Loss of the Acetyltransferase NAA50 Induces Endoplasmic Reticulum Stress and Immune Responses and Suppresses Growth.

Stress signaling in plants is carefully regulated to ensure proper development and reproductive fitness. Overactive defense signaling can result in dwarfism as well as developmental defects. In addition to requiring a substantial amount of energy, plant stress responses place a burden upon the cellular machinery, which can result in the accumulation of misfolded proteins and endoplasmic reticulum (ER) stress. Negative regulators of stress signaling, such as ENHANCED DISEASE RESISTANCE 1 (EDR1), ensure that stress responses are properly suspended when they are not needed, thereby conserving energy for growth and development. Here, we describe the role of an uncharacterized N-terminal acetyltransferase, NAA50, in the regulation of plant development and stress responses in Arabidopsis (Arabidopsis thaliana). Our results demonstrate that NAA50, an interactor of EDR1, plays an important role in regulating the tradeoff between plant growth and defense. Plants lacking NAA50 display severe developmental defects as well as induced stress responses. Reduction of NAA50 expression results in arrested stem and root growth as well as senescence. Furthermore, our results demonstrate that the loss of NAA50 results in constitutive ER stress signaling, indicating that NAA50 may be required for the suppression of ER stress. This work establishes NAA50 as essential for plant development and the suppression of stress responses, potentially through the regulation of ER stress.

As sessile organisms, plants frequently encounter and respond to stress conditions such as drought, salinity, heat, and microbial infection. Various adaptations enable plants to defend themselves against these stresses; however, they often come at a significant cost (Cipollini et al., 2014). Plant defense responses require sacrifices by infected cells and tissues, which can negatively impact plant growth. The hypersensitive response, a form of programmed cell death, is a primary mode of defense for infected plant cells (Greenberg and Yao, 2004). Thus, plants must carefully tailor their defense responses to conserve energy for growth and reproduction (Huot et al., 2014). This trade-off is exhibited by enhanced resistance mutants such as suppressor of npr1 constitutive1-1 (snc1-1) and constitutive pr1 (cpr1), which have constitutively active defense responses and are dwarfed (Bowling et al., 1994;Li et al., 2001).
Stress responses place strain upon the cellular machinery, which can result in endoplasmic reticulum (ER) stress (Bao and Howell, 2017). ER stress can occur during biotic or abiotic stress, as well as during normal developmental processes that place increased demands on the protein translation and protein secretion machinery (Vitale and Boston, 2008). Response to ER stress is mediated by the unfolded protein response (UPR), which occurs in two phases. The first phase aims to alleviate ER stress through increased expression of chaperones, removal and degradation of misfolded proteins from the ER, and reduction of protein translation (Liu and Howell, 2010;Williams et al., 2014). If these attempts are unsuccessful, the UPR transitions into a proapoptotic phase (Walter and Ron, 2011;Woehlbier and Hetz, 2011;Srivastava et al., 2018). Recent studies have demonstrated that UPR genes are required for plant growth and development (Kim et al., 2018;Bao et al., 2019). On the other hand, mutations that constitutively activate the UPR cause dwarfism (Iwata et al., 2018). Just as stress responses to external stimuli must be regulated to ensure proper growth and development, so must responses to internal stress and the UPR.
We have previously identified and characterized the ENHANCED DISEASE RESISTANCE1 (EDR1) gene and demonstrated its role in negatively regulating plant stress response signaling (Frye and Innes, 1998;Christiansen et al., 2011;Serrano et al., 2014). In particular, EDR1 negatively regulates the salicylic acid and ethylene pathways (Frye et al., 2001;Tang et al., 2005). Mutant edr1 plants display enhanced sensitivity to a variety of stimuli, including drought, pathogen infection, abscisic acid, and ethylene (Frye and Innes, 1998;Frye et al., 2001;Tang et al., 2005;Wawrzynska et al., 2008). The variety of edr1-related phenotypes implies that EDR1 function impacts a diversity of plant stress responses. Interestingly, edr1 plants appear phenotypically wild type in the absence of external stresses. This transitory requirement of EDR1 indicates that it may be functionally active only after a stress response has been induced.
There remain many unanswered questions regarding EDR1 function. EDR1 encodes a protein kinase bearing similarity to Raf-like MEK kinases (Frye et al., 2001), yet no in vivo substrates of EDR1 have been identified. EDR1 is believed to negatively regulate KEEP ON GOING, an E3 ubiquitin ligase required for postembryonic development and endomembrane trafficking (Wawrzynska et al., 2008;Gu and Innes, 2011;Gu and Innes, 2012). However, it is unclear whether EDR1 itself is a regulator of development or endomembrane trafficking. Interestingly, EDR1 primarily localizes to the ER, yet no ER-associated function of EDR1 has been demonstrated (Christiansen et al., 2011).
To gain a greater understanding of EDR1 function, we performed a yeast two-hybrid screen to identify potential substrates of EDR1. These screens yielded a particularly interesting hit, At5g11340, a predicted N-terminal acetyltransferase (NAT) that bears similarity to the human Naa50 protein (Fig. 1).
Most NATs function as the catalytic component of larger complexes, designated as NatA-F in humans, to modify substrate proteins in a cotranslational manner (reviewed in Polevoda et al., 2009;Aksnes et al., 2016). However, human Naa80 (also known as NatH) appears to function by itself to modify the N terminus of actin subunits using a posttranslational mechanism (Drazic et al., 2018). Similarly, the Arabidopsis (Arabidopsis thaliana) protein AtNAA70 (NatG), which localizes to plastids, does not appear to require additional plant proteins for function, because overexpression of this protein inside Escherichia coli induces N-terminal acetylation (NTA) of over 120 different E. coli proteins (Dinh et al., 2015). Human Naa50 serves as the catalytic component of the NatE complex, which also includes the Naa10 and Naa15 subunits (Arnesen et al., 2006). Naa10, Naa15, and Naa50 are also found in the NatA complex, for which Naa10 provides catalytic function. NAT complexes mediate NTA, a widespread cotranslational protein modification believed to affect the majority of eukaryotic proteins (Brown and Roberts, 1976;Polevoda and Sherman, 2003;Arnesen et al., 2009). These complexes target unique N-terminal sequences. NAT specificity is largely influenced by the two most N-terminal residues of a given peptide . Human Naa50 preferentially targets N-termini that have retained their initiator Met and have a hydrophobic residue in the second position Van Damme et al., 2011).
Based on work in yeast (Saccharomyces cerevisiae) and humans (Homo sapiens), there is a solid biochemical understanding of how NATs function; however, the purpose of NTA is not well understood. Emerging evidence suggests that NTA serves various functions. In humans, the Golgi-localized Naa60 specifically targets transmembrane proteins and is required for the maintenance of Golgi integrity (Aksnes et al., 2015). In yeast, NTA mediated by NatC is required for the recruitment of the Arl3p GTPase to the Golgi (Behnia et al., 2004;Setty et al., 2004). Recent work in plants has implicated NTA in the regulation of stress responses and development. Both NAA10 and NAA15 are essential for plant embryonic development, and knockdown of either results in morphological defects and drought resistance (Feng et al., 2016;Linster et al., 2015). Differential NTA of the SUPRRESSOR OF npr1 CONSTITUTIVE1 (SNC1) protein was found to affect its accumulation, demonstrating a role for NTA in the regulation of defense signaling (Xu et al., 2015). In addition, NatBmediated NTA of the transcriptional coregulator SIGMA FACTOR-BINDING PROTEIN1 stabilizes this protein, leading to enhanced expression of a subset of defense-related genes, and repression of photosynthesisassociated genes (Li et al., 2020). Plant NATs bear strong similarity to their nonplant orthologs; however, the discovery of the plant-specific, plastid-localized NatG indicates that NTA in plants may serve unique purposes (Dinh et al., 2015). This early work demonstrates that NTA plays an important role in plant physiology and stress responses. However, many aspects of plant NATs have yet to be investigated.
Here, we demonstrate a role for the uncharacterized Arabidopsis NAA50 gene in regulating plant growth and stress responses. Using knockout and transgenic knockdown lines, we show that NAA50 is indispensable for normal plant growth and development. Loss of NAA50 triggers defense response pathways in Arabidopsis, implicating NAA50 in the negative regulation of defense signaling. Loss of NAA50 also induces constitutive ER stress, whereas loss of EDR1 leads to enhanced sensitivity to ER stress. Thus, EDR1 and NAA50 may be involved in the negative regulation of ER stress. This work demonstrates the importance of NAA50 in plant stress responses and development, as well as a potential link between NAA50 and ER stress.

NAA50 Interacts with EDR1
To verify the initial yeast two-hybrid screen that identified NAA50 as a potential interactor of EDR1, we performed additional assays to detect protein-protein interaction. To test for physical interactions between EDR1 and potential substrates, we used a "substrate-trap" Figure 1. NAA50 physically interacts with EDR1. A, Naa50 is conserved in Arabidopsis. Amino acid alignment depicting Arabidopsis NAA50 and human Naa50. This alignment was generated using Clustal Omega (https://www.ebi.ac.uk/Tools/msa/clustalo/) and visualized in Jalview (Waterhouse et al., 2009). Individual residues are colored based upon the Clustal color scheme, which assigns color to residues where amino acid category is conserved. B, EDR1 ST interacts with NAA50 in yeast two-hybrid. AD, GAL4 activation domain fusion; BD, GAL4 DNA binding domain fusion; LAM, lamin; T, SV40 large T antigen. C, Immunoblot analysis of yeast strains from B. EDR1-BD accumulated poorly in yeast, and a significant accumulation of degraded EDR1-BD (*) was visible. D, NAA50 coimmunoprecipitates with EDR1. The indicated constructs were transiently expressed in N. benthamiana and then immunoprecipitated using GFP-Trap beads. sYFPtagged MYC was used as a negative control. E, NAA50 colocalizes with the ER marker SDF2. mCherry-tagged NAA50 and GFP-tagged SDF2 were transiently coexpressed in N. benthamiana. Bars 5 50 mm. These experiments were repeated three times with similar results. mutant of EDR1, EDR1 ST . EDR1 ST contains a D810A substitution in the phosphotransfer domain, which is necessary for substrate phosphorylation, thus stabilizing the potential interaction between EDR1 and potential substrates (Gibbs and Zoller, 1991). Our initial yeast two-hybrid screen was carried out using EDR1 ST as bait. In the yeast two-hybrid assay, NAA50 was found to physically interact with EDR1 ST , but not with wild-type EDR1 (Fig. 1B). This result indicates that NAA50 may be a substrate of EDR1. However, immunoblotting demonstrated that wildtype EDR1 accumulation is substantially lower than that of EDR1 ST in yeast, potentially explaining the absence of an interaction (Fig. 1C). Coimmunoprecipitation using proteins expressed transiently in Nicotiana benthamiana demonstrated that NAA50 physically associates with both EDR1 and EDR1 ST in vivo, contrasting with our yeast two-hybrid results (Fig. 1D). We did not observe NAA50 coimmunoprecipitating with the super Yellow Fluorescent Protein2 (sYFP)-tagged MYC negative control, demonstrating that this interaction is specific (Fig. 1D). EDR1 has been previously demonstrated to localize to the ER (Christiansen et al., 2011). We similarly observed an ER localization of NAA50 tagged with mCherry when transiently expressed in N. benthamiana (Fig. 1E). NAA50 was found to colocalize with the GFP-tagged ER marker STROMAL-DERIVED FACTOR-2 (SDF2; Nekrasov et al., 2009), indicating that NAA50 at least partially localizes to the ER. These experiments indicate that EDR1 and NAA50 physically interact and that both proteins partially localize to the ER.

Arabidopsis NAA50 Is Highly Conserved and Essential for Development
The discovery that NAA50 physically interacts with EDR1 prompted us to investigate its potential functions in Arabidopsis. There is a 51.25% identity match between Arabidopsis and human NAA50 proteins (Fig. 1A). This high degree of sequence similarity indicates that NAA50 function is likely conserved between plants and animals.
To investigate the role of NAA50 in plants, we characterized two T-DNA insertion mutants (SAIL_1210_A02 and SAIL_1186_A03), which we designated naa50-1 and naa50-2, respectively. The naa50-1 mutant contains a T-DNA insertion in intron 2, while naa50-2 contains a T-DNA insertion in intron 1, and neither produces detectable full-length transcript or NAA50 protein (Armbruster et al., 2020). We confirmed the presence of homozygous T-DNA insertions in NAA50 in each line using PCR-based genotyping (see "Materials and Methods"). Both mutant lines were found to be severely dwarfed compared to wild-type plants (Fig. 2, A and B). Knockout naa50 seedlings displayed abnormal and dwarfed growth ( Fig. 2A). As they developed, naa50 plants remained dwarfed and were sterile, although stems and flowers did form (Fig. 2C).
We were able to fully complement the naa50-1 mutant phenotypes by transformation of a transgene carrying NAA50 tagged with sYFP under the control of the native NAA50 promoter, demonstrating that loss of NAA50 is responsible for the dwarf phenotype and sterility (Fig. 2B). These observations establish that NAA50 is essential for normal plant growth and development.

Loss of NAA50 Alters Plant Growth
In addition to being dwarfed, naa50 seedlings displayed a variety of developmental phenotypes. Root morphology was found to be altered in naa50 plants, with root hairs and root cell morphology appearing irregular (Fig. 3A). This led us to hypothesize that loss of NAA50 may result in altered vacuole development. Loss of KEEP ON GOING, another EDR1-interacting protein, has been shown to result in altered vacuolar development (Gu and Innes, 2012). In naa50-1 seedlings expressing the tonoplast marker GAMMA TONO-PLAST INTRINSIC PROTEIN (gTIP; Nelson et al., 2007), altered vacuole shape was observed (Fig. 3B). Many naa50-1 vacuoles appeared fractured and contained many "blebs", similar to those observed in keg mutant seedlings (Gu and Innes, 2012). Additionally, naa50-1 root cells were larger and irregularly shaped. This could indicate that NAA50 is involved in vacuole maturation.
The severe dwarfing and sterility of naa50-1 homozygous mutant plants compromised our ability to study the role of NAA50 in later stages of plant development. To overcome this limitation, we generated transgenic plants in which NAA50 expression can be reduced based on the expression of an artificial micro-RNA (amiRNA) driven by a dexamethasone-inducible promoter (DEX:NAA50-ami). As a control, we utilized a scrambled amiRNA line (DEX:Scrambled-ami), which contains a dexamethasone-inducible amiRNA with no predicted targets. We generated ten transgenic lines carrying this DEX:NAA50-ami construct, and observed that six of these displayed obvious chlorotic and necrotic leaves in the T2 generation by 7 d post treatment with dexamethasone (Supplemental Fig. S1). We selected one of these lines (no. 7) for further analyses. Quantification of NAA50 mRNA levels in rosette leaves of homozygous T3 plants from this line revealed an approximately 85% reduction in NAA50 transcripts by 16 h post amiRNA induction (Fig. 3C).
Reduced expression of NAA50 in the DEX:NAA50ami plants caused severe morphological changes. Growth of DEX:NAA50-ami seedlings on Murashige and Skoog (MS) media supplemented with dexamethasone increased the length of root hairs, recapitulating the naa50 root hair phenotype (Fig. 3D). Additionally, dexamethasone treatment caused DEX:NAA50-ami seedlings to grow slower than the control lines, resulting in shorter roots (Fig. 3E). NAA50 knockdown also elicited changes in stem growth. 24 h after dexamethasone treatment, the stems of DEX:NAA50-ami plants bent approximately 90° (Fig. 3F). As in the roots, dexamethasone treatment completely halted elongation of the primary stem in DEX:NAA50-ami plants (Fig. 3G). Interestingly, this shoot bending phenotype was suppressed by removal of the shoot apical meristem before dexamethasone treatment (Fig. 3H), suggesting that it is dependent on auxin redistribution. Our observations of knockout and knockdown plants confirm that NAA50 is essential for normal plant growth and development.

Loss of NAA50 Triggers Cell Death
As well as inducing growth changes, reduced expression of NAA50 resulted in early senescence in leaves. Leaves of adult DEX:NAA50-ami plants turned yellow and became necrotic following dexamethasone treatment ( Fig. 4A; Supplemental Fig. S1). Senescence also occurred in DEX:NAA50-ami seedlings after transfer to MS plates supplemented with dexamethasone ( Fig. 4B). In both adults and seedlings, the senescence phenotype developed about 4 d after the initial dexamethasone treatment, long after the changes in growth rate and stem bending occurred.
The discovery that knockdown of NAA50 induces cell death prompted us to investigate whether loss of NAA50 results in cell death in naa50-1 seedlings. Indeed, roots of naa50-1 and naa50-2 seedlings were readily stained by trypan blue dye, indicating that loss of NAA50 leads to the accumulation of dead cells in roots (Fig. 4C). Trypan blue staining of naa50 roots was spotty and irregular, indicating that only a subset of naa50 root cells died (Fig. 4D). Taken together, these results demonstrate that in addition to being essential for plant development, NAA50 is also required for the repression of cell death and senescence.
Given the interaction between EDR1 and NAA50, we hypothesized that introduction of the edr1-1 allele into DEX:NAA50-ami plants may affect the senescence phenotype. However, we did not observe any major change in the senescence phenotype when edr1-1 was introduced (Fig. 4E). That the edr1-1 mutation did not enhance or suppress the senescence phenotype indicates that NAA50 and EDR1 may regulate senescence through a shared mechanism.

Loss of NAA50 Represses Growth and Induces Stress Signaling
The discovery that reduced expression of NAA50 triggers changes in plant growth and senescence prompted us to investigate the transcriptional changes taking place in these plants. We therefore conducted an RNA sequencing-based analysis of the DEX:NAA50ami transcriptome. Four-week-old plants were treated with dexamethasone, and RNA was collected 0, 12, and 24 h later. Dexamethasone-treated DEX:Scrambled-ami plants were utilized as a control, with RNA samples collected at the same time points. Transcripts from dexamethasone-treated DEX:NAA50-ami plants were deemed significantly altered only if they differed significantly (adjusted P-value , 0.05) from the dexamethasonetreated Scrambled-ami plants at the same time point. To determine whether these transcripts were significantly up-or down-regulated, we compared expression levels in the 12 and 24 h DEX:NAA50-ami transcriptomes to the time zero DEX:NAA50-ami transcriptome. Transcripts were considered significantly up-or downregulated if they differed from the time zero DEX:NAA50-ami expression level (adjusted P-value , 0.05 and log 2 fold-change . 1.5). By comparing the DEX:NAA50-ami and DEX:Scrambled-ami transcriptomes at various time points, we were better able to exclude potential off-target effects of dexamethasone treatment, amiRNA overexpression, or circadian-influenced expression changes.
Our RNA sequencing analysis indicated that reduced expression of NAA50 resulted in altered expression of approximately 2,000 genes by 12 h postdexamethasone application (Supplemental Dataset S1). To determine the biological processes most impacted by loss of NAA50, we analyzed the biological gene ontology (GO) term enrichment in the 12 and 24 h DEX:NAA50-ami datasets. This analysis demonstrated that NAA50 knockdown leads to up-regulation of genes involved in stress hormone signaling (Fig. 5A). Responses to the stress hormones abscisic acid, jasmonic acid, and salicylic acid were induced 12 h after dexamethasone treatment. On the other hand, NAA50 knockdown resulted in the downregulation of a variety of plant growth and photosynthetic processes (Fig. 5A). In particular, transcripts of genes involved in photosynthesis, light responses, and  . Loss of NAA50 results in developmental changes. A, naa50 seedlings have altered root morphology. The seedling roots depicted are from 1-week-old seedlings. B, Vacuole and cell morphology are altered in naa50 seedling roots. Shown are fluorescence micrographs taken of 7-d-old wild type and naa50-1 seedlings expressing mCherry-tagged gTIP. Scale bars 5 50 mm. C, Dexamethasone (DEX) treatment induces knockdown of NAA50 in DEX:NAA50-ami plants. qRT PCR was performed on cDNA generated from multiple adult DEX:NAA50-ami plants following dexamethasone treatment. Displayed are the averages of three replicates with error bars indicating SD. Asterisk denotes P , 0.05 (by Student's t test). Expression values were normalized to ACTIN2. This experiment was repeated three independent times with similar results. D, NAA50 knockdown induces changes to root cell morphology. Five-day-old seedlings were transferred from MS plates to MS plates supplemented with DEX. Images were taken 3 d after dexamethasone exposure. E, NAA50 knockdown slows root elongation. Seven-day-old seedlings were transferred to MS plates supplemented with ethanol or dexamethasone. Images were taken 3 d after transfer to ethanol-or DEX-supplemented media. F, NAA50 knockdown induces stem bending. Images were taken 24 h after dexamethasone treatment. Numbers indicate proportion of all stems which displayed the given morphology. G, NAA50 knockdown stalls stem growth. Stem measurements were taken on DEX:Scrambled-ami (n 5 8) and DEX:NAA50-ami (n 5 10) immediately before and 6 d after dexamethasone treatment. No stem growth was detected in DEX:NAA50-ami plants subsequent to dexamethasone application. Error bars indicate growth hormone responses were negatively impacted. These changes in expression correlate with the altered development and induced senescence phenotypes observed during NAA50 knockdown. For instance, we found that auxin signaling was significantly reduced following NAA50 knockdown, potentially explaining the meristem-dependent stem-bending that occurs following NAA50 knockdown (Fig. 3, F and H). The increase in the expression of defense genes and salicylic acid signaling correlates with the cell death and senescence that results from NAA50 knockdown (Fig. 4).
To further analyze our transcriptome data, we searched for studies that had identified similar transcriptional changes using the Genevestigator Signature tool (Hruz et al., 2008). We selected the 330 most significantly altered transcripts (greatest log 2 fold-change relative to DEX:NAA50-ami at 0 h) from the 12 h DEX:NAA50-ami dataset, and searched for studies that displayed similar expression profiles. We found that the most similar expression profiles were those of studies investigating plant-pathogen interactions, or light stress (Fig. 5B). This overlap demonstrates that NAA50 knockdown elicits stress signaling in plants.

NAA50 and EDR1 Repress ER Stress
The observation that edr1-1 mutants have enhanced sensitivity to diverse abiotic and biotic stresses (Frye et al., 2001;Tang et al., 2005) led us to question whether this mutant might also have enhanced sensitivity to agents that induce ER stress. We therefore tested edr1-1 plants for sensitivity to tunicamycin (TM), an inhibitor of protein glycosylation that induces ER stress. Injection of rosette leaves induced chlorosis and necrosis of the injected region more rapidly in edr1-1 plants than in wild-type plants (Fig. 6A). This observation suggests that EDR1 is required for proper execution of the unfolded protein response, or that loss of EDR1 results in enhanced cell death signaling during ER stress signaling.
Given the potential role of EDR1 in regulating NAA50 function, and the observation that NTA can alter protein stability, localization, and transport (Arnesen, 2011), we hypothesized that loss of NAA50 function might induce the ER stress response. Indeed, many of the observed naa50-mediated developmental phenotypes, such as stunted growth and cell death, can be caused by ER stress. Treatment with TM or dithiothreitol (DTT), which reduces disulfide bonds and induces ER stress, resulted in shorter roots, increased root hair length, and altered cell morphology in wild-type seedlings (Fig. 6, B and C). Additionally, TM and DTT treatments resulted in root cell death like that observed in naa50 seedlings (Fig. 6D). These results demonstrate that ER stress treatment and loss of NAA50 produce similar physiological changes.
To test whether naa50 seedlings display constitutive ER stress responses, we measured the transcription of ER stress marker genes by reverse transcription quantitative PCR (RT-qPCR). naa50-1 seedlings were found to have significantly higher levels of BINDING PRO-TEIN3 (BIP3) and SECRETORY31A (SEC31A) expression in the absence of any treatment (Fig. 6E). However, BIP3 and SEC31A expression did not significantly differ between wild-type and naa50 seedlings that had been treated with TM. During ER stress, the transcription factor BASIC REGION/LEU ZIPPER MOTIF60 (bZIP60) undergoes splicing, leading to its activation (Deng et al., 2011). Thus, detection of the spliced form of bZIP60 indicates an active ER stress response. Untreated naa50-1 seedlings were found to contain significantly higher levels of spliced bZIP60 relative to the wild type (Fig. 6F). However, similar levels of bZIP60 splicing occurred in TM-treated naa50-1 and wild-type seedlings. These results demonstrate that loss of NAA50 leads to constitutive ER stress, but not an increase in maximum ER stress response signaling. Thus, EDR1 and NAA50 both appear to play important roles in regulating ER stress in plants.

NAA50 Enzymatic Activity Is Required for Development
Given the high sequence conservation between Arabidopsis and human NAA50 proteins (Fig. 1A), we hypothesized that the enzymatic activity of NAA50 would be conserved. In addition to functioning as a NAT, human Naa50 has been shown to be capable of auto-acetylation . We therefore tested NAA50 for auto-acetylation activity using recombinant NAA50 protein. In vitro auto-acetylation assays using recombinant NAA50 protein demonstrated that Arabidopsis NAA50 is indeed capable of auto-acetylation ( Fig. 7A; Supplemental Fig. S2). Human Naa50 has previously been shown to associate with the NatA complex, which includes the Naa10 subunit (Arnesen et al., 2006). Transient expression of sYFP-tagged AtNAA50 with mCherry-tagged AtNAA10 indeed demonstrated that these proteins colocalize in plants (Fig. 7B).
Based on the sequence conservation between Arabidopsis NAA50 and human Naa50, as well as the colocalization of AtNAA50 with AtNAA10, we hypothesized that AtNAA50 likely functions as a NAT. To determine whether NAA50 is active in NTA, we tested whether various loss-of-function NAA50 mutants could complement naa50-2 mutant phenotypes. naa50-2 plants were transformed with NAA50 Y34A -hemagglutinin (HA) and NAA50 I145A -HA. It has been demonstrated that the comparable Y31 and I142 residues in human Naa50 mediate N-terminal peptide binding, and substitution with Ala in these positions reduces NTA activity to below 10% and 42.2% of wild-type levels, respectively (Liszczak et al., 2011). Thus, if NAA50-mediated NTA is indeed required for plant development, these mutations may prevent full rescue of the naa50-2 allele.
We identified numerous transgenic lines expressing both the Y34A and I145A proteins (Fig. 7C). Following transformation with the NAA50 I145A -HA transgene, we observed that the naa50 root phenotype was not fully complemented in the transgenic lines, as roots retained their dwarf phenotype and altered cell morphology (Fig. 7, D and E). Despite retaining the naa50 root phenotypes, some NAA50 I145A lines had wild-type-sized rosettes (Fig. 7F). I145A line 2 plants displayed the greatest expression of the I145A transgene and had a more wild-type-like rosette size than I145A line 3 plants (Fig. 7, C and F). However, although line 2 NAA50 I145A transgenic plants had wild-type-sized rosettes, normal siliques did not develop and viable seed was not produced (Fig. 7G). Four-week-old plants were sprayed with dexamethasone. Images were taken immediately before and 7 d after treatment. B, NAA50 knockdown induces senescence in seedlings. Seedlings were grown on MS plates for 7 d, and then transferred to MS plates supplemented with ethanol or dexamethasone. Images were taken 7 d after transfer to ethanolor dexamethasone-supplemented media. C, naa50 seedling roots contain dead cells. Seven-day-old seedlings were stained with trypan blue dye. D, Cell death staining in naa50 roots is spotty and irregular. Images depict trypan blue-stained roots from 7-d-old seedlings. E, Loss of EDR1 does not alter senescence in NAA50 knockdown plants. Images were taken of 4-week-old plants immediately before and 7 d after dexamethasone treatment. WT, wild type.
The Y34A mutation, which has been shown to have a more significant impact on human Naa50 NTA activity than I145A (Liszczak et al., 2011), also prevented full rescue of naa50-2 plants. NAA50 Y34A transgenic plants did not have normal roots or rosettes and were infertile (Fig. 7, D-G). As in the I145A transgenic lines, we observed a correlation between NAA50 Y34A protein accumulation and rosette size (Fig. 7, C and F). However, unlike NAA50 I145A , NAA50 Y34A was unable to fully complement the rosette dwarfism phenotype (Fig. 7F). Even in Y34A line 3 plants, which displayed a similar level of NAA50 accumulation to I145A line 2 plants, rosette size was not fully restored. Figure 5. Reduction in NAA50 expression induces changes to growth and defense signaling. A, Reducing NAA50 expression results in a downregulation of growth signaling, and an up-regulation of defense signaling. Significantly altered transcripts were identified by comparing expression levels in dexamethasone-treated DEX:NAA50-amiRNA plants at the given time point to Scrambled-amiRNA plants at the same time point. Furthermore, expression levels in DEX:NAA50-amiRNA plants at 12 and 24 h were compared with NAA50-amiRNA plants at 0 h after DEX treatment. To eliminate circadian effects, all samples were harvested at the same time (12 and 24 h samples were sprayed with dexamethasone 12 and 24 h before harvest). GO term enrichment analysis was performed using the BiNGO application to determine whether the DEX:NAA50-amiRNA transcriptome was enriched for specific biological processes. NS, not statistically significant. B, The DEX:NAA50-amiRNA transcriptome bears similarity to biotic and abiotic stress studies. The 330 most significantly altered transcripts were compared were previous studies using the Genevestigator Signature tool. The five most related transcriptomes based on the calculated Relative Similarity scores are shown. A heatmap was generated using Heatmapper (http://www2.heatmapper. ca/expression/) to display the relative log 2 foldchange for each of the 330 transcripts for each study. These 330 transcripts were selected based on log 2 fold-change when comparing expression levels in the DEX:NAA50-amiRNA 12-h dataset to the DEX:NAA50-amiRNA time 0 dataset; all 330 were also significantly different when comparing the DEX:NAA50-amiRNA 12-h dataset to the Scrambled-amiRNA 12-h dataset.
The inability of NAA50 I145A and NAA50 Y34A transgenes to fully rescue naa50-2 plants demonstrates the importance of NAA50-mediated NTA in plant growth and development. That the NAA50 I145A mutant was able to complement the rosette dwarfism, but not the root phenotypes or sterility demonstrates that NAA50mediated NTA may be especially required for the growth and development of roots as well as fertility. Furthermore, the inability of the NAA50 Y34A transgene to rescue the rosette dwarfism phenotype indicates that this mutation is indeed more deleterious than the I145A mutation. Notably, an in vitro Lys acetyl transferase assay using recombinant NAA50 Y34A and NAA50 145A confirmed that these mutations do not block this activity (Supplemental Fig. S3), which further suggests that it is the NTA activity of NAA50 that is required for normal development.

NAA50 Is Required for Growth and the Suppression of Stress Responses
The investigation of NTA in regulating cell signaling in eukaryotes is still in its infancy, and identification Figure 6. Loss of EDR1 and NAA50 result in changes to ER stress signaling. A, edr1-1 mutants display heightened ER stress sensitivity. Leaves from 6-week-old plants were infiltrated with various concentrations of TM using a needleless syringe. Leaves were removed, and images taken 3 d after injection. B, ER stress induces naa50-like root dwarfism. Seedlings were germinated on MS plates or MS supplemented with TM or DTT. Representative 10-d-old seedlings are shown. C, ER stress induces naa50-like root cell morphology. Roots of 10-d-old seedlings are depicted. Seedlings were grown on regular MS plates or MS plates supplemented with TM or DTT. D, ER stress induces cell death in roots. Tenday-old seedlings were stained with trypan blue after growth on MS or MS supplemented with TM or DTT. E, naa50-1 seedlings display heightened ER stress signaling in the absence of TM treatment. qRT-PCR was performed on cDNA generated from wildtype and naa50-1 seedlings. Seedlings were germinated on MS plates and 5 d later transferred to regular MS or MS supplemented with 1 mg/mL TM. RNA was collected 20 h after transfer to new plates. Gene expression values were normalized to AC-TIN2. Values depict the averages of three biological replicates, each consisting of twenty individual seedlings. Error bars represent SD between three independent biological replicates. Asterisks denotes P-value , 0.05 (by Student's t test). F, bZIP60 splicing is induced in naa50-1 seedlings. RT-PCR was performed on the same cDNA used in E. Each lane represents a unique biological replicate derived from 20 seedlings. WT, wild type. and characterization of all plant NATs is incomplete. Our understanding of how NATs function comes primarily from work in human cell culture and yeast. However, recent work in plants has demonstrated a role for NTAs in regulating diverse processes Xu et al., 2015;Huber et al., 2020). Posttranslational modification of proteins has long been appreciated as a mechanism by which cell signaling and crosstalk is regulated (Hunter, 2007). NTA may provide a mechanism by which plants regulate responses to external and internal stress signals at the translational level.
With this work, we have begun to characterize the role of Arabidopsis NAA50 in regulating plant growth and development. Complete loss of NAA50 results in severely dwarfed and sterile plants, as well as altered root morphology. By using hormone-inducible amiRNA transgenic plants, we demonstrated that NAA50 knockdown results in reduced expression of developmental processes and inhibits growth. Taken together, these results indicate that NAA50 is required for plant growth and development.
We initially identified NAA50 as an interactor of EDR1 in a yeast two-hybrid screen. Our results indicate that NAA50 and EDR1 do indeed interact (Fig. 1). However, it is unclear what the basis for this interaction is. It is possible that NAA50 is a substrate of EDR1, however, we were unable to detect EDR1-mediated phosphorylation of NAA50 during coexpression in N. benthamiana. If NAA50 is indeed a substrate of EDR1, Figure 7. NAA50 enzymatic activity is required for plant development. A, Recombinant NAA50 displays auto-acetylation activity in vitro. Recombinant HIS-tagged NAA50 was expressed and purified from E. coli. In vitro reactions were performed at 30°C for the indicated time points. Samples were then boiled and subjected to gel electrophoresis and immunoblotting using an antiacetyl-Lys antibody. This experiment was repeated three times with similar results. This figure was generated from one immunoblot derived from a single experiment with irrelevant lanes removed. B, NAA50 colocalizes with NAA10. sYFP-tagged NAA50 was transiently coexpressed with mCherrytagged NAA10 in N. benthamiana. Bars 5 50 mm. C, Immunoblotting demonstrates that HA-tagged NAA50 mutant transgenes are expressed in transgenic plants. Leaf tissue from hygromycinresistant T3 plants was subjected to gel electrophoresis and immunoblotting using an anti-HA antibody. D, Mutant NAA50 transgenes do not complement naa50 root dwarfism. Representative 10-d-old seedlings are depicted. E, Mutant NAA50 transgenes do not complement naa50 root cell morphology defects. Images were taken of representative 10-d-old seedlings. F, NAA50 I145A can complement naa50mediated rosette dwarfism. The image depicts representative 6-week-old plants. G, NAA50 I145A and NAA50 Y34A do not complement naa50-mediated sterility. Stems were removed from 7-week-old plants. WT, wild type. then phosphorylation by EDR1 may require an unidentified cofactor or signal.
Our work adds to growing evidence that NATs are required for plant development. NAA10 and NAA15 have previously been demonstrated to be essential for development. Loss-of-function mutations in NAA10 or NAA15 are embryonic lethal Feng et al., 2016), while partial loss of NAA15 results in dwarfism and enhanced defense signaling (Xu et al., 2015). Reduced expression of NAA10 and NAA15 alters root morphology, enhances the growth of the primary root and inhibits the growth of lateral roots . NatB also appears to be required for development, as partial loss of NatB components NAA20 and NAA25 result in dwarfism (Ferrández-Ayela et al., 2013;Xu et al., 2015;Huber et al., 2020). Loss of NAA30, the catalytic component of NatC, results in minor dwarfism as well as defects in photosystem II efficiency (Pesaresi et al., 2003). The range of developmental phenotypes resulting from mutations in NATs demonstrate that NATs differ in their involvement in plant development.
Our results demonstrate that, in addition to altering plant development, loss of NAA50 results in the activation of plant stress signaling. Reduced expression of NAA50 elicits senescence in adults as well as seedlings, while the roots of naa50 seedlings contain an abundance of dead cells. Gene expression analysis confirmed that knockdown of NAA50 results in an up-regulation of defense signaling. These results are consistent with the observation that loss of NAA50 results in an accumulation of stress response proteins (Armbruster et al., 2020).
NATs appear to play unique roles in the regulation of plant stress responses. Loss of NatA has been shown to increase drought tolerance . The NatA and NatB complexes have been previously implicated in the regulation of the nucleotide binding Leu rich-repeat protein SNC1 (Xu et al., 2015). Partial loss of NAA15 results in increased stability and accumulation of SNC1, as well as enhanced defense signaling and resistance. Interestingly, loss of NatB leads to decreased accumulation of SNC1, and suppression of snc1-induced dwarfism (Xu et al., 2015). Partial loss of NAA20 and NAA25 results in enhanced sensitivity to osmotic stress and altered expression of stress-related transcripts (Huber et al., 2020). Our observations add to the growing list of evidence that NATs play an important role in plant stress signaling.
Our results suggest that NAA50-mediated NTA may be required for plant development. Our complementation experiments demonstrate that the NAA50 I145A and NAA50 Y34A mutations, previously shown to inhibit human Naa50 NTA activity, prevent the NAA50 transgene from fully complementing the dwarfism and infertility of naa50 plants (Fig. 7, D-G). These results are consistent with the recent observation that human Naa50, but not the enzymatically inactive yeast Naa50, can rescue naa50-mediated dwarfism (Armbruster et al., 2020). Taken together, these observations suggest that NAA50 enzymatic activity is essential for plant development and fertility.
The NTA activity of human Naa50 has been demonstrated previously (Liszczak et al., 2011;Van Damme et al., 2011;Reddi et al., 2016;Evjenth et al., 2009). In the companion paper from Armbruster et al. (2020), Arabidopsis NAA50 is shown to have NTA activity in vitro on synthetic peptides, and in vivo when expressed in E. coli. It is thus likely that enzymatic function is conserved between Arabidopsis and humans. In addition, we were able to detect auto-acetylation on Lys residues of recombinant Arabidopsis NAA50 in vitro, confirming that NAA50 is indeed a functional acetyltransferase (Fig. 7A). Because NAA50 possesses both NTA activity and Lys acetyltransferase activity, it is a formal possibility that both activities contribute to normal development. However, our observation that a Y34A substitution retains Lys acetyltransferase activity (Supplemental Fig. S3), but blocks the ability of NAA50 to complement a naa50 knockout mutation (Fig. 7), indicates that it is the NTA activity of NAA50 that is essential. Consistent with this conclusion, Arabidopsis NAA50 localizes at least partially to the ER, where NTA of nascent proteins is likely to occur (Figs. 1 and 7B).

Loss of NAA50 Induces ER Stress
In addition to its role in negatively regulating defense signaling, our results implicate NAA50 in the repression of ER stress. The developmental defects observed in naa50 plants can be recapitulated by TM and DTT treatment, indicating that they may result from constitutive activation of ER stress responses. In support of this hypothesis, we observed increased expression of ER stress genes and bZIP60 splicing in untreated naa50 seedlings (Fig. 6). Following TM treatment, naa50-1 seedlings displayed wild-type levels of ER stress signaling. Therefore, loss of NAA50 induces ER stress signaling, but does not lead to greater induction during TM treatment. Additionally, the expression of BIP3 and SEC31A in naa50-1 seedlings was significantly lower in the absence of TM than during TM treatment. This indicates that the level of constitutive ER stress that occurs in naa50-1 plants is significantly lower than that elicited by chemical treatment. A recent proteome analysis of naa50-2 plants demonstrated an increase in the accumulation of proteins involved in ER stress responses (Armbruster et al., 2020). Based on these results, we hypothesize that NAA50 is required for the prevention of protein misfolding and aggregation, which contribute to ER stress. Plant NATs have not previously been demonstrated to play a role in the regulation of ER stress. Although NatE seems to be required for the repression of ER stress, it is possible that other NAT complexes are required as well.
There is evidence from human and yeast systems for the involvement of NATs in responding to ER stress and protein aggregates. NTA is known to contribute to protein stability, trafficking, and translocation to the ER (Arnesen, 2011;Forte et al., 2011). The NatA complex has been implicated in the regulation of protein aggregation (Arnesen et al., 2010). HUNTINGTIN YEAST PARTNER K HYPK, a NatA component, has chaperone activity, and has been shown to inhibit the formation of protein aggregates (Raychaudhuri et al., 2008). We have demonstrated that loss of NAA50 in plants results in constitutive ER stress, providing additional evidence that NATs may be involved in the repression of protein aggregation and ER stress.
There is a well-established link between ER stress, the UPR, and defense signaling in plants. Plants carrying loss-of-function mutations in the stearoyl-ACP desaturase SUPPRESSOR OF SA INSENSITIVE2 exhibit dwarfism, enhanced accumulation of the ER stress marker BiP3, and higher expression of PATHOGENE-SIS-RELATED GENE1 (Kachroo et al., 2001;Iwata et al., 2018). This mirrors the increased biotic and ER stress signaling observed in naa50 plants. Mutants lacking UPR regulators INOSITOL-REQUIRING1 and bZIP60 display enhanced susceptibility to bacterial pathogens, demonstrating a link between the UPR and salicylic acid-based defense signaling (Moreno et al., 2012). If ER stress cannot be properly maintained, the UPR shifts into a cell death phase (Walter and Ron, 2011;Woehlbier and Hetz, 2011). A recent investigation of the transcriptional changes that occur during a prolonged UPR demonstrated that transcripts associated with biotic stress responses are elicited during the UPR (Srivastava et al., 2018). Biotic stress signaling is also impacted by the ER quality control (ERQC) pathway. The membrane-bound receptors upon which plant defense signaling relies undergo maturation through the ERQC pathway. Impairment of ERQC machinery can result in enhanced susceptibility, as the receptors required for pathogen recognition are unable to function (Tintor and Saijo, 2014). Thus, compromised ER integrity can hinder plant pathogen responses. Unsurprisingly, plant pathogens have been found to attack the host ER stress pathway for their own benefit. The mutualistic fungus Piriformospora indica induces cell death using an ER stress-dependent mechanism, enabling its colonization of the Arabidopsis root (Qiang et al., 2012). The high degree of overlap between ER stress and biotic stress responses opens the possibility that the observed increase in stress signaling in NAA50 knockout and knockdown plants results from changes to ER stress, rather than direct regulation of stress responses by NAA50.
A link between NTA and osmotic stress in plants has been recently proposed Aksnes et al., 2016). NatA knockdown plants display enhanced drought tolerance, and levels of NatA-mediated NTA were shown to fluctuate in response to abscisic acid treatment . Loss of NatB was also shown to affect plant osmotic stress responses. Mutations affecting NAA20 and NAA25 result in reduced germination in the presence of osmotic stress (Huber et al., 2020). Here, we have demonstrated that loss of a plant NAT induces ER stress. The observed osmotic stress phenotypes of NatA and NatB mutant plants may be an indirect result of induced ER stress. Consistent with this hypothesis, overexpression of the chaperone BiP in tobacco (Nicotiana tabacum) and soybean (Glycine max) results in enhanced drought tolerance (Valente et al., 2009). BiP expression in soybean was found to inhibit both ER-and osmotic stressinduced cell death (Reis et al., 2011). In wheat (Triticum aestivum), treatment with tauroursodeoxycholic acid alleviates osmotic stress-induced cell death and ER stress signaling (Zhang et al., 2017), and strong osmotic stress alters root architecture and induces cell death through an ER stress-dependent mechanism (Duan et al., 2010).

NAA50 Tissue-Specificity
Most work on NTA has been performed in unicellular organisms, making it impossible to determine whether NATs display tissue-specific functions. There are likely to be differences in the expression patterns of NATs in different tissues. According to the Bio-Analytic Resource for Plant Biology (BAR) ePlant browser (http://bar.utoronto.ca/eplant), root expression of NAA50 and NAA10 is predicted to be the highest in the meristematic region. It is likely that some NAT complexes are specifically active at certain developmental periods, or in specific tissues. If other plant NAT complexes are indeed able to fulfill some functions of NAA50, it is also possible that they are only able to do so in certain tissues or at certain points in development, based upon their own expression profiles. The study of plant NATs has the potential to expose tissue-and development-specific NAT activity.
Loss of NAA50 especially affected certain cell types and tissues. In NAA50 knockdown plants, loss of NAA50 led to reduced growth of both roots and stems. Furthermore, stem bending and altered root morphology were observed. The use of an inducible knockdown line enabled us to compare the effects of reducing NAA50 knockdown in new and old cells. Interestingly, phenotypes resulting from NAA50 knockdown were mainly exhibited in newly developed cells. The sterility of naa50 plants demonstrates that NAA50 is required for reproductive as well as vegetative development. The inability of NAA50 I145A and NAA50 Y34A transgenes to fully rescue naa50-induced root defects and sterility highlight the necessity of NAA50 for the development of these tissues. These observations demonstrate that NAA50 activity may be especially required by developing cells, or cells undergoing rapid growth and division.
If NAA50 is indeed required for the regulation of ER stress, it follows that roots, shoots, and anthers would be especially impacted by its loss. There is evidence that plant vegetative and reproductive development require an intact UPR to manage ER stress. Roots have been shown to be particularly sensitive to ER stress (Cho and Kanehara, 2017). Substantial changes in root and shoot development result from mutations in UPR genes, indicating that a functional UPR is essential for vegetative development (Deng et al., 2013;Kim et al., 2018;Bao et al., 2019). Mutations in UPR genes also have a strong impact on plant reproductive development (Deng et al., 2013;Deng et al., 2016). In fact, the UPR is constitutively active in anthers (Iwata et al., 2008). The requirement for UPR signaling in unstressed plants implies that ER stress occurs during normal development and must be managed by the UPR, or that UPR genes are involved in the direct regulation of developmental genes (Kim et al., 2018). A requirement for the UPR in development has long been demonstrated in animals. The rapid production of immunoglobulins by B cells is preceded by an up-regulation of the UPR, which manages potential ER stress (van Anken et al., 2003). Roots, shoots, and anthers may rely upon UPR signaling due to the high level of protein translation which occurs in these tissues during development. That these tissues were indeed particularly affected by the loss of NAA50 demonstrates that NAA50 may be required for the management of ER stress which occurs during development.

Plant Material and Growth Conditions
Arabidopsis (Arabidopsis thaliana) accession Col-0 and Col-0 mutants edr1-1 (Frye and Innes 1998), naa50-1 (SAIL_1210_A02), and naa50-2 (SAIL_1186_A03) were used in this study. The presence of homozygous T-DNA insertions in NAA50 in the SAIL lines was confirmed by PCR-based genotyping using NAA50 gene-specific primers and a T-DNA left border primer (Supplemental Table S1) following the primer design criteria and protocol given by the T-DNA Express Primer Design tool (http://signal.salk.edu/ tdnaprimers.2.html).
For growth on MS plates, seeds were surface sterilized with a solution of hydrogen peroxide and ethanol (1:19) and planted on one-half-strength MS plates supplemented with 0.8% (w/v) agar. For soil-grown plants, seed was directly sown onto Pro-Mix PGX Biofungicide plug and germination mix supplemented with Osmocote 14-14-14 fertilizer (ICL Fertilizers). Plates and flats were placed at 4°C for 48 h for stratification before being transferred to a growth room set to 23°C and a 12-h light (150 mEm-2s-1)/12-h dark cycle. For transient expression experiments, Nicotiana benthamiana was grown under the same growth room conditions as Arabidopsis.

Plasmid Construction and Generation of Transgenic Arabidopsis Plants
NAA50 clones were derived by PCR amplification using complementary DNA (cDNA) from Col-0. Site-directed mutagenesis was used to introduce the I145A and Y34A mutations into NAA50 (Qi and Scholthof, 2008). All primers used in this study for cloning and site-directed mutagenesis are listed in Supplemental Table S1.
For yeast-two hybrid assays, the full-length open reading frames of EDR1, EDR D810A , and lamin were cloned into the DNA-binding domain vector pGBKT7 (Clontech Matchmaker System). The full-length open reading frame of NAA50, and the SV40 Large T Antigen (T) were cloned into pGADT7. EDR1 full-length wild-type cDNA and EDR1 ST (D810A) were cloned into pGBKT7 using SmaI and SalI restriction sites. NAA50 was cloned into pGADT7 using ClaI and XhoI restriction sites.
For the generation of amiRNA transgenic plants, a NAA50-specific amiRNA construct was created by PCR amplification following the procedures of Schwab et al. (2006), which included insertion of the NAA50 sequence flanked by regions of the MIR319 microRNA. The resulting amiRNA construct was cloned into pBAV154 using Gateway cloning.
To generate transgenic plants containing NAA50-sYFP under the control of a native promoter, the 297 nucleotides upstream of the NAA50 start site were cloned into PMDC32 (Qi and Katagiri, 2009) using KpnI and HindIII restriction sites. NAA50 I145A and NAA50 Y34A were generated using site-directed mutagenesis (Qi and Scholthof, 2008) and cloned into pEarlyGate100 (Earley et al., 2006) with a C-terminal 33HA tag using multisite Gateway cloning.
Transgenic plants were generated using the floral dip method (Clough and Bent, 1998). Plasmids were transformed into Agrobacterium tumefaciens strain GV3101 (pMP90) by electroporation with selection on Luria-Bertani plates containing 50 mg/mL kanamycin sulfate (Sigma-Aldrich) and 20 mg mL 21 gentamicin (Gibco). Selection of transgenic plants carrying a BASTA resistance cassette was performed by spraying 1-week-old seedlings with 300 mM BASTA (Finale), or by selection on MS plates supplemented with 300 mM BASTA. Selection of plants carrying a hygromycin resistance cassette was performed by germinating seed on MS plates supplemented with 20 mg mL 21 hygromycin (Fischer Scientific). For transgenes transformed into naa50 mutant lines, we confirmed the presence of homozygous T-DNA insertions in NAA50 to rule out possible seed contamination events.
For expression in Escherichia coli, NAA50 was cloned into pDEST17 using Gateway cloning. The resulting plasmid was transformed into E. coli strain BL21.AI (Invitrogen).

Yeast Two-Hybrid Assays
For yeast two-hybrid assays between EDR1 and NAA50, pGBKT7 and pGADT7 clones were transformed into haploid yeast strain AH109 (Clontech) by electroporation and selected on SD-Trp-Leu medium. Successful transformants were selected after 48 h of growth at 30°C and then struck onto fresh SD-Trp-Leu medium and allowed to grow for another 48 h. Before carrying out yeast two-hybrid assays, yeast was grown in liquid SD-Trp-Leu medium for 16 h at 30°C. Cultures were resuspended in water to an optical density at 600 nm of 1.0, serially diluted, and plated on appropriate SD media. Plates were grown for up to 4 d at 30°C.
antibodies. Visualization of immunoblots from yeast strains used in the two-hybrid assay were performed using the KwikQuant Imager (Kindle Biosciences).

Fluorescence and Light Microscopy
Confocal laser scanning microscopy was performed on a Leica TCS SP8 confocal microscope (Leica Microsystems) equipped with a 633, 1.2-numerical aperture water objective lens and a White Light Laser. sYFP fusions were excited at 514 nm and detected using a 522 to 545 nm band-pass emission filter. mCherry fusions were excited at 561 nm and detected using a custom 595 to 620 nm band-pass emission filter.
To capture detailed images of Arabidopsis roots, images were captured using a Stemi 305 compact Greenough stereo microscope (Zeiss). Digital images were captured using Labscope software (Zeiss).

Reverse Transcription Quantitative PCR
For RT-qPCR experiments, RNA was extracted using the Spectrum plant total RNA kit (Sigma-Aldrich) according to the manufacturer's instructions. cDNA was produced from 1 mg total RNA using the Verso cDNA synthesis kit (Thermo Fisher Scientific). Relative RNA amounts were determined by quantitative PCR using the Power Up SYBR Green Master Mix (Thermo Fisher Scientific). A comparative Ct method was used to determine relative quantities (Schmittgen and Livak, 2008). ACTIN2 was used for normalization.

NAA50 Knockdown Transcriptome Profiling
For RNA sequencing, plants were first sprayed with a solution containing 50 mM dexamethasone and 0.02% (v/v) Silwet L-77 (OSi Specialties) 24, 12 h, and immediately before tissue collection. This experimental design avoided changes in gene expression caused by circadian effects because all samples were harvested at the same time. Three biological replicates were performed per genotype per treatment, each consisting of approximately 0.4 g leaf tissue taken from the fourth leaf of four unique plants. RNA was extracted from 4-week-old Arabidopsis leaves using the Spectrum plant total RNA kit (Sigma-Aldritch) according to the manufacturer's instructions.
Total RNA was prepared into equimolar pools for each sample submitted to Indiana University's Center for Genomics and Bioinformatics for cDNA library construction using a TruSeq Stranded mRNA LT Sample Prep Kit (Illumina) following the standard manufacturing protocol. Sequencing was performed using an Illumina NextSeq500 platform with 75 cycle sequencing kit generating 84-bp single-end reads. After the sequencing run, demultiplexing was performed with bcl2fastq v2.20.0.422. RNA sequencing data were deposited in the National Center for Biotechnology Information Gene Expression Omnibus (GEO) database and is accessible through the accession number GSE145580.
Identification of significantly altered transcripts was performed by comparing DEX:NAA50-ami and DEX:Scrambled-ami datasets. Transcripts that differed significantly (adjusted P-value , 0.05) between the DEX:NAA50-ami and DEX:Scrambled-ami datasets were then analyzed to determine whether expression had increased or decreased relative to the DEX:NAA50-ami 0 h dataset. Those transcripts that were significantly (adjusted P-value , 0.05 and log 2 fold-change . 1.5) up-or down-regulated relative to the DEX:NAA50-ami 0 Hr dataset were then used for GO term enrichment analysis. GO term enrichment analysis was performed in Cytoscape using the BiNGO app (Maere et al., 2005).
Transcriptome similarity analysis was performed using the Genevestigator Signature tool (https://genevestigator.com/gv/doc/signature.jsp). For this analysis, a list of the 330 most significantly altered (greatest log 2 fold-change) transcripts at 12 h was used as input. A heatmap comparing this input to similar transcriptomes was generated using the Heatmapper Expression tool (http:// www2.heatmapper.ca/expression; Babicki et al., 2016).

Trypan Blue Staining
Trypan blue staining of Arabidopsis roots was performed by soaking seedlings in a solution of 10 mg mL 21 trypan blue (Sigma) in water for 20 min. Seedlings were then washed three times with deionized water.

ER Stress Treatments
ER stress treatments of Arabidopsis seedlings were performed by growing seeds directly on MS plates supplemented with TM (Sigma) or DTT (Bio-Rad). For treatment of adult plants, TM was injected directly into one half of an Arabidopsis leaf using a needleless syringe.
In Vitro Acetylation Assays E. coli strain BL21.AI was transformed with a pDEST17 vector carrying NAA50. 5xHIS-tagged NAA50 was purified from E. coli using a Nickel-His column (Sigma). A 5 mL culture was incubated at 37°C for 16 h, and then subcultured to a final volume of 100 mL. The culture was grown until the optical density at 600 nm reached 0.5. Expression of NAA50 was induced by adding 1 mM isopropylthio-b-galactoside and 0.2% (w/v) arabinose to the culture. The culture was then incubated at 30°C for 3 h. Cells were then harvested and resuspended in 8 mL of native purification buffer (50 mM NaH 2 PO 4 , 500 mM NaCl) supplemented with 8 mg lysozyme and a protease inhibitor tablet (Roche). The suspension was then incubated on ice for 30 minutes, and then sonicated. After sonication, the cell debris was pelleted by centrifugation (5,000g, 15 min) at 4°C. The Ni-NTA resin was washed twice with native purification buffer and then incubated with the lysate for 1 h at 4°C. The resin was then washed four times with wash buffer (native purification buffer supplemented with 6 mM imidazole). Fractions were eluted with elution buffer (native purification buffer supplemented with 250 mM imidazole).

Supplemental Data
The following supplemental materials are available.
Supplemental Figure S1. Transgenic Arabidopsis plants carrying the DEX:NAA50-amiRNA transgene display chlorosis and necrosis following dexamethasone treatment.
Supplemental Table S1. Primers used in this study.