Nucleotide limitation results in impaired photosynthesis, reduced growth and seed yield together with massively altered gene expression

Nucleotide limitation and imbalance is a well described phenomenon in animal research but understudied in the plant field. A peculiarity of pyrimidine de novo synthesis in plants is the complex subcellular organization. Here, we studied two organellar localized enzymes in the pathway, with chloroplast aspartate transcarbamoylase (ATC), and mitochondrial dihydroorotate dehydrogenase (DHODH). ATC knockdowns were most severely affected, exhibiting low levels of pyrimidine nucleotides, a low energy state, reduced photosynthetic capacity and accumulation of reactive oxygen species (ROS). Furthermore, altered leaf morphology and chloroplast ultrastructure were observed in ATC mutants. Although less affected, DHODH knockdown mutants showed impaired seed germination and altered mitochondrial ultrastructure. Transcriptome analysis of an ATC-amiRNA line revealed massive alterations in gene expression with central metabolic pathways being downregulated and stress response and RNA related pathways being upregulated. In addition, genes involved in central carbon metabolism, intracellular transport and respiration were mainly downregulated in ATC mutants, being putatively responsible for the observed impaired growth. ONE-SENTENCE SUMMARY Impaired pyrimidine nucleotide synthesis results in nucleotide limitation and imbalance, resulting in impaired photosynthesis, reduced growth, reproduction, and seed yield together with massively altered gene expression


INTRODUCTION
Pyrimidine nucleotides are essential components of all living cells. They serve as building blocks for DNA and RNA and participate in metabolic processes ranging from sugar interconversion and polysaccharide metabolism to biosynthesis of glycoproteins and 20 phospholipids (Kafer et al., 2004;Garavito et al., 2015). Most of nucleotides are incorporated 21 into ribosomal RNA and thus influence translation and growth (Busche et al., 2020). Whereas the levels of free nucleotides are kept constant and balanced between purines and 23 pyrimidines, ribosomal RNA pools dynamically respond to growth signals and during 24 acclimation to cold (Busche et al., 2020, Garcia Molina et al., 2021. This response in RNA

Results
We recently provided a preliminary characterization of a set of ATC mutants which contained 90 two lines denominated as atc#1 and atc#2 which contained 17% and 12% residual amounts 91 of transcript leading to 34% and 12% residual protein, respectively (Bellin et al., 2021a). This 92 study was now complemented by analysis of two additional knock-down lines for DHODH, Especially atc#2 exhibits increased levels of the purine breakdown products 2-ureido glycine, allantoate, and allantoin, the latter are suggested to function in attenuating ROS stress moderate reductions (light blue) and 301 DEGs with an FC < -1 had significantly greater  chloroplast (14/49), plastid transcription (6/11), protein refolding (13/49), rhythmic process processes", "cell wall organization", and "photosynthesis" (Supplemental Figure S3A). synthesis with reduced expression of ADSL. However, IMPDH and GMPS leading to GMP 162 synthesis were increased. Genes of salvage pathway enzymes showed reduced expression 163 except for the plastidic NDPK2 (Table 1). Genes of purine and pyrimidine catabolism were all 164 reduced in expression (Table 1) (Kirch et al., 2004), HKL, serving as a negative growth regulator (Karve and Moore, mitochondrial respiration not coupled to ATPproduction (Table 3). Marked reduction was 177 observed for PDC3 (Log2FC -2.8), one of the two main pyruvate dehydrogenase complexes and cytosolic fumarase 2 (FUM2, Log2FC-1.49) involved in carbohydrate partitioning and 179 adaptation to abiotic stress, for example cold stress, (Dyson et al., 2016;Pracharoenwattana 180 et al., 2010).

182
To examine whether the observed differences between ATC and DHODH mutants can be 183 attributed to differences in tissue-specific expression, transgenic plants were produced. 187 DHODH (Supplemental Figure S4I-P) are highly expressed during seed germination and early 188 seedling development (Supplemental Figure S4 A-C and I-K). In two-week, old seedlings GUS 189 signal could be detected over the whole cotyledons (Supplemental Figure 4C, K). Furthermore, the vasculature showed intense staining in cotyledons. Staining of the leaf weakened with increasing age of the leaves (Supplemental Figure S5). Moreover, intensive staining was also visible in primary and secondary roots as well as in root tips (Supplemental Figure S4D, E, l, M).

197
When developing siliques of mutant lines were inspected, empty positions with aborted seeds 198 (red asterisks) and less colored seeds (lacking embryos; white arrows) were visible in all 199 knock-down lines, but not in Col-0 controls ( Figure 4A). The number of seeds per silique was 200 found to be reduced in all knock-down lines, but strongest in atc#2 with only 22.1% of residual 201 viable seeds ( Figure 4B).

202
Silique length was reduced in all mutant lines. Compared to control plants the silique 203 length in ATC knock-down lines was reduced to 60.5% and 42.6%, and for DHODH knock-204 down lines down to 80.7% and 60.7% ( Figure 4C). Shorter siliques as well as increased 205 numbers of aborted seeds per silique in knock-down lines resulted in reduced yield of mature 206 seeds per plant. The weight of seeds per plant was reduced by 66% and 91% for atc#1 and 207 atc#2 and by 24% and 65% for dhodh#1 and dhodh#2 in comparison to the Col-0 ( Figure 4D).

208
Analysis of the 1.000-seed weight of mature, dried seeds revealed a reduced seed weight in 209 both ATC knock-down lines, whereas in DHODH knock-down plants the 1.000-seed weight 210 was comparable to the wild type ( Figure 4E).

211
To determine whether the seed development impacts mature seed properties, the seed 212 germination was analyzed. Whereas only small alterations were observed between Col-0 and 213 ATC knock-down lines, surprisingly both DHODH knock-down lines showed a significant delay 214 in germination ( Figure 5A). 30 hours after transfer of seeds to ambient growth conditions, 97% lines only 44% (dhodh#1) and 30% (dhodh#2) of the seeds were germinated after the same 217 time ( Figure 5A). Rescue experiments with uridine and uracil (1mM each) did not support 218 germination in DHODH mutants, but uracil provoked delayed germination in Col-0 and ATC 219 mutants ( Figure 5B, C).

220
Monitoring of root growth revealed that DHODH knock-down plants had compensated 221 their germination delay within five days and appeared similar to wild-type plants, whereas the 222 development of atc#1 and atc#2 was nearly arrested 5 days after germination ( Figure 5D).

Knock-down lines in pyrimidine de novo synthesis reveal altered ultrastructure of chloroplasts and mitochondria
Since all mutant lines showed a severely reduced growth, the leaf morphology and cellular 229 ultrastructure were analyzed in detail by means of histology and transmission electron 230 microscopy. For this, we focused on the most affected lines atc#2 and dhodh#2. Light microscopy analysis of cross sections from mature leaves revealed that leaf thickness was 232 reduced by approximately 26% and 5% in atc#2 and dhodh#2 compared to corresponding 233 control plants ( Figure 6A-F) (Supplemental Figure S6A). However, whereas the leaf 234 architecture was wild-type like in dhodh#2, atc#2 knock-down lines showed an altered

246
Since the chloroplast ultrastructure was largely unchanged in dhodh#2; we intended 247 to determine if the observed phenotype of DHODH knock-down lines might be based on 248 defects in the mitochondrial ultrastructure. Thereby, no significant differences in mitochondrial 249 sizes were observed between wild-type and mutant lines (Supplemental Figure S6C).

250
However, compared to Col-0 and atc#2 plants about 16% of the mitochondria from dhodh#2 251 showed an altered ultrastructure in which the granules were less abundant, and the cristae 252 formed by the inner membrane were reduced. Additionally, the formation of ring like structures 253 has been observed ( Figure 6O; black arrows).

287
In this work, we analyzed the function of two enzymes of the de novo pyrimidine biosynthesis 288 pathway located in the chloroplast (ATC) and mitochondria (DHODH) by employing 289 corresponding knock-down lines. ATC transcript levels were reduced by 84 and 90% relative 290 to wild type in the two analyzed, representative knock-down lines, which corresponded to the 291 lower levels of ATC protein (Bellin et al., 2021a) consistent with previous reports (Chen and Slocum, 2008). These lines showed severe growth limitations throughout development.

293
Notably, DHODH lines had much weaker phenotypes than ATC lines, although both sets of 294 lines were characterized by a comparable reduction in transcript levels ( Figure 1A). There is 295 no information available that would suggest different protein amounts or enzyme activity in 296 response to transcript reduction in both mutants. However, it is not surprising that mutation in 297 the first committed step of a pathway, here ATC, results in more pronounced phenotypes as 298 mutations in later steps, because these might be compensated by upregulation of early steps, 299 resulting in increased substrate availability or the possibility that compensatory salvage 300 reactions rescue phenotypes. In support of this view, previous reports on mutants in de novo pyrimidine synthesis from Arabidopsis or solanaceous species, identified more pronounced 302 phenotypic alterations in ATC mutants, compared to those encoding later pathway reactions.

306
The expression pattern of the ATC and DHODH genes did not reveal substantial     (Table 2, Table 3). This is reflected by reduced amounts of isocitrate in atc#1, demand for nucleotides to provide ribosomal RNA which can consume up to 50% of all cellular DHODH were shown to be upregulated by the glucose-TOR complex (Xiong et al., 2013).

364
Conversely, nucleotide limitation negatively affected TOR activity. It is thus likely that 365 nucleotide limitation in ATC mutants, but not in DHODH causes the large reprograming of 366 metabolism via altered gene expression, partially regulated by the TOR pathway.

Plant growth
For DNA isolation, tissue collection and phenotypic inspection, wild-type and transgenic s −1 , temperature 22°C, humidity 60%) Illumination was done with LED light (Valoya NS1,    Table S1) were used, and subsequently the PCR product was introduced into

565
The batch data was processed using the untargeted metabolomics workflow of the Compound   Table 4. CO2-Assimilation-and Respiration rate. Determination of assimilation and respiration rate by Gas-exchange measurements. Plotted are mean values of eight biological replicates ± standard error. For statistical analysis one way ANOVA was performed followed by Dunnett's multiple comparison tests (* = p < 0.05, ** = p < 0.005, *** = p < 0.001).

SUPPLEMENTAL DATA
Supplemental Table S1. Primers used in this study Supplemental Table S2.
Protocol for preparation of leave cuttings for histological and ultrastructural analysis Supplemental Table S3.
Chromatographic and mass spectrometry conditions for the untargeted metabolite analysis Supplemental Figure S1.
Phenotype of ATC and DHODH mutant plants at flowering time Supplemental Figure S2.
Heatmap of relative changes in quantities of selected metabolites Supplemental Figure S3.
Lists of DEGs sorted to selected pathways Supplemental Figure S4.
Analysis of leaf ultrastructure: leaf thickness and organelle area    Reduced expression was detected for 1401 genes (blue) and increased expression for 1356 genes (red). Thereby 301 DEGs showed a Log2FC < -1 and 330 DEGs a Log2FC > +1. Detected DEGs were subdivided by GO-terms analysis into different biological processes (BP) that were either (C) enriched or (D) repressed in atc#1 mutants compared to Col-0.  For statistical analysis in A-C One way ANOVA was performed followed by Dunnett's multiple comparison tests (*** = p < 0.001). Different letters in G denote significant differences according to two-way ANOVA with post-hoc Turkey HSD testing (p < 0.5). Scale bar in D-F = 1 cm.