Two distinct phases of chloroplast biogenesis during de-etiolation in Arabidopsis thaliana

Light triggers chloroplast differentiation whereby the etioplast transforms into a photosynthesizing chloroplast and the thylakoid rapidly emerges. However, the sequence of events during chloroplast differentiation remains poorly understood. Using Serial Block Face Scanning Electron Microscopy (SBF-SEM), we generated a series of chloroplast 3D reconstructions during differentiation, revealing chloroplast number and volume and the extent of envelope and thylakoid membrane surfaces. Furthermore, we used quantitative lipid and whole proteome data to complement the (ultra)structural data, providing a time-resolved, multi-dimensional description of chloroplast differentiation. This showed two distinct phases of chloroplast biogenesis: an initial photosynthesis-enabling ‘Structure Establishment Phase’ followed by a ‘Chloroplast Proliferation Phase’ during cell expansion. Moreover, these data detail thylakoid membrane expansion during de-etiolation at the seedling level and the relative contribution and differential regulation of proteins and lipids at each developmental stage. Altogether, we establish a roadmap for chloroplast differentiation, a critical process for plant photoautotrophic growth and survival.


186
In angiosperms, chlorophyll synthesis arrests in the dark but starts immediately upon seedling

209
After 24 h of illumination (T24), the density of lamellae per chloroplast was higher than that at 210 8 T4 due to an increase in lamellar length and number. Appressed regions corresponding to 211 developing grana stacks also appeared by T24 ( Figure 2C and G). These early grana stacks 212 consisted of 2-6 lamellae with a thickness of 13 nm each (Figure 2-figure supplement 1). In 213 addition, starch granules were present at T24, supporting the notion that these chloroplasts 214 are photosynthetically functional and able to assimilate carbon dioxide (CO2). At T96, thylakoid 215 membrane organisation was visually similar to that at T24, but with more layers per grana (up 216 to 10 lamellae per grana; Figures 3D and H). In addition, singular lamella thickness at T96 217 increased by 2-3 nm compared to that at T24 (Figure 2-figure supplement 1). The major 218 differences observed between T24 and T96 were increases in starch granule size and number 219 and overall chloroplast size. Etioplast average length (estimated by measuring the maximum 220 distance on individual slices) was 2 µm (± 0.9, n=10) in the dark (T0), whereas chloroplast 221 average length was 6 µm (± 1.62, n=10) at T96 (Table1). Collectively, these data show that     Table 1). The 248 total chloroplast volume increased about 11-fold from T4 (9.4 µm 3 ) to T96 (112.14 µm 3 ) ( Table   249 1). In parallel, the thylakoid surface area increased about 30-fold reaching 2,086 (± 393) µm 2 250 per chloroplast at T96 ( Figure 4A and Table 1). The surface area increased drastically between 251 T4 and T24 (about 22-fold) and much less (about 1.4-fold) between T24 and T96. Accordingly, 252 quantification of the envelope surface area indicated that the ratio of the thylakoid to envelope 253 surface area increased drastically from T4 to T24, but decreased slightly between T24 and 254 T96 (Table 1).

255
Our observations indicated that chloroplast development during the first 96 hours of de-256 etiolation could be separated into two phases: a first phase reflected by qualitative changes 257 (i.e. structure establishment and reorganisation of the thylakoid network architecture) and a 258 second phase (starting before T24) during which thylakoid surface increased due to the 259 expansion and stacking of lamellae. We further analysed these temporal processes at the 260 molecular level focusing on proteins and lipids that constitute the thylakoid membrane.

262 263
Dynamics of plastid proteins related to thylakoid biogenesis 264

265
We analysed the full proteome to reveal the dynamics of protein accumulation during de-266 etiolation. Total proteins were prepared from 3-day-old etiolated seedlings exposed to light for 267 0-96 h (eight time points; Figure 1A) and quantified by label-free shot-gun mass spectrometry.

268
For relative quantification of protein abundances between different samples, peptide ion 269 abundances were normalized to total protein (see Materials and Methods). We considered 270 further only those proteins that were identified with a minimum of two different peptides (with 271 at least one being unique; see Methods for information on protein grouping), resulting in the 272 robust identification and quantification of more than 5,000 proteins. Fold changes of protein 273 abundances between two time points were regarded as significant if their adjusted p-value (i.e.

275
The first 12 h of illumination (T12) saw very few significant changes in protein abundance

287
To monitor the dynamics of the plastidial proteome, we selected proteins predicted to localize showed a categorization into six main clusters. Cluster 1 (purple) contained proteins whose 292 relative amounts decreased during de-etiolation. Clusters 2, 5, and 6 (pink, light green, and 293 dark green, respectively) contained proteins whose relative amounts increased during de-

304
To analyse the dynamics of proteins related to thylakoid biogenesis, we selected specific 305 proteins and represented their pattern of accumulation during de-etiolation ( Figure 5). We 306 included proteins constituting protein complexes located in thylakoids (complexes constituting 307 the electron transport chain and the ATP synthase complex) and proteins involved in 308 chloroplast lipid metabolism, chlorophyll synthesis, and protein import into the chloroplast. In  11 subsequently decreasing at the protein level upon activation and degradation following light 320 exposure (Blomqvist et al., 2008;Runge et al., 1996;Von Wettstein et al., 1995). In agreement, 321 illumination resulted in increased amounts of all detected proteins of the chlorophyll 322 biosynthesis pathway, except PORA, which clearly decreased and was separated from other 323 chlorophyll-related proteins ( Figure 5C). We also selected proteins involved in protein import 324 in chloroplasts, focusing on the TOC-TIC machinery ( Figure 5D) that is the major route for 325 plastid protein import and essential for chloroplast biogenesis (Kessler and Schnell, 2006).

326
Past studies identified several TOC preprotein receptors that are proposed to display 327 differential specificities for preprotein classes (Bauer et al., 2000;Bischof et al., 2011). The

331
Accordingly, the TOC receptors TOC120 and TOC132, which are important for the import of        We observed a massive increase in the accumulation of photosynthesis-related proteins and 389 galactolipids between T24 and T96, corresponding to FC>2 in the levels of all major chloroplast 390 proteins and lipids (Figures 6 and 7). Intriguingly, the total thylakoid surface per chloroplast 391 increased by only 41 % between these two time points ( Figure 4A and Table 1). We reasoned 392 that the increase in chloroplast proteins and lipids between T24 and T96 could be explained 398 (26 ± 6); however, in parallel with cell expansion ( Figure 8A and B), chloroplast number 399 increased sharply (4-fold increase) between T24 (26 ± 6) and T96 (112 ± 29), indicating that  the portion of lipids incorporated into the envelope rather than present in the thylakoids (Table   434 1, Table 2

436
corresponding to stroma-exposed surface) were retrieved from the literature (Table 3).

438
To quantify the surface area occupied by the galactolipids and photosynthetic complexes in 439 thylakoids per seedling, the number of molecules per seedling of galactolipids was multiplied 440 by the corresponding molecular surface area, whereas the number of molecules per seedling 441 of PsbA, PetC, and PsaC (subunits of PSII, Cyt b6f, and PSI, respectively) were multiplied by 442 the surface area of the corresponding complex (see Table 3).

443
We    Table 1). As shown in Figure 9, the two approaches showed very similar total thylakoid 483 surface area per seedling at T4 and T24 and differences in this parameter by T96.   Table 1

802
where a spectrum report was created using a false discovery rate (FDR) of 10% and 0.5% at 803 the protein and peptide level, respectively, and a minimum of one identified peptide per protein.

804
After loading the spectrum report into ProgenesisQI, samples were normalized using the                   ' red line at T0, T4, T8, T12, T24, T48, T72, and T96, and   999   Table 1  Error bars indicate ± SD (n=3). The total thylakoid surface indicated in A corresponds to the thylakoid surface exposed to the stroma, calculated in Amira software, in addition to the percentage of the grana surface (%Gs) calculated as described in Figure 3-figure supplement 1.
AT5G08280 HEMC Error bars indicate ± SD (n=3). The total thylakoid surface indicated in A corresponds to the thylakoid surface exposed to the stroma, calculated in Amira software, in addition to the percentage of the grana surface (%Gs) calculated as described in Figure 3-figure supplement 1.  nmol/seedling (T96)    Thylakoid surface per seedling was estimated using quantitative data from 3View analysis ('MORPHO' black dots at T4, T24, and T96; and see Figure  4 and Table 1) and model generated using the quantitative data from proteomics and lipidomics ('MODEL' red line at T0, T4, T8, T12, T24, T48, T72, and T96, and Table 1). Further details are provided in Figure 9-figure supplement 1 and 2.  The 'Structure Establishment Phase' is correlated with disassembly of the PLB and gradual formation of the thylakoid membrane as well as an initial increase of eukaryotic (after 8 h) and prokaryotic (after 24 h) galactolipids and photosynthesis-related proteins (PSII subunits at 4 h, PSI and cyt b6f at 12 h). The subsequent 'Chloroplast Proliferation Phase' is associated with an increase in chloroplast number in concomitance with cell expansion, a linear increase of prokaryotic and eukaryotic galactolipids and photosynthesis-related proteins, and increased grana stacking. The red curve (retrieved from the Figure 9) shows thylakoid surface/seedling dynamics during the de-etiolation process.