Proteomic survey of the thylakoid subfractions from stn7 and pph1 mutants
In order to perform LC-MS/MS analysis, proteins contained in the sam-ples were concentrated on a SDS-PAGE between the stacking and the sepa-rating gels. LC-MS/MS analysis were performed as reported in Tomizioli et al., 2014  and spectral counting was used for relative quanti cation in order to determine the sub-thylakoid localisation of the identi ed proteins. According to conclusions drawn in Tomizioli et al., 2014 , grana-margins, even if analysed in LC-MS/MS were not considered for the further statistical analysis (see section 2.5.2 at page 88 for further details).
Proteomic analysis was carried out in order to highlight proteins char-acterised by a di erent distribution between stroma-lamellae and BBY in the two Arabidopsis stn7 and pph1 mutants. To perform this we focused on the values of protein logarithmic folding change (LogFC). As previously seen, value of LogFC can be associated to the enrichment of a given protein in either BBY or stroma-lamellae fractions. In particular, negative values of LogFC characterise protein enriched in the BBY, while positive values of LogFC characterise protein enriched in the stroma-lamellae. Moreover, for a given protein, the absolute value of the LogFC also allows to esti-mate in which extent the protein is shared between the two thylakoid sub-compartments (i.e BBY and stroma-lamellae). For example, if the protein A is found in stn7 (state-1) with a LogFC = -2 and the same protein A is found in the pph1 (state-2) with a LogFc = -1.8, it means that this protein A is enriched in the BBY fraction in both the genotypes, but in the case of pph1 (state-2) is more shared with stroma-lamellae and their di erence in LogFC ( LogFC stn7-pph1) is 0.2. We exploited this information to put in evidence proteins characterised by a varied sub-thylakoid repartition be-tween stn7 and pph1 mutants. Among the identi ed proteins, speci c Lhcb isoforms are present and their identi cation is summarised in table 3.2.
LC-MS/MS analysis evidenced 10 di erent LHCII isoforms, namely Lhcb 1.4/1.5, Lhcb2, Lhcb 3.1, Lhcb 4.1/4.2 Lhcb 4.3, Lhcb 5 and Lhcb 6. Each antenna is characterised by a di erent enrichment in the stroma-lamellae and BBY sub compartments within the stn7 and pph1 mutants. Indeed these results reveal a clear trend in the antenna distribution, which is agreement with the traditional view of state transitions. In fact because of the lack of the speci c kinase and phosphatase, Lhcb isoforms are more accumulated in stack membranes (i.e. BBY) in stn7 than in pph1. However, when the statistical analysis is taken in consideration, only three antenna show a p-value lower than 0.05 (our con dence threshold). These three Lhcb isoforms are Lhcb 1.5, Lhcb 1.4 and Lhcb 2.2 (see tab.3.2).
In Arabidopsis, LHCII is a trimeric complex composed of a combination of three highly homologous Lhcb isoforms Lhcb1-3 encoded by a large multigene family . LHCII trimers are normally distinguished in function of their (S)trong , (M)oderate or (L)oose association with the photosystem complex. In a recent work Galka et al., 2012  performed a detailed biochemical and mass-spectrometry analysis of the isoform composition of these three trimers and concluded that the trimers mainly involved in state transitions are the speci c (L)oosely bound to PSII (also referred to as mobile). In our data (tab.3.2), isoforms Lhc1.5 and Lhc2.2 display a high di erential accu-mulation between state-1 and state-2. This nding is in agreement with data reported in literature where it was demonstrated that Lhcb 1 and Lhcb 2 iso-forms reach their higher content in the mobile subset of LHCII associated trimers (L) . At the same time, Lhcb 1.5, Lhcb 1.4 ( LogFC 2.4) and Lhcb 2 (tab.3.2) are poorly represented in the fraction M which is thought to have a very limited role in state transitions . One Lhcb isoform that we found with the highest LogFC is Lhcb5 (also referred to as CP26). Studies in Chlamydomonas (where PSII is largely reorganized under state-2 conditions) have shown the participation of monomeric antennae CP26 and CP29 ( LogFC 2.2) in state transitions ; nevertheless involvement of their counterparts in plants has never been evidenced . One interest-ing case is the Lhcb 3.1 controversy. According to our results, its di erential accumulation between stroma-lamellae and BBY in state-1 and state-2 is not negligible ( LogFC 2.3, tab.3.2). Bassi et al., 1988 analysed stroma-lamellae under state-2 conditions and concluded that the Lhcb 3 isoform was essentially absent from this thylakoid sub-compartment . Indeed, Lhcb 3 is a highly enriched in trimer (M)oderate  which are not involved in state transitions , and lacks of the N -terminal phosphorylation site. However, more recent results suggested the contrary and that Lhcb 3 might play a key role in modulating the rate of state transitions. Moreover, Galka et al., 2012  puri ed PSI-LHCI-LHCII supercomplex in maize and per-formed an electrophoresis analysis. After staining with Sypro Ruby, they were capable to identify not-negligible quantity of Lhcb 3. This can be ex-plained by the participation of a few M trimers in state transitions as was proposed before  although the presence of some Lhcb 3 isoform in other trimers than M is still debated .
In addition to this we also compared antenna repartition in mutants with respect to WT (data not shown). We observed that, when compared, WT and stn7 display an almost equivalent antenna distribution. Therefore, according to our proteomic analysis Arabidopsis WT seems to be in a state which is quite similar to that of stn7 (i.e. state-1). This seems in contrast with data reported in gure 3.3 in which intact leaves of Arabidopsis WT were found in a state more close to state-2 (i.e. pph1 mutant). Reasons for such a discrepancy can be due to the fact that we did not add any phosphatase inhibitor during the preparation and the conditions in which our fractions are puri ed essentially promote dephosphorylation of the LHCII and transition to state-1 in Arabidopsis WT plants.
In conclusion, our results about the distribution of the di erent Lhcb isoforms, partially support data previously reported and con rm that Lhcb1 and Lhcb2 are probably the two Lhcb isoforms mostly involved in state tran-sitions. Nevertheless, as shown in table 3.2 p-values do not allow to draw de nitive conclusions for most of the Lhcb isoforms.
Conclusion and perspectives
Other candidate proteins
To move a step forward, we proceeded to a preliminary data mining and we highlighted further interesting proteins. In order to analyse the data we arbitrarily chose not to consider proteins for which LogFC stn7-pph1 was lower than that of Lhcb 4.1 (the antenna isoform with the lowest LogFC stn7-pph1, see tab.3.2). Results of the di erentially abundance analysis evidenced that stn7 and pph1 have 81 proteins for which the sub-thylakoids repartition was di erent between the two genotypes. Among them, 11 proteins are statistically signi cant and are summarised in gure 3.9 together with their curated function (Tomizioli et al., 2014 ) and thy-lakoid sub-compartment distribution (LogFc). Unfortunately, when we look to the description and to the function (lanes curated protein description and curated function) there is not a clear and common link between them. One interesting protein is subunit PsaH (At3g16140) of the PSI. As previously seen in section 184.108.40.206, PsaH represents the docking site of LHCII to PSI during state transitions. In the stn7 mutant, this protein is found to be par-ticularly enriched in stroma-lamellae with respect to pph1 mutant. Given that, at the moment of the dissertation, semiquantitative proteomic data is not yet available, we can only speculate that plants devoid of stn7 might somehow accumulate more PsaH in the stroma-lamellae as direct/indirect response to the missing migration of LHCII from PSII. Moreover, among the 11 proteins in table 3.9, one other interesting candidate seems the PSII subunits PSBP1. However, the presence of a PSBP-Like protein in the list (At3g63540) might reflect a possible problem in distinguishing between true PSII subunits (PsbP) and PsbP-like subunits. These are proteins probably associated to the NDH complex (stroma-lamellae).
Figure 3.9 – List of the 11 proteins found to have a statistically diﬀerent distribution in thylakoid sub-compartments between Arabidopsis stn7 and pph1 mutants. Moreover these 11 proteins are characterised by a ΔLogFC stn7-pph1 greater than that of Lhcb1.4 (1.8). A negative value of LogFC indicates particular protein enrichment in the BBY fraction. A positive value of LogFC indicates particular protein enrichment in the stroma-lamellae fraction.
To conclude, as a whole, the data currently available indicate that around 80 chloroplast proteins are probably characterised by a trend in thylakoid repartition between stroma-lamellae and BBY during state transitions. How-ever as mentioned above, statistical analysis does not corroborate this trend for most of them. Indeed, giving the limited extent of state transition in plant (max 15-20 %) together with the background noise of the measure, the number of biological/technical replicates that we performed might not be suﬃcient to extract fully reliable information. Hence, following steps of the study should firstly focus on increasing the number of biological/technical replicates. More replicates would allow to obtain an enlarged pool of sta-tistically significant results. This will be also supported by the forthcoming generations of more-sensitive mass spectrometers.
Alternatively, a tempting hypothesis to explain the lack of clear diﬀer-ences in protein distribution between stn7 and pph1 would invoke the role of grana-margins. As mentioned in section 220.127.116.11 of this thesis (page 41), it was recently proposed an alternative mechanism to account for state transitions : instead of the sole LHCII being transferred between stroma-lamellae and BBY, there would be protein migrations in two directions: (i) movement of PSI-LHCI towards the grana by attractive electrostatic forces and/or by van der Waals forces and (ii) P-LHCII-PSII movement towards the stroma lamellae due to charge repulsion between adjacent P-LHCII proteins. The meeting point of these complexes would be the margin fraction. (see sec-tion 1.5.1 at page 38). According to this model, in fact, we would not be able to see any di erence in protein composition of our fractions between BBY and stroma-lamellae in stn7 with respect to pph1. This is explained by the fact that, the site where protein movement takes place during state transitions, is the margin fractions (and not the inner part of the grana disc i.e. BBY or the stroma-lamellae). At this purpose, we can hypothesize that analysis of our margin fractions would reveal a di erential accumulation of LHCI-PSI/LHCII-PSII when puri ed from stn7 and pph1 mutants. How-ever, we are conscious that this phenomena, even if occurring, might not be detectable in our margins fractions giving that their purity was found to be quite low. To successfully accomplish the analysis, it would be con-ceivable the development of an improved protocol to obtain highly puri ed grana-margin fractions. For this purpose successfully implementation of the yeda-press in Spinacia oleracea is already reported in literature  and would need to be adapted to our model organism Arabidopsis.
LTR: photosystem stoichiometry adjustment
Transcript levels of several protein subunits involved in photosynthesis are known to be regulated by photosynthetic redox signals (i.e. PQ). For example Brautigam et al. demonstrated that, following a change to PSII light, a clear di erence is observable in the expression of genes involved in photosynthesis or metabolism between WT and stn7 plants. Up to 85% of the genes responding to the light shift do not longer respond in stn7 mu-tant. This indicates that most of these genes are under STN7-mediated redox control. In this context, our semiquantitative approach based on the spec-tral counting, is expected to deliver signi cant insights into this phenomena. During the following steps of the research it will be interesting to compare the stoichiometry of the photosynthetic complexes between WT plants and plants devoid of STN7. It is expected that lack of the speci c sensor of the redox-signal would translate into a varied composition of the photosynthetic chain. This varied composition can comprise not only changes at the level of the whole complex stoichiometry but also at the level of the single subunit composition (i.e. LHC).
Seeking for the chloroplast-redox signal messenger toward the nu-cleus STN7 kinase is of a key importance in LTR and acts as a sensor and signal transducer . In particular, data on phosphorylation of chloro-plast protein kinases (PKs) by other kinases have been obtained from phos-phoproteomics studies , which support the existence of phosphorylation signalling cascades in the organelle. This would promote the activation/ de-activation of proteins involved in post-transcriptional events both in nucleus and chloroplast. At the moment the mechanisms by which redox signals are transduced to the nucleus are largely not understood. Nevertheless it is conceivable that an unknown protein that is phosphorylated in an STN7-dependent manner could relay the required signal for the LTR. Sensing of the redox signal by STN7, would promote the phosphorylation of this protein kinase messenger out of the chloroplast in order to reach the cellular cytosol and to phosphorylate another protein in a phosphorylation-cascade mediated mechanism ( gure 3.10).
Table of contents :
1.1 The plant cell
1.2 The Chloroplast
1.2.3 Thylakoids and their biogenesis
1.3 Oxygenic photosynthesis
1.3.1 The Z-scheme
1.4 Major photosynthetic complexes
1.4.1 Photosystem I
1.4.2 Photosystem II
1.4.3 Cytochrome b6f
1.4.4 ATP synthase
1.5 Lateral heterogeneity
1.5.1 Grana margins
1.6.1 Non-photochemical quenching
2 Deciphering thylakoid sub-compartments
2.2 Purication of the fractions
2.2.1 Intact chloroplasts purication
2.2.2 Thylakoid subfractions purication
2.2.3 Spectroscopic evaluation of the purity of the samples
2.3 Introduction to proteomics
2.3.1 Organelle purication
2.3.2 Assessment of the quality of the puried subfractions
2.3.4 Mass spectrometry-based protein quantication
2.3.5 The whole chloroplast experimental proteome
2.4 Previous proteomic investigations
2.4.1 Proteomics of the chloroplast envelope
2.4.2 Proteomics of the chloroplast stroma
2.4.3 Proteomics of the thylakoids and thylakoid lumen
2.5.5 Materials and methods
3 Looking for new potential actors in state transitions
3.2 Results and discussions
3.2.1 Measure of state transitions in WT, stn7 and pph1
3.3 Purity of the sample
3.3.1 Proteomic survey of the thylakoid subfractions from stn7 and pph1 mutants
3.4 Conclusion and perspectives
3.5 Materials and methods
4 TPK two-pore K+ channel
4.2.1 The proton motive force : pH and
4.2.2 Ion channels are involved in pmf control
4.4 Further aspects
4.4.2 K+ channels are involved in plant responses to high-light199
5 Concluding remarks