Nucleotide Imbalance, Provoked by Downregulation of Aspartate Transcarbamoylase Impairs Cold Acclimation in Arabidopsis

Aspartate transcarbamoylase (ATC) catalyzes the first committed step in pyrimidine de novo synthesis. As shown before, mutants with 80% reduced transcript and protein levels exhibit reduced levels of pyrimidine metabolites and thus nucleotide limitation and imbalance. Consequently, reduced photosynthetic capacity and growth, accompanied by massive transcriptional changes, were observed. Here, we show that nucleotide de novo synthesis was upregulated during cold acclimation of Arabidopsis thaliana (ecotype Columbia, Col-0) plants, but ATC knockdown mutants failed to acclimate to this condition as they did not accumulate neutral sugars and anthocyanins. A global transcriptome analysis revealed that most of the transcriptional changes observed in Col-0 plants upon cold exposure were also evident in ATC knockdown plants. However, several responses observed in cold-treated Col-0 plants could already be detected in knockdown plants when grown under standard conditions, suggesting that these mutants exhibited typical cold responses without prior cold stimulation. We believe that nucleotide signaling is involved in “cold-like priming” and “cold acclimation” in general. The observed transcript levels of genes involved in central carbon metabolism and respiration were an exception to these findings. These were upregulated in the cold but downregulated in warm-grown ATC mutants.


Introduction
Due to their sessile lifestyle, plants are forced to acclimate changing environmental conditions. Over the last decades, extensive research on various plant species has already led to the discovery of various metabolic pathways important for acclimation of the plant's metabolism to low temperatures. This complex process of acclimation to low but nonfreezing temperatures has been described in plants as "cold acclimation" (CA) and can be divided into three stages involving a variety of molecular, biochemical, and physiological changes [1,2]. In general, CA involves the rearrangement of membrane composition, adjustment of water contents in different tissues, global modification of gene expression, and regulation of protein activity [2][3][4][5].
The first phase of CA involves the direct cellular response to cold, including responses to an energy imbalance between photochemistry, electron transport, and metabolism. State transitions and enhanced non-photochemical quenching can minimize excitation energy at photosystem II (PSII) and therefore reduce the risk of damage [6].
To further counteract the effects of low temperatures, in the second phase of cold acclimation, extensive reprogramming of the transcriptome, proteome, and metabolome results in alterations in primary and specialized metabolism [7]. Further, cold treatment

Cold Leads to Increased Expression of Genes Involved in Nucleotide De Novo Synthesis
Although the precise balancing of nucleotides and de novo biosynthesis play an important role in plant development, little research on the role of this metabolic pathway in the abiotic stress response has been performed in the past. Nevertheless, a recent study showed that the biosynthesis of pyrimidines and purines was greatly increased during cold acclimation, resulting in significantly larger nucleotide pools [21]. In-depth analysis of a previously published transcriptome dataset obtained from cold-treated wild-type (Col-0) plants re-vealed that major transcriptional changes in pyrimidine and purine biosynthesis occur during the early phases of cold acclimation (1-3 days) ( Figure 1A,B) [7]. Among all analyzed pyrimidine de novo synthesis genes, dihydroorotate dehydrogenase (DHODH) exhibited the greatest increase in expression (1.8-fold) during cold treatment. Further transcriptional changes could be observed for the plastid-localized enzymes carbamoyl phosphate synthase (CPS SSU/LSU) and aspartate transcarbamoylase (ATC, red) ( Figure 1B). In line with this, genes of purine biosynthesis also showed an increase in transcript levels. The most pronounced change was seen in inosine monophosphate dehydrogenase 1 (IMPDH1) ( Figure 1A,B). During the de-acclimation phase, transcript levels for both pyrimidine and purine de novo synthesis genes decreased to or even below the initial levels before cold acclimation ( Figure 1B).
Molecules 2023, 28, x FOR PEER REVIEW 3 of 22 in the abiotic stress response has been performed in the past. Nevertheless, a recent study showed that the biosynthesis of pyrimidines and purines was greatly increased during cold acclimation, resulting in significantly larger nucleotide pools [21]. In-depth analysis of a previously published transcriptome dataset obtained from cold-treated wild-type (Col-0) plants revealed that major transcriptional changes in pyrimidine and purine biosynthesis occur during the early phases of cold acclimation (1-3 days) ( Figure 1A,B) [7]. Among all analyzed pyrimidine de novo synthesis genes, dihydroorotate dehydrogenase (DHODH) exhibited the greatest increase in expression (1.8-fold) during cold treatment. Further transcriptional changes could be observed for the plastid-localized enzymes carbamoyl phosphate synthase (CPS SSU/LSU) and aspartate transcarbamoylase (ATC, red) ( Figure 1B). In line with this, genes of purine biosynthesis also showed an increase in transcript levels. The most pronounced change was seen in inosine monophosphate dehydrogenase 1 (IMPDH1) ( Figure 1A,B). During the de-acclimation phase, transcript levels for both pyrimidine and purine de novo synthesis genes decreased to or even below the initial levels before cold acclimation ( Figure 1B).

ATC Abundancy Is Strongly Increased Up on Cold Acclimation and Adaptation
To confirm the increased expression of ATC in response to cold in our laboratory, corresponding analyses were carried out using targeted gene expression analysis (qRT-PCR). Plants were grown under control conditions (22 • C, 14 h light/10 h dark) for 21 days and subsequently transferred to 4 • C for defined timespans. The greatest changes in ATC transcript levels were seen after 2 and 3 days of cold treatment, with increases up to 3.5-fold compared to the onset of cold treatment ( Figure 1C). During prolonged cold exposure (days 8 and 63), the transcript levels again decreased significantly to the initial level before the cold treatment ( Figure 1C). Comparable results could be obtained by quantifying protein levels using an ATC antiserum raised against the recombinant protein [19]. Thereby, Arabidopsis wild type (Col-0) plants revealed an almost 5-fold increase in ATC protein levels after 3 days of cold treatment, which declined to initial levels at 8 days after transfer to 4 • C ( Figure 1D).
After seven days at 4 • C, Col-0 plants showed 40.8% less fresh weight compared to their growth at ambient temperature. In contrast, the fresh weight of the ATC mutants was not impaired after 7 days of incubation in the cold, compared to untreated plants of the same genotype ( Figure 2A). However, it became obvious that atc#1 and atc#2 failed to accumulate anthocyanins ( Figure 2B,D). While Col-0 plants accumulated 147-fold more anthocyanins in the cold, the atc#2 line only showed a 3.6-fold increase compared to the control plants ( Figure 2D). Col-0 plants that were exposed to 4 • C for 63 days showed a significant reduction in their maximum growth height, but no differences in plant height could be observed in ATC mutants between the control and cold conditions ( Figure 2C). It was also found that the chlorophyll content in the ATC mutants was reduced, thereby explaining the slightly pale green appearance of the mutant leaves after cold treatment ( Figure 2E).

Cold Acclimation Imposes Global Adaptations of the Transcriptome
To investigate the impact of an altered nucleotide metabolism on the transcriptional changes induced during cold acclimation, RNA-Seq analysis was performed. To this end Col-0 and atc#1 grown at ambient temperature for six weeks (control) were transferred to cold (4 • C) for one and three days, before harvesting the aboveground tissue for RNA isolation. A comparison of transcriptional changes between Col-0 and atc#1 (grown under ambient temperature) was shown previously [20]. Where necessary for a better understanding, we will also refer to these results in the following. Data points and means +/− standard deviation are shown. Different letters denote significant differences according to a two-way ANOVA with post-hoc Tukey HSD testing (p < 0.05).

Cold Acclimation Imposes Global Adaptations of the Transcriptome
To investigate the impact of an altered nucleotide metabolism on the transcriptional changes induced during cold acclimation, RNA-Seq analysis was performed. To this end Col-0 and atc#1 grown at ambient temperature for six weeks (control) were transferred to cold (4 °C) for one and three days, before harvesting the aboveground tissue for RNA isolation. A comparison of transcriptional changes between Col-0 and atc#1 (grown under ambient temperature) was shown previously [20]. Where necessary for a better understanding, we will also refer to these results in the following.
Overall, when analyzing the differentially expressed genes (DEGs) that were significantly (padj. < 0.05) altered by cold treatment and not filtered by a certain fold change threshold, massive changes in the global transcriptome could be observed in both genotypes.
Compared to the control conditions, one day of cold treatment resulted in 12,564 DEGs in Col-0 ( Figure 3A, Supplemental Data 1), and the number of DEGs decreased to Overall, when analyzing the differentially expressed genes (DEGs) that were significantly (p adj. < 0.05) altered by cold treatment and not filtered by a certain fold change threshold, massive changes in the global transcriptome could be observed in both genotypes.
Compared to the control conditions, one day of cold treatment resulted in 12,564 DEGs in Col-0 ( Figure 3A, Supplemental Data S1), and the number of DEGs decreased to 8608 transcripts after 3 days compared to day 1 at 4 • C ( Figure 3B, Supplemental Data S1). For atc#1, 10,878 genes with altered expression were observed after 1 day of cold compared to the control at 22 • C ( Figure 3C), and 7603 after three days of cold compared to the same genotype after one day of cold ( Figure 3D). Though there was a significant overlap of DEGs between Col-0 and atc#1 after one to three days of cold treatment, approximately 13% of the DEGs observed for Col-0 were not altered in expression in atc#1.  A category only appearing in the top 10 list in atc#1 (one day and three days of cold) is "response to Karrikin," a compound synthesized during wildfire. Upregulated genes in this category belong to sugar metabolism, photosynthesis, chaperones, and transcription factors. Gene ontology analysis showed that similar groups of transcripts were cold-responsive in both genotypes. After one day of cold, functional groups of DEGs exhibiting the highest significance values were "ribonucleotide-complex biogenesis," "response to cold," "response to cadmium ions," and "photosynthesis" (Figure 3E,F). These are also typical categories found previously in the analysis of plant cold acclimation [7]. When comparing atc#1 to Col-0 grown at 22 • C, most GO terms differ, with only "photosynthesis" and "response to metal ions" overlapping in both lines [20].
After three days of cold, new categories appear among the ten most significantly regulated functional groups, with "response to water," "response to bacterium," and "response to light intensity" in both genotypes except for "response to bacterium," which only appeared in atc#1 ( Figure 3G,H). To sum up, for atc#1, we found a lower number of DEGs, but in general, changes were part of the same pathways when comparing both genotypes. After 1 day of cold, both genotypes exhibited more upregulated and less downregulated genes, whereas after 3 days of cold, the majority of transcripts were downregulated compared to 1 day of cold in both genotypes.
A category only appearing in the top 10 list in atc#1 (one day and three days of cold) is "response to Karrikin," a compound synthesized during wildfire. Upregulated genes in this category belong to sugar metabolism, photosynthesis, chaperones, and transcription factors.
When comparing Col-0 to atc#1 at 22 • C, a total of 2757 DEGs were found. After 1 day of cold, 9303 genes showed differential expression in both Col-0 and atc#1, whereas 3261 were only found in Col-0 and 1575 only in atc#1 ( Figure 4A, middle panel). Up-and downregulated genes (red and blue arrows) were similar in numbers, with slightly more upregulation in the intercept but slightly more downregulation in the individual groups ( Figure 4A, middle panel). After 3 days of cold, the number of DEGs in the intercept was reduced to 59%, whereas the number of DEGs in the atc#1 group increased by 37% and was constant in Col-0 ( Figure 4A, right panel). When focusing on those DEGs found in both comparisons (1 day of cold and 3 days of cold), the higher number of upregulated genes after one day of cold is reversed after 3 days of cold. In fact, as can be seen from the analysis of the functional groups, genes upregulated after 1 day of cold were often downregulated after 3 days of cold and vice versa. In other words, the global expression after 3 days of a cold is more like the situation before the onset of a cold.

Transcriptional Response in Functional Groups "Ribosomal Protein" and "Transcription Factor"
A new hub in acclimation to cold was identified with "Translation". In this category, a high number of ribosomal proteins dominate, and most were found to be upregulated upon cold exposure [7]. From 207 genes encoding ribosomal proteins altered in expression, we identified 154 in Col-0 after one day in the cold but only 135 genes in atc#1 (Supplemental Data S2) [7]. Among these, 13 genes were only altered (upregulated) in atc#1, whereas 32 genes were found altered (upregulated) exclusively in Col-0. This could indicate a reduced cold responsiveness of the atc#1. In several cases, we observed that the cold response was already seen in atc#1 grown at ambient temperature, compared to Col-0 ( Figure 4B). This phenomenon is similar to "cold priming" [17], but without the requirement of a brief cold stimulus. We also observe this phenomenon in genes involved in nucleotide metabolism (Table 1) and, to a lesser extent, intracellular transport (Table 2), whereas in genes involved in carbohydrate metabolism and respiration for atc#1, contrasting expression in ambiently grown plants compared to cold-grown plants was observed (Table 3).    Figure 4C), several TF with massively altered expression upon cold were identified. Among these are BHLH100 (At2g41240) involved in iron homeostasis and leaf growth [22,23], DREB2C (at2g40340) of the DREBA-2 subfamily, involved in drought response [24]; COR47 (At1g20440) involved in cryoprotection [25]; COR314 (At1g29395), cold regulated protein of the chloroplast inner envelope [26]; MYB77 (At3g50060) involved in light dependent auxin signaling [27] and WRKY30 (At5g24110), and WRKY58 (At3g01080) [28]. For ribosomal proteins and TF, we observed that changes after 1 day of cold were reversed almost completely after 3 days of cold ( Figure 4B,C). This was also observed in the next category, "nucleotide metabolism".
Nucleotide metabolism is directly targeted by knocking down ATC (encoded by PYRB). Most genes in de novo synthesis and salvage were upregulated after 1 day cold and partially this is already seen in the comparison of atc#1 against Col-0 at 22 • C (AIR carb., GMPS, IMPDH1, DNK) ( Table 1). Nucleotide catabolic genes (UAH, ALN, PYD2, PYD3) show downregulation, as already seen in atc#1 grown at ambient temperature compared to Col-0. Furthermore, the pastidic uracil importer PLUTO is upregulated in cold conditions, whereas the nucleotide transporters PMANT1 and ENT3 are downregulated (Table 1).
In a comparison of DEG between atc#1 and Col-0 in the warm, a downregulation of glycolysis, tricarboxyclic acid (TCA) cycle and respiration was observed [20]. When inspecting the same set of genes for DEGs in cold-grown plants, 11 genes were found upregulated and only four downregulated in Col-0 and atc#1 (Table 3). Among those upregulated were AOX1A (alternative oxidase) and NDA1 (alternative NAD(P)H dehydrogenase) both already upregulated in atc#1 at 22 • C and less responsive in atc#1 in the cold. Both corresponding gene products function by avoiding mitochondrial respiration's energy-saving reactions. AOX1A has a function in acclimation to low temperatures and can attenuate ROS production [31,32], a limitation Of ROS production was also observed in NDA1 overexpressor lines [33]. Cytosolic fumarase 2 (FUM2) was downregulated in Col-0 after 1 day of cold exposure and further repressed after 3 days of cold exposure (Table 3). Interestingly, atc#1 already shows downregulation of FUM2 at 22 • C, and most likely as a consequence, less downregulation after 1 day of cold and no further response after 3 days of cold. Again, this is another example of apparent cold priming at 22 • C.

ATC Knockdown Showed Diminished Activation of Flavonoid Biosynthesis in the Cold
As we observed reduced anthocyanin levels in cold-grown ATC knockdown mutants ( Figure 2), we had a closer look at the corresponding genes in the pathway. We observed reduced expression of early biosynthetic genes and a corresponding TF (e.g., MYB111) after 3 days of cold treatment ( Figure 5A-C). However, a more pronounced effect was observed for the late biosynthetic genes involved in anthocyanin biosynthesis. While strongly induced in Col-0 at 4 • C, DFR, LDOX, UGT79B1, and the TF TT8 were markedly reduced in expression in atc#1 after 3 days of cold ( Figure 5D,E).
As we observed reduced anthocyanin levels in cold-grown ATC knockdown mutants (Figure 2), we had a closer look at the corresponding genes in the pathway. We observed reduced expression of early biosynthetic genes and a corresponding TF (e.g., MYB111) after 3 days of cold treatment ( Figure 5A-C). However, a more pronounced effect was observed for the late biosynthetic genes involved in anthocyanin biosynthesis. While strongly induced in Col-0 at 4 °C, DFR, LDOX, UGT79B1, and the TF TT8 were markedly reduced in expression in atc#1 after 3 days of cold ( Figure 5D,E).

Alterations in Photosynthetic Rate Led to ROS Accumulation & Reduced Sugar
Accumulation in atc#1 and 2 ATC knockdown mutants are characterized by impaired photosynthesis [20]. When plants were grown for 28 days in the cold, an increase in photosynthetic efficiency (YII) compared to 22 °C was observed in Col-0; however, this increase was more pronounced in atc#1 and atc#2 ( Figure 6A). For easier comparison, we only show results for PAR 134; full light curves are given in Supplemental Figure S1. At the same time, reductions in NPQ

Alterations in Photosynthetic Rate Led to ROS Accumulation & Reduced Sugar Accumulation in atc#1 and 2
ATC knockdown mutants are characterized by impaired photosynthesis [20]. When plants were grown for 28 days in the cold, an increase in photosynthetic efficiency (YII) compared to 22 • C was observed in Col-0; however, this increase was more pronounced in atc#1 and atc#2 ( Figure 6A). For easier comparison, we only show results for PAR 134; full light curves are given in Supplemental Figure S1. At the same time, reductions in NPQ levels were detected in all genotypes, again more pronounced in both ATC mutants ( Figure 6B, Supplemental Figure S1). Reduced photosynthetic efficiency was accompanied by increased levels of reactive oxygen species, as revealed by NBT and DAB staining ( Figure 6C). Surprisingly, levels of ascorbate and dehydro-ascorbate were massively reduced in atc#2 ( Figure 6D). Accumulation of the neutral sugars glucose, fructose, and sucrose is a typical response during cold acclimation, which was corroborated with our analysis of Col-0 exposed to the cold for 28 days ( Figure 6E). However, both ATC knockdown mutants entirely failed to accumulate sugars in the cold and essentially showed the same carbohydrate level at 4 • C and 22 • C.
increased levels of reactive oxygen species, as revealed by NBT and DAB staining ( Figure  6C). Surprisingly, levels of ascorbate and dehydro-ascorbate were massively reduced in atc#2 ( Figure 6D). Accumulation of the neutral sugars glucose, fructose, and sucrose is a typical response during cold acclimation, which was corroborated with our analysis of Col-0 exposed to the cold for 28 days ( Figure 6E). However, both ATC knockdown mutants entirely failed to accumulate sugars in the cold and essentially showed the same carbohydrate level at 4 °C and 22 °C.

Discussion
Here we showed that upon cold exposure, transcripts encoding enzymes involved in purine and pyrimidine de novo synthesis were induced, while those involved in catabolism were repressed on a global scale (Figure 1). This finding was further corroborated by the transcriptional induction of aspartate transcarbamoylase (ATC) in qPCR analysis and

Discussion
Here we showed that upon cold exposure, transcripts encoding enzymes involved in purine and pyrimidine de novo synthesis were induced, while those involved in catabolism were repressed on a global scale (Figure 1). This finding was further corroborated by the transcriptional induction of aspartate transcarbamoylase (ATC) in qPCR analysis and the accumulation of ATC protein in Col-0 (wild type) plants after 3 days of growth at 4 • C (Figure 1). In line with these observations, upregulation of purine and pyrimidine de novo synthesis was observed by increased amounts of pathway intermediates in a cold-time course experiment as well [21].
These observations agree well with a general upregulation of components for translation (synthesis of ribosomal proteins and ribosome assembly) upon cold exposure, facilitating reprogramming of metabolism during cold acclimation [7]. As ribosomes consist of proteins and RNA, the latter of which is built from nucleotides, it is conceivable that increased translation, requiring ribosome biosynthesis, is demanding increased nucleotide provision. Interestingly, the target of rapamycin (TOR), a central growth regulator, was shown to activate nucleotide de novo synthesis upon growth signals, and vice versa, nucleotide limitation impairs TOR activity [23]. Together with the observation of markedly altered expression of transcription factors involved in the abiotic stress (including cold) response, it becomes likely that nucleotide signaling is at least partially responsible for the massive alterations in gene expression between Col-0 and atc#1 in warm and cold.
Downregulation of the first committed step in pyrimidine de novo synthesis in plants, catalyzed by ATC, leads to impaired growth and development [19,34,35]. Reduced chlorophyll contents, impaired photosynthetic activity, and accumulation of reactive oxygen species (ROS) are very likely causative of the growth phenotype [20]. ATC knockdown mutants atc#1 and atc#2, exhibiting a 3-and 20-fold reduction in ATC protein abundance compared to Col-0 [19], were unable to accumulate sugars and anthocyanins as usually observed in plants undergoing cold acclimation [13,36] (Figures 2 and 6). Surprisingly, photosynthetic yield (YII) in atc#1 and 2 was improved in the cold, whereas this was hardly observed in Col-0. However, ATC knockdowns still perform worse than controls. When looking at central carbohydrate metabolism, genes upregulated upon cold in both genotypes were downregulated in atc#1 vs. Col-0 at ambient temperature (Table 3). This may hinder cold acclimation and explain why ATC mutants are not able to accumulate neutral sugars in the cold (Figure 6), which is most likely related to the inability of atc#1 to accumulate anthocyanins ( Figure 2) [14]. Similar to carbohydrate metabolism, reduced expression of genes active in anthocyanin metabolism was also observed. A downregulation of early biosynthetic genes CHS and CHI as well as MYB111 transcription factor but even more pronounced all late biosynthetic genes and the TT8 transcription factor was observed in the mutants (Figure 5), explaining the low anthocyanin content in cold treated ATC knockdown plants.
In contrast, increased expression of genes encoding ribosomal proteins was observed in atc#1 plants grown at ambient temperature. When these plants were cold treated, the same set of genes was induced, however, at a lower level as compared to Col-0 at day 1 of cold treatment. For transcripts corresponding to "nucleotide metabolism," the sum of changes in atc#1 vs. Col-0 and atc#1 4 • C vs. 22 • C equals the changes observed in Col-0 4 • C vs. 22 • C (i.e., the first and third columns of Table 1 with respect to the second columns (e.g., ADSL, AIRCAR, GMPS, and so on). This behavior in a way resembles a phenomenon known as "cold priming," where a short cold treatment, normally too short to permit the full induction of the acclimation response, affects the expression of a certain set of genes essential to cope with low growth temperatures. Among these are the ROS markers ZAT10 and BAP1 [37]. These were also found upregulated in warm-grown atc#1 [19]. As a result, we conclude that atc#1 and the current nucleotide imbalance mimic cold priming via an as-yet unknown mechanism.
Two further putative scenarios arise from the transcriptomic data: 1. reprogramming to support chloroplast metabolism; and 2. actions to remove excess reducing power and thus limit ROS accumulation.
Upregulation of TST2 upon cold allows for increased import of sucrose into the vacuole, which represents the main site of sucrose cleavage by invertase [38]. Thus, sucrose and monosaccharides can accumulate in the vacuole, a typical response to cold acclimation also seen in Col-0 in our study ( Figure 6). Alternatively, glucose might be exported from the vacuole by monosaccharide transporters [39], to allow for hexose-phosphate synthesis and subsequent reimport into the chloroplast by GPT2, which was highly upregulated in the cold in Col-0 but even more in atc#1 (Table 2). This apparent futile cycle of chloroplast triose phosphate export and reimport in the form of glucose-6-phosphate could prevent the depletion of Calvin cycle intermediates. Moreover, strongly impaired photosynthesis in atc#1 means less energy (ATP and reducing equivalents) for growth and development. In line with this observation, in addition to what was seen in Col-0, NTT was already upregulated at 22 • C in atc#1 and could allow for ATP import into the chloroplast to replenish the energy pool. Reimport of sugars is further supported by reduced expression of TCA cycle genes in atc#1, which may further explain the inability of atc#1 and 2 to accumulate sugars in the cold.
ROS accumulating in atc#1 and 2 under warm and cold conditions ( Figure 6C) may drive the removal of reducing power by the malate valve, converting NADPH and oxaloacetate to malate in the chloroplast, and subsequently exporting this metabolite to the cytosol [31]. We observed higher malate contents in atc#2 at 22 • C [20], together with increased alternative oxidases AOXA1 and NADPH dehydrogenase ND1, and both can dissipate excess energy without producing more ATP [40]. Under cold conditions, these effects are more pronounced as DIC transporters mediating the exchange of dicarboxylates between mitochondria and cytosol are upregulated, as well as tonoplastidic TDT and ALMT carriers that might balance malate levels between the cytosol and vacuole ( Table 2). In addition, AOX1a shows a further marked increase in expression upon exposure to cold. Downregulation of cytosolic fumarase 2 (FUM2) represents a typical cold response [41] and is seen in both genotypes (Table 3). In addition, this response of FUM2 is seen in atc#1 grown at ambient temperature, together with reduced levels of succinate and fumarate but wild-type-like levels of 2-oxoglutarate [20]. This may point to an export of oxoglutarate to the chloroplast to serve as the carbon skeleton for amino acid synthesis by refixing ammonia released by photorespiration [42,43]. High glyoxylate levels are indicative of high photorespiration, as glyoxylate is a key intermediate of this pathway [43]. In addition, glyoxylate is a product of purine catabolism [44]. Together with the observed accumulation of other intermediates in purine catabolism, this points to purine breakdown to balance purine to pyrimidine pools, latter being low in atc#1 [20].
When oxoglutarate is exported from mitochondria, resulting in low succinate and fumarate, as discussed above, this will lead to low malate generation in the TCA cycle. At the same time, this might support malate production from oxaloacetate in the cytosol, thereby reducing NADH levels.

Plant Growth
For RNA isolation, tissue collection, and phenotypic inspection, wild-type (Col-0) and transgenic Arabidopsis thaliana (L.) Heynh. plants (ecotype Columbia) were used throughout. Plants were grown on standardized ED73 soil (Einheitserde and Humuswerke Patzer) under a short day regime (10 h light and 14 h darkness). Light conditions were set to an intensity of 120 µmol quanta m −2 s −1 Illumination was done with LED light (Valoya NS1, Valoya, Finland). The ambient temperature was either set to 22 • C or, for cold treatment, to 4 • C with a constant humidity of 60%. If not stated otherwise, plant material was harvested in the middle of the light period and frozen in liquid nitrogen for further use. The used ATC ami-RNA knockdown lines were described previously [19].

RNA Extraction
RNA was extracted from the entire leaf material of soil-grown plants, which was homogenized in liquid nitrogen prior to the extraction of RNA with the Nucleospin RNA Plant Kit (Macherey-Nagel, Düren, Germany) according to the manufacturer's advice. RNA purity and concentration were quantified using a NanoPhotometer (Implen, Munich, Germany).

cDNA Synthesis and Gene Expression Analyses
Total RNA was transcribed into cDNA using the qScript cDNA Synthesis Kit (Quantabio, Beverly, MA, USA). QPCR was performed using the quantabio SYBR green quantification kit (Quantabio) on the PFX96 system (BioRad, Hercules, CA, USA) using specific primers. Actin (At3g18780) was used as a reference gene for transcript normalization. All used primers are listed in Supplementary Table S1.

RNA Seq and Transcriptome Analyses
All leaf material from at least three plants was sampled for one biological replicate. All conditions were analyzed with three biological replicates. RNA library preparation and transcriptome sequencing were performed by Novogene (Beijing, China). Before library preparation, the isolated RNA was first analyzed for quality by agarose gel electrophoresis and the Agilent 2100 Bioanalyzer (Agilent, Palo Alto, CA, USA). To remove ribosomal RNA and generate the cDNA libraries, poly(A)+ RNA enrichment and mRNA fragmentation followed by random-prime cDNA synthesis was performed. This was followed by pairedend sequencing of the reads of 150 base pairs in length, which were captured using an Illumina sequencer (NovaSeq 6000). To verify and quantify the quality of the reads, the program Hisat2 was used in bioinformatic analyses, and the reads were mapped to the Arabidopsis reference genome. The further bioinformatic evaluation was also performed by Novogene. To investigate which biological processes were highly altered in expression, DESeq2 [45] was used. Therefore, all genes with an adjusted p-value < 0.05 are included. We compared the following groups: Col-0 1d_4 • C vs. Col-0_22 • C; Col-0 3d_4 • C vs. Col-0_1d_4 • C; atc#1 1d_4 • C vs. atc#1 _22 • C; atc#1 3d_4 • C vs. atc#1 _1d_4 • C. In the publication Bellin et al. (2021b), we compared atc#1_22 • C vs. Col-0_22 • C, and these data were also shown in Figure 4B

Protein Extraction
Leaf extract of wild-type and knockdown mutants was prepared by homogenizing leaf material in an extraction buffer (50 mM HEPES-KOH, pH 7.2, 5 mM MgCl 2 , 2 mM DTT, 0.01% DDM, 1 mM EDTA, and 2 mM phenylmethylsulfonyl fluoride (PMSF)) on ice. This homogenous extract was centrifuged for 10 min, 20,000× g and 4 • C. The supernatant was collected and stored on ice until use.

Immunoblotting and Western Blot Detection
For immunoblotting 15 µg of a protein extract were separated in a 15% SDS-PAGE gel were transferred onto a nitrocellulose membrane (Whatman, Germany) by blotting. The membrane was blocked in phosphate-buffered saline plus 0.1% [v/v] Tween 20 (PBS-T) with 3% milk powder for 1 h at room temperature, followed by three washes of 10 min in PBS-T. Then, the membrane was incubated with a rabbit polyclonal antiserum raised against recombinant ATC (Eurogentec, Belgium) for 1 h, followed by three washes with PBS-T. Next, the membrane was incubated for 1 h with a horseradish peroxidase (HPR)-conjugated anti-rabbit antibody (Promega, Walldorf, Germany) diluted in PBS-T with 3% milk powder. The result was visualized by chemiluminescence using the ECL Prime Western blotting reagent (GE Healthcare) and a Fusion Solo S6 (Vilber-Lourmat) imager.

Chlorophyll and Anthocyanin Extraction and Quantification
Chlorophyll pigments were extracted from 50 mg of ground leaf tissue in 80% ethanol. After boiling at 95 • C for 10 min and subsequent sedimentation of insoluble contents (5 min, 12,500 rpm), chlorophyll was measured by the absorbance of the supernatant at 652 nm. The calculation was performed as described by [46].
Anthocyanin pigments were extracted from 50 mg in an extraction buffer (H 2 O: 1propanol: HCL (18:1:81)) on ice. After boiling at 95 • C for 3 min while being shaken at 650 rpm, samples were incubated on RT in darkness for 24 h. After centrifugation (15 min, 12,500 rpm), the extinction of the supernatant was measured at 535 nm and 650 nm. Extinction was corrected by Rayleigh's formula (E 535 -2.2E 650 ).

Pulse-Amplitude-Modulation (PAM) Fluorometry Measurements
A MINI-IMAGING-PAM fluorometer (Walz Instruments, Effeltrich, Germany) was used for in vivo chlorophyll measurements. A light curve assays intact, six-week-old plants that were grown for three weeks under standard conditions (22 • C) before being transferred to cold conditions (4 • C). After 10 min of dark adaptation, measurements were performed in a standard light curve setting [47]. Therefore, incrementally increasing light pulses with intensities from PAR (µmol photons m −2 s −1 ) 0 to PAR 726 in 14 steps were applied to the plants.

NBT (O 2 − ) and DAB (H 2 O 2 ) Staining
Superoxide (O 2 − ) was detected by nitroblue tetrazolium (NBT) staining [48]. For this, leaves from six weeks old plants (three weeks 22 • C and three weeks 4 • C) were used. In brief, leaves were vacuum-infiltrated with 0.1% NBT. 50 mM potassium phosphate buffer (pH 7.8) and 10 mM sodium azide for 20 min and incubated for 1 h at room temperature. Stained leaves were boiled in 95% ethanol for 15 min and photographed. The detection of H 2 O 2 was performed by treating plants with DAB-HCl. Leaves were vacuum-infiltrated with 5 mM DAB-HCl, pH 3, for 20 min and incubated in the same solution for at least 8 h overnight. Stained leaves were boiled in an ethanol: acetic acid: glycerol (3:1:1) solution under the hood until they turned transparent and were later documented.

Whole-Leaf Ascorbate Determination
Whole-leaf samples from six-week-old plants were harvested six hours after the onset of light. These plants were grown for three weeks under standard conditions (22 • C) and three weeks in cold (4 • C). Ascorbate and dehydroascorbate levels were determined calorimetrically as described previously [49].

Sugar Extraction
For sugar extraction 50 mg of frozen, grounded leaf sample were boiled in 500 µL deionized water for 15 min at 95 • C. After centrifugation (20,000× g for 10 min at 4 • C), the supernatant was used for further sugar analysis. Extracted sugars were quantified using a NAD + -coupled enzymatic assay [50].

Quantification of Aminoacids
For the quantification of amino acids, the supernatant derived from sugar extraction was used. 20 µL of the probe were mixed with 60 µL 0.2 M pH 8.8 (NaOH) borat and 20 µL 1.28 mM AQC (6-aminoquinolyl-N-hydroxysuccinimidyl carbamate, Synchem UG & Co., KG-S041, dissolved in acetonitrile), and the probe was directly vortexed for 10 s. The measurement was done using high-pressure liquid chromatography (HPLC).

Supplementary Materials:
The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/molecules28041585/s1, Figure S1. Photosynthetic parameters of Col-0 and ATC knockdown mutants grown in warm and cold. Plants were grown for three weeks at ambient temperature und then transferred to 4 • C for 28 days. Yield (upper panel) and NPQ (lower panel) were recorded by imaging PAM. Error bars show the standard deviation of the mean. n = 5. Supplemental Data S1, DEG after 1 and 3 days of cold, for Tables 1-3