ppGpp influences protein protection, growth and photosynthesis in Phaeodactylum tricornutum

Summary Chloroplasts retain elements of a bacterial stress response pathway that is mediated by the signalling nucleotides guanosine penta‐ and tetraphosphate ((p)ppGpp). In the model flowering plant Arabidopsis, ppGpp acts as a potent regulator of plastid gene expression and influences photosynthesis, plant growth and development. However, little is known about ppGpp metabolism or its evolution in other photosynthetic eukaryotes. Here, we studied the function of ppGpp in the diatom Phaeodactylum tricornutum using transgenic lines containing an inducible system for ppGpp accumulation. We used these lines to investigate the effects of ppGpp on growth, photosynthesis, lipid metabolism and protein expression. We demonstrate that ppGpp accumulation reduces photosynthetic capacity and promotes a quiescent‐like state with reduced proliferation and ageing. Strikingly, using nontargeted proteomics, we discovered that ppGpp accumulation also leads to the coordinated upregulation of a protein protection response in multiple cellular compartments. Our findings highlight the importance of ppGpp as a fundamental regulator of chloroplast function across different domains of life, and lead to new questions about the molecular mechanisms and roles of (p)ppGpp signalling in photosynthetic eukaryotes.


Introduction
Diatoms are a group of unicellular eukaryotic photosynthetic organisms that form a major part of phytoplankton, and are responsible for up to one fifth of net global carbon fixation (Falkowski et al., 2000). Beside their ecological importance, diatoms are also studied because they have a wide range of potential applications that include drug delivery in chemotherapy, biofuels, and as environmental indicators for monitoring water quality (Levitan et al., 2014;Lavoie et al., 2017;Uthappa et al., 2018).
Diatoms appeared relatively recently in evolutionary history, around 200 million years ago (Medlin, 2016;Benoiston et al., 2017). The diatom chloroplast was acquired through complex endosymbiotic events, where it is thought that a red algal ancestor was engulfed by a eukaryote that already possessed green algal genes from a previous endosymbiosis (Dorrell et al., 2017). The acquisition of additional genes by lateral gene transfer from bacteria is also likely to have been an important driving force in the evolution of diatoms (Bowler et al., 2008). These complex origins confer a unique cellular physiology to diatoms that allows them to adapt to multiple environments (Vardi et al., 2008;Gruber & Kroth, 2017).
The diatom chloroplast differs in a number of aspects from the chloroplast of plants and green algae. At the level of membrane architecture, the diatom chloroplast is surrounded by four membranes, and the thylakoids are loosely stacked with three interconnected membranes (Flori et al., 2017). Moreover, the light-harvesting complex (LHC) of diatoms is known as the Fucoxanthin Chlorophyll Protein complex (FCP), and is composed of tetramers containing chlorophyll c and the carotenoid fucoxanthin (Lepetit et al., 2012;Roding et al., 2018;Nagao et al., 2019;Pi et al., 2019). Diatoms have other specific features such as a functional urea cycle (Allen et al., 2011), a eukaryotic Entner-Doudoroff glycolytic pathway (Fabris et al., 2012) and they lack the plastid oxidative pentose phosphate pathway found in plants (Gruber & Kroth, 2017).
The regulation of diatom metabolism is also distinct from that in plants (Jensen et al., 2017).
Like plants, diatoms possess signalling pathways with prokaryotic origins such as the bacterial histidine-kinase-based two-component systems (Bowler et al., 2008), as well as pathways of eukaryotic origins such as the Target Of Rapamycin (TOR) kinase (Prioretti et al., 2017;Prioretti et al., 2019) and the G protein-coupled receptor signalling pathway (Port et al. 2013).
Another signalling pathway that could play an important role in diatom stress acclimation is the pathway mediated by the nucleotides guanosine tetraphosphate and guanosine pentaphosphate or (p)ppGpp (Field, 2018;Avilan et al., 2019). In bacteria, ppGpp and to a lesser extent pppGpp accumulate in response to a range of different stresses, and specifically target transcription and translation to slow growth and promote stress acclimation (Hauryliuk et al., 2015;Steinchen & Bange, 2016). (p)ppGpp is also found in plants and green algae (Takahashi et al., 2004), and chloroplast-localized RelA SpoT Homolog (RSH) enzymes for (p)ppGpp synthesis are widespread among the photosynthetic eukaryotes (Atkinson et al., 2011;Ito et al., 2017;Avilan et al., 2019).
Currently, the function of ppGpp is best characterised in the flowering plant Arabidopsis, where it inhibits chloroplast gene expression, reduces chloroplast size, and reduces photosynthetic capacity (Maekawa et al., 2015;Sugliani et al., 2016;Honoki et al., 2018). While the mechanisms are still uncertain, ppGpp may act by downregulating the transcription of chloroplast encoded genes via the inhibition of chloroplastic RNA polymerases or GTP biosynthesis (Nomura et al., 2014;Yamburenko et al., 2015;Sugliani et al., 2016). Levels of ppGpp increase in response to different abiotic stresses, as well as in response to treatment with stress-associated plant hormones (Takahashi et al., 2004). Increased ppGpp levels can affect growth under nitrogen limiting conditions (Maekawa et al., 2015;Honoki et al., 2018) and plant immune signalling (Abdelkefi et al., 2018).
Genes coding for RSH enzymes are present in all fully sequenced photosynthetic eukaryotes (Field, 2018;Avilan et al., 2019). However, very little is known about the role of (p)ppGpp in algae, and in particular in algae of the red lineage. A recent study in the extremophile red alga Cyanidioschyzon merolae showed that overexpression of CmRSH4b, a functional (p)ppGpp synthetase, results in a reduction in chloroplast size and decreased chloroplast rRNA transcription, although (p)ppGpp levels were not measured (Imamura et al., 2018). The situation in diatoms, with their mixed red-green heritage and specific lifestyles, is even less clear. A recent analysis showed that the nuclear genome of the marine diatom Phaeodactylum tricornutum encodes three functional RSH enzymes from red-lineage specific clades: PtRSH1, a bifunctional (p)ppGpp synthetase/hydrolase; and PtRSH4a and PtRSH4b, which act exclusively as (p)ppGpp synthetases (Avilan et al., 2019). Here, we studied (p)ppGpp signalling in P. tricornutum by manipulating endogenous ppGpp levels using transgenic lines that express a heterologous (p)ppGpp synthetase under the control of an inducible promoter.
We found that ppGpp accumulation led to a reduction in photosynthetic capacity and an inhibition of ageing and growth. Strikingly, a proteomic analysis revealed that ppGpp accumulation also leads to the robust activation of a protein protection response involving chaperones and proteases. Our findings demonstrate that core ppGpp signalling is highly conserved across the photosynthetic eukaryotes, and that ppGpp has species specific roles that may be linked to adaptation to particular environments and life-styles.

Materials and methods
Cell culture P. tricornutum Bohlin (strain name Pt1_8.6; RCC 2967) was obtained from Roscoff Culture Collection. Cells were grown in f/2 medium (Guillard & Ryther, 1962) without silica and adjusted to pH 8. The medium contained either 0.88 mM NaNO 3 (f/2-NO 3 ) or NH 4 Cl (f/2-NH 4 ) as a nitrogen source. Cultures, in liquid medium in Erlenmeyer flask or on agar plates (2% bacto-agar), were maintained at 18 °C under continuous illumination (30 μmol photon m -2 s -1 ) with LED lamps. Samples for cell counting (Malassez counting chamber) and protein extraction were withdrawn from liquid cultures.

Plasmid construction and transformation of P. tricornutum
The vector pPha-NR carrying the nitrate reductase promoter and the ble gene for zeocin resistance (Chu et al., 2016) (GenBank accession number JN180663, kindly provided by Pr. P. Kroth), was used as backbone for all plasmid constructions. The DNA fragments used for gene fusion and cloning in the final constructs were amplified by PCR using Q5 DNA polymerase (New England Biolabs) using standard conditions. The primers are specified in Table S1. The amplified fragments were assembled in the vector, previously digested with EcoRI and HindIII, by Sequence and Ligation Independent Cloning (SLIC) method (Jeong et al., 2012).
The genes coding for a (p)ppGpp synthetase (SYN) and an inactive mutant form (D275G) of this synthetase (SYN D>G ) were obtained by PCR, using plasmids described in Sugliani et al., (2016) as templates. SYN corresponds to the (p)ppGpp synthetase Rel A from E. coli (residue 1 to 386). To target the enzymes to the chloroplast, the genes were fused by PCR to the sequence coding for the bipartite targeting sequence of the chloroplastic gamma ATP synthetase of P. tricornutum (GeneBank accession number U29898) whose ability to target proteins to the chloroplast is well characterised (Apt et al., 2002). A Kozak sequence (AAG) was included in the primer before the ATG translation start site to facilitate translation. The final constructions were used to transform cells by particle bombardment.
Transformation of P. tricornutum was performed using a Bio-Rad Biolistic PDS-1000/He particle delivery system with rupture disc of 1350 psi as previously described (Falciatore et al., 1999;Kroth, 2007). Prior to the bombardment, 1 x 10 7 cells, in the exponential growth phase, were spread in the center (5 cm diameter) of the f/2-NH4 agar plate without antibiotic, dried under sterile hood and placed under illumination for 24 h. Ten µl of M17 tungsten microcarriers (1.1 μm diameter, Bio-Rad) equivalent to 600 μg particle were coated with 1 μg of the specific plasmid constructs in the presence of CaCl 2 and spermidine (Kroth, 2007) and used to bombard the diatom cells. After 24 h illumination cells were resuspended and re-plated onto f/2-NH 4 agar plates containing 80 µg mL -1 zeocin (Invitrogen). The plates were maintained under illumination for 3 weeks and the resistant colonies were screened by PCR for the presence of SYN genes as previously described (Falciatore et al., 1999) (primers in Table S1). A cell suspension from each colony was re-plated on agar medium and the screening process was repeated. The final colonies were also screened for the presence of the proteins SYN by Western blotting. Interestingly, we obtained fewer SYN lines (n=9) than SYN D>G (n=66), suggesting that the presence of SYN negatively affects the recovery of transgenics. Clones of SYN lines were prone to losing their phenotype after several multiplications. The transformants were therefore periodically re-plated from cells stored at 4 o C and re-screened based on their growth phenotype on solid agar plates.
Multiple independent transgenics lines were used for each construction in subsequent experiments.

Nucleotide determination
GTP and ppGpp levels were determined using stable isotope labelled internal standards as described (Bartoli et al., 2019). Briefly, 8 mg (dry weight equivalent) of P. tricornutum cells were harvested from cultures 2 days after induction and nucleotides directly extracted with 3 ml 2M formic acid containing 12.5 pmoles 13 C-labelled ppGpp and 125 pmoles 13 C labelled GTP. Extracts were incubated on ice for 30 minutes, 3 ml 50 mM ammonium acetate at pH 4.5 was then added, and then samples were loaded onto pre-equilibrated Oasis WAX SPE cartridges and nucleotides eluted with a mixture of methanol/water/NH 4 OH (20:70:10). The eluates were lyophilized and resuspended in water before analysis by HPLC-MS/MS with multiple reaction monitoring. The quantification of (p)ppGpp in Figure 1C was hindered by poor detection of the 13 C-ppGpp internal standard during HPLC-MS/MS analysis. ppGpp concentrations were therefore calculated using the 13 C-GTP internal standard adjusted for the response factor to estimate recovery. Figure S1 shows absolute quantification of ppGpp based on the 13 C-ppGpp internal standard.

Growth parameters and pigment quantification
For the analysis of growth phenotypes cells were grown in 25 ml liquid f/2-NH 4 in 250 ml Erlenmeyer flasks, without agitation, in triplicate for five different transformants.
Optical density (OD) at 750 nm and number of cells were monitored every 24 hours.
Cell number showed a linear relationship with OD 750 nm up to an OD of 1. Cultures were inoculated from a preculture to an initial OD 750 nm of 0.05. After 48 hours, cells were centrifuged at 5000 g (Allegra® X-15R Centrifuge, Beckman Coulter) for 10 minutes at 16 °C and the medium was substituted with f/2-NO 3 to induce the expression of SYN or SYN D>G . For control cultures, the medium was substituted with fresh f/2-NH 4 . Pigments were extracted on ice from cell pellets (1.6 x 10 7 cells) with 1 ml 96% (v/v) ethanol, maintained in the dark for 30 minutes and centrifuged at 12.000 rpm for 10 minutes at 4°C. The spectra of the supernatant from 350 to 750 nm were recorded using a Perkin Elmer spectrophotometer (PTP-6 Peltier System). Chlorophyll a and c concentrations were calculated according to (Ritchie, 2006) and fucoxanthin concentration according to (Wang et al., 2018). For extended darkness, 24 hours after induction cells were transferred to darkness for 21 days.

Light microscopy and cell measurement
Light microscope images of live cells were taken using a MoticBA410 microscope equipped with a Moticam 1080 camera. Images were used to determine length of the individual cells using the Motic image 3 plus software after calibration of the microscope. For fluorescence microscopy, cells were incubated with the fluorescent dye AC202 as described in (Harchouni et al., 2018). Fluorescence was visualized in an epifluorescence microscope Eclipse 80i (Nikon) using the DAPI filter cube (excitation filter: 360BP40; emission filter 460BP50). Images were acquired using MS Elements imaging software (Nikon). Images were merged using ImageJ software.

Measurement of photosynthetic activity
Chlorophyll fluorescence parameters were measured using a Fluorcam FC 800-O imaging fluorometer (Photon System Instruments, Drasov, Czech Republic). Cells plated on solid medium were recovered in f2/NH 4 medium and 10 μl of this cell suspension (between 5 and 9 x 10 7 cells mL -1 ) was dotted onto f2/NO 3 or f2/NH 4 medium and grown under standard growth conditions. Before measurement, the plates were kept in dark for 20 minutes and the chlorophyll fluorescence imaged to obtain minimum fluorescence yield (F o ) and maximum fluorescence yield (F m ). PSII maximum efficiency was calculated as F v /F m = (F m -F o )/F m . To calculate relative electron transfer rate (ETR), samples were exposed to different photon flux densities (PPFD) in a stepwise fashion. Relative ETR was then calculated as the product of the photochemical Cell extract, SDS-PAGE and immunoblotting P. tricornutum cells (2 x 10 7 ) were harvested from liquid culture by centrifugation (5000 g for 15 minutes at 18°C) and resuspended in 100 μl of rupture buffer (10 mM Tris, pH 8 containing 2% n-dodecyl β-D-maltoside) and then broken by sonication using 6 pulses using an ultrasonicator (Sonics & Materials Inc, Vibracell, Bioblock, Danbury, Connecticut, USA). After incubation at 4°C for 30 minutes, the cell extract was centrifuged at 12000 g for 30 minutes at 4°C. The supernatant was submitted to SDS-PAGE and gels were stained with Coomassie blue R-250. For Western blot analysis, proteins were transferred onto nitrocellulose membrane as described (Sambrook et al., 1989). Loading control gels were run in parallel. The membrane was blocked with TBS (Tris 10 mM, pH 8, 150 mM NaCl) containing 5% nonfat milk and incubated with the specified primary antibodies in the same buffer containing 1% nonfat milk. The antibodies used were: anti-E.coli RelA (dilution 1/2000, kindly provided by M. Cashel), anti-hemagglutinin (HA) (monoclonal, dilution 1/10000, Sigma-Aldrich, catalog number:H9658, clone HA-7), anti-PsbA (polyclonal, dilution 1/10000, Agrisera, catalog number AS05 084) and anti-P. tricornutum LHCf1-LHCf11 (Juhas and Buchel, 2012) (polyclonal, dilution 1:5000, kindly provided by C. Büchel). After washing with TBS, the membranes were incubated either with horseradish peroxidase coupled antirabbit IgG (GE Healthcare, catalog number NA934V) for the polyclonal antibodies or anti-mouse IgG (Sigma-Aldrich) for the monoclonal antibodies. Immunodetection was performed using the enhanced chemiluminescence method (ECL substrates, GE Healthcare).

Lipid and chrysolaminarin determination
The fluorescent probe Nile red was used to detect neutral lipids by flow cytometry as specified in (Prioretti et al., 2017). The cells at different growth phases were fixed with 2% (v/v) glutaraldehyde (Prioretti et al., 2017) and 1 ml of cell suspension was incubated with 2 µl of Nile red solution (0.25 mg ml -1 ) for 5 minutes before the analysis with a bench flow cytometer (BD Accuri C6, BD Biosciences). To determine Nile red fluorescence, the trigger signal was set to FL2 fluorescence and combined with the side scatter (SSC) signal. Ten thousand events were analysed in each case.
For lipid analysis, cells of P. tricornutum (50 x 10 6 cells) were harvested by centrifugation (5000 g, for 10 minutes) and the pellet was resuspended with 1 mL of hot isopropanol (preheated at 85°C) and incubated at this temperature for 10 minutes to quench endogenous lipases. Detailed lipid extraction protocol can be found in (Legeret et al., 2016). Extracted total lipids were then dried under a stream of N 2 , and then dissolved in a solvent mixture of chloroform: methanol (2:1 vol/vol) for quantification by Thin Layer Chromatography (TLC). Two types of TLC were run: one for triacylglycerol (TAG) quantification, and a second for polar membrane lipid analysis.
Detailed TLC procedures, standard used and quantifications have been described, together with total fatty acid (FA) content and composition analysis in (Siaut et al., 2011).
Extraction and determination of the β-1,3-glucan chrysolaminarin was performed following the method of (Granum & Myklestad, 2002) from pellets containing 2 x 10 7 cells. A calibration curve was obtained using glucose as a standard.

Mass spectrometry analysis
Proteins extracts (50 µg protein) from 3 SYN lines, 2 SYN D>G lines and 1 WT, made on equal number of cells, were loaded on a gel (SDS-PAGE) and the band from the stacking gel corresponding to the total proteins was excised and submitted to in gel trypsin digestion for proteomic analysis as previously described in (Santin et al., 2018) with minor modifications. In parallel, total proteins were also separated to analyse the protein profile in the gel (Fig. S1). Tryptic peptide samples were quantified by a colorimetric peptide assay (Pierce, Thermo Fisher) and aliquots of 200 ng were injected  et al., 2016). LFQ normalized intensities were uploaded from the proteinGroups.txt and converted to base 2 logarithms to obtain a normal distribution. Quantifiable proteins were defined as those detected in at least 70% of samples in at least one condition. Missing values were replaced using data imputation by randomly selecting from a normal distribution centred on the lower edge of the intensity values. To determine whether a given detected protein showed differential accumulation, a two-sample t-test was applied using a permutation based FDR-set at a conservative threshold of 0.0001 (250 permutations), and the p value was adjusted using a scaling factor s0 set to 1. The results are illustrated in a volcano plot (Fig. S2), and the list of identified proteins and differentially accumulating proteins in SYN versus the control lines is available in Table S2.

An inducible (p)ppGpp synthetase efficiently increases ppGpp levels in P.
tricornutum As a first step towards understanding the function of (p)ppGpp in diatoms, we developed a strategy similar to that reported by (Sugliani et al., 2016) to modulate (p)ppGpp levels in the chloroplast. For this purpose, we transformed P. tricornutum with a synthetic gene encoding a chloroplast-targeted fragment of a bacterial (p)ppGpp synthetase (SYN) under the control of an NO 3 inducible promoter (Fig. 1a). Control transgenic lines were also generated that express catalytically inactive forms of the same enzyme, SYN D>G . Positive transformants were identified by PCR and analysed to confirm protein expression following induction by transfer from NH 4 -containing medium to NO 3 -containing medium (Fig. 1b). Importantly, induction of SYN lines also caused an increase in ppGpp levels to 1.30 ± 0.34 nmol mg -1 dry cell weight (n=5 independent biological replicates) (Fig. 1c). pppGpp was not detected, suggesting that SYN preferentially acts as a ppGpp synthetase, and/or that pppGpp can be converted to ppGpp by endogenous enzymes as in bacteria. No ppGpp was detected in SYN D>G lines or the wild type under inducing conditions. The absence of basal levels of ppGpp is likely to be a consequence of the nutrient shift used for induction because we were able to detect low quantities of ppGpp in wild type cells grown in standard media in the light ( Fig. S1). Interestingly, we also observed that incubation of cells in the dark for prolonged periods led to the accumulation of ppGpp. We simultaneously quantified GTP levels and found that concentrations were similar to the wild-type control regardless of ppGpp levels ( Fig. 1c, Fig. S1).

ppGpp accumulation strongly inhibits cell division and has a major effect on photosynthesis
We first examined the effects of ppGpp accumulation on growth and proliferation. Five independent SYN lines showed a severe reduction in proliferation upon induction in liquid culture ( Fig. 2a and Fig. S2). In contrast, induction did not affect the proliferation of control SYN D>G lines or the WT. The reduced proliferation of SYN lines was also clearly visible on agar plates following induction (Fig. 2b). Observation by light microscopy showed that SYN cells were significantly longer than those of the controls ( Fig. 2c).
We then analysed pigment content at different times after induction (Fig. 3, Fig.   S3). In the WT, the level of chlorophyll a (Chla) and fucoxanthin decreases as the cells enter stationary phase, five days post induction. To our surprise, we found that this drop in pigment content does not occur in induced SYN lines: five days after induction levels of Chla and fucoxanthin are significantly higher in SYN than in SYN D>G or WT (Fig.   3). In contrast, the drop in Chlc levels was similar in all lines, leading to a higher Chla:Chlc ratio in the SYN lines (Fig. 3). These results suggest that ppGpp accumulation inhibits cell division while stabilizing Chla and fucoxanthin levels in the chloroplast.
In Arabidopsis, ppGpp accumulation has major effects on photosynthetic activity (Maekawa et al., 2015;Sugliani et al., 2016). We therefore examined the maximal efficiency of photosystem II (F v /F m ) ( Fig. 4a-c). F v /F m decreased rapidly in induced SYN transformants following induction, reaching a minimum of 0.12 ± 0.004 (SE) two days post induction (Fig. 4a,b and Fig. S4). The F v /F m of the SYN D>G lines was 0.6 at the same timepoint. We performed immunoblots to determine whether the decrease in F v /F m in induced SYN lines was due to changes in PSII architecture. Indeed, we found that the levels of the chloroplast-encoded protein D1, that constitutes the core of the PSII reaction centre, decreased in SYN lines (Fig. 4d). In contrast, the levels of the light-harvesting protein, LHCf, that forms a major part of the FCP antenna, remained relatively constant (Fig. 4d). These results indicate that the composition of PSII undergoes major changes in response to ppGpp accumulation. In agreement with the reduced PSII efficiency, the relative electron transfer rate (rETR) of the entire photosynthetic chain was also lower in SYN lines than in the WT or SYN D>G controls

ppGpp affects the accumulation and distribution of lipids and other reserve compounds
Using light microscopy, we observed a prominent spot near, but clearly separated from, the chloroplast in all SYN lines at two days post induction (Fig. 5a). This spot was absent in non-induced SYN (Fig. 5b), and was not detected in SYN D>G lines regardless of the condition (Fig. 5c, Fig. 5d). The prominent spot that was clearly visible under light microscopy was strongly stained by the neutral lipid specific fluorophore AC202, indicating that it is a lipid droplet (LD) ( To further characterize the prominent LD in SYN cells, we analysed neutral lipid content at different stages of the growth using flow cytometry and Nile red staining and then confirmed by quantification using TLC (Fig. 6a). A strong correlation between Nile red fluorescence and TAG content is established in diatoms (Greenspan et al., 1985;Prioretti et al., 2017). Nile red fluorescence, determined using flow cytometry, increased with the age of the culture in WT and SYN D>G cells, but not in SYN cells, in agreement with our microscopic observations ( Fig. 5g, h). However, flow cytometry was not sensitive enough to detect differences in lipid content at two days post induction. Therefore, we directly quantified TAG by extraction of total lipids followed by TLC quantification (Fig. 6b, c). Two days post induction, the amount of TAG per cell was almost two-fold higher in the SYN lines than in the controls (P =0.213) (Fig.   6b). Although not significant, this finding tends to support our microscopic observation of higher neutral lipid levels in SYN lines two days after induction (Fig. 5a). However, five days post induction, SYN lines contained much lower (6-fold) levels of neutral lipids than the control lines, in agreement with the flow cytometry results (Fig. 6c) and microscopy images (Fig. 5g, h). Chrysolaminarin (β-1,3-glucan), the main carbohydrate storage compound in diatoms which is located in the vacuole (Suzuki & Suzuki, 2013), was also less abundant in induced SYN lines than in the controls five days post induction (Fig. S6).
To investigate the impact of altered ppGpp levels on membrane lipids, we also  .

Proteome analysis reveals that ppGpp accumulation leads to the coordinated upregulation of the protein protection response
The analysis of the protein profiles of induced SYN, SYN D>G and WT by SDS-PAGE revealed major changes in total proteins in SYN lines, including the appearance of additional protein bands (Fig. S7). In order to understand the extent of these changes, we performed an untargeted proteomic analysis to compare protein accumulation between induced SYN lines and induced SYN D>G lines two days after induction. Of the 1046 proteins that were identified, 145 proteins showed a significant difference in accumulation in SYN compared to the control lines (Table S1). The expression of known housekeeping proteins (Siaut et al., 2007) was not affected by ppGpp accumulation, with the exception of actin 12 (Table S1). The major groups of proteins showing differential accumulation in SYN are listed in Table I. Unexpectedly, many chaperones and co-chaperones showed higher levels of expression in SYN, and the chaperone HSP20 (accession number B7G195) showed the greatest increase. The differentially expressed chaperones in SYN are encoded by nuclear genes and have different predicted cellular localizations. Four chaperones including HSP40 and one isoform of HSP90 contain the ASAFAP hexapeptide motif that predicts likely chloroplast localization (Gruber et al., 2015). Other chaperones have signal peptides for the secretory pathway without an ASAFAP motif, suggesting localization in the endoplasmic reticulum. Some chaperones such as the Lon protease, that functions as both a chaperone and protease, are predicted to be mitochondrial.
We found that many transporters and enzymes involved in amino acid metabolism were among those proteins showing the greatest differential accumulation in the SYN lines (Table 1). Notably, the synthesis of glutamate appeared to be upregulated due to greater accumulation of the chloroplastic glutamate synthase and glutamate dehydrogenase, and lower levels of glutamine synthase. Glutamate is involved in nitrogen assimilation, and is also a precursor for chlorophyll biosynthesis.
Interestingly, the chloroplast protoporphyrin IX magnesium chelatase, a key enzyme of the chlorophyll biosynthetic pathway, was also more abundant. Proteins of lipid metabolism were also affected. The levels of two enzymes involved in fatty acid degradation increased: the acyl-CoA dehydrogenase and a homologue of enoyl-CoA hydratase, a key enzyme of fatty acid beta-oxidation. Phage shock protein A, a protein that is involved in managing extra-cytoplasmic stress responses in bacteria (Joly et al., 2010), and glutathione S-transferase, an enzyme important in the redox homeostasis of the cell were also more abundant.
Many chloroplast-encoded proteins with links to photosynthesis and chloroplast translation were less abundant in SYN. Seven proteins from the PSII complex accumulated to lower levels, including the PsbA / D1 protein, in agreement with our immunoblotting experiments (Fig. 4d). Some proteins from PSI and the Calvin cycle were also less abundant, as well as a bicarbonate transporter and two carbonic anhydrase proteins that were just outside the selection cut-off (Table S1), together indicating that photosynthetic capacity is diminished at several different levels. The photosynthetic machinery requires large amounts of iron, and consistent with the downregulation of photosynthesis we also observed the downregulation of two ironstarvation induced proteins. Interestingly, two light harvesting antenna proteins were also less abundant, Lhcf15 and Lhcx2. Therefore, although our immunoblotting experiment showed that Lhcf1-11 remained unchanged (Fig. 4d), SYN induction also appears to cause changes in the accumulation of specific antenna isoforms. Also striking was the reduced levels of twelve chloroplast-encoded ribosomal proteins in SYN cells, suggesting that ppGpp accumulation has a major effect on chloroplast translation capacity. The ribulose-1,5-bisphosphate carboxylase/oxygenase (Rubisco) remained unaltered in our proteomic analysis, as was also directly visible on Coomassie stained protein gels (Fig. S7).

Discussion
In the present work, we studied the function of (p)ppGpp using transgenic P.
tricornutum SYN lines containing an inducible system for the expression of a bacterial (p)ppGpp synthetase (Fig. 1a). We demonstrate that increased ppGpp levels in SYN Iron starvation induced protein (ISIP2B) -5.27 SP C: Chloroplast; C*chloroplast-encoded; M: mitochondria; SP: presence of a signal peptide lines alter the architecture of the photosynthetic machinery and reduce photosynthetic efficiency (Fig. 4). In addition, ppGpp accumulation dramatically decreases the growth rate (Fig. 2), stabilizes the level of chlorophyll and fucoxanthin (Fig. 4), and affects accumulation of the reserve molecules TAG and chrysolaminarin (Fig. 5-7, Fig. S6).
Finally, ppGpp accumulation causes a striking increase in the levels of chaperones and other stress-related proteins in multiple cellular compartments (Table 1, Fig. 8).
The independent SYN lines produced different quantities of SYN and accumulated different quantities of ppGpp (Fig. 1). The differences we observed in SYN protein accumulation are likely due to differences in the copy number or the insertion site(s) of the exogenous DNA as previously shown (Falciatore et al., 1999).
The average levels of ppGpp reached in induced SYN lines (1.3 nmol mg -1 dry cell weight) are similar to those reached in E. coli (1.2 to 3 nmol mg -1 dry cell weight) during the stringent response (Riesenberg, 1985;Rodionov & Ishiguro, 1995). If we assume that total GTP content is similar between species, then these levels of ppGpp (representing 600% of total GTP) are considerably higher than those observed in Arabidopsis overexpressing a similar (p)ppGpp synthetase (9% of total GTP) (Sugliani et al., 2016). This may be due to differences between unicellular and multicellular organisms, or the capacity of the cell to control ppGpp levels. Interestingly, the levels of We also detected ppGpp in WT cells and discovered conditions that cause ppGpp levels to increase (Fig. S1). Low ppGpp levels could be detected in WT cells under normal growth conditions. These ppGpp levels correspond to 0.25% of total GTP, which is similar to Arabidopsis where ppGpp accumulates to 0.36% of total GTP under normal growth conditions (Bartoli et al., 2019). While nitrogen deprivation did not affect ppGpp levels, extended dark treatment caused a 4-fold increase in WT and SYN control lines (Fig. S1). Dark treatment also causes an increase in ppGpp levels in   In plants, ppGpp accumulation inhibits photosynthesis, in particular causing a large decrease in the abundance of Rubisco, and in the ratio of PSII reaction centre (RC) to PSII LHC (Maekawa et al., 2015;Sugliani et al., 2016). Similar results were obtained in this study (Fig. 4). We found that photosynthesis was strongly inhibited in induced SYN lines, and that this was accompanied by a large drop in the PSII RC to FCP antenna ratio and no obvious change in Rubisco levels (Fig. 4). Notably, these changes occur despite the significant differences in the structure and organisation of the diatom FCP antenna system compared to the equivalent light harvesting system in plants (Nagao et al., 2019;Pi et al., 2019). The overall rate of photosynthesis was also reduced by ppGpp accumulation (Fig. 4). This is consistent with reduced protein abundance across all parts of the photosynthetic machinery including PSII, PSI, the Calvin cycle and the bicarbonate transport through Solute Carrier family 4 transporter, SLC4 (Nakajima et al., 2013), as well as two carbonic anhydrases involved in the CO 2 concentrating mechanism (Table I, Table S1). PSI was less widely affected than PSII, with a particularly strong reduction in levels of the PSI acceptor flavodoxin, while in the We also found evidence that ppGpp accumulation may slow cellular ageing in P.
tricornutum by restricting growth, preventing chloroplast senescence, and blocking reserve compound accumulation. ppGpp accumulation caused a striking and sustained inhibition of growth ( Fig. 2a) that is considerably stronger than that observed in plants (Maekawa et al., 2015;Sugliani et al., 2016). The arrest in growth was accompanied by the appearance of a single prominent LD, together with a slight increase in TAG (Fig. 5-6). The appearance of the prominent LD could be a consequence of a metabolic overshoot following growth arrest where the CO 2 fixation rate exceeds requirements and the resulting carbon is diverted into the synthesis of reserve compounds. In the stationary phase, nutrients become limiting, and in wild type cells, this induces the synthesis of storage compounds and the degradation of photosynthetic pigments.
However, we found that ppGpp over-accumulation prevented the degradation of chlorophyll and carotenoids (Fig. 3), and even led to increased levels of enzymes for chlorophyll biosynthesis (Table I). The storage compounds TAG and chrysolaminarin also did not increase in lines over-accumulating ppGpp, and we additionally observed higher levels of MGDG and SQDG, consistent with a more stable chloroplast (Fig. 5-6).
Together these results suggest that ppGpp over-accumulation prevents cellular ageing and promotes a quiescent-like state. The anti-ageing effect of ppGpp in in P.
tricornutum contrasts with the situation in plants where ppGpp accumulation leads to reduced chlorophyll levels, reduced chloroplast size and accelerated senescence (Maekawa et al., 2015;Sugliani et al., 2016). Together these results point to fundamental differences in (p)ppGpp signalling between multicellular and unicellular organisms (i.e. Arabidopsis versus P. tricornutum).
One of the most striking consequences of ppGpp accumulation in P. tricornutum was the strong induction of a wide range of chaperones and proteases (Table 1).
Together these proteins can play an important role in protein protection by preventing wrote the manuscript. All authors provided critical feedback and helped shape the research, analysis and manuscript.     The scale bar corresponds to 10 μm length.

Figure Legends
: c  i  a  t  o  r  e  A  ,  C  a  s  o  t  t  i  R  ,  L  e  b  l  a  n  c  C  ,  A  b  r  e  s  c  i  a  C  ,  B  o  w  l  e  r  C  .  1  9  9  9  .   T  r  a  n  s  f  o  r  m  a  t  i  o  n  o  f  N  o  n  s  e  l  e  c  t  a  b  l  e  R  e  p  o  r  t  e  r  G  e  n  e  s  i  n  M  a  r  i  n  e  D  i  a  t  o  m  s  .  M  a  r  B  i  o  t  e  c  h  n  o  l  (  N  Y  )   1   (  3  )  :  2  3  9  -2  5  1  .   F  a  l  k  o  w  s  k  i  P  ,  S  c  h  o  l  e  s  R  J  ,  B  o  y  l  e  E  ,  C  a  n  a  d  e  l  l  J  ,  C  a  n  f  i  e  l  d  D  ,  E  l  s  e  r  J  ,  G  r  u  b  e  r  N  ,  H  i  b  b  a  r  d  K  ,  H  o  g  b  e  r  g  P  ,  L  i  n  d  e  r  S  ,  e  t  a  l  .  2  0  0  0  .   T  h  e  g  l  o  b  a  l  c  a  r  b  o  n  c  y  c  l  e  :  a  t  e  s  t  o  f  o  u  r  k  n  o  w  l  e  d  g  e  o  f  e  a  r  t  h  a  s  a  s  y  s  t  e Z  h  a  o  S  ,  W  a  n  g  W  ,  L  i  u  D  ,  X  u  C  ,  H  a  n  G  ,  K  u  a  n  g  T  ,  S  u  i  S  F  ,  S  h  e  n  J  R  .  2  0  1  9  .   T  h  e  p  i  g  m  e  n  t  -p  r  o  t  e  i  n  n  e  t  w  o  r  k  o  f  a  d  i  a  t  o  m  p  h  o  t  o  s  y  s  t  e  m  I  I  -l  i  g  h  t  -h  a  r  v  e  s  t  i  n  g  a  n  t  e  n  n  a  s  u  p  e  r  c  o  m  p  l  e  x  .  S  c  i  e  n  c  e   3  6  5   (  6  4  5  2  )  .   P  r  i  o  r  e  t  t  i  L  ,  A  v  i  l  a  n  L  ,  C  a  r  r  i  e  r  e  F  ,  M  o  n  t  a  n  é  M  ,  F  i  e  l  d  B  ,  G  r  e  g  o  r  i  G  ,  M  e  n  a  n  d  B  ,  G  o  n  t  e  r  o  B  .  2  0  1  7  .   T  h  e  i  n  h  i  b  i  t  i  o  n  o  f  T  O  R  i  n  t  h  e  m  o  d  e  l  d  i  a  t  o  m  P  h  a  e  o  d  a  c  t  y  l  u  m  t  r  i  c  o  r  n  u  t  u  m  p  r  o  m  o  t  e  s  a  g  e T  y  a  n  o  v  a  S  ,  T  e  m  u  T  ,  S  i  n  i  t  c  y  n  P  ,  C  a  r  l  s  o  n  A  ,  H  e  i  n  M  Y  ,  G  e  i  g  e  r  T  ,  M  a  n  n  M  ,  C  o  x  J  .  2  0  1  6  .   T  h  e  P  e  r  s  e  u  s  c  o  m  p  u  t  a  t  i  o  n  a  l  p  l  a  t  f  o  r  m  f  o  r  c  o  m  p  r  e  h  e  n  s  i  v  e  a  n  a  l  y  s  i  s  o  f  (  p  r  o  t  e  )  o  m  i  c  s  d  a  t  a  .  N  a  t  u  r  e  M  e  t  h  o  d  s   1  3   :  7  3  1  .   U  t  h  a  p  p  a  U  T  ,  B  r  a  h  m  k  h  a  t  r  i  V  ,  S  r  i  r  a  m  G  ,  J  u  n  g  H  Y  ,  Y  u  J  ,  K  u  r  k  u  r  i  N  ,  A  m  i  n  a  b  h  a  v  i  T  M  ,  A  l  t  a  l  h  i  T  ,  N  e  e  l  g  u  n  d  G  M  ,  K  u  r  k  u  r  i  M  D  .  2  0  1  8  .   N  a  t  u  r  e  e  n  g  i  n  e  e  r  e  d  d  i  a  t  o  m  b  i  o  s  i  l  i  c  a  a  s  d  r  u  g  d  e  l i