In vivo evidence for a regulatory role of phosphorylation of Arabidopsis Rubisco activase at the Thr78 site

Significance

Rubisco activase (Rca) regulates the activation state of Rubisco, the carboxylating enzyme of photosynthesis. Regulation of Rca by redox status of cysteine residues in species such as Arabidopsis is well recognized, but the role of recently identified phosphorylation of Rca at threonine-78 was uncertain. We now show a regulatory role of Arabidopsis Rca phosphorylation. Surprisingly, we also observed that the conservative substitution of serine for threonine-78 results in impaired functionality of Rca in vivo that is associated with retention of phosphorylation well into the light period and with reduced plant growth. This likely reflects in part the specificity of the requisite protein kinase(s) for serine versus threonine and may explain the absence of serine at this position in terrestrial plants.

Abstract

Arabidopsis Rubisco activase (Rca) is phosphorylated at threonine-78 (Thr78) in low light and in the dark, suggesting a potential regulatory role in photosynthesis, but this has not been directly tested. To do so, we transformed an rca-knockdown mutant largely lacking redox regulation with wild-type Rca-β or Rca-β with Thr78-to-Ala (T78A) or Thr78-to-Ser (T78S) site–directed mutations. Interestingly, the T78S mutant was hyperphosphorylated at the Ser78 site relative to Thr78 of the Rca-β wild-type control, as evidenced by immunoblotting with custom antibodies and quantitative mass spectrometry. Moreover, plants expressing the T78S mutation had reduced photosynthesis and quantum efficiency of photosystem II (ϕ_PSII) and reduced growth relative to control plants expressing wild-type Rca-β under all conditions tested. Gene expression was also altered in a manner consistent with reduced growth. In contrast, plants expressing Rca-β with the phospho-null T78A mutation had faster photosynthetic induction kinetics and increased ϕ_PSII relative to Rca-β controls. While expression of the wild-type Rca-β or the T78A mutant fully rescued the slow-growth phenotype of the rca-knockdown mutant grown in a square-wave light regime, the T78A mutants grew faster than the Rca-β control plants at low light (30 µmol photons m⁻² s⁻¹) and in a fluctuating low-light/high-light environment. Collectively, these results suggest that phosphorylation of Thr78 (or Ser78 in the T78S mutant) plays a negative regulatory role in vivo and provides an explanation for the absence of Ser at position 78 in terrestrial plant species.

Rubisco is the CO₂-fixing enzyme of the reductive pentose phosphate pathway and can be one of the major limitations to the rate of leaf photosynthesis (1), which is an important component of crop productivity (2, 3). The activity of Rubisco is dependent on its dedicated AAA⁺ helper protein, Rubisco activase (Rca), which hydrolyzes ATP to induce a conformational change at Rubisco active sites to allow release of a variety of tightly bound inhibitory sugar phosphates prior to rapid activation of Rubisco by carbamylation (4–6). In Arabidopsis (Arabidopsis thaliana), Rca is encoded by 1 gene that is alternatively spliced to generate 2 protein isoforms: the full-length α-isoform and the shorter β-isoform (7). The C-terminal extension of the α-isoform contains 2 redox-active cysteine residues that form a disulfide at low light and in the dark that down-regulates Rca activity, and as a result the Rubisco activation state is reduced (8). Phosphorylation of Rca at the Thr78 site also occurs under the same conditions that result in disulfide formation (e.g., at low light and in darkness) (9, 10) and thus has the potential to contribute to the light/dark regulation of Rca activity and thereby the Rubisco activation state. A previous study (10) concluded that Rca phosphorylation is not essential for its activity, but could not rule out the possibility of a regulatory role that is redundant with that produced by redox status.

Consequently, the overall objective of the present study was to evaluate the role of Rca phosphorylation at the Thr78 site when redox regulation is strongly reduced. To do this, we transformed a strong rca-knockdown mutant with cDNAs encoding wild-type Rca-β, which lacks the redox-active cysteine residues, or Rca-β with site-directed mutations (T78A and T78S) of the phosphosite. As expected, the T78A site–directed mutation prevented phosphorylation but, unexpectedly, the T78S–directed mutant was hyperphosphorylated relative to the wild-type Rca-β. Thus, we could compare plants without phosphorylation as well as plants with greater phosphorylation relative to wild-type Rca-β, and we examined the impact on photosynthetic parameters and plant growth under different light conditions. Collectively, the results obtained provide genetic evidence to establish a negative regulatory role for Rca phosphorylation at the Thr78 site.

Results

Transgenic Expression of Rca-β Rescued the Slow-Growth Phenotype of an rca-Knockdown Mutant.

We obtained a Salk line (SALK_003204c) with a T-DNA insertion in the promoter of the Rca gene (At2g39730) that produces a strong knockdown of both the α- and β-isoforms of Rca, which severely reduces plant growth (Fig. 1A). However, the plants survive and grow slowly at ambient [CO₂], unlike the full rca⁻ knockout that requires high [CO₂] (11). We transformed the Salk rca-knockdown lines with a cDNA encoding Rca-β that restored normal growth in T1 plants (Fig. 1B). Likewise, transformation of the rca-knockdown line with the phospho-null T78A-directed mutant also restored growth and, as expected, only the residual wild-type Rca-β protein in the knockdown mutant was phosphorylated at the Thr78 site when plants were darkened. Dark-harvested Col-0 and rca-knockdown plants were used as controls (Fig. 1C). Rca-β and T78A homozygous plants were established. These transgenic plants provided a convenient experimental system to determine whether phosphorylation at the Thr78 site has a regulatory role in vivo when redox regulation of Rca activity is strongly reduced. Several independent lines of each transformation event were selected (See SI Appendix, Fig. S1; 2 of each are shown) and used for further characterization in the present study.

Fig. 1.

Wild-type Rca-β and the T78A site–directed mutant rescue the slow-growth phenotype of the strong rca-knockdown mutant (rca_SALK). (A, Top) Gene map showing the position of the T-DNA insertion in the promoter of the Rca gene (At2g39730). (Middle) Comparison of the rca⁻ mutant and Col-0 plants grown in a 16-h photoperiod for 24 d. (Bottom) Immunoblot showing the large reduction in Rca isoforms in the knockdown mutant and the Coomassie Brilliant Blue (CBB)–stained blot showing the Rubisco large subunit as a loading control. (B, Top) T1 plants of Rca-β or T78A rescue the rca-knockdown dwarf phenotype. Plants were photographed 21 d after germination and were grown in a long-day photoperiod. (Bottom) Immunoblot showing increased levels of Rca-β in T1 transgenic plants. (C) Transgenic homozygous plants expressing Rca-β or the T78A site–directed mutation were harvested in the dark; phosphorylation of Rca-β at the Thr78 site, but not the phospho-null T78A mutant, was confirmed using sequence- and phosphorylation-specific (anti-pT78) antibodies. Dark-harvested Col-0 and rca_SALK were used as controls. The asterisk identifies an off-target protein that is recognized by anti-pT78 antibodies (10).

Prevention of Phosphorylation of Rca-β Enhanced Photosynthetic Induction and Growth at Low Light.

Rca is one of the major determinants of the rate of photosynthetic induction during low- to high-light transition (12, 13). Accordingly, we monitored induction kinetics in the transgenic plants with and without the Thr78 phosphosite. As shown in Fig. 2A, plants expressing the phospho-null Rca (T78A) mutation had higher rates of net CO₂ assimilation following transition from low to high light compared with plants expressing the wild-type Rca-β. Photosystem II operating efficiency (ϕ_PSII) was correspondingly higher in the T78A plants following the transition, which is consistent with higher rates of CO₂ assimilation. In fully light-adapted plants, the maximum rate of carboxylation, V_cmax, and the maximum rate of electron transport, J_max, were not significantly different in the transgenic lines (Fig. 2A, Inset). This suggests that all lines reached similar final levels of Rubisco activation and is consistent with the Rca function limiting Rubisco activation to a greater extent during inductive, compared with steady-state, photosynthesis. Phosphorylation of Rca at the Thr78 site occurs in the dark and at low light (10), and thus the results in Fig. 2A suggest that the phosphorylation of Rca that occurred in the transgenic plants expressing wild-type Rca-β was sufficient to negatively impact photosynthetic induction profiles. In steady-state light, Rca is required to sustain Rubisco activation, but wild-type Rca activity is not limiting, which is consistent with dephosphorylation of the protein after 15–30 min of exposure to high light.

Fig. 2.

Plants expressing T78A have faster low- to high-light induction kinetics of photosynthesis and increased growth at low light compared with plants expressing wild-type Rca-β. (A) CO₂ assimilation (Left) and PSII quantum efficiency (Right) during photosynthetic induction (n ≥ 11; error bars represent ±SE). Measurements were performed at 22 °C on leaves from mature rosettes prior to bolting. Leaves were dark-adapted for at least 30 min prior to the start of each experiment, which consisted of 20 min at low light followed by 30 min at high light. Horizontal bars above the chart indicate experimental light levels. Plotted points are the average of at least 11 biological replicates, ± SE. (Inset) Parameters estimated from CO₂ response curves (n = 3, ±1 SE) conducted during steady-state photosynthesis. (B) Growth of Rca-β and T78A plants as a function of light intensity. The difference between Rca-β and T78A plants at PAR 30 was statistically significant based on a 1-tailed, unpaired t test (P value < 0.001). All plants were 4 wk old at harvest for shoot fresh-weight determination. (C) Leaf area accumulation of plants growing at 30 µmol photons m⁻² s⁻¹. Ten plants were used to generate values of average and SE for B and C.

Because phosphorylation of Rca was retained at low light, we speculated that the transgenic plants expressing wild-type Rca-β or the phospho-null T78A mutation would differ in growth at low photosynthetic photon flux density (PPFD). As expected, plant growth (monitored as shoot fresh weight) decreased progressively as light decreased from 150 to 30 µmol photons m⁻² s⁻¹ in both genotypes, but a different shoot mass between genotypes was observed only for plants grown at the lowest light tested, with lower mass in the wild-type Rca-β plants compared with the phospho-null T78A plants (significance: 1-tailed, unpaired t test, P value < 0.001; Fig. 2B). The difference in growth between the 2 genotypes at 30 µmol photons m⁻² s⁻¹ was also documented by continuous monitoring of the increase in total rosette leaf area as a function of time (Fig. 2C).

Substitution of Ser for Thr78 in Rca-β Resulted in Hyperphosphorylation in the Light and Reduced Photosynthesis.

While the functional significance of a phosphosite is usually tested by generating a phospho-null (e.g., the T78A mutant) or by substituting an acidic residue that sometimes acts as a phosphomimetic, another approach is to substitute a serine for a threonine (or vice versa). The Thr-to-Ser approach is a more subtle modification that has the potential to impact phosphosite function because the 2 residues are similar but often not equivalent at the molecular level (14–16). Accordingly, we generated transgenic plants expressing the T78S site–directed mutant in the rca-knockdown mutant background (SI Appendix, Fig. S1). Fortunately, the custom phospho-specific antibodies that recognize phospho-Thr78 also recognized the cpCK2-mediated phosphorylated form of the T78S site–directed mutant (SI Appendix, Fig. S2A). Indeed, recombinant cpCK2 phosphorylated a synthetic peptide containing the T78S substitution more rapidly than the corresponding T78 peptide (SI Appendix, Fig. S2 B and C).

It is interesting that, while phospho-Thr78 in transgenic plants expressing wild-type Rca-β was completely dephosphorylated 1 h after the transfer of plants from dark to light (125 µmol photons m⁻² s⁻¹ in these experiments), phosphorylation at the Ser78 site in the T78S-directed mutant was removed much more slowly and persisted for several hours into the light period (Fig. 3A). In the dark samples, phosphorylation of Thr78 in Rca-α, which is present at low levels in the rca-knockdown mutant, was also readily apparent in the immunoblots, and that signal was completely removed within 1 h of illumination (Fig. 3A), as expected (10). Thus, phospho-Thr78 was rapidly dephosphorylated in marked contrast to phospho-Ser78 in the transgenic plants expressing the T78S mutant protein.

Fig. 3.

Plants expressing T78S retain phosphorylation into the light period and have impaired photosynthesis relative to plants expressing wild-type Rca-β. (A) Two-week-old seedlings were harvested after 8 h of darkness (D8) or after transfer to light for 1–8 h (L1–L8), as indicated, prior to harvest for extraction of protein and immunoblot assay. Anti-pT78 antibodies were used to detect phosphorylation of both Rca-β-T78 (Left) and T78S (Right); the histogram shows densitometry of the immunoblots where the pT78 (or pS78) phospho-signal associated with Rca-β was normalized to the Rca-β protein signal. (B) CO₂ assimilation (Left) and PSII quantum efficiency (Right) during photosynthetic induction (n ≥ 5, ±SE). Plants were grown in Cone-tainers at 140 µmol photons m⁻² s⁻¹ for ∼3 wk. Induction curves were obtained with a whole-rosette chamber (6400-40 Leaf Chamber Fluorometer, LI-COR) and the RGB light source. Fluorescence was measured on a separate day with the Technologica Fluorimager to assess changes in quantum efficiency of ϕ_PSII. Horizontal bars above the chart indicate experimental light levels. (Inset) parameters estimated from CO₂ response curves (n = 6, ±SE) conducted during steady-state photosynthesis.

These results were confirmed and extended by quantitative phosphoproteomic analysis of transgenic plants expressing Rca-β or the T78S mutation, specifically targeting Rca-β peptides (identified in Dataset S1) and, in particular, the phosphosite at residue 78. Leaves of plants were harvested after 8 h in the dark or after subsequent exposure to light (125 µmol photons m⁻² s⁻¹) for 1 h. By quantifying the amount of unphosphorylated tryptic peptide containing residue 78 and the corresponding phosphopeptide, we could estimate site-specific phosphorylation stoichiometries. In plants expressing wild-type Rca-β, phosphorylation stoichiometry was ∼51% in the dark and reduced to near zero levels after 1 h of light (Fig. 4A). The transgenic plants expressing the T78S-directed mutant also contained some residual wild-type Rca protein isoforms (α, β1, and β2), and thus phosphorylation at Thr78 could also be monitored in these plants (as noted previously in relation to immunoblot analysis in Fig. 3). A similar phosphorylation stoichiometry at the Thr78 site in the rca-knockdown background of 50% was observed, but again the phosphosite was completely dephosphorylated after exposure to 1 h of light (Fig. 4A). In the T78S mutant, phosphorylation at the engineered Ser78 site was confirmed with high confidence by mass spectrometric analysis (Fig. 4 C–E); the phosphorylation stoichiometry was high in the dark (∼76%) and reduced in the light but only to ∼52% after 1 h (Fig. 4B).

Fig. 4.

Quantitative phosphopeptide analysis of the phosphosite at position 78 in the wild-type Rca-β and the T78S-directed mutant. (A) The wild-type form of Rca-β was completely dephosphorylated at position T78 in Rca-β–complemented and T78S-complemented knockdown plants. Significance: ***P < 0.001. (B) The engineered T78S was not significantly dephosphorylated at position 78 in the light. Average log₂ intensities of phosphopeptide and unmodified peptide were calculated (n = 3, Rca-β2_dark n = 2). A Student’s t test was applied for light–dark comparison. Phosphosite occupancy on the Rca protein is stated above the corresponding peptide and sample. (C and D) Fragmentation spectra for GLAYD(p)TSDDQQDITP and GLAYD(p)STSDDQQDITP, respectively. The b (in blue) and y (in red) represent ion series of the fragmented peptides (E) Scores of phosphorylated Rca-β2 peptides.

We then monitored photosynthetic induction kinetics following transition of whole rosettes from low to high light. As shown in Fig. 3B, plants expressing T78S had substantially slower photosynthetic rates relative to plants expressing wild-type Rca-β in the rca-knockdown background. As expected, there was a corresponding reduction in ϕ_PSII in plants expressing T78S that was consistent with differences in CO₂ assimilation (Fig. 3 B, Right). AC_i curves performed on fully light-adapted T78S plants yielded V_cmax = 24.6 ± 1.4 µmol m⁻² s⁻¹ and J_max = 61.6 ± 2.4 µmol m⁻² s⁻¹, neither of which was significantly different from the values for the wild-type Rca-β. The AC_i measurements were performed on different days than the low- to high-light induction profiles and were also performed from 3 to 8 h into the photoperiod, at which time phosphorylation of T78S at the engineered S78 phosphosite was likely to have been substantially reduced (Fig. 3B). As was the case for T78A, this was consistent with the notion that Rca functions primarily during inductive photosynthesis but is also essential for maintaining Rubisco activation under steady-state conditions.

One explanation for the reduced growth and photosynthetic performance of transgenic plants expressing the T78S-directed mutant could be that the conservative substitution of Ser for Thr directly affects the ability of the protein variant to activate Rubisco even in the unphosphorylated state. To test this possibility, we prepared wild-type Rca-β and the T78A- and T78S-directed mutants as His₆-tagged recombinant proteins for use in Rubisco activation assays in vitro. While both the T78A- and T78S-directed mutants were less active in vitro compared with wild-type Rca-β, there was not a statistically significant difference between the T78A and T78S proteins (SI Appendix, Fig. S3A). This observation, coupled with the fact that maximum Rca activity is not limiting in vivo during steady-state photosynthesis, suggests that the reduced growth and reduced photosynthetic activity of T78S plants cannot be attributed to the intrinsic activity of the fully unphosphorylated T78S mutant protein. It is also worth noting that, while there is considerable variation among terrestrial plants in the amino acid found at position 78 in Rca, Ser is not known in nature; the most common residue is Thr or a hydrophobic residue (Ile, Met, Val, or Phe; See SI Appendix, Fig. S3B and Dataset S2).

Rca-β Phosphorylation Reduced Plant Growth in a Rapidly Fluctuating Light Environment.

If phosphorylation of Rca reduces the rate at which photosynthesis increases following a low- to high-light transition, then growing plants with wild-type Rca in a fluctuating light regime should negatively impact plant growth. To test this notion, we monitored the increase in rosette leaf area as a function of time for transgenic plants differing in phosphorylation of Rca-β at the Thr78 site and growing in fluctuating-light or a square-wave light regime. In the square-wave regime, transgenic plants expressing wild-type Rca-β and the phospho-null T78A-directed mutant grew similarly and had higher total leaf areas compared with the hyperphosphorylated T78S mutant (Fig. 5A). Inflorescence induction occurred at a similar rosette size in all lines, despite being slightly delayed in T78S. In contrast, when plants were grown in the fluctuating-light regime, the transgenic plants expressing the phospho-null T78A-directed mutant had the highest total leaf area, followed by wild-type Rca-β and the T78S mutant (Fig. 5B). Rosette dry weights at terminal harvest showed similar patterns as leaf area, indicating that the differing growth was not due to differences in specific leaf area (Fig. 5C). Under square-wave light, the dry weights of T78A and wild-type Rca-β were identical, while T78A plants were ∼30% heavier than the wild-type Rca-β plants under fluctuating light. This difference was statistically significant at a 10% confidence interval based on a 1-tailed, unpaired t test (P value = 0.099).

Fig. 5.

Growth curves of wild-type Rca-β, T78A, and T78S plants under square-wave light or fluctuating-light regimes. (A) Accumulation of leaf area in the square-wave light regime (n ≥ 16, ±SE). Light was constant at 125 µmol photons m⁻² s⁻¹ for the entire 16-h light period. Arrows indicate the average plant age at inflorescence initiation. (B) Accumulation of leaf area in the fluctuating light regime consisting of low light (25 µmol m⁻² s⁻¹ PPFD) and high light (125 µmol m⁻² s⁻¹ PPFD; 10 min each) for the entire 16-h photoperiod (n ≥ 11, ±SE). Data in A and B were normalized (Rca-β mean = 1 at 38 d after planting) to account for variation in average growth rate between experiments. (C) Rosette dry weight comparisons 40 d after sowing (n ≥ 4, ±SE). Significant differences (t test, P ≤ 0.1) in fluctuating light are indicated by different uppercase letters.

Altered Gene Expression in the Rca-β T78S-Directed Mutant Was Consistent with Reduced Photosynthetic Activity and Slow Growth.

In order to explore the underlying causes of reduced growth in plants expressing the T78S-directed mutation, we performed microarray analysis comparing transgenic seedlings expressing wild-type Rca-β or the T78S-directed mutation (all in the rca-knockdown mutant background). After the results were sorted to include only samples with expression values greater than 6.5 and significant P values, false discovery rate (FDR) correction was applied, which yielded 752 genes up-regulated in T78S compared with wild-type Rca-β, with 80 of those genes up-regulated more than 2.5-fold. A total of 460 genes were down-regulated, and 43 of those were down-regulated with fold changes less than 0.5. To confirm the changes in gene expression identified by microarray analysis, we quantified the expression level of several up- and down-regulated genes using qRT-PCR (SI Appendix, Fig. S4A). Although we only examined a total of 8 transcripts from the list of differentially expressed genes, there was a positive correlation (R² = 0.71) for the fold changes obtained by the 2 methods and the linear regression line passed through the origin (SI Appendix, Fig. S4B). These results demonstrate the overall reliability of the microarray results.

A selected list of genes that were more than 2-fold up- or down-regulated is presented in Table 1. Interestingly, a number of the up-regulated genes are DARK-INDUCED (DIN) genes (17, 18) and include DIN2 (β-glucosidase; AT3G60140), DIN6 (asparagine synthetase ASN1; AT3G47340), and DIN11 (2-oxoacid–dependent dioxygenase; AT3G49620). Other DIN genes were significantly up-regulated but with fold changes below the cutoff used for Table 1, including DIN1 [AT4G35770; senescence-associated gene 1 (SEN1): 1.87-fold], DIN3 [AT3G06850; dihydrolipoamide branched chain acyltransferase (BCE2): 1.71-fold], DIN4 (AT3G13450; branched chain α-keto acid dehydrogenase E1 β: 1.59-fold), DIN9 [AT67070; phosphomannose isomerase 2 (PMA2): 1.89-fold], and DIN10 [AT5G20250; raffinose synthase 6 (RS6): 1.40-fold]. While DIN gene transcripts accumulate in darkness, their levels in darkened leaves can be repressed by sucrose feeding, suggesting a role for cellular sugar levels. Many of the DIN genes are also induced in senescing leaves (19, 20), as are several other genes that were up-regulated in the T78S plants, including a senescence-associated gene (SAG13) (21), a nitrilase gene (NIT4) (22), and an alternative oxidase 3 gene (AOX1D) (23) (Table 1). It is important to note, however, that the T78S plants did not show signs of senescence and, in fact, did not flower or mature earlier than the control plants expressing wild-type Rca-β (SI Appendix, Fig. S1). Because the DIN genes are considered to be low-energy status markers, their up-regulation in T78S plants is consistent with reduced photosynthetic activity in these plants (Fig. 5A).

Table 1.

Differentially regulated genes in transgenic plants expressing the T78S-directed mutation relative to wild-type Rca-β in the rca⁻ knockdown mutant background

			Fold Change
The Arabidopsis Information Resource	Symbol	Gene Product	(T78S/RCA-β)
Up-regulated genes
Genes induced by low carbon, dark, or senescence
AT3G49620	DIN11	Encodes 2-oxoacid–dependent dioxygenase.	8.38
AT3G60140	DIN2	Beta-glucosidase 30 (SEN2)	7.44
AT2G29350	SAG13	Senescence-associated gene SAG13	3.67
AT5G22300	AtNIT4	Encodes a nitrilase isomer	3.12
AT1G32350	AOX1D	Alternative oxidase 3: up-regulated in senescence	2.95
AT1G35140	EXL1	Required for growth at low carbon availability	2.88
AT3G47340	DIN6	Glutamine-dependent asparagine synthetase	2.66
Down-regulated genes
Hydrolases and expansins
AT2G32990	AtGH9B8	Cellulase; GUN11	−3.22
AT1G64390	AtGH9C2	Glycosyl hydrolase; GUN6	−2.03
AT3G53190	PLL17	Pectate lyase–like	−2.01
AT1G04680	PLL26	Pectate lyase–like	−2.50
AT2G40610	ATEXP8	Alpha-expansin gene family	−1.99
Lignin-related
AT5G62360	NA	Pectin methylesterase inhibitor	−2.32
AT5G62350	NA	Pectin methylesterase inhibitor	−2.07
AT4G21960	PRXR1	Cell wall peroxidase, possibly involved in lignin	−2.04
Development
AT3G25717	DVL6	ROTUNDIFOLIA-like 16	−2.15
Cell wall–binding proteins
AT5G24105	AGP41	Arabinogalactan protein (AGP41)	−2.24
AT5G65390	AGP7	Arabinogalactan protein 7: involved in development	−2.04
AT1G55330	AGP21	Arabinogalactan protein	−1.99
AT4G18280	NA	Glycine-rich cell wall–like protein	−2.16
AT3G16670	NA	Extensin-like protein	−2.14
AT1G12090	ELP	Extensin-like protein (ELP)	−1.98
Hormone-related
AT1G52400	ATBG1	Glycosyl hydrolase family 1 protein generates active form of ABA	−3.58
AT3G22231	PCC1	Circadian clock–regulated gene; flowering and ABA responses	−2.57
AT3G22240	NA	PCC1-like protein	−1.99
AT1G19050	ARR7	Arabidopsis response regulator (ARR) protein, cytokinin signaling, and meristem stem cell maintenance	−2.30
AT5G18010	SAUR19	Positive effector of cell expansion	−2.13
AT5G18030	SAUR21	Positive effector of cell expansion	−2.10
AT5G18060	SAUR23	Positive effector of cell expansion	−2.18
AT5G18050	SAUR22	Positive effector of cell expansion	−2.16

As noted previously, a number of genes were down-regulated in the T78S plants, and many of those genes are associated with growth. For example, cell growth involves several different cell wall–loosening proteins that facilitate movement of cell wall polymers to accommodate water uptake and cell expansion, including glycosyl hydrolases, α-expansins, and various small cell wall–binding proteins (24, 25). Thus, it is noteworthy that 2 glycosyl hydrolases, GUN6 and GUN11, were down-regulated in the T78S plants relative to the Rca-β control plants (Table 1). Likewise, the α-expansin AtEXP8 (26) was down-regulated, as were several arabinogalactan proteins (AGP7, AGP21, and AGP41), which play multiple roles in plant development, including promotion of growth (27). Pectins are major components of the cell wall and often involve complex structures that can be modified or degraded to alter cell wall properties. Accordingly, pectin-related enzymes are important in cell expansion, and 2 pectate lyase–like genes and 2 genes encoding pectin methylesterase inhibitors were down-regulated in the present study. Pectate lyases are associated with depolymerization of pectins and cell separation, whereas pectin lyase–like (PLL) proteins are thought to be involved in numerous aspects of growth potentially independent of cell separation (28); 2 PLL genes were down-regulated in T78S plants. Methyl-esterified pectin is secreted in cell walls, and removal of methyl-esters from pectin by pectin methylesterases (PMEs) can generate negative charges that coordinate calcium ions to form cross-links that affect cell wall extensibility. Because PME inhibitors block this action, down-regulation of PME inhibitors could result in cell walls with reduced capacity for expansion (29). Interestingly, genes encoding 2 PME inhibitors were down-regulated in T78S plants. Additional growth-related genes include the SMALL AUXIN UP RNA (SAUR) genes that are auxin-responsive and often correlated with auxin-mediated cell expansion. In particular, genetic evidence implicates the SAUR19–24 subfamily as positive effectors of cell expansion (30); the down-regulation of this subfamily of SAUR genes would be expected to contribute to reduced growth of the T78S transgenic plants.

Functional interactions within the differentially expressed genes were explored using the STRING database (31). Analyses were carried out using a medium confidence threshold and all lines of evidence except neighborhood and gene fusion. Within the up-regulated genes, there was a large network of interactors containing 56 nodes and 136 edges (SI Appendix, Fig. S5A). AT1G26380 and CYP71A12 were the most central nodes, based on weighted closeness centrality (32). These genes were implicated recently in the production of cyanogens during pathogen response (33). Up-regulation of this network may therefore suggest activation of stress response pathways in T78S. Among the down-regulated genes, no large interaction networks were found (SI Appendix, Fig. S5B), only a small cluster of the aforementioned SAUR genes.

Discussion

The results of the present study identify a regulatory role for phosphorylation of Rca and implicate the Rca Thr78 phosphosite in the evolution of terrestrial plants. We tested the regulatory role of Rca phosphorylation by expressing wild-type or mutated versions of the Rca-β isoform, which lacks the C-terminal redox regulatory domain found in the α-isoform, in an Arabidopsis rca-knockdown mutant. Relative to transgenic plants expressing wild-type Rca-β with the phosphorylatable Thr78 site, results obtained with the phospho-null T78A-directed mutant and the hyperphosphorylated T78S-directed mutant strongly suggest that phosphorylation of Rca plays a negative regulatory role in vivo that inhibits photosynthesis and growth. The phospho-null T78A mutant exhibited more rapid induction kinetics of photosynthesis and higher ϕ_PSII during low- to high-light transitions (Fig. 2A), whereas the hyperphosphorylated T78S mutant exhibited slower photosynthesis induction and lower ϕ_PSII (Fig. 3B) compared with the wild-type control. In addition, changes in photosynthetic parameters noted for the site-directed mutants were associated with altered growth phenotypes compared with the transgenic plants expressing wild-type Rca-β. For example, T78A plants were similar to the control plants when grown in a typical square-wave light regime, but they grew faster than controls in a fluctuating low-light/high-light regime where more of the CO₂ assimilation occurred under inductive (non-steady-state) conditions. In addition, the T78A plants had a small but significant growth advantage at low light (30 µmol photons m⁻² s⁻¹), a condition that retains phosphorylation of wild-type Rca-β (10). Thus, in the absence of the Rca phosphorylation that occurs in wild-type Rca-β, the T78A mutant plants utilized light energy more efficiently and grew slightly faster (Fig. 2B). Conversely, the hyperphosphorylated T78S plants had reduced growth in both the square-wave light and fluctuating-light regimes (Fig. 4) consistent with the retention of Rca phosphorylation observed well into the light period (Fig. 3B). Reduced growth of the T78S-directed mutant would be expected given the reduced CO₂ assimilation rate (per unit leaf area; Fig. 3B) and reduced expression of numerous growth-related genes (Table 1). Collectively, these results support the notion that phosphorylation of Rca residue-78 negatively regulates photosynthesis, with its impact more pronounced with the T78S-directed mutant compared with the wild-type Rca-β.

Our finding that the T78S-directed mutant was hyperphosphorylated into the light period was based on immunoblotting with sequence- and site-specific antibodies (which recognized the phosphorylated form of Thr78 and Ser78; Fig. 3B) and on quantitative mass spectrometry–based phosphoproteomic analysis (Fig. 4B); it has important implications that are worth noting. First, the result was unexpected because substitution of serine for a threonine residue is the most conservative substitution possible; however, the change can have surprising effects because, while the 2 residues are similar, they are often not equivalent. In our system, this result suggested that the protein kinase(s) and/or protein phosphatase(s) acting on Rca discriminated slightly between serine and threonine at the phosphosite of Rca. Indeed, it is generally recognized that while many Ser/Thr kinases prefer serine as the phosphoacceptor, most protein phosphatases prefer phosphothreonine (15, 34, 35). Although the protein phosphatases acting on phospho-Rca have not been identified, cpCK2 has been established as a major protein kinase that phosphorylates Rca at the Thr78 site (10). Plastid-localized cpCK2 is a member of the nearly ubiquitous CK2 kinase family, and previous work with mammalian and yeast CK2 kinases determined that Ser is strongly preferred over Thr as the phosphoacceptor residue (34, 36). Consistent with earlier studies with nonplant CK2 kinases, our results with cpCK2 and synthetic peptides indicated a similar preference for Ser over Thr at position 78 (SI Appendix, Fig. S2B and C). A major determinant of Ser vs. Thr phosphoacceptor specificity appears to reside in a specific residue of the kinase activation segment, termed “DFG + 1” (37). The residue corresponding to DFG + 1 in cpCK2 is a leucine, which is consistent with a preference for serine and may contribute to the increased phosphorylation of the T78S-directed mutant of Rca-β in vivo. However, it is likely that the specificity of protein phosphatase(s) also plays a role in the hyperphosphorylation of the T78S mutant, but these studies await identification of the requisite phosphatases acting on phospho-Rca in vivo. Second, these results also have relevance to our understanding of the variation in residues found in Rca at the position corresponding to residue 78 in the Arabidopsis protein. Overall, the Rca protein is highly conserved in terms of sequence, but there is a surprising amount of variation among terrestrial species at position 78 (SI Appendix, Fig. S3B). However, among the 59 terrestrial species for which the National Center for Biotechnology Information Reference Sequence Database Rca sequences are available, none have a serine at the position corresponding to Thr78 of Arabidopsis Rca (SI Appendix, Fig. S3B). We speculate that the lack of serine may be explained by natural selection against this residue because of its hyperphosphorylation and negative impact on photosynthesis.

While present studies have established a regulatory role for phosphorylation of Thr78 in Rca-β in vivo, a number of important questions remain for future studies. First, it will be important to examine the role of Thr78 phosphorylation of the α-isoform of Rca, which has redox regulation capability, independently as well as in plants expressing both Rca isoforms. Previous studies with a cpck2⁻ knockout mutant concluded that phosphorylation in wild-type plants has no regulatory role (10) but, because of the potential for additional protein kinases to phosphorylate Rca at the Thr78 site, this conclusion needs further examination with directed mutants lacking the phosphosite residue. Second, and as noted earlier, it will be important to identity the protein phosphatase(s) that dephosphorylate phospho-Thr78 and determine whether there is a preference for phospho-Thr over phospho-Ser. If so, then substrate preferences of both kinase(s) and phosphatase(s) may contribute to the hyperphosphorylation of the T78S-directed mutant. Third, and perhaps most important, future efforts need to determine the mechanism by which phosphorylation of Rca affects photosynthesis (Figs. 2 and 3) and plant growth (Fig. 5). The simplest explanation is that phosphorylation of Rca-β directly inhibits Rubisco activation. To test this possibility, we performed in vitro assays of Rubisco activation using phosphorylated or nonphosphorylated Rca proteins, and found no significant inhibition by phosphorylation (SI Appendix, Fig. S6B). The phosphorylation stoichiometries of the recombinant proteins used in the Rubisco activation experiments were roughly 20% for Rca-β and 45% for T78S (SI Appendix, Fig. S6A). The phosphorylation stoichiometries of Rca-β and T78S in vivo in darkened leaves were determined by quantitative phosphoproteomic analysis (Fig. 4) to be roughly 50% for Rca-β and 75% for T78S, which were only slightly higher than the phosphorylation stoichiometries tested in vitro. Rubisco contents were not changed in T78S in light conditions (SI Appendix, Fig. S7). Thus, we tentatively conclude that phosphorylation of Rca does not directly inhibit Rubisco activation activity but rather is more complex. For example, it may not be possible to mimic the phosphorylation impact in vitro if the phospho-Thr78 Rca recruits cofactors (e.g., phosphoprotein-binding proteins) to sequester phospho-Rca or inhibit Rubisco activation in vivo. Fourth, if phosphorylation of Rca is broadly important in vivo, then understanding how phosphorylation stoichiometry is controlled warrants examination; that is, why is Rca slowly phosphorylated in the dark but rapidly dephosphorylated in the light? A role for redox in the light/dark control of Rca phosphorylation was established (10) which was consistent with increased phosphorylation in the dark; however, how the redox signal was mediated was not elaborated. Moreover, once all of the protein kinases and protein phosphatases are identified, it will be possible to determine if the changes in stromal pH and [Mg²⁺] that accompany light/dark changes are additional factors that regulate Rca phosphorylation. Last but not least, it will be important to unravel the full complexity of the posttranslational modifications of Rca. Next to phosphorylation and redox regulation, lysine acetylation was recently discovered on Lys438 of Rca-β2 and was found to be under the control of a plastid lysine deacetylase in low-light conditions (38). Thus, the potential for control of Rca activity by phosphorylation, lysine acetylation, and redox is apparent, and it will be important to explore these interactions further in future studies.

Materials and Methods

Detailed descriptions of all materials and methods are available in SI Appendix. Briefly, the Arabidopsis rca-knockdown mutant (Salk_003204C) containing a T-DNA insertion in the Rca gene promoter was transformed with cDNAs encoding wild-type Rca-β, Rca-β (T78A), or Rca-β (T78S). Plants were grown under both nonfluctuating and fluctuating light. Sodium dodecyl sulfate polyacrylamide gel electrophoresis and immunoblotting were used to confirm Rca protein expression. Peptide phosphorylation and Rubisco activation assays were conducted in vitro. Liquid chromatography–tandem mass spectrometry was used to identify phosphorylated protein fractions. Leaf photosynthesis was measured using portable gas exchange systems equipped with fluorometers. RNA was isolated for microarray expression analyses, which were confirmed using qRT-PCR.