Article
Reference
State transition7-dependent phosphorylation is modulated by changing environmental conditions, and its absence triggers
remodeling of photosynthetic protein complexes
BERGNER, Sonja Verena, et al.
Abstract
In plants and algae, the serine/threonine kinase STN7/STT7, orthologous protein kinases in Chlamydomonas reinhardtii and Arabidopsis (Arabidopsis thaliana), respectively, is an important regulator in acclimation to changing light environments. In this work, we assessed STT7-dependent protein phosphorylation under high light in C. reinhardtii, known to fully induce the expression of light-harvesting complex stress-related protein3 (LHCSR3) and a nonphotochemical quenching mechanism, in relationship to anoxia where the activity of cyclic electron flow is stimulated. Our quantitative proteomics data revealed numerous unique STT7 protein substrates and STT7-dependent protein phosphorylation variations that were reliant on the environmental condition. These results indicate that STT7-dependent phosphorylation is modulated by the environment and point to an intricate chloroplast phosphorylation network responding in a highly sensitive and dynamic manner to environmental cues and alterations in kinase function. Functionally, the absence of the STT7 kinase triggered changes in protein expression and photoinhibition of photosystem [...]
BERGNER, Sonja Verena, et al. State transition7-dependent phosphorylation is modulated by changing environmental conditions, and its absence triggers remodeling of photosynthetic protein complexes. Plant Physiology, 2015, vol. 168, no. 2, p. 615-634
DOI : 10.1104/pp.15.00072 PMID : 25858915
Available at:
http://archive-ouverte.unige.ch/unige:86430
Disclaimer: layout of this document may differ from the published version.
1 / 1
1 Running head: STT7 is tuning acclimation 1
2
Corresponding Author: Prof. Dr. Michael Hippler 3
4
Address:
5
Institut für Biologie und Biotechnologie der Pflanzen 6
Westfälische Wilhelms-Universität Münster 7
Schlossplatz 8 8
48143 Münster 9
10
Telephone: +49 251 83-24790 11
12
E-mail: mhippler@uni-muenster.de 13
14
Research Area: Membranes, Transport and Biogenetics: Associate Editor Anna Amtmann 15
(Glasgow) 16
17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32
2 Title of article:
33
State transition7-dependent phosphorylation is modulated by changing environmental 34
conditions and its absence triggers remodeling of photosynthetic protein complexes 35
36
Sonja Verena Bergnera, Martin Scholza, Kerstin Trompelta, Johannes Bartha, Philipp Gäbeleina, 37
Janina Steinbecka, Huidan Xuea, Sophie Clowezb, Geoffrey Fucilec, Michel Goldschmidt- 38
Clermontc, Christian Fufezana and Michael Hipplera*
39
40 a Institute of Plant Biology and Biotechnology, University of Münster, Münster 48143, Germany 41
b Institut de Biologie Physico-Chimique, UMR 7141 CNRS-UPMC, 13 rue P et M Curie, 75005 42
Paris, France.
43
c Department of Botany and Plant Biology and Institute of Genetics and Genomics in Geneva 44
University of Geneva, CH-1211 Geneva 4, Switzerland 45
46
S.V.B. and M.S. contributed equally to this work 47
48 49
One sentence summary:
50 51
Lack of STT7-dependent protein phosphorylation in Chlamydomonas triggers stress-induced 52
changes in protein expression, photosystem II core phosphorylation and association of LHCSR3 53
with photosystem I 54
55 56 57 58 59 60 61 62 63 64 65 66 67 68 69
3
Footnotes: Support by the Swiss National Foundation (grant 31003A_146300) to M.G.C. and 70
from the “Deutsche Forschungsgemeinschaft” (DFG, HI 739/7.2 in FOR964 and HI 739/8.1) to 71
M.H. is acknowledged.
72 73
*Corresponding author: Michael Hippler (mhippler@uni-muenster.de) 74
75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115
4 Abstract
116 117
In plants and algae the Ser/Thr kinase STN7/STT7 is an important regulator in acclimation to 118
changing light environments. In the present work we assessed STT7-dependent protein 119
phosphorylation under high light, known to fully induce expression of LHCSR3 and non- 120
photochemical quenching mechanism, in relationship to anoxia where activity of cyclic electron 121
flow (CEF) is stimulated. Our quantitative proteomics data revealed numerous novel STT7 122
protein substrates and STT7-dependent protein phosphorylation variations that were reliant on the 123
environmental condition. These results indicate that STT7-dependent phosphorylation is 124
modulated by the environment and point to an intricate chloroplast phosphorylation network 125
responding in a highly sensitive and dynamic manner to environmental cues and alterations in 126
kinase function. Functionally, the absence of the STT7 kinase triggered changes in protein 127
expression, photoinhibition of PSI and resulted in remodeling of photosynthetic complexes. This 128
remodeling initiated a pronounced association of LHCSR3 with PSI-LHCI-FNR-supercomplexes.
129
Lack of STT7 kinase strongly diminished PSII-LHCII supercomplexes while PSII core complex 130
phosphorylation and accumulation was significantly enhanced. In conclusion, our study provides 131
strong evidence that regulation of protein phosphorylation is critical for driving successful 132
acclimation to high light and anoxic growth environments and gives new insights into 133
acclimation strategies to these environmental conditions.
134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154
5 Introduction
155 156
Oxygenic photosynthesis converts solar energy into chemical energy. This energy is utilized for 157
carbon dioxide assimilation, allowing the formation of complex organic material. Plant 158
photosynthesis is performed by a series of reactions in and at the thylakoid membrane resulting in 159
light-dependent water oxidation, NADP reduction and ATP formation (Whatley et al., 1963).
160
These light reactions are catalyzed by two photosystems (PSI and PSII). A third multi-protein 161
complex, also embedded in the thylakoid membrane, is the cytochrome b6f complex (cyt b6f 162
complex) that links photosynthetic electron transfer processes between the two photosystems and 163
functions in proton translocation. The ATP synthase takes advantage of the proton-motive force 164
that is generated by the light reactions (Mitchell, 1961) to produce ATP. ATP and NADPH, 165
generated through linear electron flow from PSII to PSI, drive the Calvin Benson Bassham cycle 166
(Bassham et al., 1950) to fix CO2. Alternatively cyclic electron flow (CEF) between PSI and the 167
cyt b6f complex solely produces ATP (Arnon, 1959).
168
Under normal growth conditions, CEF provides additionally required ATP for CO2
169
fixation (Lucker and Kramer, 2013), counteracts over-reduction of the PSI acceptor side under 170
stressful environmental cues and re-adjusts the ATP poise leading to increased lumen 171
acidification important for photo-protection (Alric, 2010; Peltier et al., 2010; Leister and 172
Shikanai, 2013; Shikanai, 2013). In microalgae and vascular plants, CEF relies on the NAD(P)H 173
dehydrogenase (NDH)-dependent and/or PROTONGRADIENT REGULATION5 (PGR5)- 174
related pathways (Munekage et al., 2002; Munekage et al., 2004; Petroutsos et al., 2009; Tolleter 175
et al., 2011; Johnson et al., 2014). For both pathways, supercomplexes consisting of PSI-LHCI 176
and components of the respective electron transfer routes have been identified. In Arabidopsis 177
thaliana, a novel NDH-PSI supercomplex with a molecular mass of >1000 kDa was discovered 178
(Peng et al, 2008). From Chlamydomonas reinhardtii, Iwai et al. (Iwai et al, 2010) isolated a 179
protein supercomplex composed of PSI-LHCI, LHCII, the cytochrome b6/f complex, ferredoxin 180
(Fd)-NADPH oxidoreductase (FNR), and PGRL1.
181
PGRL1 and PGR5 interact physically in A. thaliana and associate with PSI to allow the 182
operation of CEF (DalCorso et al, 2008). Functional data suggest that PGRL1 might operate as a 183
ferredoxin-plastoquinone reductase (Hertle et al., 2013). The PGRL1 containing CEF- 184
supercomplex isolated from C. reinhardtii is capable of CEF under in vitro conditions in the 185
presence of exogenously added soluble plastocyanin and ferredoxin (Iwai et al, 2010). Terashima 186
6
et al. isolated a CEF supercomplex of similar composition from anaerobic growth conditions that 187
was active in vitro and contained proteins such as the chloroplast localized Ca2+ sensor CAS and 188
ANR1 (anaerobic response 1), which were also shown to be functionally important for efficient 189
CEF in the alga (Terashima et al., 2012). Notably it was suggested that the onset of CEF in C.
190
reinhardtii is redox-controlled (Takahashi et al., 2013).
191
It has been demonstrated that efficient CEF is crucial for successful acclimation to excess 192
light (Munekage et al., 2004; Dang et al., 2014; Johnson et al., 2014; Kukuczka et al., 2014). The 193
most rapid response to excess light, however, relies on a mechanism called non-photochemical 194
quenching (NPQ). The fastest constituent of NPQ is qE, which operates at a time scale of seconds 195
to minutes and regulates the thermal dissipation of excess absorbed light energy, thereby 196
providing effective photo-protection. In vascular plants, the PSII protein PSBS is essential for qE 197
(Li et al, 2000), whereas qE induction in the green alga Chlamydomonas reinhardtii is mediated 198
by LHCSR3, an ancient light-harvesting protein that is missing in vascular plants (Peers et al., 199
2009). CEF and qE are complementary for acclimation to excess light as double mutants deficient 200
in both mechanisms possess additive phenotypes and are highly sensitive to light (Kukuczka et 201
al., 2014). Another constituent of NPQ is qT, which is related to state transitions. State transitions 202
are important to balance the excitation energy between PSI and PSII (Bonaventura and Myers, 203
1969; Murata, 1969). Under light conditions where PSII is preferentially excited, both PSII core 204
and LHCII proteins become phosphorylated (Lemeille and Rochaix, 2010). As a consequence 205
phosphorylated LHCII proteins detach from PSII and partly connect to PSI (state 2). Under 206
conditions where PSI excitation is predominant, this process is reversed. LHCII proteins are de- 207
phosphorylated and associate with PSII (state 1). The extent of state transition between vascular 208
plants such as A. thaliana and C. reinhardtii differs significantly. The proportion of mobile 209
LHCII antenna is about 80 % in the alga, whereas in A. thaliana only 15-20 % of LHCII are 210
transferred to PSI under state 2 conditions (Lemeille and Rochaix, 2010). However, the large 211
increase in PSI antenna size in C. reinhardtii has recently been challenged (Nagy et al., 2014;
212
Unlu et al., 2014): While 70 to 80 % of mobile LHCII detached from PSII in response to 213
transition to state 2 conditions, only a fraction of about 20 % functionally attached to PSI.
214
Phosphorylation of LHC proteins requires the function of the STT7 kinase or its ortholog 215
STN7 in C. reinhardtii or A. thaliana, respectively. In the absence of the STT7/STN7 kinase, the 216
initiation of state transitions is blocked (Depege et al., 2003; Bellafiore et al., 2005). The mobile 217
LHCII fraction of Chlamydomonas includes the two monomeric LHCII proteins CP26 and CP29 218
7
and the major LHCII protein LHCBM5 (Takahashi et al., 2006), but also the LHCSR3 protein 219
was suggested to migrate during state transitions (Allorent et al., 2013). Takahashi and co- 220
workers (Takahashi et al., 2014) suggested that only CP29 and LHCBM5 directly associate with 221
PSI to form the PSI-LHCI-LHCII supercomplex, while the binding of CP26 could occur 222
indirectly or via the other two proteins. However, it is not yet known whether STT7 directly 223
phosphorylates the LHCII proteins, or if this takes place as part of a kinase cascade (Rochaix, 224
2007). Nevertheless, the direct interaction between STT7 and the LHCII proteins is quite likely, 225
since none of the other chloroplast kinases was found to be specifically required for LHCII 226
phosphorylation (Rochaix, 2014). The activity of the STT7 kinase is mainly determined by the 227
redox status of the plastoquinone pool (Vener et al., 1997; Zito et al., 1999). The identification of 228
a PP2C-type phosphatase responsible for the dephosphorylation of the LHCII proteins in 229
Arabidopsis has been described by two studies in parallel pointing to the fact that this enzyme 230
called PPH1/TAP38 acts directly on phosphorylated LHCII proteins, in particular when they are 231
associated with the PSI-LHCI supercomplex (Pribil et al., 2010; Shapiguzov et al., 2010).
232
Moreover, it is not known whether these phosphatases are constitutively active or if they are 233
regulated by other means, for example through the redox state of the PQ-pool. Nonetheless, both 234
enzymes are conserved in land plants and exhibit orthologous proteins in C. reinhardtii (Rochaix 235
et al., 2012).
236
Another kinase related to STN7/STT7 is encoded in the Arabidopsis and Chlamydomonas 237
genomes and named STN8 and STL1, respectively. STN8 is involved in PSII core subunit 238
phosphorylation and influences repair of PSII after photodamage (Bonardi et al., 2005; Vainonen 239
et al., 2005). Remarkably, the disassembly of the PSII holocomplex is inhibited in STN7/STN8 240
double mutants (Tikkanen et al., 2008; Fristedt et al., 2009; Dietzel et al., 2011; Nath et al., 241
2013), suggesting that phosphorylation of core subunits is required for PSII disassembly. It was 242
further suggested that STN8 controls the transition between LEF and CEF by phosphorylation of 243
PGRL1 in Arabidopsis (Reiland et al., 2011). As described for STN7, activity of STN8 is 244
probably regulated via the redox state of the PQ pool (Bennett, 1991; Fristedt et al., 2009).
245
Notably, the action of STN8 is counteracted by a chloroplast PP2C phosphatase (PBCP) (Samol 246
et al., 2012), allowing for fast reversibility of STN8 mediated acclimation responses. Thus it 247
appears that an intricate regulatory network of chloroplast protein kinases and phosphatases 248
evolved in vascular plants and algae that drives the acclimation response to various 249
environmental cues including excess and changing light settings (Rochaix et al., 2012). As 250
8
STN7/STT7 and STN8/STL1 kinase activities appear to be controlled by the redox poise of the 251
PQ pool, the PQ pool would be a central player in these acclimation responses. On the other hand 252
the kinases themselves are subjected to phosphorylation (Reiland et al., 2009; Lemeille et al., 253
2010; Reiland et al., 2011; Wang et al., 2013). However, the functional consequences of this 254
phosphorylation are unknown.
255
Recent comparative analyses revealed the presence of at least 15 distinct chloroplast protein 256
kinases, suggesting an intricate kinase phosphorylation network in the chloroplast (Bayer et al., 257
2012). Generally phosphorylation of proteins is one of the most abundant post-translational 258
modifications (PTM). In complex eukaryotic systems protein phosphorylation occurs most 259
frequently on serine (Ser, pS) followed by threonine (Thr, pT) residues, whereas protein 260
phosphorylation of tyrosine (Tyr, pY) residues (1800:200:1) is comparatively rare (Hunter, 1998;
261
Mann et al., 2002). Protein phosphorylation is a general phenomenon in vivo; it is assumed that 262
about one third of all proteins are phosphorylated at a given time (Cohen, 2000; Ahn and Resing, 263
2001; Venter et al., 2001; Manning et al., 2002; Knight et al., 2003). A recent large-scale 264
quantitative evaluation of human proteomic data strengthened the importance of protein 265
phosphorylation for cellular function and human biology (Wilhelm et al., 2014). Chlamydomonas 266
and Arabidopsis genomes encode large kinase families (Arabidopsis, 2000; Kerk et al., 2002;
267
Merchant et al., 2007), supporting the view that protein phosphorylation also plays an important 268
role in a plant’s life cycle. It is thus evident that the understanding of protein phosphorylation, 269
including the specificity of residues phosphorylated or dephosphorylated in response to cellular 270
as well as environmental factors, is one key to the understanding of the complex functional 271
biological networks at a whole system level. Likewise, it is crucial to design experimental setups 272
allowing the linkage between phosphorylation events and particular physiological consequences 273
to be elucidated.
274
In this regard we designed experiments to investigate STT7 kinase-dependent 275
phosphorylation dynamics in C. reinhardtii in response to high light and anoxia, employing 276
quantitative proteomics in conjunction with in-depth physiological characterization. These 277
conditions are particularly interesting, as high light conditions are known to fully induce 278
LHCSR3 protein expression and qE while anoxia promotes CEF activity. Recently it was 279
demonstrated, that qE and CEF are complementary and crucial in acclimation to these 280
environmental cues (Kukuczka et al., 2014). Notably, LHCSR3 phosphorylation was suggested to 281
depend on STT7 function (Bonente et al., 2011), while CEF-supercomplex formation was found 282
9
to be independent of STT7 kinase function (Takahashi et al., 2013), indicating that STT7 function 283
might impact acclimation to high light and anoxia in different ways. However, our quantitative 284
proteomics and physiological data reveals that STT7-dependent variations in protein 285
phosphorylation profiles have similar dramatic phenotypic consequences in both conditions, 286
strongly suggesting that regulation of protein phosphorylation is critical for driving successful 287
acclimation to high light and anoxic growth environments.
288 289 290 291
10 Results
292 293
Acclimation responses of plants towards adverse environmental conditions require and correlate 294
with phosphorylation of chloroplast proteins, implying functional importance of this post- 295
translational modification for a successful survival strategy. The STT7 kinase is an important 296
chloroplast kinase involved in protein phosphorylation events driving acclimation processes. To 297
gain deeper insights into STT7-dependent chloroplast protein phosphorylation in acclimation to 298
high light and anaerobic growth conditions, we analyzed two STT7 mutants (stt7-1 (Depege et 299
al., 2003) and stt7-9 (Cardol et al., 2009)) and wild type for protein phosphorylation by 300
employing phosphoproteomics and quantitative mass spectrometry.
301
11
To this end, thylakoid membranes were isolated from the three strains under the 302
respective conditions and digested with trypsin (Figure 1, Supplement Tables 1 and 2) (Turkina et 303
al., 2006). Resulting peptides were subjected to SIMAC and using titanium dioxide (TiO2) 304
12
chromatography for the enrichment of phosphopeptides (Thingholm et al., 2008). In parallel, 305
thylakoid membranes from wild type (14N-labeled) and stt7-1 (15N-labeled) were isolated from 306
cells grown for 24h under high light. These membranes were solubilized with detergent and 307
13
fractionated by sucrose density centrifugation according to previously published work (Tokutsu 308
and Minagawa, 2013). In total six distinct sucrose fractions were collected (see below, Figure 309
4B). These fractions contained LHCII monomers and trimers, PSII core complexes, PSI-LHCI 310
14
and PSII-LHCII complexes as well as putative CEF supercomplexes. Sucrose gradient fractions 311
were tryptically digested according to the FASP method (Wisniewski et al., 2009) followed by 312
TiO2 phosphopeptide enrichment. Subsequently, peptides were analyzed by tandem mass 313
15
spectrometry (MS/MS). For peptide and phosphopeptide identification, OMSSA and X!
314
TANDEM algorithms were used to evaluated the MS/MS data incorporated as pipeline in 315
Proteomatic (Specht et al., 2011). In total, 884 phosphoproteins and 2,434 distinct 316
16
phosphopeptides were identified from the thylakoid membranes (Supplemental Table 1). To 317
“enhance” identification and quantification, LC-MS/MS runs of enriched (after SIMAC 318
enrichment) and non-enriched peptide samples (after tryptic digestion) were aligned in intensity 319
17
and retention time domains using piqDB (Hohner et al., 2013; Barth et al., 2014). This 320
database/platform integrated information of peptide identification and quantification qualities into 321
the definition of accurate isotopologue retention time windows (IRTW), thereby allowing high 322
18
quality alignments to be performed. As a result confident quantifications of all peptides in all LC- 323
MS/MS runs is possible, an approach very similar to that of SuperHirn (Mueller et al., 2007).
324
Subsequent label-free quantitation was performed using pyQms, a tool for MS1-based 325
19
quantitation (Hohner et al., 2013; Barth et al., 2014). Protein and peptide quantitation and peptide 326
phosphorylation events are listed in Supplemental Table 2. For quantitative arguments, these data 327
will be solely considered in the following. Although these criteria are very stringent, hundreds of 328
peptide phosphorylations can be described for the first time in a quantitative way and can be 329
partially related to STT7 kinase dependent action. Correspondingly, the focus of this work was to 330
dissect the phosphorylation events dependent on STT7 kinase activity.
331
Surprisingly, quantitative data presented in Figure 1A demonstrated that the STT7 kinase 332
is present in strain stt7-9, since an STT7 phosphopeptide, corresponding to a previously 333
recognized STT7 phosphopeptide (Lemeille et al., 2010), could be identified from stt7-9 334
thylakoid membranes. The presence of STT7 in stt7-9 is verified by SDS-PAGE and immunoblot 335
analyses using STT7-specific antibodies, showing that stt7-9 accumulates as a ~100 kDa fusion 336
protein of STT7 and, probably, ARG7, which can be explained by the in-frame insertion in the C- 337
terminal domain of the arg7 cassette used for genetic disruptions of the stt7 gene (Depege et al., 338
2003) (Supplemental Figure 1). Thus stt7-9 can be considered a leaky STT7 mutant. Due to the 339
presence of the phosphopeptide DAGLA(p)S533MEEAILK and peptide 340
VNQGALTEAQLMEELGLQEPAPVAPR we conclude that the kinase domain in stt7-9 possibly 341
intact while the C-terminus with yet unknown function might be absent. In contrast, no STT7 342
peptides could be identified from stt7-1 neither from thylakoids (Figure 1A) nor in sucrose 343
gradient fractions (Supplemental Table 2), in line with the fact that no STT7 protein could be 344
observed in the immunoblot assay (Supplemental Figure 1). Taking advantage of the label free 345
quantitation event of peptide 16 (VNQGALTEAQLMEELGLQEPAPVAPR) present in wild type 346
and stt7-9 thylakoids from high light conditions (Supplemental Table 2), we estimate that the 347
STT7 kinase is about 6-fold more abundant in wild type as compared to stt7-9. Notably, several 348
proteins in this dataset show STT7-dependent phosphorylation (Supplemental Table 1, 2 and 3).
349
In many cases proteins that lacked phosphorylation in stt7-1 were phosphorylated in stt7-9. In 350
more rare cases protein phosphorylation was present in wild type but absent in both stt7-1 and 351
stt7-9.
352
Here we identified and quantified for the first time LHCSR3 peptides that were 353
phosphorylated in an STT7 dependent fashion. Quantification results for LHCSR3 demonstrated 354
the lack of phosphorylation of S26, S28, T32 and T33 in the stt7-1 mutant (Figure 1B), whereas 355
phosphorylation was observed in stt7-9 (phosphopeptide enriched fraction) and wild type (non- 356
enriched and SIMAC enriched fractions). The absence of phosphorylation in LHCSR3 from stt7- 357
20
1 was confirmed by analyses of the sucrose gradient fractions. More importantly, it became 358
evident that non-phosphorylated peptides harbouring S26, S28, T32 and T33 were more abundant in 359
the mutant compared to the wild type (see peptide 22, 23, 29 and 30). At the same time the 360
average abundance of peptides, which do not contain known phosphorylation sites and which 361
were quantified in both strains is slightly higher in the wild type (see peptide 4, 5, 11 and 16).
362
Consequently, the absence of detectable phosphopeptides in the mutant is not due to a lack of 363
LHCSR3 expression. Moreover, the differences between wild type and mutant with respect to 364
abundance of the non-phosphorylated peptides allowed the estimation of the degree of 365
phosphorylation. Accordingly we could detect 60 % more non-phosphorylated peptides 366
harbouring the phosphorylatible residues S26, S28, T32 and T33 for the mutant compared to the wild 367
type, likely reflecting the fraction of wild type-specific phosphorylated peptides. While the 368
STT7-dependent phosphorylation was observed in the N-terminal part of the protein, there was a 369
STT7-independent phosphorylation found in a peptide closer to the C-terminal that is not 370
proteotypic and shared between LHCSR1 and LHCSR3. Thus it remains unclear whether this 371
phosphorylation is present in only one LHCSR protein or in both but it demonstrates that a 372
second chloroplast kinase, other than STT7, is phosphorylating LHCSR protein(s) at their C- 373
terminal part. Besides the above mentioned non-proteotypic phosphorylation site, LHCSR1 was 374
not found to be phosphorylated at other proteotypic sites.
375
Other noteworthy proteins that showed STT7 dependent phosphorylation as revealed by 376
MS/MS analyses of both, thylakoid and sucrose density fractions, were PETO, LHCBM5 (Figure 377
1C, D) and TEF23 (Supplemental Table 2, 3). Phosphorylation of LHCBM5 has been reported 378
before (Lemeille et al., 2010). Interestingly, phosphorylation events of PETO and TEF23 were 379
also detected in stt7-9 whereas LHCBM5 phosphorylation could neither be detected in stt7-1 nor 380
in stt7-9. For Cre12.g540700.t1.2, a protein predicted to localize to the chloroplast by PredAlgo 381
(Tardif et al., 2012), phosphorylated peptides could be detected in the wild type and stt7-9 but not 382
in stt7-1 in non-enriched high light thylakoid samples, while phosphorylated peptides could be 383
observed in stt7-1 after SIMAC based enrichment, indicating that the amount of phosphorylation 384
of Cre12.g540700.t1.2 was drastically reduced in the stt7-1 mutant particularly under high light 385
growth conditions. This is supported by the fact that only one phosphorylation quantitation event 386
(peptide 5) could be detected for Cre12.g540700.t1.2 after TiO2 based phosphopeptide 387
enrichment in sucrose density fractions stemming from high light thylakoids of stt7-1 in contrast 388
to the wild type, where four phosphorylation events were recognized. In the case of STL1 (Figure 389
21
2), LHCA6 and PGRL1 (Supplemental Table 2, 3), no phosphorylation was seen under high light 390
conditions in stt7-1 (thylakoids and sucrose density fractions) in contrast to the wildtype.
391
However, phosphorylation was present in stt7-1 under anaerobic conditions as observed in the 392
wild type, implying that phosphorylation of these proteins is independent of STT7 under anoxia.
393
On the other hand, STT7 is involved in phosphorylation under high light settings. Also PSAF 394
was not phosphorylated in high light thylakoids from stt7-1, while it was in wild type. However 395
under anaerobic conditions phosphorylation of PSAF was enhanced in stt7-1 in respect to wild 396
type. The absence of phosphorylation for STL1 and PGRL1 under high light is confirmed by the 397
MS/MS analyses of the sucrose fractions (Figure 2, Supplemental Table 2, 3). In contrast, PSAF 398
was found phosphorylated in sucrose density fractions of stt7-1. Subunit IV of the cyt b6f 399
complex appeared to be phosphorylated in a STT7 dependent manner. In thylakoids from high 400
light conditions, its phosphorylation was three-times less abundant in stt7-1 as compared to wild 401
type, while it was missing in the mutant under anaerobic conditions. After sucrose density 402
centrifugation of detergent solubilized thylakoids treated with high light, no phosphorylation of 403
subunit IV was detectable in stt7-1 but present in wild type, confirming the STT7-dependent 404
phosphorylation. Other proteins, that were phosphorylated in a STT7-dependent way 405
(Supplemental Table 2, 3) and only found in phosphopeptide enriched and/or non-enriched 406
22
fractions from isolated thylakoids are CDS1 (Cre12.g561550.t1.1), HSP70B 407
(Cre06.g250100.t1.1), Cre17.g737450.t1.2, Cre06.g284100.t1.1, Cre09.g398400.t1.1 and 408
Cre17.g724150.t1.1. Cre01.g071450.t1.2 shares similarity with STN8 and STN7 kinases from 409
Arabidopsis. This protein was found phosphorylated in enriched fractions of high light thylakoids 410
from wild type and stt7-9 but not in stt7-1. As observed for STL1, this putative kinase was found 411
phosphorylated in stt7-1 under anaerobic conditions. However, according to PredAlgo (Tardif et 412
al., 2012) prediction, it is not chloroplast-localized.
413
In contrast to proteins showing lack of phosphorylation in the absence of the STT7 kinase, 414
the reaction center building PSII subunit PSBD (D2) displayed increased phosphorylation in the 415
absence of the kinase (Figure 3, Supplemental Table 2). The amounts of phosphorylated PSBD 416
peptides stemming from thylakoids isolated from high light conditions increased in stt7-1 5.6 and 417
4.4 fold as compared to wild type and stt7-9, respectively (peptides 21, 22, 25, 26). At the same 418
time the overall amount of PSBD non-phosphorylated peptides only slightly augmented 1.08 and 419
1.03 fold in stt7-1 in comparison to wild type and stt7-9, respectively. In thylakoids from 420
anaerobic conditions, the differences in amounts of phosphorylated PSBD peptides between stt7- 421
1 versus wild type and stt7-9 were even more pronounced. In this regard, a 73 and 44 fold 422
induction of phosphorylation was observed in stt7-1 while the quantities of non-phosphorylated 423
PSBD increased 1.6 fold in relation to wild type and stt7-9. A similar picture was observed for 424
phosphorylation of PSBC (CP43) (Supplemental Table 2, 3). Phosphopeptides (38, 39) were 2.3 425
and 1.7 fold more abundant in stt7-1 thylakoids isolated from high light conditions in comparison 426
to wild type and stt7-9, respectively. These differences further increased to 10.4 and 2.4 for wild 427
type and stt7-9, respectively, when thylakoids stemming from anaerobic conditions were 428
analyzed. Comparing the amounts of non-phosphorylated peptides of stt7-1 under high light and 429
anaerobic conditions revealed that PSBC was slightly enriched with ratios of 1.2 and 1.07 and 1.1 430
and 1.4 for wild type and stt7-9, respectively. In conclusion, deletion of the STT7 kinase caused a 431
significant increase in PSBD and PSBC phosphorylation, whereas depletion of the kinase in stt7- 432
9 led to only slightly higher PSBD and PSBC phosphorylation as compared to wild type, 433
suggesting that the amount of PSII core phosphorylation is not linked to STT7 in a strict 434
concentration dependent manner but rather to a threshold in STT7 activity. Notably, other core 435
subunits of PSII such as PSBA (D1) and PSBB (CP47) were not found phosphorylated.
436
Beside differences in protein phosphorylation in stt7-1 versus wild type and stt7-9, many 437
proteins were not altered in their phosphorylation status. As examples, it can be affirmed that 438
23
photosynthetic proteins such as PSAH, PSBR and PSBO showed a similar protein 439
phosphorylation profile in the three strains as already reported for PSAH and PSBR in stt7-1 and 440
wild type (Lemeille et al., 2010). In case of LHCB4 no qualitative changes in phosphorylation 441
events could be detected between the three strains. Moreover, LHCB4 peptide VATSTGTR, 442
which was previously found to be differentially phosphorylated between stt7-1 and wild type 443
(Lemeille et al., 2010), was detected to be phosphorylated in stt7-1 (Table 1, Supplemental Table 444
1), demonstrating that this peptide can be phosphorylated in an STT7 independent manner.
445
However quantitative changes in LHCB4 phosphorylation were observed. Phosphopeptides 20 446
and 21 (NNKGS103VEAIVQATPDEVSSENR, Charge 2 and 3, respectively) of LHCB4 in stt7-1 were 447
diminished three-fold in regard to wild type and stt7-9, whereas phosphopeptides 5 and 6 and the 448
overall amount of non-phosphorylated peptides in stt7-1 were slightly more abundant as 449
compared to wild type and stt7-9. MS/MS analysis of non-enriched samples derived from FASP- 450
digested sucrose density gradient fractions showed that the same phosphopeptide (18 and 19 451
corresponding to phosphopeptides 20 and 21 found in thylakoids) was present in wild type 452
(monomeric LHCII) but not in stt7-1, therefore underpinning the notion that this phosphopeptide 453
is diminished in stt7-1 under high light growth conditions (Supplemental Table 2).
454
24
In the quantitative analyses of STT7 dependent phosphorylation, the comparison of 455
phosphopeptides from thylakoids and sucrose density fractionated thylakoids determined 456
differences independently. However, the sucrose density fractionation of detergent solubilized 457
thylakoids offered another advantage, namely the possibility to analyze co-migration of proteins 458
with known multi-protein complexes such as PSII and PSI, the cyt b6f complex, the ATPase or 459
the CEF-supercomplex. Thus the aim of these experiments was to determine whether STT7 460
deficiency alters the dynamic association of proteins in the thylakoid membrane. To define 461
amounts of PSII, PSI, the cytochrome b6f complex and the ATPase in thylakoids isolated after 24 462
h high light treatment, SDS-PAGE fractionation of wild type and stt7-1 thylakoids and 463
immunoblotting experiments were performed (Figure 4A). Using anti-ATPB antibodies 464
demonstrated that ATPB, a subunit of the chloroplast ATPase, was similar in amounts in wild 465
type and stt7-1, while the use of anti-PSBA and anti-CYTF antibodies showed that the amounts 466
of PSBA and CYTF are increased in stt7-1. In contrast, LHCSR3, PSAD and FNR were slightly 467
decreased in stt7-1 as revealed with specific antibodies directed against these proteins. Overall 468
these differences were also reflected in quantitative mass spectrometric data stemming from 469
isotopic labeling experiments (see below, Figure 5). In a next step, thylakoid membranes isolated 470
from wild type and stt7-1 cells grown in high light were solubilized with detergent and 471
fractionated by sucrose density centrifugation according to previously published work (Tokutsu 472
and Minagawa, 2013). Figure 4B shows the corresponding sucrose gradients after ultra- 473
centrifugation. Six green bands could be distinguished. These green bands corresponded, as 474
revealed by SDS-PAGE fractionation and immunoblotting experiments (Figure 4 C, D), to LHCII 475
monomers, LHC trimers, PSII core complexes, PSI-LHCI, PSII-LHCII complexes and putative 476
CEF supercomplexes. The data clearly showed that the intensity of the LHCII monomer and the 477
PSII core band increased in stt7-1 versus wild type. On the other hand, PSI-LHCI and PSII- 478
LHCII supercomplexes were diminished in stt7-1. The immunoblotting experiments after SDS- 479
PAGE fractionation of the sucrose gradients fractions using various specific antibodies revealed 480
some very interesting results. The amounts of PSBA in PSII-LHCII- (fractions 9-11) and PSAD 481
in PSI-LHCI- supercomplexes (fractions 11-13) were slightly diminished in stt7-1 compared to 482
the wild type, as seen in the band pattern of the sucrose gradients. Fraction five in wild type and 483
stt7-1 corresponded to the peak fraction of a putative CEF supercomplex, where PSAD, ANR1, 484
CYTF and FNR co-migrated as previously reported (Iwai et al., 2010; Terashima et al., 2012;
485
Takahashi et al., 2013). However, the CYTF signal peaked in stt7-1 in fraction 7 and was not 486
25
detectable for wild type. Thus we designated the putative protein complex in fraction 5, present in 487
wild type and stt7-1, as PSI-LHCI-FNR supercomplex as described previously (Takahashi et al., 488
2014). Moreover, the intensity of the LHCSR3 signal in fraction 5 was also stronger in stt7-1 489
26
than in wild type. The fact that LHCSR3 co-migrated with this PSI and FNR containing 490
supercomplex has not been described before. Association of LHCSR3 with PSI-LHCI in state 2 491
and high light had been described instead (Allorent et al., 2013; Xue et al., 2015). Notably, the 492
amounts of LHCSR3 in the LHCII monomeric and trimeric fractions in stt7-1 decreased in 493
comparison to wild type (Figure 4D). Thus the increase of LHCSR3 in the PSI-LHCI-FNR- 494
supercomplex fraction of stt7-1 is likely due to a change in association caused by a lack in STT7 495
rather than an overall increase in LHCSR3 abundance. This result is further underpinned when 496
considering the LHCSR3 amount in stt7-1 thylakoids (Figure 4A). On the other hand, the 497
amounts of CYTF significantly increased in fractions 4-16 in stt7-1 as compared to the 498
corresponding fractions in wild type.
499
To independently assess the amounts of thylakoid membrane proteins in the distinct 500
protein complexes in stt7-1 and wild type we performed quantitative mass spectrometric analyses 501
based on isotopic labeling. To this end, thylakoid membranes from 14N-labeled wild type and 502
15N-labeled stt7-1 cells grown for 24h in high light were isolated and subjected to detergent- 503
mediated solubilization and sucrose density gradient centrifugation and investigated by mass 504
spectrometry. The data confirmed that CYTF was more abundant in stt7-1 than in wild type in all 505
sucrose fractions analyzed (Figure 5, Supplemental Table 2). Moreover, FNR was enriched in 506
stt7-1 PSI-LHCI-FNR supercomplex fractions. PSAD amounts were found to be diminished in 507
stt7-1, in agreement with immunoblot data shown in Figure 4C. Furthermore, these data showed 508
that LHCSR3 was not only significantly more abundant in PSI-LHCI-FNR supercomplexes in 509
stt7-1 but also in PSII core complexes. At the same time the abundance of PSBD in stt7-1 was 510
increased considerably in PSII core complexes and decreased in PSII-LHCII supercomplexes.
511
The latter notion is in line with the fact that PSII-LHCII supercomplexes were decreased in the 512
stt7-1 sucrose density band pattern (Figure 4B), in agreement with the observation that PSBD and 513
PSBC phosphorylation was higher in stt7-1 PSII core complexes (Supplemental data Table 2).
514
To provide an experimental prove for a link between LHCSR3 and the PSI-LHCI-FNR 515
supercomplex, we took advantage of a C. reinhardtii His6-tagged PSAA strain (Gulis et al., 2008) 516
and purified the supercomplex from cells grown for 24 h in high light (Figure 6A) using 517
detergent solubilization of isolated thylakoids and sucrose density ultracentrifugation. In a next 518
step, the PSI supercomplex from sucrose density fractions 5-7 were further purified by the tagged 519
PSI-subunit (PSAA) via Ni-NTA chromatography. Fractions 5-7 before and after Ni-NTA 520
purification were separated by SDS-PAGE and analyzed by immunoblotting, employing specific 521
27
antibodies directed against LHCSR3, PSAD and FNR (Figure 6B). The immunoblotting data 522
revealed that PSAD and LHCSR3 were similarly co-purified while FNR was lost via Ni-NTA 523
separation. Therefore we can conclude that LHCSR3 is associated with PSI in the supercomplex.
524
Now, what is the function of LHCSR3 associated with PSI within the PSI-LHCI-FNR 525
supercomplex? A possible scenario might be that LHCSR3 is involved in photo-protection of PSI 526
within such supercomplexes. As recently reported, PSI from pgr5, pgrl1 and pgrl1/npq4 mutant 527
strains (npq4 is lacking LHCSR3 (Peers et al., 2009)) was susceptible to photoinhibition in high 528
light (Johnson et al., 2014; Kukuczka et al., 2014). To investigate whether the PSI 529
photoinhibition occurred faster in pgrl1/npq4 due to a lack of LHCSR3, we performed a time 530
course experiment under high light conditions. As reported earlier (Kukuczka et al., 2014), the 531
amount of oxidizable P700 was comparable in the wild type, pgrl1, npq4 and pgrl1/npq4 at low- 532
28
light intensities (Figure 7A). Interestingly after 4 h high light the extent of oxidizable P700 was 533
already further diminished in pgrl1/npq4 as compared to pgrl1. The same was true after 24 h of 534
high light, while the amounts of oxidizable P700 in the wild type and npq4 remained comparable 535
to the low light control. Fractionation of wild type, pgrl1, npq4 and pgrl1/npq4 cells by SDS- 536
PAGE and immunoblotting using anti-PSAD and anti-ATPB antibodies revealed that the amount 537
29
of PSAD diminished in pgrl1 and pgrl1/npq4 after 4 h and even more after 24 h high light, 538
whereas the amount of ATPB remained similar over the time course (Figure 7B). In line with the 539
observation made for the decrease in amounts of oxidizable P700, the immunoblotting data 540
confirmed that the decrease in PSAD was more pronounced in pgrl1/npq4 than in pgrl1. This 541
supports the notion that LHCSR3 has a functional impact, either direct or indirect, on photo- 542
protection of PSI under conditions where PGRL1 function is absent. The deficiency of PGRL1 543
affects CEF and will alter the redox poise and lumen ΔpH of the chloroplast, implying that PSI is 544
particularly prone to photoinhibition under conditions when CEF is impaired as suggested before 545
(Johnson et al., 2014; Kukuczka et al., 2014). We next investigated PSI photoinhibition under 546
30
high light conditions in stt7-1. Wild type and stt7-1 cells were shifted from low light to high light 547
conditions and the amount of oxidizable P700 was measured after 1, 4 and 24h high light and 24h 548
low light. The data shown in Figure 8A revealed that the amounts of oxidizable P700 decreased for 549
stt7-1 gradually to about 50% oxidizable P700 after 24h in high light in comparison to cells grown 550
for 24h in low light. SDS-PAGE fractionation of wild type and stt7-1 cells followed by 551
immunoblotting using anti-PSAD- anti-LHCSR3 and anti-ATPB antibodies revealed that 552
amounts of PSAD diminished in stt7-1 after 24 h high light to about 50% of the low light control 553
cells (Figure 8B). In addition the comparison to PSAD amounts between stt7-1 cells harvested 554
after 1 h and after 24 h high light, revealed a 50% reduction of PSAD after 24 h, while no change 555
in abundance was observed for PSAD when comparing wild type cells after 1 and 24 h high light.
556
Thus the conclusion that PSI in stt7-1 is also prone to photoinhibition, although not to the same 557
extent as observed for pgrl1 and pgrl1/npq4, was further supported. Moreover, the immunoblot 558
clearly revealed that in stt7-1 LHCSR3 expression is, already present in low light and faster 559
induced in high light in contrast to wild type (Figure 8B), underpinning the assumption that stt7-1 560
is already stressed under conditions where the wild type is not and possibly explaining why the 561
PSI-LHCI-FNR complex contains more LHCSR3 in stt7-1.
562 563 564 565
31 Discussion
566 567
In the present work we investigated STT7 related protein phosphorylation dynamics in C.
568
reinhardtii in response to high light and anoxia. We intended to assess STT7-dependent protein 569
phosphorylation under conditions where LHCSR3 and qE were fully induced in comparison to 570
conditions known to promote CEF activity. Our quantitative proteomics data revealed that the 571
protein phosphorylation status in STT7-deficient cells was dependent on the environmental 572
condition examined. Furthermore, the phosphoproteome differed between STT7-deficient cells 573
(stt7-1) and cells with diminished expression of a mutant form of STT7 kinase (stt7-9). These are 574
indications that the intricate chloroplast phosphorylation network is highly sensitive and dynamic 575
towards environmental cues and alterations in STT7 kinase function. Variations in chloroplast 576
protein phosphorylation profiles due to the absence of STT7 kinase caused changes in protein 577
expression, photoinhibition of PSI and resulted in remodeling of photosynthetic complexes and 578
triggering a pronounced association of LHCSR3 with PSI-LHCI-FNR supercomplexes.
579
Moreover, absence of STT7 kinase strongly diminished PSII-LHCII supercomplexes while PSII 580
core complex phosphorylation and formation was enhanced. In conclusion, our study provides 581
strong evidence that regulation of protein phosphorylation is critical for driving successful 582
acclimation to high light and anoxic growth environments. Furthermore, numerous new potential 583
STT7 kinase substrates were identified and quantified.
584 585
Expression of cyt b6f complex subunits and association of LHCSR3 with PSI-LHCI-FNR is 586
enhanced in the absence of STT7 function 587
The formation of CEF-supercomplexes has been described for state 2 and anoxic conditions (Iwai 588
et al., 2010; Terashima et al., 2012; Takahashi et al., 2013). Herein we provide evidence that such 589
supercomplexes might be formed in high light under aerobic phototrophic conditions as CYTF 590
co-migrates with PSI, FNR, and ANR1 in stt7-1 (Figure 4C). However, the peak fractions of 591
CYTF in comparison to ANR1, PSAD and FNR are different and it is currently unknown 592
whether the putative CEF-supercomplex is functional under high light conditions. The co- 593
migration of PSI, FNR, and ANR1 in wild type and stt7-1 points to the formation of a PSI-LHCI- 594
FNR supercomplex in high light as previously described (Takahashi et al., 2014). Notably, 595
expression of cyt b6f complex subunits in stt7-1 was enhanced (Figure 4A, C, Supplemental 596
Table 2, 3). This increased expression was also observed under anoxic conditions (Supplemental 597
32
Table 2, Supplemental Figure 3). Interestingly, as observed for high light, under anoxic 598
conditions more CYTF and PETO co-migrated with the CEF-supercomplex peak fraction (Figure 599
4C, Supplemental Table 3, Supplemental Figure 2). It is currently unclear how this increase of 600
expression is achieved. As the STT7 kinase is associated with the cyt b6f complex (Lemeille et 601
al., 2009), its absence might impact the assembly-controlled biosynthesis of the complex 602
(Choquet et al., 2001). On the other hand, the increase in complex abundance could be indirectly 603
modulated via e.g. changes in the physiological status of the cells. Depletion in CEF 604
supercomplex associated proteins such as ANR1 or PGRL1 combined with a reduced CEF 605
capacity caused a severe growth defect under high light and/or anoxic conditions (Terashima et 606
al., 2012; Kukuczka et al., 2014), indicating a requirement of efficient CEF for acclimation to 607
these conditions.
608
At the same time a remodeling of PSI-LHCI-FNR supercomplexes occurred by increasing 609
the amount of LHCSR3 associated with PSI within the supercomplex, as revealed by affinity 610
purification using the His6-tagged PSAA strain (Figure 6). The function of LHCSR3 bound to 611
PSI is likely to protect the complex from photoinhibition (Figures 7 and 8). The support for this 612
idea is that a pgrl1/npq4 mutant strain displayed a more pronounced PSI photoinhibition in high 613
light as compared to pgrl1 (herein, (Kukuczka et al., 2014)). However, these data cannot exclude 614
an indirect protective function of LHCSR3 in regard to shielding PSI from photoinhibition under 615
high light stress.
616
Nonetheless these data suggest that non-phosphorylated LHCSR3 was bound 617
preferentially to PSI-LHCI-FNR in stt7-1. In agreement only a small amount of phosphorylated 618
LHCSR3 was found associated with the respective supercomplex in the wild type. In conclusion, 619
the binding of LHCSR3 to PSI-LHCI-FNR is negatively correlated with the activity of the STT7 620
kinase. In this scenario, the inactivation of STT7 dependent phosphorylation via reduction of the 621
luminal STT7 disulfide bridge due to prolonged high light, as outlined by Lemeille and co- 622
workers (Lemeille et al., 2009) would increase the amount of non-phosphorylated LHCSR3, 623
which in turn would increase its binding to PSI-LHCI-FNR, thereby protecting PSI and 624
resembling the situation as in stt7-1.
625 626
Absence of STT7 function causes HL-induced photoinhibition of PSI and increased PSII 627
core phosphorylation and formation 628
33
Despite the binding of LHCSR3 to PSI, a sensitivity of PSI to photoinhibition was observed in 629
stt7-1. The extent of photoinhibition was not as pronounced as seen for the mutant strains pgrl1 630
and pgr5 (Figure 7, 8 (Johnson et al., 2014; Kukuczka et al., 2014)) but significant. It has been 631
reported that STT7/STN7 function is essential for state transitions in algae and vascular plant, 632
respectively (Depege et al., 2003; Bellafiore et al., 2005), respectively. State transitions are 633
important for an efficient balancing of excitation energy between PSI and PSII (Bonaventura and 634
Myers, 1969; Murata, 1969). They involve the phosphorylation of LHCII proteins by 635
STT7/STN7, which then detach from PSII and in part migrate to PSI (state 2). As reported 636
before, LHCBM5 is phosphorylated in state 2 settings under anaerobic and high light conditions 637
in C. reinhardtii. Moreover, LHCBM5 phosphorylation was shown to be STT7-dependent 638
(Lemeille et al., 2010). Notably, as found for stt7-1 in our study, no LHCBM5 phosphorylation 639
was detectable in stt7-9. Like stt7-1, stt7-9 is unable to perform state transitions (Cardol et al., 640
2009). So far phosphorylation of LHCBM5 has only been detected when the protein was 641
associated with PSI-LHCI complexes (Takahashi et al., 2006). Hence, the inability to 642
phosphorylate LHCBM5 might explain the fact that both mutants are locked in state 1. On the 643
other hand, no increased expression of cyt b6f complex subunits, augmented PSII core formation 644
or enhanced binding of LHCSR3 to PSI-LHCI-FNR was observed in stt7-9 (Supplemental Table 645
2, Supplemental Figure 2). Moreover, phosphorylation profiles of stt7-9 and stt7-1 showed clear 646
differences, likely explaining these phenotypic differences. These findings indicate that triggering 647
the association of LHCII with PSI requires far more active STT7 kinase than triggering the 648
acclimation process, which permits full PSII-LHCII supercomplex formation and circumvents 649
PSI photoinhibition in high light conditions. We cannot exclude, that STT7 from stt7-9 is 650
differently regulated due to its altered protein sequence (see also below). However, the lack of a 651
functional interconnection between PSI and LHCII cannot explain the phenotypic differences 652
between stt7-1 and stt7-9. Requirement of STN7 for effective retrograde signaling has been also 653
described for Arabidopsis, as growth phenotypes and diminished relative amounts of PSI had 654
been found in stn7 Arabidopsis plants grown in fluctuating light conditions (Tikkanen et al., 655
2010). Along the same line, growth phenotypes were also observed in plants treated with 656
fluctuating PSII and PSI lights (Bellafiore et al., 2005).
657
Recently the process of state transitions was revisited in C. reinhardtii (Nagy et al., 2014;
658
Unlu et al., 2014). Interestingly these new data indicated that although 70 to 80 % of LHCII 659
detached from PSII in state 2 conditions, only a fraction of about 20 % was attached to PSI. It 660
34
was postulated that the disconnected antenna complexes form energetically quenched complexes 661
that are protected against photodamage and exhibit a shortened excited-state fluorescence lifetime 662
(Unlu et al., 2014). In addition it was suggested that phosphorylated LHCII may also remain with 663
PSII but become quenched (Nagy et al., 2014). Importantly, these studies were conducted under 664
low light in photoheterotrophic conditions where expression of LHCSR3 is not expected. Thus 665
the observed quenching should be independent of LHCSR3 and qE. In this regard it is important 666
to note that phosphorylation of LHCB4 phosphopeptides 20 and 21 667
(NNKGS103VEAIVQATPDEVSSENR, Charge 2 and 3, respectively) were diminished two to 668
three-fold in stt7-1 thylakoids isolated from high light conditions in comparison to stt7-9 and 669
wild type (Supplemental Table 2), respectively. However, herein a postulated STT7-dependent 670
LHCB4 phosphorylation (Lemeille et al., 2010) was found to be present in the absence of STT7.
671
In total 12 different phosphorylation sites were detected, five more than described previously 672
(Turkina et al., 2006). Our data indicate that under high light and photoautotrophic conditions, 673
phosphorylation of LHCB4 is changed merely quantitatively by the absence of STT7. LHCB4 is 674
a minor core antenna protein of PSII that is suggested to be a linker between PSII core and 675
LHCBM proteins and likely involved in excitation energy transfer between the trimeric LHCBM 676
proteins and PSII core (Dainese and Bassi, 1991; Dekker and Boekema, 2005; van Amerongen 677
and Croce, 2013) as well as in state transitions in C. reinhardtii (Tokutsu et al., 2009). Moreover 678
it has been suggested that phosphorylation of LHCB4 and LHCB5 is responsible for the 679
detachment of all LHCBM polypeptides from PSII in the transition from state 1 to state 2.
680
Notably, we were unable to detect LHCB5 phosphorylation in the quantitative data. Since 681
changes in LHCB4 phosphorylation status between wild type and stt7-9 were similar, it is 682
possible that the larger reduction in LHCB4 phosphorylation in stt7-1 directly affected unbinding 683
of LHCBM proteins from PSII. Additionally, phosphorylated LHCB4 was not detected bound to 684
PSII core and suggested to be associated with other free LHCBM (Iwai et al., 2008), thereby 685
decreasing the amount and possibly the organization of unbound LHCII polypeptides. It is 686
tempting to speculate that this altered phosphorylation status might affect the formation of 687
energetically quenched LHCII complexes, since differences in LHCII complex formation became 688
apparent as more monomeric LHCII and less trimeric LHCII was observed in stt7-1 as compared 689
to wild type (Figure 4B).
690
Moreover, the amounts of LHCSR3, which co-migrated with the other free LHCII 691
polypeptides, was severely diminished in stt7-1 (Figure 4D) due to enhanced co-migration with 692
35
PSI in the PSI-LHCI-FNR fraction, which in turn would certainly have an impact on the 693
formation of LHCII complexes, that are energetically quenched by LHCSR3 as it has been shown 694
to quench excited LHC protein bound chlorophyll molecules (Bonente et al., 2011). Therefore a 695
lack in LHCSR3 association with aggregated LHCII complexes would likely promote 696
photooxidative stress, because quenching capacity should be lower.
697
It had been suggested that PSII core subunit phosphorylation functions as a mechanism to 698
protect photodamaged PSII complexes under high light conditions from proteolytic degradation 699
(Koivuniemi et al., 1995; Kruse et al., 1997; Turkina et al., 2006). Additionally, PSII core 700
phosphorylation has been implicated in the regulation of the repair cycle of photodamaged PSII 701
(Aro et al., 1993; Tikkanen et al., 2008; Fristedt et al., 2009; Dietzel et al., 2011; Nath et al., 702
2013) by facilitating the migration of damaged PSII centers from grana to stroma thylakoid 703
where damaged PSBA is replaced with a newly synthesized copy. Phosphorylation of PSBD and 704
PSBC had been also proposed to facilitate dissociation of peripheral light-harvesting antenna 705
(Turkina et al., 2006; Iwai et al., 2008). Our quantitative data clearly showed that PSBD and 706
PSBC were significantly phosphorylated to a higher degree in stt7-1 thylakoids and particularly 707
in PSII core complexes compared to the wild type. This is in line with recent anti-phospho- 708
threonine immunoblot data on isolated stt7-1 and stn7 thylakoids (Fristedt et al., 2010; Lemeille 709
and Rochaix, 2010). Our quantitative data also demonstrated that the formation of PSII core 710
complexes was highly enhanced in stt7-1 and that at the same time the amount of PSII-LHCII 711
supercomplexes was significantly decreased in comparison to wild type. Thus indicating that the 712
absence of STT7 was responsible for this transition and suggesting that PSII-LHCII complexes 713
were prone to photoinhibition in HL-acclimated stt7-1.
714
It is conceivable that strong PSII core phosphorylation in stt7-1, possible due to higher 715
requirement in PSII repair cycle, is linked to enhanced PSII core formation, as it is suggested to 716
trigger dissociation of LHCII proteins from PSII. Like in Arabidopsis (Bonardi et al., 2005;
717
Vainonen et al., 2005), it is proposed that STL1, the C. reinhardtii STN8 ortholog, is required for 718
phosphorylation of PSII core proteins PSBD and PSBC (Rochaix et al., 2012). Notably STL1 719
phosphorylation was absent in stt7-1 under high light but present in anoxic conditions. Yet PSBD 720
and PSBC phosphorylation was detected under high light and anoxic conditions. Thus STL1 721
phosphorylation is not required for PSII core phosphorylation in C. reinhardtii and STL1 722
phosphorylation is not strictly dependent on STT7.
723 724
36
The intricate chloroplast phosphorylation network is highly dynamic and sensitive towards 725
environmental cues and alterations in kinase function 726
Phosphorylation of STL1, LHCSR3 and PETO, for the latter in particular under anoxia, were 727
present in stt7-9, but reduced in regard to wild type, suggesting that the level of decline 728
corresponds and correlates to the diminishment in amounts of STT7 kinase in stt7-9. On the other 729
hand, no phosphorylation was observed for LHCBM5, another abundant substrate, in stt7-1 and 730
stt7-9. Here it is tempting to speculate that beside the depletion in kinase abundance, the 731
proposed regulatory C-terminal part of STT7, which is missing in stt7-9, is also involved in 732
substrate recognition and phosphorylation efficiency and could also be responsible for changes in 733
phosphorylation detected between stt7-1 and stt7-9.
734
As revealed for STL1, phosphorylation of PGRL1 was absent under high light but present 735
under anoxic conditions in stt7-1, thus pointing to highly dynamic chloroplast protein 736
phosphorylation that is dependent on environmental cues and alterations in kinase function. It has 737
been suggested that STN8 is responsible for PGRL1 phosphorylation (Reiland et al., 2011).
738
Phosphorylation of PGRL1 by STN8 slowed down the transition from CEF to LEF in 739
Arabidopsis, thereby providing a link between phosphorylation and control of CEF activity 740
(Reiland et al., 2011). In C. reinhardtii phosphorylation of PGRL1 was observed under anoxia in 741
stt7-1 but it was not increased compared to wild type or stt7-9. In contrast under the same 742
conditions the capacity of CEF was not diminished in stt7-1 in regard to the two other strains 743
(Supplemental Figure 3), implying that the CEF activity per se is not dependent on PGRL1 744
phosphorylation, as CEF rates and PGRL1 phosphorylation are not proportional. In the same line, 745
CEF activity is also not dependent on PETO phosphorylation, as it is absent in stt7-1 under 746
anoxia. It is tempting to speculate that there is a link between PGRL1and STL1 phosphorylation, 747
since their phosphorylation patterns coincide. Currently it is unknown which other chloroplast 748
kinase could phosphorylate STL1 or STN8. The protein Cre01.g071450.t1.1 shares similarity 749
with STN7/STN8 from Arabidopsis and showed the same phosphorylation dynamics as STL1.
750
However, this protein is probably not chloroplast localized (Tardif et al., 2012). Interestingly 751
phosphorylation of subunit IV of the cyt b6f complex was strongly diminished in stt7-1 under 752
high light but completely missing in anoxia. Yet, this is another example showing that protein 753
phosphorylation is not only kinase-dependent but also related to the respective environmental 754
conditions. TEF23 was the other STT7-dependent substrate that was not phosphorylated under 755
high light or anoxia in stt7-1. TEF23 is an ortholog of EGY1 (ethylene-dependent gravitropism- 756
37
deficient and yellow-green 1), a zinc metalloprotease required in Arabidopsis and tomato 757
(Solanum lycopersicum) for proper chloroplast development (Chen et al., 2005; Barry et al., 758
2012). Plants that are deficient in the protease possessed reduced grana stacks, poorly developed 759
thylakoid lamella networks and low amounts of chlorophyll a/b binding proteins. Furthermore it 760
has been suggested that EGY1 participates in abscisic acid/reactive oxygen species (ROS)- 761
dependent plastid retrograde signaling networks, in particular with regard to the regulation of 762
ammonium stress (Li et al., 2012; Li et al., 2013). Mutant plants deficient in EGY1 are 763
characterized by diminished H2O2 accumulation in guard cell chloroplasts, likely caused by an 764
impaired turnover and assembly of membrane-associated components of PSI and PSII (Li et al., 765
2012). In Arabidopsis STN7 was proposed to be involved in ROS homoeostasis and ROS-based 766
retrograde signaling through controlling LHCII distribution and the redox balance in the electron 767
transfer chain (Tikkanen et al., 2012). Hence, the STT7 dependency of TEF23 phosphorylation is 768
striking and may reveal that STT7 affects ROS-based retrograde signaling on multiple levels.
769
Beside the control of redox balance, STT7 might also be involved in the regulation of transport 770
processes in the chloroplast and/or the cell. Accordingly, another STT7 substrate is the protein 771
Cre13.g571500.t1.1 that is phosphorylated in wild type and stt7-9 but not in stt7-1. This protein 772
shows similarity to various transporters and permeases (Merchant et al., 2007), though its 773
location is unknown. Other transport proteins that were found phosphorylated in wild type and 774
stt7-9 are CDS1 (Cre12.g561550.t1.1), RHP1 (Cre06.g284100.t1.1) and TRP5 775
(Cre09.g398400.t1.1). In stt7-1, however, there was no evidence for protein phosphorylation nor 776
presence of non-phosphorylated peptides, indicating that either phosphorylation was absent 777
and/or protein expression vastly diminished. CDS1, a half-size ABC transporter with similarity to 778
Arabidopsis ATM3, suggested to be involved cadmium resistance (Kim et al., 2006) and iron- 779
homeostasis (Chen et al., 2007) is supposedly mitochondrion-localized, although proteomic data 780
also suggest chloroplast localization in Chlamydomonas (Terashima et al., 2011). RHP1 on the 781
other hand is a CO2 channel (Soupene et al., 2002). TRP5 is a member of the transient receptor 782
potential (TRP) multigene superfamily encoding integral membrane proteins that function as ion 783
channels. Notably, this multigene family of channels is absent in vascular plants and represented 784
by 14 potential TRP genes in C. reinhardtii (Fujiu et al., 2011). Currently it is unknown whether 785
TRP5 is chloroplast-localized. However, the absence of STT7-dependent phosphorylation 786
suggests direct or indirect control of TRP5 function and/or presence via STT7. The same holds 787
for the two other discussed transport proteins, since indirect impact cannot be excluded. As TRP 788