Multiple strategies for heat adaptation in rice endosperms revealed by on-site cell-specific analysis

Plant cells have multiple strategies to adapt to environmental stresses. Rice endosperms form chalkiness in a part of the tissue under heat conditions during the grain-filling stage, although nitrogen supply reduces chalky rice. Air spaces formed in the cells cause an irregular light reflection and create chalkiness, yet what exactly occurs remains unclear at cell level. Through on-site cell-specific analysis, we show that heat-treated cells adjust osmotically and retard protein synthesis to preserve protein storage vacuoles in the cytosol, resulting in air space formation. Application of nitrogen enhances heat tolerance to sustain protein body and amyloplast development during strong osmotic adjustment, which diminishes air spaces to avoid chalkiness. Hence, we conclude that rice endosperm cells could alter organelle compartments spatially during the heat adaptation, depending on the available nitrogen level. Our findings provide new insight into the cellular mechanism of rice chalky formation as a strategy for heat acclimation.

remarkably (see deep-colored bars in Figure 4E). In N+34°C treatment, the area of CysR10P 202 in PBI was similar to that in 34°C and prone to be higher than 26°C treatment. Consequently, 203 the area of CysR10P (number x area) in the chalky cells increased due to N application, 204 compared to that in 34°C treatment (Supplementary file 3). There was little effect on the 205 accumulation of glutelin precursors (pro-glutelin) and acidic ()-glutelin to 34°C, although the 206 content of CysR16P, Cys-poor 13-kDa prolamin (CysP13P), -globulin, and basic ()-glutelin 207 decreased in 34°C treatment ( Figure 4G). When N was supplied to the soil prior to 34°C 208 treatment, the content of -glutelin and CysR16P increased to the level of 26°C treatment 209 and greater than that of 34°C treatment ( Figure 4G). The accumulation of CysR10P was 210 specifically observed in 34°C treatments ( Figure 4G). Analysis of the time course of changes 211 in PSV and PBII (stored in PSV) volumes in the chalky zone showed that the expansion of 212 PSVs in 34°C treatment progressively occurred, reaching 26 fL at maturation, 2.6-fold larger 213 than that in 26°C treatment ( Figure 5A). PB accumulation rate retarded in 34°C treatment, 214 compared to that in 26°C treatment ( Figure 5B). Additionally, in 26°C and N+34°C treatments 215 increase in heat-induced chalky rice has been a serious concern in global rice production 224 under the advancing global warming. N supply decreases heat-induced chalky formation 225 (Perez et al., 1996;Wakamatsu et al., 2008), although the underlying mechanisms of heat-226 induced chalky formation and the N-enhanced adaptation have never been systematically 227 examined mostly due to the technical difficulties at the cellular level. As pointed previously 228 (Hakata et al., 2017), another difficulty for studying heat-related damages in crop plants 229 would be due to the low reproducibility of high-temperature treatment under field conditions, 230 which implies the need of environmental controls. We have developed a new on-site cell 231 metabolomics performable in the controlled environments to make such a cellular analysis 232 possible (Supplementary file 1). By using this method, we carried out a cell-specific analysis 233 in the putative chalky zone in the attached kernels grown under high-temperature 234 environments, in a combination with a transmission electron microscopy. It has been 235 uncovered that the preservation of relatively large PSVs among loosely-packed starch 236 granules in the cytosol was attributed to the air spaces formed in the dorsal chalky zone in 237 the kernels. Although 34°C-treated cells disturbed protein synthesis, N supply enhanced the 238 protein synthesis rate to sustain normal development of amyloplast and PBs in the zone, 239 leading to the reduction in air spaces under heat conditions. Therefore, we conclude that rice 240 endosperm cells could alter vacuolar morphology by regulating vacuolar trafficking and 241 protein processing as a heat stress adaptation, depending on the available nitrogen level. 242 We also propose that rice chalky formation is a form of heat acclimation. 243

244
Establishing an on-site cell-specific analysis to investigate heat-induced damages in 245

rice plants 246
Recently, single cell metabolomics has been extended mainly due to the introduction of 247 Orbitrap mass analyzer, and this approach has been applied to a number of biological studies 248 including plant cell analysis (Gholipour et al., 2013;Fujii et al., 2015). A cell pressure probe, 249 originally invented by Steudle's group (Hüsken et al., 1978) and long-used to measure the 250 cellular water status in plant sciences, has been adopted as a picolitre pipette to establish a 251 new type of in-situ analytical method by combining with an Orbitrap mass spectrometer in the analytical method have been improved by introducing an internal electrode in the capillary 254 holder and adopting a mixture of an ionic solution and silicone oil in the quartz capillary 255 (Nakashima et al., 2016). This method, called 'picolitre pressure-probe-electrospray-256 ionization mass spectrometry (picoPPESI-MS)', is performable in intact plants. This type of 257 cell metabolomics appears to be a powerful tool. However, to the best of our knowledge, 258 most existing cell metabolomics, including this method, has been confined to the laboratory 259 use to study cellular metabolisms at room temperature. To date, no attempts have been made 260 on further improvement of the methods to investigate metabolic responses to some are known to be localized (Ellis et al., 1987;Hoshikawa, 1989) (Figure 3). As discussed below, 274 it is noteworthy that our cell-specific analysis clarified that the fate of the endosperm cells 275 was determined, in which there were no obvious morphological treatment differences ( Figure   maturation was similar to that of 26°C treatment, 14.8% lower than that of white-back rice 290 harvested in 34°C treatment ( Figure 1I). As the results indicated, approximately 70% 291 (∆perfect rice 40.7% . ∆white-back rice 58.8% -1 ) of white-back rice simply turned into the 292 perfect rice without causing any reduction in kernel weight (Table 1). 293

294
It has been shown that heat stress affected the PB development, and the volume of PSVs 295 progressively increased over time, reaching to 2.5-fold larger than PBII in 26°C treatment at 296 maturation ( Figures 4F and 5A  Osmotic adjustment occurs in the growing cells at moderately low water potential (Morgan, 312 1977;Cutler et al., 1980aCutler et al., , 1980bMeyer and Boyer, 1981). During the osmotic adjustment, 313 turgor pressure can be maintained by accumulating osmotically active solutes, such as 314 sugars and amino acids, into the cells (Morgan, 1977;Meyer and Boyer, 1981). As reported 315 previously (Wada et al., 2011;Wada et al., 2014), osmotic adjustment also occurred in the 316 rice endosperms cells growing under dry wind conditions prior to chalky ring formation. On-317 site cell metabolomics and in situ turgor assay conducted in outer endosperms indicated 318 clear treatment differences in the heat adaptive responses ( Figure 3B-E). Different from the 319 drought-related studies, there seem to be few studies regarding osmotic adjustment under 320 heat conditions (e.g., Jiang and Huang, 2000 in turfgrass). In this study, the determination of 321 cell osmotic pressure was not attainable using a freezing point osmometer, because of the 322 interference of starches contained in the sap. However, based on the treatment differences 323 in metabolites in the target zone, it was reasonably interpreted that N-treated cells strongly 324 under low N level, the cells adjusted less osmotically and the protein synthesis rate slowed 326 down, and therefore the energy requirement in cytosol could be maintained low.  corresponds to the occurrence of white-back kernels (Table 1). Hence, it is expected that 338 higher Cys accumulation would have occurred in the cytosol in 34°C treatment, as well as 339 sugars and other major amino acids, which might be caused by a low assimilate input under 340 long-term heat conditions. It appeared that N application increased amyloplast development into the storage proteins during the adaptation process. In the comparative analysis between 347 34°C and N+34°C treatments, our TEM image analysis indicated that the treatment difference 348 in the spatial ratio of PBI and PBII occupied in the cells was 0.52% and 3.86% on average, 349 respectively. The percentage of the chalky area to the transversal area corresponds to 17.4 350 ± 0.8% (mean ± SD, n = 18). The mean area of the transversal section measured in 34°C-351 and N+34°C-treated kernels was 0.051 and 0.050 cm 2 , respectively, and the averaged 352 longitudinal kernel length in these treatments was 5.15 and 5.16 mm, respectively. With the 353 specific gravity of PBI and PBII, 1.27 and 1.29 g . mL -1 , respectively (Tanaka et al., 1980), the 354 weight difference in the chalky zone between those treatments corresponded to 0.32 355 mg . chalky zone -1 . Because the treatment difference in the protein content was 0.25 356 mg . kernel -1 ( Figure 4B) and the difference in the occurrence of white-back rice was 58.8% 357 (Table 1) conditions. PSVs treated at 34°C were found to be expanding over time, but with a retardation 376 of PB accumulation ( Figure 5A and B). Importantly, the volumetric increase in PSV could be 377 explained by increasing PSV matrix (see Figure 5C), indicating that substantial amount of 378 water had been entering PSVs towards maturation to increase the vacuolar volume. Thus, it 379 is quite unlikely that tonoplast membrane lipids were degraded under heat conditions. One 380 plausible explanation is that a partial degradation of PBII in PSV would occur through the 381 activation of proteases or the autophagy-like process, leading to an increase in the vacuolar 382 osmotic pressure. This would promote water entry into PSVs to sustain the vacuolar volume, 383 and maintain the cell volume. The source of water accumulated in the PSV matrix remains 384 unknown. However, based on the increases in the content of ascorbate, glutathione, and 385 monodehydroascorbate detected at the cellular level (Supplementary file 2), we speculate 386 that the accumulation might be a consequence of an increased activity of ascorbate 387 peroxidase catalyzing the conversion of hydrogen peroxide into water. PB development 388 sustained during the N-enhanced adaptation process supports our conclusion that disulfide 389 bond formation and tonoplast denaturation would be facilitated by strong osmotic adjustment 390 ( Figure 2). As observed in Figure 2J-L, 34°C-treated kernels exhibited higher water content 391 than 26°C treatment, consistent with previous studies (Ishimaru et al., 2009;Iwasawa et al., 392 2013;Hayashi et al., 2015). Given the fact that water is a major compound in both PSV matrix 393 and vacuoles in the gap spaces, it is not surprising that chalky zone (or even chalky rice) 394 exhibited relatively high moisture content under heat conditions, compared to 26°C treatment. 395 Storing water in the endosperms along the dorsal vasculature may be an essential event to 25.1% in N+34°C treatment and 34°C treatment, respectively (see Figure 1I). In another type 405 of chalkiness, called milky-white rice, chalkiness was observed at 13.3% on average (Wada 406 et al., 2014). Taken together, significant transparency loss would occur in a range of 10.3 to 407 13.3%. The preservation of PSVs expanded in cytosol was responsible for the formation of 408 heat-induced white-back rice ( Figure 6). In the chalky zone, a small number of cells, More recently, it has been possible to trace metabolites in the target zone using the stable 420 isotope(s) in mass spectrometry (Wada et al., 2017). The use of this analytical method may 421 extend our understanding of heat adaptation mechanisms in rice endosperms. 422 423

Conclusion and remarks 424
In this work, the on-site cell-specific analysis was utilized to investigate the cellular 425 mechanisms of heat-induced chalky formation and N-enhanced adaptation response. 426 Hampshire, the UK) by two-days polymerizing at 60°C. Semi-thin sections (appropriately 900 514 nm) for light microscopy were stained with 0.1% (w/v) Coomassie Brilliant Blue for 1h 515 followed by potassium iodide for 1min, and ultra-thin sections (appropriately 80−100 nm) for 516 electron microscopy were stained with lead citrate. After the staining, ultra-thin sections were 517 observed with a transmission electron microscope (TEM; JEM-1010, JEOL Ltd., Tokyo, 518 Japan). For the organelle arrangement image analysis, the outline of all amyloplasts, protein 519 bodies, and other area (referred to as 'gap') on the light microscopic images, and the area of 520 both PBs on TEM images were traced by using ImageJ software (US National Institutes of 521 Health, Bethesda, MD, the US). The outline of PSV and PBII in PSV was also traced. By 522 assuming that they were spherical with a same r, the ratio of volume (V=4/3r 3 ) to the area 523 (A=r 2 ) was 4/3r. This value was regarded as the representative ratio to calculate V from A 524 with the conversion, V= 4/3rA. The volume of PSV matrix was calculated as the difference 525 between PSV and PBII volumes. 526 527

Protein Extraction from Rice Kernels and SDS-PAGE 528
In each treatment, one-third of the kernel containing the dorsal side of the matured kernels, 529 corresponding to the chalky zone of white-back rice, was removed using a razor blade. Total Matsuba (1991). The appearance of dehulled grains (the numbers of white-back kernels) 537 was visually evaluated with >1.8mm thickness of grains, according to the standard evaluation 538 method of The Ministry of Agriculture, Forestry and Fisheries of Japan (http://www.maff.go.jp/j/seisan/syoryu/kensa/pdf/genmai_kaisetsu.pdf). The dry weight of 540 kernel samples was determined during the treatment and at harvest, as described previously 541 (Wada et al., 2011). 542 543

Nitrogen and Starch Content Assay 544
The protein content of kernels is conventionally estimated using an N-protein conversion 545 factor, 5.95 from the N content determined by the Kjeldahl method. 546 547

Statistical Analysis 548
Statistical analysis of all other data was conducted using a Tukey's test in a general linear 549 model (GLM) procedure in JMP software (version 12.1.0, SAS Institute Inc., Cary, NC). 550    the TEM images at 40 DAH (see Figure 2G, H, I). Each parameter in C-F indicates mean±SEs for 25-59 PBs from at least three kernels. In E, 898 each deep-gray bar embedded in the light gray bars showing the area of PBI indicates the area of Cys-rich 10-kDa layer, which locates at the 899 core of PBI (see inset). Different letters in A-F represents significant differences among treatments (Tukey's HSD test, p < 0.05) within the 900 protein content, protein weight, and the area and number of both PBs (letter only) and area of Cys-10-kDa layer of PBI ('). In G, SDS-PAGE 901 analysis of one-third of dorsal side of kernel, corresponding to the chalky zone in 34°C treatment. CysR10P, Cys-rich 10-kDa prolamin; CysR16P, 902 Cys-rich 16-kDa prolamin; CysP13P, Cys-poor 13-kDa prolamin; pGT, proglutelins; GT, glutelin acidic subunit; GT, glutelin basic subunit; 903 Glb, -globulin. treatment. In C, the PSV volume as a function of the corresponding PSV matrix volume. 935 Closed circle, opened square, and gray downward-pointing triangle indicate 26°C, 34°C, and 936 N+34°C treatments, respectively. Each point indicates mean±SEs calculated from areas of 937 4-38 PSVs or PBIIs. In A and B, nitrogen application at 4 DAH was shown, and the thick 938 black bars at the x-axis indicate the duration of 34°C treatment. Different letters in A and B 939 indicate significant differences among treatments (Tukey's HSD test, p <0.05). In C, the 940 regression line between the PSV matrix volume (x) and the entire PSV volume (y) in 34°C 941 treatment was y = 1.08x +0.41 with r 2 = 0.99 (p < 0.01). Inset in C shows the relationship 942 between PSV volume (y) and PBII volume (x) in 26°C and N+34°C treatments, and the 943 regression lines in 26°C and N+34°C were y=0.89x with r 2 =0.99 (p < 0.01) and y=0.98x+0.01 944 with r 2 =0.99 (p<0.05), respectively. Each dashed line in C indicate a 1:1 line. to the reduction in the cytosol occupied with mainly vacuoles, resulting in the adequate development of amyloplasts and PBs to diminish 966 cytosolic space (see Supplementary file 4C, F and I). When N level is low prior to high temperature (34°C treatment), the cells adjust less 967 osmotically and reduce protein synthesis rate to preserve PSVs and vacuoles (see Discussion), resulting in air space formation during kernel 968 dehydration, which turns to chalk. When N supply is adequate in the cells prior to high temperature (N+34°C treatment), cells sustain protein 969 synthesis rate by strong osmotic adjustment to promote normal PB and amyloplast development, which diminishes cytosolic space, resulting 970 in a substantial increase in perfect rice. 971