in the dark and spread on 1% sucrose MS agar plates complemented with vitamins (Duchefa).Seed germination was induced by exposing plates to 100 mmol photons m 22 s 21 light, 22uC for 12 hours.Plates were transferred to darkness for 3 days and then shifted to growth cabinet in continuous light during 24 hours (100 mmol photons m 22 s 21 , 22uC).Mature plants were grown on soil for 6 weeks (100 mmol photons m 22 s 21 , 9 hr light/15 hr darkness, 22uC).

Morphological and physiological analysis
For chlorophyll content analysis, 6-weeks-old plants grown on soil at 100 mmol photons m 22 s 21 , 9 hr light/15 hr darkness, 22uC.Samples were ground with liquid nitrogen and 1 mL of buffered acetone (80% acetone, 25 mM HEPES pH 7.5) was added to 100 mg of material and incubated over night at 4uC.Chlorophyll content was determined according to Porra [29].For TEM, 6-weeks-old plants grown on soil and seedlings grown on MS 16 agar plates supplemented with 1% sucrose were prepared, embedded and cut according to Keskitalo et al. [30].Flowering time was determined by counting the number of leaves and days when floral buds were visible at the centre of the rosette.

cDNA synthesis and Real-Time PCR analysis
Total RNA was isolated using Total RNA Miniprep kit (EZNA) according to the manufacturer's instructions.Total RNA concentration was determined with a Nanodrop ND-100 spectrophotometer.cDNA was synthesized from 0.5 mg of total RNA using the iScript cDNA synthesis kit (Bio-Rad) according to the manufacturer's instructions.2 ml of 10-fold diluted cDNA was used in 10 ml iQ SYBR Green Supermix reaction (BioRad).All reactions were performed in triplicates.The different sets of primers used are described in Table S1.RT-PCR was run in a CFX96 real time system (BioRad) and analyzed using LinRegPCR [31,32].

ALA/Mg-ProtoIX feeding experiment and tetrapyrrole quantification
Seedlings were grown on MS agar plates supplemented with 1% sucrose for 2 weeks under long day conditions (100 mmol photons m 22 s 21 , 15 hr light/9 hr darkness, 22uC).Liquid medium containing 50 mM of Mg-ProtoIX (Frontier Scientific Inc.) in 16 MS (Murashige and Skoog medium) supplemented with 1% sucrose was spread on the plates before the dark period and samples were collected the next day in the middle of the light period.Mock plates were treated with same medium but without Mg-ProtoIX.For ALA feeding treatment, 2 weeks old seedlings grown in 16 MS plates 1% sucrose under long day conditions (100 mmol photons m 22 s 21 , 15 hr light/9 hr darkness, 22uC) were incubated overnight in darkness in 10 mM potassium phosphate, pH 7.5, 5 mM MgCl 2 , 10 mM 5-aminolevulinic acid (SIGMA-Aldrich) and transferred back to light conditions for one hour before harvesting.Mock samples were incubated in the same phosphate buffer but without 5-aminolevulinic acid.To determine tetrapyrrole content, samples were homogenized in liquid nitrogen and extracted with acetone/1 M NH 4 OH (99:1; v/v) under dim green light at 4uC.Samples were continuously shaken during 40 min and spun for 10 min at 14000 rpm, 4uC.The supernatant was transferred to a fresh microcentrifuge tube and extracted twice with equal volume of hexane.Mg-ProtoIX/Mg-ProtoIX-ME fluorescence emission spectra from 460 to 720 nm of the acetone fraction was recorded at room temperature by using the excitation wavelength 416 nm.Final data analysis was made using Igor Pro software (Ver.6.22A).

Okadaic acid treatment
Seedlings were grown on filter paper soaked with MS medium and 1% sucrose.The plates were transferred to 100 mmol photons m 22 s 21 constant light and 22uC for 12 hours to induce germination.Plates were transferred to darkness for 3 days and then shifted from darkness to light (24 hours at 100 mmol photons m 22 s 21 , 22uC).During the light incubation, the seedlings were treated with 10 nM okadaic acid (SIGMA) diluted in 16 MS medium supplemented with 1% sucrose medium.For the mock treatment, okadaic acid was excluded.

Binding assay
Immunoprecipitation of native PAPP5 complex was performed using 4-weeks old Petita havana tobacco plants transfected with 35S:PAPP5-cMyc expression construct.Tobacco leaves were collected and grinded in liquid nitrogen.Proteins from the leaf tissue were extracted in 1000 mL immunoprecipitation buffer (25 mM Tris-HCl, pH 7.8, 75 mM NaCl, 10 mM MgCl 2 , 2 mM dithiothreitol (DTT), 5 mM EGTA, 0.2% Triton X-100, 10% glycerol, 0.2 mM PMSF) for 60 min, 4uC.Extracts were incubated with 5 mg of anti-cMYC monoclonal antibody (Bio-Site) bound to protein G coated magnetic beads (Dynabeads Protein G Immunoprecipitation, Invitrogen) for 1 h at 4uC.All the washing steps were performed according to the manufacturer's recommendations.The immunoprecipitated protein complex was resuspended in 200 mL of PBS buffer (pH 7.4) and incubated for 1 h with 2.5 mM Mg-ProtoIX at 4uC.Beads were washed with 1 ml PBS three times and proteins were eluted with Native elution buffer.Mg-ProtoIX bound to the PAPP5 complex was quantified by spectrofluorometry with the excitation wavelength at 416 nm.Sample of Mg-ProtoIX incubated with immunoprecipitated proteins from tobacco plants transfected with pGWB16 control plasmid was used as the negative control.

Results
The papp5 mutation reverts the pale crd phenotype in papp5crd double mutant We used a T-DNA insertion mutant (crd) of the CHL27/CRD gene encoding a potential subunit of the cyclase enzyme complex involved in chlorophyll biosynthesis downstream of Mg-ProtoIX [28,35].The crd mutant has a pale phenotype and reduced growth (Figure 1A-B, Figure S2) [28,35,36].The strong reduction in chlorophyll b content compared to chlorophyll a results in a higher chlorophyll a/b ratio in crd compared to wild type (Figure 1B).The T-DNA insertion line for PAPP5, papp5 demonstrated wild type phenotype under our growth conditions.In the papp5crd double mutant the pale phenotype observed in the crd single mutant is lost and the chlorophyll content is recovered to wild type levels (Figure 1A-B) together with the recovery of the morphology and the flowering time (Figure S3).
In addition to the pale phenotype, the chloroplast structure in the crd mutant is different compared to wild type and papp5.The crd chloroplasts are poorly developed containing only stromaexposed thylakoids and abnormal grana thylakoids (Figure 1C) [28].Coinciding with the defect in chloroplast structure is the difference in the arrangements of the photosynthetic complexes in the crd mutant compared to wild type and the papp5 mutant (Figure 1D).Similarly to what was observed for chlorophyll content (Figure 1A-B), the papp5crd double mutant reverts the chloroplast phenotype observed in the crd single mutant (Figure 1C-D).Thus, in contrast to the crd single mutant the papp5crd double mutant displays normal growth and chlorophyll content, fully developed chloroplast with proper thylakoid membranes and correctly assembled photosynthetic complexes (Figure 1A-D).
We hypothesized that the obtained phenotype in the papp5crd double mutant could be linked to a defective PHYA/PHYB signalling pathway.However, the phyBcrd double mutant maintained the pale phenotype of crd compared to wild type (Figure S2A-B).These results suggest that impairment in the phyB mediated light-signalling pathway is not involved in the restoration of papp5crd double mutant phenotype.We also tested the possibility that introducing the papp5 mutation would revert the phenotype of any pale mutant unrelated to the accumulation of Mg-ProtoIX by generating the papp5gun5 double mutant.The gun5 mutant has a pale phenotype but it does not accumulate Mg-ProtoIX.This pale phenotype was maintained also in the papp5gun5 double mutant (Figure S2C).

The papp5crd double mutant demonstrates normal chloroplast development
Depletion of PAPP5 restored the crd mutant phenotype in mature plants.To investigate if PAPP5 plays a role during chloroplast development and seedling establishment we analyzed the chloroplast morphology at different time points during the first 24 hours of seedling development in the light.In the dark grown samples, we observed no differences in either the prolamelar bodies (PLB) or the plastid structures between wild type and any of the mutants (Figure 2A-D).In wild type the first signs of chloroplast development was detected after 4 hours of illumination when the PLBs start to disappear and the first thylakoid membranes were formed (Figure 2E).A similar development of the chloroplasts was observed in the papp5 mutant but in papp5 it was also possible to detect formation of grana membranes already after 4 hours of light exposure (Figure 2F).In contrast, chloroplast development is significantly delayed in crd (Figure 2G).However, in the papp5crd double mutant the chloroplasts develop normally (Figure 2H).After 12 hours of illumination, the chloroplasts in wild type, papp5 and papp5crd are already fully developed and the basic internal thylakoid membranes and grana stacks are formed.The chloroplasts in crd display less evolved thylakoid membranes that lack grana structures (Figure 2I-L).The same observation was made after 24 hours of illumination where the crd seedlings showed minimal thylakoid membrane structures without grana complexes (Figure 2O).However, this feature was restored in the papp5crd double mutant.These studies confirm that the changes that restore the crd phenotype in papp5crd start during the deetiolation process.
Similar changes were observed when we analyzed the chlorophyll content in the different genotypes during the first hours of seedlings development upon illumination.The crd accumulated significantly less chlorophyll compared to wild type following both 12 and 24 hours illumination (Figure 3A-B).In addition, the chlorophyll a/b ratio was elevated in the crd seedlings similarly to what was observed in the mature plants (Figure 3C and 1B).However, papp5crd double mutant demonstrated similar chlorophyll a and b levels and chlorophyll a/b ratio to wild type (Figure 3A-C).Interestingly, the papp5 mutant demonstrated higher levels of chlorophyll following both 12 and 24 hours illumination suggesting that PAPP5 acts as a negative regulator of chlorophyll biosynthesis during the early light response (Figure 3A-B).
Normal induction of PhANG expression in response to light is obtained in the papp5crd double mutant The different physiological features observed in papp5crd compared to the crd single mutant encouraged us to investigate expression of genes encoding components essential for chloroplast development.We analysed expression of the PhANGs, LHCB2.4,GLK1 and GLK2 during the first 12 and 24 hours of illumination.LHCB2.4 is a nuclear gene encoding a chloroplast LHCII antenna apoprotein that is expressed during the light-induced chloroplast development.As expected, LHCB2.4 expression increased following exposure to light compared to the dark samples in wild type (Figure 4A).In papp5, the LHCB2.4 transcript levels increased more strongly following 12 and 24 hours light exposure compared to wild type.Also when expression levels were compared to wild type at each time point PhANG expression was higher in the papp5 mutant (Figure S4).This coincides with the increased accumulation of chlorophyll that was also observed in the papp5 mutant compared to wild type (Figure 3).In contrast, the crd mutant demonstrated impaired induction of LHCB2.4 expression compared to wild type and papp5.However, the papp5 mutation restores the suppressed PhANG expression in crd and the papp5crd double mutant showed significantly stronger induction of LHCB2.4 expression compared to the crd single mutant (Figure 4A).The expression data was supported by the LHCB2.4 protein levels in the different genotypes as demonstrated by Western blot (Figure 4A).
The GLK1 and GLK2 are transcription factors required for chlorophyll biosynthesis and PhANG expression during chloroplast development [25,37].Similarly to the expression of LHCB2.4,expression of GLK1/2 is induced by exposure to light in wild type (Figure 4B-C).Again, a stronger induction of GLK1 and GLK2 expression in response to light was observed in papp5 compared to wild type whereas in crd expression of GLK1 did not change significantly during the first 24 hours of light compared to the dark sample.Also the induction of GLK2 expression was significantly suppressed in crd compared to what was observed in wild type (Figure 4B-C).The suppressed PhANG expression in crd coincides with the observed arrest of chloroplast development in the mutant (Figure 2).Finally, the papp5crd seedlings displayed GLK1 and GLK2 expression similar to wild type (Figure 4B-C).Also when expression levels were compared to crd at each time point PhANG expression was restored in the papp5crd double mutant (Figure S4).In summary, restored induction of LHCB2.4,GLK1 and GLK2 expression could explain the wild type chloroplast morphology in papp5crd during the early light response.

Significant tetrapyrrole accumulation observed also in papp5crd
Due to impaired Mg-ProtoIX-ME cyclase activity in the crd mutant, the crd plants accumulate large pools of tetrapyrrole intermediates and the accumulation has been shown to correlate with a negative effect on the PhANG expression [28].The accumulation of Mg-ProtoIX and its methyl ester, Mg-ProtoIX-ME was determined in the different genotypes during the first 24 hours of illumination.In wild type, Mg-ProtoIX/Mg-ProtoIX-ME accumulated slightly in the light samples compared to the dark (Figure 5A).In papp5, the accumulation of tetrapyrroles was slightly higher compared to wild type which also fits with the faster accumulation of chlorophyll in response to light in papp5 (Figure 3).However, in crd and papp5crd a massive accumulation Mg-ProtoIX/Mg-ProtoIX-ME was observed (Figure 5A).Although the accumulation of tetrapyrroles was lower in papp5crd compared to crd, the papp5 mutation does not rescue the cyclase activity or restores the amount of CRD protein (Figure 5A).
Tetrapyrroles are photoreactive molecules and in response to light, ROS could be generated [38].Induction ASCORBATE PEROXIDASE 2 (APX2) expression is one of the most commonly used markers for the H 2 O 2 response and expression of APX2 was therefore investigated in the different genotypes following exposure to light for 12 and 24 hours.In wild type and papp5, APX2 expression was slightly induced following 12 hours light exposure but following 24 hours illumination the expression levels dropped to control levels (Figure 5B).After 12 hours light exposure the chloroplasts are still immature (Figure 2) and most likely not able to completely quench the light.However, after 24 hours in the light the chloroplasts appear mature and the different photosynthetic complexes are able to efficiently absorb the light and as a consequence less ROS is produced (Figure 5B).The crd seedlings displayed much stronger induction of APX2 expression compared to wild type and papp5.Furthermore, the APX2 expression level did not drop following 24 hours illumination in the crd mutant.This coincides with the undeveloped chloroplasts observed in the crd mutant at this time point (Figure 2).In contrast to the crd single mutant, the papp5crd double mutant did not show elevated APX2 expression (Figure 5B).This is despite the fact that the papp5crd double mutant maintains very high pools of accumulated tetrapyrroles (Figure 5A).Thus, the ROS mediated signal triggering APX2 induction is not correlated with the accumulated pools of tetrapyrroles but rather to the developmental stage of the chloroplasts.Relative expression levels of A) LHCB2.4 (At3g27690) complemented with Western blot analysis of LHCBII protein and b-Tubulin as protein loading control, B) GLK1 (At2g20570) and C) GLK2 (At5g44190) in seedlings grown for three days in dark and exposed to 12 h and 24 h of illumination.Expression levels were compared to the respective dark control for each genotype and relative expression was calculated using Ubiquitin-protein ligase (At4g36800) as a reference gene.Data represents the mean (6 SD) from three independent biological replicates.Significant differences relative to Col0 (crd and papp5) and to crd (papp5crd) were calculated according to t-test (a, P#0.001; b, P#0.005; c, P#0.01).The bands were quantified using ImageJ software and the relative band intensities were obtained and related to Col0 12 h samples.doi:10.1371/journal.pone.0060305.g004) and samples were collected following 12 h and 24 h exposure.Each data point represents the mean (6 SD) of four independent biological replicates.Fluorescence data is complemented with Western blot analysis of the CRD protein levels in the different genotypes.b-Tubulin was used as protein loading control.B) Relative expression of APX2 (At3g09640) in Col0, papp5, crd, and papp5crd.Expression levels were compared to the respective dark control for each genotype and relative expression was calculated using Ubiquitin-protein ligase (At4g36800) as a reference gene.Each bar represents the mean (6 SD) of at least three independent biological samples.Significant differences relative to Col0 (crd and papp5) and to crd (papp5crd) were calculated according to t-test (a, P,0.001; b, c, P,0.01) doi:10.1371/journal.pone.0060305.g005 PAPP5 is a component of a cytosolic protein complex that binds Mg-ProtoIX In contrast to the crd single mutant, papp5crd demonstrated wild type expression of PhANGs and normal development of chloroplast in response to light (Figure 2-4) despite significant accumulation of the chlorophyll intermediates Mg-ProtoIX/Mg-ProtoIX-ME.Thus, the results suggest that PAPP5 is involved in the perception and/or the mediation of the tetrapyrrole triggered retrograde signal.To determine if there is a direct interaction between PAPP5 and Mg-ProtoIX, recombinant full-length PAPP5 and TPR-PAPP5, lacking the TPR domain, was expressed and Mg-ProtoIX bound by the proteins was quantified by spectrofluorometry (Figure S5A).Using this method, no direct interaction between Mg-ProtoIX and PAPP5 could be detected.However, we also tested in vivo if PAPP5 is a component of a distinct protein complex that is able to bind Mg-ProtoIX.We transiently expressed PAPP5 fused to a cMyc-tag in tobacco plants, immunoprecipitated PAPP5 containing protein complexes and incubated the isolated complexes with Mg-ProtoIX as it was done for the recombinant proteins (Figure S5A).The experimental procedure used in this assay requires elution of the protein complex with acidic buffer in order to break the Immunoglobulin-to-antigen binding.Under these acidic conditions, the Mg-ProtoIX molecule is unstable and the Mg 2+ ion is released from the tetrapyrrole ring (Figure S5B) and as a consequence we detected ProtoIX instead of Mg-ProtoIX.It is clear from this in vivo approach that protein(s) immunoprecipitated with PAPP5 can bind Mg-ProtoIX (Figure 6).Possibly another protein partner(s) is required for PAPP5 to be able to bind the tetrapyrrole or PAPP5 is associated with another tetrapyrrole binding protein in a complex.

PAPP5 responds to the tetrapyrrole mediated plastid signal and acts as a negative regulator of PhANG expression
In order to test the link between PAPP5 and the tetrapyrrolemediated retrograde signalling pathway we analysed PhANG expression in wild type and papp5 following feeding with Mg-ProtoIX or 5-aminolevulinic acid (ALA).ALA is the first metabolite in the tetrapyrrole biosynthetic pathway that leads to chlorophyll.Mg-ProtoIX feeding results in increased tetrapyrrole levels in the cytosol whereas ALA treatment increases tetrapyrrole content in the chloroplast which better represents a biological situation when flux through the tetrapyrrole pathway is altered.Mg-ProtoIX (Figure 7A) and ALA (Figure 7B) feeding resulted in a significant accumulation of Mg-ProtoIX both in wild type and papp5 compared to the mock samples.Following the Mg-ProtoIX and ALA feeding expression of LHCB2.4,GLK1 and GLK2 was repressed in wild type (Figure 7C-D).This is in agreement with several published reports on the negative effects of tetrapyrroles on PhANG expression [15,16,17,18,28].However, in the papp5 mutant, LHCB2.4,GLK1 and GLK2 expression was insensitive to Mg-ProtoIX feeding and no repression in expression levels could be detected (Figure 7C).Following ALA feeding repression of GLK1 and GLK2 was observed in the papp5 mutant but the repression was significantly less compared to wild type (Figure 7D).Thus, these results suggest that PAPP5 is required to respond correctly to accumulation of tetrapyrroles in the cell.

Phosphatase activity is required to mediate the tetrapyrrole related plastid signal to the nucleus
To address whether PAPP5 phosphatase activity is required to transmit the tetrapyrrole-mediated plastid signal to the nucleus, we used okadaic acid to inhibit the phosphatase activity during the first 12 hours of illumination.Okadaic acid is a cytotoxin derived from algae that specifically blocks the PP5-2A type phosphatases [39] and it has been shown to reduce PAPP5 activity in vitro [40].LHCB2.4,GLK1 and GLK2 expression was significantly induced following okadaic acid treatment in crd compared to the untreated control (Figure 8).Thus, blocking phosphatase activity phenocopied the papp5crd phenotype in the crd single mutant and reverts the suppression of PhANG expression.These data demonstrates that phosphatase activity is important to mediate the retrograde signal to regulate PhANG expression in crd.

Discussion
The regulation of PhANG expression is very complex and involves signals from multiple signalling pathways, such as those triggered by light, circadian clock and signals originating in the plastids [41].Although light and plastid signals trigger distinct signalling pathways [7], it has been shown that plastid signals and light signals can regulate PhANG expression using common or adjacent promoter elements [16,42].Plastid signals have been suggested to play an essential role in the regulation of PhANG expression when proplastids develop into chloroplasts [7,41,43].One of the plastid signals described to regulate the expression of PhANGs is linked to the tetrapyrrole biosynthesis and we have demonstrated that PAPP5 phosphatase activity is required to transmit the tetrapyrrole-mediated plastid signal to the nucleus during chloroplast development.Our results further suggest that PhANG expression is controlled by a balance between inductive light signalling pathways and a repressive plastid signal triggered by impaired flux through the chlorophyll biosynthesis (Figure 9).
In the double mutant papp5crd, the crd phenotype is restored and in contrast to the crd single mutant, papp5crd accumulated wild type levels of chlorophyll, developed proper chloroplasts and showed normal induction of PhANG expression in response to light (Figures 1-4).The recovery of the pale phenotype is observed in papp5crd even though papp5crd showed a massive accumulation Mg-ProtoIX/Mg-ProtoIX-ME in the light, similar to what was shown for crd (Figure 5A).Furthermore, the papp5 mutation does not rescue the cyclase activity in the crd background or restores the amount of CRD protein (Figure 5A).Thus, PAPP5 is most likely a component in the plastid signalling pathway down stream of the tetrapyrrole Mg-ProtoIX/Mg-ProtoIX-ME.In support of this, the papp5 single mutant demonstrated higher levels of chlorophyll following both 12 and 24 hours light exposure compared to wild type suggesting that PAPP5 acts as a negative regulator of chlorophyll biosynthesis during the early light response.In addition, PhANG transcript levels following 12 and 24 hours light exposure were significantly higher in papp5 compared to wild type (Figure S4).Thus, our result suggests that PAPP5 receives an inbalance in the tetrapyrrole biosynthesis through the accumulation of Mg-ProtoIX and acts as a repressor during chloroplast biogenesis and development.
The deduced amino acid sequence of PAPP5 has two distinctive domains, the N-terminal domain containing the three tetratricopeptide repeats (TPRs) responsible for the protein-protein interaction and the C-terminal domain containing the highly conserved signature motifs of a type 2A serine/threonine protein phosphatase (PP2Ac).This domain structure is a characteristic feature of members of the type 5 serine/threonine protein phosphatase (PP5) subfamily [44].PAPP5 phosphatase activity was shown to be essential to mediate the retrograde signal and to suppress PhANG expression in the crd mutant (Figure 8).Inhibition of phosphatase activity by okadaic acid phenocopied the papp5crd phenotype in the crd single mutant and reverted the suppression of PhANG expression in crd (Figure 8).Furthermore, a direct link between the tetrapyrrole Mg-ProtoIX and PAPP5 was demonstrated in a co-immunoprecipitation assay performed in tobacco leaves where PAPP5 was shown to interact with protein(s) that binds Mg-ProtoIX (Figure 6).Thus, in vivo PAPP5 appears to be a component of a distinct protein complex that is able to bind Mg-ProtoIX.Possibly another protein partner(s) is required for PAPP5 to be able to bind the tetrapyrrole or PAPP5 is associated with Samples were collected 4 h and 1 h into the light period following the Mg-ProtoIX and ALA treatment, respectively.Each bar represents the mean (6 SD) of three independent biological replicates.Relative expression of LHCB2.4 (At3g27690), GLK1 (At2g20570) and GLK2 (At5g44190) in Col0 and papp5 plants following C) Mg-ProtoIX or D) ALA feeding is shown.Expression levels were compared to the respective mock control for each genotype and relative expression was calculated using Ubiquitin-like protein (At4g36800) as internal standard.Each bar represents the mean (6 SD) of at least three independent biological replicates.Significant differences relative to untreated control or Col0 were calculated according to t-test (a, P,0.001; b, P,0.005; d, P,0.05).doi:10.1371/journal.pone.0060305.g007another tetrapyrrole binding protein in a complex.Besides the interaction with PHYA/PHYB [24], PAPP5 has been demonstrated to interact in vivo with HSP90, PP2A and ASK1 [40,45,46,47].HSP90 was recently shown to be required for the Mg-ProtoIX/GUN5 mediated plastid signal and the interaction between HSP90 and Mg-ProtoIX was shown to inhibit ATP-ase activity of HSP90 [17].Possibly, HSP90 and PAPP5 are components in the same complex responding to and mediating the tetrapyrrole signal from the plastids.
Feeding experiments with Mg-ProtoIX or 5-aminolevulinic acid (ALA) further demonstrated a connection between PAPP5 and the tetrapyrroles.Feeding resulted in a significant accumulation of Mg-ProtoIX both in wild type and papp5 compared to the mock samples and expression of PhANGs was repressed in wild type (Figure 7C-D).In contrast, LHCB2.4,GLK1 and GLK2 expression was insensitive to Mg-ProtoIX feeding and no repression of the PhANGs could be detected in the papp5 mutant.Following ALA feeding the repression of GLK1 and GLK2 was observed but was significantly less compared to wild type (Figure 7C-D).To act as signalling molecule and to affect the activity of PAPP5 and later the expression of the PhANGs, the chlorophyll intermediate must reach the cytosol.Numerous porphyrins synthesized in the chloroplast, e.g.chlorophyll catabolites, heme and heme precursors have been found to exit the chloroplast [36,48,49,50].Mg-ProtoIX/Mg-ProtoIX-ME has also been shown to accumulate in the cytosol during stress conditions and it was proposed that the tetrapyrroles are transported across the membrane acting as putative signalling metabolites [15,36].However, the route(s) for transport of any tetrapyrroles from the chloroplast or the components involved in the transport are still unknown.It has also been suggested that ROS accumulation is in fact the origin of the tetrapyrrole mediated plastid signal instead of the specific accumulation of Mg-ProtoIX/Mg-ProtoIX-ME [22,23].The different ROS species activate distinct signalling pathways and the release of ROS could be an alternative explanation for the role of tetrapyrrole intermediates in retrograde signalling because many porphyrins are photoreactive and generate ROS in the presence of light [38].Under our experimental conditions, papp5crd accumulated high levels of Mg-ProtoIX/Mg-ProtoIX-ME but in contrast to the crd single mutant where APX2 expression was strongly induced, no induction of APX2 expression was detected in papp5crd (Figure 5).Thus, despite the tetrapyrrole accumulation, PhANG expression was restored and APX2 expression was similar in papp5crd to what was observed in wild type.Possibly PAPP5 could be required to fully induce APX2 expression in response to ROS or more likely, the ROS levels are correlated with the degree of chloroplast development in crd and papp5crd (Figure 2).APX2 expression was induced following 12 hours light exposure in wild type and the papp5 mutant (Figure 5).However, the expression levels were reduced after 24 hours light exposure that correlates with the time point when functional chloroplasts have developed in wild type and papp5 (Fig. 2 and 5).APX2 expression level did not drop following 24 hours illumination in the crd mutant which coincides with the undeveloped chloroplasts observed in the crd mutant at this time point (Figure 2).Thus, the ROS triggered APX2 induction does not appear to be correlated with the accumulated pools of tetrapyrroles but rather to the functional stage of the chloroplasts.Furthermore, specific ROS eliminators were shown to only partly reverse the tetrapyrroletriggered repression of LHCB [15] and expression of marker genes for ROS were not different in gun5 mutant, impaired in the tetrapyrrole-mediated pathway, compared to wild type [51].Taken together, the tetrapyrrole mediated signal is most likely not related to an altered accumulation of ROS.
Our results suggest that expression of GLK1 and GLK2 is repressed by the tetrapyrrole-PAPP5 mediated plastid signal during the early light response (Figure 4, 9).Higher expression levels of GLK1 and GLK2 were observed in papp5 compared to wild type whereas in crd the induction of GLK1/2 following light exposure was significantly suppressed (Figure 4B-C).The papp5crd seedlings on the other hand displayed GLK1 and GLK2 expression similar to wild type (Figure 4B-C).GLK1 and GLK2 are transcription factors shown to be required for chlorophyll biosynthesis and photosynthesis related gene expression during chloroplast development [25,37].Our results suggest that tetrapyrrole accumulation generates a PAPP5-mediated plastid signal involving a cytosolic protein that suppresses GLK1/2 expression.The two GLK genes have been shown previously to respond to plastid retrograde signals and expression of GLK1/2 was shown to be sensitive to the feedback signalling from the chloroplast suggesting that they may operate downstream of plastid retrograde signalling [25].Furthermore, the crd and glk1glk2 mutants share many common features, similarly to crd, glk1glk2 is pale and contain thylakoid membranes without proper grana.In addition, the glk1glk2 mutant exhibits reduced transcript and protein levels for nuclear-encoded photosynthetic genes, especially those associated with chlorophyll biosynthesis and light harvesting [28,52].This suggests that the phenotype observed in the crd mutant could partly be explained by the mis-regulation of the GLK transcription factors.The recovery of the crd phenotype in papp5crd would then be established through higher levels of chlorophyll binding proteins in the chloroplast and improved photosynthetic performance.In summary, our results demonstrate that the master regulators of PhANG expression and chloroplast development, the GLKs, are regulated by signals originating in the plastids communicating an imbalance in the biosynthesis of the photosynthetic pigments.This signal is transmitted by an unknown factor probably activated upon PAPP5-mediated dephosphorylation.This factor acts as a negative regulator of GLK expression and consequently also of chloroplast biogenesis (Figure 9).Thus, chloroplast development is controlled by a delicate interplay between light and plastid signalling pathways.Figure S2 Characterization of phyB, gun5, phyBcrd and papp5gun5 plants.6-week-old plants of wild type, crd, phyB and phyBcrd grown on soil under short day conditions (9 hours light/ 15 hours dark).A) Representative images from 6-week-old plants.Scale bar = 1 cm.B) Hypocotyl length in seedlings grown in constant red light (630 nm at 20 mmol cm 22 s 21 ) for 5 days.The data is presented as mean (6 SD) where n = 100 seedlings.C) Representative images from 6-week-old plants of wild type, gun5 and papp5gun5 grown in the same condition as described above.Chlorophyll data represents the mean (6 SD) of three independent biological replicates.Significant differences relative to Col0 were calculated according to t-test (d, P#0.05).(PDF)

Supporting Information
Figure S3 Flowering time in crd and papp5crd plants.
Flowering time was determined in Col0, papp5, crd and papp5crd Arabidopsis thaliana plants grown in SD by counting A) the number of leaves when floral buds were visible at the centre of the rosette and B) the number of days from sowing to the day when floral buds appear.The results are presented as mean (6 SD) where n = 12-15 plants.Significant differences relative to Col0 (crd) and to crd (papp5crd) were calculated according to t-test (a, P#0.001).(PDF) Figure S4 PhANG and APX2 expression during chloroplast development.Relative expression levels of A) LHCB2.4 (At3g27690), B) GLK1 (At2g20570), C) GLK2 (At5g44190) and D) APX2 (At3g09640) in seedlings grown for three days in dark and exposed to 12 h and 24 h of illumination.Expression levels were compared to the Col0 level at each time point and relative expression was calculated using Ubiquitin-protein ligase (At4g36800) as a reference gene.E) Expression levels of LHCB2.4,GLK1 and GLK2 in papp5crd compared to crd.Data represents the mean (6 SD) from three independent biological replicates.(PDF) Figure S5 Test for direct interaction between Mg-ProtoIX and PAPP5 in vitro.A) LacZ was used as a control protein, PAPP5 and TPR-PAPP5 (PAPP5 lacking the TPR domain) proteins were expressed and purified.Target-6xHis proteins were mixed with the indicated concentrations of Mg-ProtoIX and then isolated with Ni-agarose beads.After elution from the beads with imidazole, protein-bound Mg-ProtoIX was quantified using spectrofluorometry.Mg-ProtoIX fluorescence intensity was then normalized to the eluted proteins quantified by immunoblot analysis with antibodies against 6xHis (upper panel).Data is expressed as a mean (6 SD) from three independent samples.B) Normalized spectra corresponding to Mg-ProtoIX dissolved in basic solution and in the acidic solution used in for the elution in the in vivo experiment (presented in Figure 6).(PDF) Table S1 Primers sequences used for the experiments presented.(PDF)

Figure 1 .
Figure 1.The papp5 mutation rescues the crd phenotype.Plants were grown on soil under short day (SD) conditions (9 hours light/15 hours darkness) to characterize wild type, crd, papp5 and papp5crd.A) Representative images from 6-week-old plants.Scale bars represent 1 cm.B) Chlorophyll a and b content and chlorophyll a/b ratio in 6-week-old plants.Significant differences relative to Col0 (crd) and to crd (papp5crd) according to t-test (a, P,0.001; b, P,0.005; c, P,0,01) are shown.C) Representative electron microscopy images of chloroplasts from Col0, crd, papp5 and papp5crd from 6-week-old plants grown under SD conditions.Scale bar is 1 mm.D) Blue native PAGE profile of thylakoid membrane protein complexes isolated from Col0, papp5, crd and papp5crd plants.Each well was loaded with the 35 mg protein.doi:10.1371/journal.pone.0060305.g001

Figure 2 .
Figure 2. Chloroplast development during de-etiolation.Sequential electron microscopy images from Col0 (A, E, I and M), papp5 (B, F, J and N), crd (C, G, K and O) and papp5crd (D, H, L and P) during the first 24 hours of illumination (100 mmol photons m 22 sec 21 ).Samples were collected 4 h, 8 h, 12 h and 24 h following transfer to light and compared to the dark sample.Arrows indicate examples of grana structures.Scale bar = 1 mm.doi:10.1371/journal.pone.0060305.g002

Figure 4 .
Figure 4. PhANG expression during chloroplast development.Relative expression levels of A) LHCB2.4 (At3g27690) complemented with Western blot analysis of LHCBII protein and b-Tubulin as protein loading control, B) GLK1 (At2g20570) and C) GLK2 (At5g44190) in seedlings grown for three days in dark and exposed to 12 h and 24 h of illumination.Expression levels were compared to the respective dark control for each genotype and relative expression was calculated using Ubiquitin-protein ligase (At4g36800) as a reference gene.Data represents the mean (6 SD) from three independent biological replicates.Significant differences relative to Col0 (crd and papp5) and to crd (papp5crd) were calculated according to t-test (a, P#0.001; b, P#0.005; c, P#0.01).The bands were quantified using ImageJ software and the relative band intensities were obtained and related to Col0 12 h samples.doi:10.1371/journal.pone.0060305.g004

Figure 5 .
Figure 5. Tetrapyrrole accumulation maintained in papp5crd.A) Relative fluorescence corresponding to Mg-ProtoIX and Mg-ProtoIX-ME in seedlings from Col0, papp5, crd, and papp5crd during the dark to light transition.3-d-old seedlings were transferred to constant light (100 mmol photons light cm 22 sec 21 ) and samples were collected following 12 h and 24 h exposure.Each data point represents the mean (6 SD) of four independent biological replicates.Fluorescence data is complemented with Western blot analysis of the CRD protein levels in the different genotypes.b-Tubulin was used as protein loading control.B) Relative expression of APX2 (At3g09640) in Col0, papp5, crd, andpapp5crd.Expression levels were compared to the respective dark control for each genotype and relative expression was calculated using Ubiquitin-protein ligase (At4g36800) as a reference gene.Each bar represents the mean (6 SD) of at least three independent biological samples.Significant differences relative to Col0 (crd and papp5) and to crd (papp5crd) were calculated according to t-test (a, P,0.001; b, c, P,0.01) doi:10.1371/journal.pone.0060305.g005

Figure 6 .
Figure 6.In planta tetrapyrrole binding assay.Tetrapyrrole fluorescence spectra from CoIP assay using pGWB16-empty (control) and 35-PAPP5-cMyc (pGWB16) transformed tobacco plants.The native PAPP5 protein complex was immunoprecipitated from PAPP5 overexpressing tobacco plants with anti c-myc antibody and incubated with Mg-ProtoIX solution.The amount of tetrapyrrole bound to the protein complex was estimated by fluorescence.The excitation wavelength used was 416 nm.The Western blot of the samples is shown in the upper panel using the anti-myc antibody.doi:10.1371/journal.pone.0060305.g006

Figure 7 .
Figure 7.The papp5 mutant is insensitive to tetrapyrrole feeding.Feeding experiments were performed to increase the levels of tetrapyrroles in the plants.During the dark period, two-week-old plants from Col0 and papp5 grown under LD conditions (15 hours light/9 hours dark) were treated with A) 50 mM Mg-ProtoIX or B) 10 mM ALA.Relative fluorescence corresponding to Mg-ProtoIX and Mg-ProtoIX-ME is shown.Samples were collected 4 h and 1 h into the light period following the Mg-ProtoIX and ALA treatment, respectively.Each bar represents the mean (6 SD) of three independent biological replicates.Relative expression of LHCB2.4 (At3g27690), GLK1 (At2g20570) and GLK2 (At5g44190) in Col0 and papp5 plants following C) Mg-ProtoIX or D) ALA feeding is shown.Expression levels were compared to the respective mock control for each genotype and relative expression was calculated using Ubiquitin-like protein (At4g36800) as internal standard.Each bar represents the mean (6 SD) of at least three independent biological replicates.Significant differences relative to untreated control or Col0 were calculated according to t-test (a, P,0.001; b, P,0.005; d, P,0.05).doi:10.1371/journal.pone.0060305.g007

Figure
Figure S1 Characterization of papp5-1 mutant allele.A) Position of T-DNA insertion in the papp5-1 allele in the At2g42810 gene and B) quantitative RT-PCR of PAPP5 transcripts in homozygous papp5-1 seedlings compared to wild type.(PDF)