1 Nonsense-mediated mRNA decay and light regulate the plant growth–defense trade-off

2 at low temperature
3
4 Running title: NMD-mediated growth–defense trade-off
5

6 Zeeshan Nasim1, Muhammad Fahim2, Katarzyna Gawarecka1, Hendry Susila1, Suhyun Jin1,
7 Geummin Youn1, and Ji Hoon Ahn1
1
8 Department of Life Sciences, Korea University, Seoul 02841, Korea,
2
9 Centre for Omic Sciences, Islamia College Peshawar, Pakistan

10
11 *Contact: Ji Hoon Ahn, email: jahn@korea.ac.kr
12

1
13
14 Abstract (should be <200 words)
15
16 Plants continuously balance growth and defense under ever-changing environmental
17 conditions, but how this trade-off is regulated at low temperature remains unclear. In
18 Arabidopsis thaliana, the nonsense-mediated mRNA decay (NMD) machinery conducts
19 global surveillance and removes aberrant transcripts. Here, we show that when NMD is
20 defective, plants prioritize defense over growth/development at low temperature. The core
21 NMD machinery includes the conserved proteins UP-FRAMESHIFT PROTEIN1 (UPF1)
22 and UPF3 and the Arabidopsis upf1-5 upf3-1 mutant underwent growth arrest, but only at low
23 temperature. As the temperature decreased, differentially expressed genes related to defense
24 were gradually upregulated in upf1-5 upf3-1 mutants, whereas genes related to
25 growth/development were downregulated. The upf1-5 upf3-1 mutant accumulated high levels
26 of salicylic acid (SA) at 16°C, leading to induction of cell death similar to the hypersensitive
27 response; inhibiting SA biosynthesis rescued this phenotype. In the upf1-5 upf3-1 mutant,
28 genes involved in defense and SA biosynthesis/signaling were upregulated in the absence of
29 pathogen attack. The growth arrest of upf1-5 upf3-1 at low temperature requires light and
30 four nucleotide-binding, leucine-rich repeat genes under temperature-dependent NMD
31 regulation are important for this response. Therefore, temperature-dependent NMD and light
32 maintain the balance between growth/development and defense responses to optimize plant
33 fitness.
34
35 Key words: NMD, Arabidopsis, immunity, NLRs, salicylic acid
36
37

2
38 Introduction

39 Plants integrate external cues, including biotic/abiotic stress, light, and temperature
40 during growth and development. Among these environmental cues, temperature affects plant
41 growth, development, and immunity (Hua, 2013). Temperature and plant growth are closely
42 interlinked in an antagonistic manner: at higher ambient temperatures, growth is favored
43 while immunity is generally compromised (Huot et al., 2017), possibly via the transcriptional
44 regulation of genes involved in plant growth (Gangappa et al., 2017). For example, plants
45 grown at 30°C are less able to respond to (and restrict the growth of) bacterial pathogens
46 compared to plants grown at 23°C (Huot et al., 2017), highlighting the negative association
47 between plant immunity and ambient temperature.

48 Eukaryotes use nonsense-mediated mRNA decay (NMD) to eliminate aberrant mRNAs

49 from cells, thus preventing the accumulation of truncated proteins. Key features of NMD
50 targets include premature termination codons (PTCs) generated by mutation, transcriptional
51 errors, or alternative splicing (AS) events (Schweingruber et al., 2013). The core NMD
52 machinery includes the conserved proteins UP-FRAMESHIFT PROTEIN1 (UPF1), UPF2,
53 and UPF3, which recognize targets of NMD (Isken and Maquat, 2008). Genome-wide
54 transcriptome analyses showed that in addition to PTC-containing transcripts, transcripts
55 containing upstream open reading frames (uORFs), long 3 untranslated regions (UTRs), and
′

56 introns in 3 UTRs are also targeted by NMD and that ~20% of endogenous transcripts
′

57 contain such signatures (Drechsel et al., 2013; Kurihara et al., 2009). Consistent with this
58 global regulation, impaired NMD causes various developmental defects, including stunted
59 growth, abnormal leaf morphology, altered flowering time, and enhanced resistance to
60 bacterial infection (Jeong et al., 2011; Riehs-Kearnan et al., 2012).

61 Plant hormones play key roles in numerous growth/developmental processes and

62 mediate defense responses against diverse biotic and abiotic stresses (Bari and Jones, 2009).
63 The plant hormones salicylic acid (SA), jasmonic acid (JA), and ethylene (ET) play critical
64 roles in plant defense against biotic stress. The levels of these hormones start to increase
65 upon pathogen infection (Bari and Jones, 2009). SA mediates defense responses against
66 biotrophic pathogens, whereas the JA and ET pathways primarily function in plant responses
67 to necrotrophic pathogens (Pieterse et al., 2012). In Arabidopsis thaliana, two distinct
68 pathways are involved in SA biosynthesis, i.e., the ICS pathway (involving

3
 69 ISOCHORISMATE SYNTHASE1 [ICS1]) and the PAL pathway (involving
70 PHENYLALANINE AMMONIA LYASE [PAL]), both of which use the same precursor,
71 chorismate. The ICS pathway predominantly contributes to overall SA biosynthesis, as the
72 ICS1-defective sid2 mutant synthesizes only ~10% of wild-type levels of total SA
73 (Wildermuth et al., 2001). However, inhibiting the PAL pathway also significantly decreases
74 SA levels (Chen et al., 2009). SA triggers the monomerization of NONEXPRESSOR OF PR
75 GENES1 (NPR1), leading to its phosphorylation (which is critical for its nuclear localization)
76 and inducing the expression of defense genes that act downstream of SA (Lee et al., 2015;
77 Tada et al., 2008).

78 The plant immune system has successive layers of defense to restrict pathogen
79 infection. One layer uses plasma membrane-localized receptors that detect conserved
80 pathogen-associated molecular patterns (PAMPs). PAMP recognition induces transcriptional
81 reprogramming, referred as PAMP-triggered immunity (PTI) (Jones and Dangl, 2006).
82 However, virulent pathogens interfere with host defense programs by excreting various
83 effectors. These effectors are encountered by ENHANCED DISEASE SUSCEPTIBILITY1
84 (EDS1) and PHYTOALEXIN DEFICIENT4 (PAD4), which regulate the basal defense layer
85 and restrict the spread of pathogens, in part via the SA defense signaling pathway (Fu and
86 Dong, 2013). Basal resistance is further reinforced by intracellular nucleotide-binding,
87 leucine-rich repeat (NLR) receptors, which detect specific pathogen effector molecules to
88 confer effector-triggered immunity (Dodds and Rathjen, 2010). Effector-triggered immunity
89 employs components of the basal resistance machinery to enhance plant defense, often
90 resulting in programmed cell death (PCD), which is also referred to as the hypersensitive
91 response (HR) (Maekawa et al., 2011).

92 In the absence of the effector, NLRs are thought to exist in equilibrium between the ON
93 and OFF states. When the effector is present, it binds to and stabilizes NLRs in the ON state,
94 thus shifting the equilibrium toward an active form to trigger defense responses (Bernoux et
95 al., 2016). NLR gene activity is also regulated at the transcriptional and posttranscriptional
96 levels (Staiger et al., 2013), and upon misexpression, autoimmunity characterized by severe
97 stunting and fitness costs is triggered (Alcázar and Parker, 2011). In addition, AS-coupled
98 NMD constitutes another layer of NLR regulation. AS can produce NLR transcripts
99 containing NMD signatures that are cleared out by NMD, which is consistent with the finding
100 that a number of NLR genes act as NMD substrates (Gloggnitzer et al., 2014).

4
101 Although NMD is a common feature of the global surveillance system in plants, little is
102 known about the importance of a functional NMD pathway in modulating plant
103 growth/development and plant immunity at different temperatures. How temperature-
104 mediated NMD affects the trade-off between growth/development and plant defense is also
105 poorly understood. Here, we demonstrate that plants with defects in NMD show a HR only at
106 a low ambient temperature and that hyperaccumulation of SA is responsible for this HR.
107 When NMD is defective, defense genes are upregulated at low temperature in the presence of
108 light, whereas growth-related genes are downregulated. During this process, a subset of NLR
109 genes is post-transcriptionally regulated by the NMD pathway in a temperature-dependent
110 manner. Our findings suggest that NMD and light are important for maintaining the balance
111 between growth and defense to optimize plant fitness, especially at lower temperatures.

5
112 Results
113
114 The NMD-deficient upf1 upf3 mutant shows arrested growth at low temperature
115 To investigate the role of NMD across a range of ambient temperatures, we analyzed
116 the phenotypes of upf mutants grown in soil and in vitro at 16°C, 23°C, and 27°C. As
117 previously reported (Arciga-Reyes et al., 2006; Nasim et al., 2017), the upf1-5 upf3-1 double
118 mutant (hereafter upf1 upf3) and upf3-1 single mutant (SALK_025175, hereafter upf3)
119 showed a growth-defective phenotype at 23°C (Figure 1A). By contrast, upf1-5
120 (SALK_112922, hereafter upf1) appeared more similar to wild-type plants, except for its
121 narrow, jagged rosette leaves and weak, late flowering. Strikingly, at 27°C, upf1 upf3 no
122 longer exhibited a growth-defective phenotype, except for narrow leaves, whereas at 16°C,
123 the growth of this mutant arrested several days after germination. At 16°C, upf1 upf3 plants
124 showed no further growth and died after generating two cotyledons and a few leaf-like
125 structures. By contrast, at 16°C, the upf1 and upf3 single mutants grew normally (as did wild-
126 type plants) and successfully completed their life cycle. The growth-arrested phenotype was
127 also observed only in upf1 upf3 plants grown on Murashige and Skoog (MS) medium
128 containing sucrose (Figure 1B) (Supplemental Figure 1), excluding the possibility that this
129 phenotype was dependent of sucrose. Analysis of a close-up view of the shoot apex region
130 revealed that an undifferentiated cell mass was produced in the apices of upf1 upf3 plants at
131 16°C (Figure 1C). To further confirm the temperature-dependent growth arrest phenotype, we
132 examined whether the arrested growth of upf1 upf3 plants could be rescued by shifting the
133 temperature from 16°C to 27°C. Two weeks after transfer to 27°C, the upf1 upf3 plants
134 showed wild-type-like, normal growth (Figure 1D), suggesting that this hypersensitivity to
135 low temperature was rescued. The fresh weight of upf1 upf3 plants that were shifted to 27°C
136 was comparable to that of wild-type plants (Figure 1E).
137 We then measured the mRNA levels of SUPPRESSOR WITH MORPHOGENETIC
138 EFFECT ON GENITALIA7 (SMG7), which is post-transcriptionally regulated by NMD
139 (Kerényi et al., 2008), in the upf mutants at different temperatures to confirm that the arrested
140 growth of the mutants is closely associated with low ambient temperature. SMG7 mRNA
141 levels were higher in the upf single mutants than in wild-type plants and even higher in the
142 upf1 upf3 mutant. Furthermore, the SMG7 mRNA levels were similar within each genotype
143 regardless of ambient temperature (Figure 1D), indicating that NMD functionality was
144 reduced in each mutant at all temperatures examined. However, we observed the growth

6
145 arrest phenotype only at 16°C, suggesting that this phenotype is caused by low ambient
146 temperature in the absence of functional NMD.
147
148 Gene Ontology (GO) terms related to defense are highly enriched in upf1 upf3 at 16°C.
149 To explore the mechanism underlying growth arrest in the upf mutants at 16°C, we
150 performed RNA-seq analysis of the mutants grown at 16°C and 27°C and compared the
151 resulting data with our published transcriptome data for these mutants grown at 23°C
152 (GSE87851) (Nasim et al., 2017). Consistent with the notion of NMD as a global
153 phenomenon, many genes were differentially expressed at different temperatures (Figure 2A-
154 C, Supplemental Figure 2). At 16°C, 3,164 and 3,351 genes were upregulated in upf1 and
155 upf3, respectively. A comparison of these genes with upregulated genes in upf1 upf3 showed
156 that 1,218 genes were commonly upregulated in the upf mutants at 16°C (Figure 2A).
157 Similarly, 1,133 and 327 genes were commonly upregulated and downregulated in the upf
158 mutants, respectively, at 23°C (Figure 2B). At 27°C, 699 and 240 genes were commonly
159 upregulated and downregulated, respectively, in the upf mutants (Figure 2C).
160 We investigated whether there were differences in NMD-coupled AS events in the
161 mutants at different temperatures. We detected an overall decrease in splicing events at lower
162 temperatures in the upf mutants (Figure 2D). For instance, the number of exon skipping,
163 intron retention, and alternative acceptor events was lower at 16°C (>2-fold) vs. 27°C.
164 Pearson correlation analysis revealed a stronger correlation between the number of splicing
165 events and temperature in the upf mutants compared to the wild type, except for alternative
166 donor events (Figure 2E).
167 Since only upf1 upf3 showed arrested growth at 16°C, we performed functional
168 classification of the differentially expressed genes (DEGs) in these plants. Interestingly, GO
169 terms related to defense were significantly enriched in upregulated genes in upf1 upf3 at
170 16°C, such as immune response, suggesting that the HR was elicited in these plants at 16°C
171 (Figure 2F). By contrast, GO terms related to growth, for instance, epidermis development
172 and organ development, were significantly enriched among downregulated genes in upf1 upf3
173 at 16°C (Figure 2G). At 23°C, among upregulated genes, highly enriched GO terms included
174 single-organism process and immune system process (Supplemental Figure 3A), and among
175 downregulated genes, highly enriched GO terms included plastid organization (Supplemental
176 Figure 3B). At 27°C, there was no enrichment for defense or cell-death-related terms among
177 upregulated genes (Supplemental Figure 3C), and among downregulated genes, highly

7
178 enriched GO terms included response to stimulus and response to chemicals (Supplemental
179 Figure 3D).
180 We listed the top 20 highly enriched GO terms at each temperature and classified them
181 into three categories: defense, growth/development, and others. For upregulated genes in the
182 upf1 upf3 mutant, of the top 20 GO terms, 14, 3, and 0 terms related to defense were
183 identified at 16°C, 23°C, and 27°C, respectively, whereas no GO terms related to
184 growth/development were identified at any temperature (Figure 2H). For downregulated
185 genes in upf1 upf3, of the top 20 GO terms, 11, 2, and 1 terms related to growth/development
186 were identified at 16°C, 23°C, and 27°C, respectively, whereas 0, 0, and 3 terms related to
187 defense were identified at 16°C, 23°C, and 27°C, respectively. These findings indicate that at
188 low temperature, GO terms related to defense were more highly enriched among upregulated
189 genes, whereas those related to growth were more highly enriched among downregulated
190 genes. These findings suggest that plants invest more energy in defense instead of
191 growth/development at low temperatures when NMD is defective.
192 To further confirm the trade-off between defense and growth, we measured the mRNA
193 levels of defense and growth/development marker genes in the upf1 upf3 double mutant. At
194 16°C, we detected a dramatic increase in transcript levels of defense marker genes (PR1,
195 PR5, PBS3, and PAD4) and a gradual decrease in transcript levels of growth/development
196 marker genes (EXPANSIN A8 [EXP8], XYLOGLUCAN ENDOTRANSGLYCOSYLASE7
197 [XTR7], and SMALL AUXIN UP RNA19 [SAUR19]) (Figure 2I). These antagonistic
198 expression patterns of genes regulating defense responses and growth/development suggest
199 that a substantial proportion of resources are allocated to defense at the expense of
200 growth/development if NMD is impaired at lower temperatures; however, an increase in
201 temperature restores the balance in resource distribution. These results highlight the
202 importance of functional NMD in regulating the balance between immunity and
203 growth/development in a temperature-dependent manner.
204
205 The roles of immunity and SA accumulation in the upf mutant phenotype
206 To investigate whether the growth arrest of the upf mutants at 16°C is associated with
207 changes in plant hormone levels, we measured the levels of phytohormones in the upf
208 mutants using gas chromatography-mass spectrometry. Compared to the wild type, SA levels
209 increased (~3-fold) in upf1 upf3 at 16°C, whereas SA levels only slightly increased in upf1
210 and upf3 (1.3- and 1.6-fold, respectively) (Figure 3A). However, at 23°C and 27°C, SA levels
211 in the upf mutants were comparable to those of wild-type plants. By contrast, the levels of

8
212 jasmonic acid (JA), abscisic acid (ABA), gibberellic acid (GA3), and the auxins phenylacetic
213 acid (PAA) and indole-3-acetic acid (IAA), which play key roles in numerous
214 growth/developmental processes as well as mediating defense responses against diverse
215 biotic and abiotic stresses (Bari and Jones, 2009), were comparable to those in wild-type
216 plants at any temperature. These results suggest that the growth arrest of upf1 upf3 at 16°C is
217 closely associated with increased SA levels.
218 To genetically verify the notion that SA is responsible for the growth arrest at 16°C,
219 we analyzed the phenotypes of the SA-hyperaccumulating mutant lesion simulating disease1
220 (lsd1) (Huang et al., 2010) and the SA-deficient mutant salicylic acid induction deficient2
221 (sid2) (Wildermuth et al., 2001). Notably, the lsd1 mutant exhibited phenotypes similar to
222 those of upf1 upf3 at 16°C, such as severe growth arrest and the inability to complete its life
223 cycle (Figure 3B). However, the lsd1 mutant showed a comparable phenotype to wild-type
224 plants at 23°C and 27°C. By contrast, sid2 mutants showed normal growth at all
225 temperatures. We also examined the SA signaling-defective npr1-2 mutant and found that it
226 exhibited normal growth at 16°C. These results confirm the notion that SA-mediated
227 immunity is involved in the growth arrest phenotype of upf1 upf3.
228 We then investigated whether the expression of plant defense genes was affected in
229 upf1 upf3. We found a dramatic increase (>29-fold) in transcript levels of PATHOGENESIS-
230 RELATED GENE1 (PR1), a marker for systemic acquired resistance (SAR) (Ward et al.,
231 1991), at 16°C (Figure 3C). In the upf1 mutant, PR1 mRNA levels were comparable to those
232 of wild-type plants, whereas the upf3 mutant showed an ~4.1-fold upregulation of this gene at
233 16°C. However, at 23°C and 27°C, PR1 mRNA levels were not significantly increased in any
234 mutant, except for upf1 upf3 (2.8-fold) at 23°C. These results are consistent with the notion
235 that PR1 is induced by SA (Uknes et al., 1992), and they suggest that the immune status of
236 upf1 upf3 is highly activated at 16°C.
237 Elicitation of defense responses leads to the HR, which is characterized by PCD at
238 infection sites (Cui et al., 2015). To investigate whether SA-induced PCD causes the
239 phenotype of upf1 upf3 observed at 16°C, we performed trypan blue staining assays. Strong
240 dark-blue staining was observed in upf1 upf3 leaves only at 16°C, suggesting that almost all
241 leaf cells were dead (Figure 3D). A similar staining pattern was observed in lsd1-2 at 16°C.
242 At 23°C, patches of dark staining were observed in the upf3 and lsd1-2 mutants. By contrast,
243 no strong staining was observed in any plant examined at 27°C. Like wild-type plants, sid2-1
244 plants did not show strong staining at any temperature. These results suggest that strong HR
245 is elicited in upf1 upf3 leaves at 16°C, eventually leading to death.

9
246 Plants respond to pathogen infection by reprogramming the transcriptional machinery
247 to activate defense responses for pathogen restriction/elimination. Most of this transcriptional
248 reprogramming is mediated by the interplay between SA and reactive oxygen species (ROS)
249 (Mammarella et al., 2015; Neuenschwander et al., 1995). To investigate whether the interplay
250 between SA and ROS occurs in the upf1 upf3 mutant, we performed nitro blue tetrazolium
251 (NBT) staining and 3,3 -diaminobenzidine (DAB) staining to detect superoxides and H2O2,
′

252 respectively. We found strongly stained leaves in upf1 upf3 only at 16°C, indicating that a
253 strong burst of both superoxides and H2O2 occurred in this mutant at 16°C (Figure 3E). To
254 further confirm the molecular basis of PCD in upf1 upf3 at 16°C, we analyzed the expression
255 levels of RING1, a gene that induces PCD in a SA-dependent manner (Lee et al., 2011).
256 Consistent with the results of cell death and ROS analyses, quantitative reverse-transcription
257 PCR (qRT-PCR) analysis revealed an ~8.2-fold upregulation of RING1 in upf1 upf3 at 16°C
258 compared to wild-type plants (Figure 3F). At 23°C and 27°C, RING1 mRNA levels were
259 reduced in upf1 upf3 (~0.4-fold and ~0.19-fold at 23°C and 27°C, respectively). The upf1 and
260 upf3 single mutants showed ~2-fold and ~2.6-fold upregulation of RING1, respectively, at
261 16°C. Collectively, these results confirm the notion that SA-mediated immunity is involved
262 in the growth arrest and subsequent constitutively active immune response of upf1 upf3 at
263 16°C, which explains its phenotype.
264
265 Both IC and PAL pathway genes for SA biosynthesis are upregulated in upf1 upf3 in a
266 temperature-dependent manner
267 To explore the molecular mechanism underlying the elevated SA levels in upf1 upf3,
268 we measured the mRNA levels of SA biosynthesis genes. SA is generated by two major
269 biosynthetic pathways: the IC pathway and the PAL pathway. We analyzed the expression
270 levels of ICS1 and ICS2, which act in the IC pathway (Garcion et al., 2008; Wildermuth et
271 al., 2001), and PAL1, PAL2, and PAL3, which act in the PAL pathway (Cochrane et al., 2004;
272 Mauch-Mani and Slusarenko, 1996), via qRT-PCR. At 16°C, compared to the wild type,
273 ICS1 was upregulated 4.7-fold in upf1 upf3, 2.6-fold in upf1, and 2.5-fold in upf3 (Figure
274 4A). However, there were no significant differences in the mRNA levels of ICS2 in the upf
275 mutants at any temperature examined (Supplemental Figure 4), which is consistent with the
276 finding that only ICS1 expression is induced upon pathogen invasion (Wildermuth et al.,
277 2001). At 16°C, PAL1 and PAL2 were upregulated ~2-fold in upf1 upf3, whereas PAL3 was
278 upregulated 1.5-fold in this mutant, compared to the wild type.

10
279 We measured the expression levels of positive regulators of SA biosynthesis, including
280 PAD4, EDS1, CAM-BINDING PROTEIN 60-LIKE G (CBP60G), AVRPPHB
281 SUSCEPTIBLE3 (PBS3), and SAR DEFICIENT1 (SARD1) (Fu and Dong, 2013). The mRNA
282 levels of these genes were significantly higher in upf1 upf3 vs. the wild type only at 16°C
283 (Figure 4B). For instance, we detected an ~3-fold increase in mRNA levels of PAD4 in upf1
284 upf3 at 16°C (and a 1.5- and 1.1-fold increase at 23°C and 27°C, respectively). Furthermore,
285 SARD1 was upregulated ~11-fold in upf1 upf3 at 16°C vs. the wild type (and 3.9-fold and
286 7.7-fold in upf1 and upf3, respectively). These results suggest that higher expression levels of
287 ICS1 and PAL are induced by elevated transcript levels of PAD4, EDS1, CBP60g, PBS3, and
288 SARD1 and that SA levels are elevated in upf1 upf3 due to the activation of both the ICS and
289 PAL pathways in a temperature-dependent manner. A subset of WRKY transcription factor
290 genes (WRKY38, 46, 54, 62, and 70) were also upregulated in the mutants; these genes act
291 downstream of SA in the plant response to biotic stress (Eulgem and Somssich, 2007).
292 WRKY38 was highly induced in upf1 upf3 at 16°C compared to 23°C and 27°C (18-, 6-, and
293 4.5-fold at 16°C, 23°C, and 27°C, respectively) (Figure 4C), which is consistent with the
294 upregulation of SA pathway genes. WRKY46 was highly induced in upf1 upf3 at 16°C
295 compared to 23°C and 27°C (4.5-, 1.8-, and 1.7-fold at 16°C, 23°C, and 27°C, respectively).
296 Similarly, WRKY54, 62, and 70 were highly upregulated in upf1 upf3 at 16°C. Overall, these
297 results indicate that SA pathway genes are upregulated in upf1 upf3 in a temperature-
298 dependent manner.
299
300 SA mediates the hypersensitive response of upf1 upf3 at 16°C
301 To investigate the molecular mechanism underlying the low-temperature-dependent
302 HR, we analyzed a previously published transcriptome dataset from wild-type plants treated
303 with SA (GSE110702) (Ding et al., 2018) and identified genes that are induced by SA
304 treatment. We compared these genes with DEGs upregulated in upf1 upf3 at 16°C and 27°C
305 to identify genes that are regulated by both SA and NMD at different temperatures. These
306 analyses identified 685 and 249 genes that were upregulated in both upf1 upf3 and SA-treated
307 plants at 16°C and 27°C, respectively (Figure 4D). GO analysis of the 685 genes showed that
308 defense and biotic-stress-related terms (for instance, response to biotic stimulus and cell
309 death) were highly enriched, whereas no defense-related GO terms were enriched among the
310 249 genes at 27°C. These findings suggest that SA-induced genes involved in the response to
311 pathogens and the HR are upregulated in upf1 upf3 only at 16°C.

11
312 To further validate the role of SA in the HR of upf1 upf3 at 16°C, we treated the plants
313 with 1-aminobenzotriazole (ABT), which inhibits the PAL pathway of SA biosynthesis (Leon
314 et al., 1995; Wang et al., 2012). Notably, ABT treatment at least partially rescued the HR of
315 upf1 upf3 plants grown at 16°C (Figure 4E). The plants produced green leaves and showed a
316 normal developmental transition to the flowering phase on MS medium supplemented with
317 ABT. The levels of chlorophyll a and b were significantly higher in upf1 upf3 plants treated
318 with ABT compared to untreated upf1 upf3 plants (Figure 4F). Consistent with their rescued
319 phenotype, the mRNA levels of PR1, SARD1, and CBP60g were reduced in ABT-treated
320 upf1 upf3 plants (Figure 4G), which is in sharp contrast to their upregulation in upf1 upf3
321 (Figures 3C and 4B), suggesting that SA is indeed a causal factor leading to the low-
322 temperature-dependent HR of these plants. Our reanalysis of public transcriptome data
323 revealed significant induction of SARD1 and CBP60g transcription upon exogenous SA
324 treatment (Supplemental Figure 5); the SA burst is characterized by the positive feedback of
325 SA to induce the expression of its upstream factors (D'Maris Amick Dempsey et al., 2011).
326 Overall, these results indicate that inhibiting the PAL pathway at least partially rescued the
327 HR of upf1 upf3 plants, confirming the NMD-dependent link between SA and the HR at
328 lower temperature.
329
330 The hypersensitive response of the upf1 upf3 mutants at a low temperature is light
331 dependent
332 In addition to ambient temperature, light is an important environmental factor that
333 affects plant growth and development. We therefore examined whether light affected the HR
334 phenotype of upf1 upf3 plants. Notably, in the absence of light, upf1 upf3 did not show the
335 HR phenotype at 16°C, as indicated by reduced growth (Figure 5A), whereas light (long-
336 day)-grown upf1 upf3 plants exhibited a strong HR. Dark-grown upf1 upf3 plants showed
337 elongated hypocotyl growth at both 16°C and 27°C, as did wild-type plants, except that upf1
338 upf3 exhibited more elongated petioles at 27°C.
339 To examine whether the phenotypic rescue of upf1 upf3 in the dark was due to the
340 reduced expression of plant defense genes and subsequently repressed immune status, we
341 measured the expression levels of PAL genes in wild-type and upf1 upf3 plants at 16°C
342 grown in the light and dark by qRT-PCR. The upf1 upf3 mutant showed significantly reduced
343 mRNA levels of PAL1, PAL2, and PAL3 in the dark, whereas the mRNA levels of these
344 genes were only slightly reduced in wild-type plants (Figure 5B), suggesting that SA
345 accumulation is reduced in dark-grown upf1 upf3 seedlings. We also measured the expression

12
346 levels of PR1 in dark-grown upf1 upf3 seedlings. The increased PR1 expression levels
347 observed in upf1 upf3 at 16°C were nearly abolished in dark-grown upf1 upf3 plants (Figure
348 5C). PR1 gene expression levels were almost undetectable in both light- and dark-treated
349 wild-type plants. Consistent with the downregulation of PAL genes, the mRNA levels of SA
350 biosynthesis genes and SA signaling genes were reduced in upf1 upf3 (Figure 5D). The
351 transcript levels of EDS1, ICS1, and PAD4 were downregulated by ~4.5-, ~8.2-, and ~12.7-
352 fold, respectively, in upf1 upf3 plants in the dark, whereas their downregulation was less
353 severe in wild-type plants. In addition, CBP60g, PBS3, and SARD1 mRNA levels were
354 reduced by ~4.0-, ~16.0-, and ~21.0-fold, respectively, in upf1 upf3 in the dark, although
355 their levels were not reduced in the dark in wild-type plants. These results are consistent with
356 the finding that light is required for the induction of SA biosynthesis and signaling genes
357 such as EDS1 and ICS1 in plants treated with the elicitor-active epitope flg22 (Sano et al.,
358 2014). These results suggest that the upf1 upf3 mutant is unable to accumulate high levels of
359 SA in the absence of light.
360
361 Defense-related genes are upregulated in upf1 upf3 even in the absence of infection at
362 16°C
363 Pathogen invasion triggers plant defense responses leading to PCD, which limits the
364 spread of pathogens (Cui et al., 2015). To gain insight into plant transcriptional
365 reprogramming upon pathogen infection and its relationship to the DEGs in upf1 upf3 at
366 different temperatures, we analyzed transcriptome datasets from plants challenged with
367 various pathogens, including a bacterial pathogen Pseudomonas syringae pv. tomato DC3000
368 (Pst DC3000) (Mine et al., 2018), a fungal pathogen (Botrytis cinerea) (Liu et al., 2015), and
369 a viral pathogen (Turnip crinkle virus [TCV]) (Wu et al., 2016) compared to our
370 transcriptome data.
371 For bacterial pathogens, 278 genes were commonly upregulated in Pst DC3000-treated
372 wild-type and upf1 upf3 plants at 16°C. GO analyses of these 278 genes revealed significant
373 enrichment of GO terms related to PCD, such as cell death (Supplemental Figure 6A),
374 whereas no significantly enriched GO term was found among the 88 genes that were
375 commonly upregulated at 27°C. By contrast, 492 fungal pathogen-induced genes were
376 upregulated in upf1 upf3 plants specifically at 16°C. GO analysis of these 492 genes revealed
377 significant enrichment for defense-related terms such as defense response to fungus, whereas
378 142 genes commonly upregulated at 27°C were not enriched for any GO term (Supplemental
379 Figure 6B). For viral pathogens, 473 common genes induced by TCV infection and

13
380 upregulated in upf1 upf3 at 16°C were significantly enriched for GO terms related to defense
381 (Supplemental Figure 6C), whereas no significantly enriched GO terms were identified
382 among the 116 genes that were upregulated at 27°C. Notably, SMG7 mRNA levels increased
383 in all transcriptome datasets analyzed in response to bacterial infection (1.8-fold), fungal
384 infection (1.6-fold), and viral infection (1.5-fold) (Supplemental Figure 6D), suggesting that
385 pathogen infection might suppress NMD. This finding is consistent with the observation that
386 NMD functionality is repressed upon bacterial infection to trigger defense responses
387 (Gloggnitzer et al., 2014). Our comparison of pathogen-induced genes and DEGs in upf1
388 upf3 at 16°C and 27°C suggested that defense-related genes that are normally induced by
389 pathogen infection are upregulated in upf1 upf3 only at 16°C, even in the absence of
390 infection, ultimately leading to the HR in this mutant.
391
392 SA pathway genes are regulated by NMD in a temperature-dependent manner
393 Genes that function in the SA pathway were upregulated in upf1 upf3 at 16°C (Figure
394 4), raising the possibility that these genes are under posttranscriptional control at low
395 temperature. To explore this possibility, we analyzed the expression levels of SA pathway
396 genes in plants treated with cycloheximide (CHX), which is widely used to inhibit NMD
397 (Carter et al., 1995), in wild-type plants at different temperatures. The qRT-PCR analysis
398 revealed increased expression of CBP60g, ICS1, EDS1, PAD4, PBS3, and SARD1 in plants
399 after CHX treatment at all temperatures (Figure 6A). Notably, these genes were expressed at
400 higher levels in seedlings treated with CHX at lower temperatures vs. the control. In addition,
401 RING1, which likely functions in PCD by facilitating the degradation of PCD activators (Lee
402 et al., 2011), was highly upregulated in CHX-treated plants at lower temperatures vs. the
403 control (Figure 6B), indicating that CHX-mediated inhibition of NMD recapitulates the
404 temperature-dependent upregulation of SA-related genes. These findings suggest that the
405 inhibition of NMD by CHX treatment can mimic the effects of the upf1 upf3 mutations. By
406 contrast, the mRNA levels of PAL1, PAL2, and PAL3 were dramatically reduced upon CHX
407 treatment (Supplemental Figure 7), possibly due to an upstream factor(s) that is not actively
408 translated in CHX-treated plants. Overall, these results confirm the notion that the expression
409 of these SA-related genes is regulated by NMD.
410
411 Temperature-dependent regulation of NLR genes by NMD
412 The defense activation signal is triggered by plant NLR proteins, which directly or
413 indirectly detect effector molecules from the pathogen (Bernoux et al., 2016; Dodds and

14
414 Rathjen, 2010; Jones and Dangl, 2006). A subset of NLR genes have been identified as direct
415 targets of NMD (Gloggnitzer et al., 2014). We therefore aimed to identify NLR genes that
416 were upregulated in upf1 upf3 only at 16°C. As the double mutant showed wild-type-like
417 phenotypes at 27°C (Figure 1A, B), we used genes upregulated in upf1 upf3 at 27°C to
418 narrow down the list of target NLR genes that act only at 16°C. Analysis of transcriptome
419 data revealed 38 NLR genes that are likely under NMD regulation (induced by CHX
420 treatment) and are upregulated in upf1 upf3 specifically at 16°C (Figure 7A, Supplemental
421 Table 3).
422 To assess the temperature dependency of the regulatory effects of NMD on these 38
423 NLR genes, we performed CHX-mediated NMD inhibition at 16°C, 23°C, and 27°C and
424 analyzed the expression levels and splice variants of the 38 NLR genes. For these
425 experiments, we confirmed CHX-induced NMD inhibition by measuring the levels of larger
426 splice variants of AT2G42500, a known NMD target (Drechsel et al., 2013), at all
427 temperatures (Supplemental Figure 8A).
428 After confirming CHX-induced NMD inhibition, we investigated the expression levels
429 of the 38 NLR genes via conventional RT-PCR at different temperatures (Supplemental
430 Figure 8B). The genes were classified into seven groups based on their changes in expression
431 in NMD-inhibited seedlings (Figure 7B). For instance, group 1 (four genes; AT1G72740,
432 ADR1-LIKE2, AT5G41550, and AT1G72910) includes genes whose expression levels
433 gradually increased with decreasing ambient temperature, and group 2 (two genes) includes
434 genes whose expression levels gradually decreased with decreasing ambient temperature. The
435 expression patterns of group 1 NLR genes (which suggest a negative correlation between
436 their expression levels and ambient temperature) at different temperatures with or without
437 CHX treatment and their read coverage graphs in CHX-treated wild-type plants (GSE41432)
438 are shown in Figure 7C. Upon CHX treatment, their expression levels increased at lower
439 temperatures. RT-PCR and read coverage graphs of the remaining genes (in groups 2–7) at
440 16°C, 23°C, and 27°C are shown in Supplemental Figure 8B.
441 As the upf1 upf3 mutant showed the HR only at 16°C, we measured the transcript
442 levels of group 1 NLR genes in the upf mutants. qRT-PCR analysis confirmed the
443 temperature-dependent accumulation of group 1 NLR transcripts in CHX-treated (and thus
444 NMD-inhibited) plants compared to control plants (Figure 7D). Consistent with their
445 expression patterns in CHX-treated plants, the upf1, upf3, and upf1 upf3 mutants also
446 exhibited similar temperature-dependent expression patterns of group 1 NLR genes, with the
447 highest expression at 16°C in upf1 upf3 (Figure 7E). To determine whether these NLRs are

15
448 directly regulated by NMD, we examined the stability of their transcripts after treatment with
449 the transcriptional inhibitor cordycepin (Johnson et al., 2000). Analysis of mRNA
450 degradation kinetics revealed increased half-lives for group 1 NLR genes in NMD-deficient
451 plants. For instance, AT1G72910 had a half-life of ~7.5 h in upf1 upf3 compared to ~0.8 h in
452 wild-type plants (Figure 7F). Similarly, AT1G72940, ADR1-LIKE2, and AT5G41550 showed
453 increased half-lives in upf1 upf3, suggesting that group 1 NLR genes are indeed direct targets
454 of NMD. For this experiment, we used SMG7, which is known to be regulated by NMD
455 (Kerényi et al., 2008), as a positive control (half-life increased by ~3.1 h in upf1 upf3).
456 The transcript levels of the SA pathway genes were significantly reduced in the dark
457 (Figure 5B–D), and darkness rescued the reduced growth of upf1 upf3 at low temperature
458 (Figure 5A). This prompted us to measure the mRNA levels of group 1 NLR genes in the
459 dark. Notably, AT1G72910 was downregulated ~73-fold in upf1 upf3 in the dark (Figure 7G).
460 In addition, the transcript levels of AT1G72940, ADR1-LIKE2, and AT5G41550 were reduced
461 by ~36-, ~3-, and ~3.7-fold, respectively, in the dark. These results indicate that light is
462 required for maintaining the mRNA levels of these group 1 NLR genes, likely to ensure the
463 timely elicitation of plant defense responses. Our results suggest that NMD fine-tunes the
464 transcript levels of a subset of NLR genes in a temperature-dependent manner to prevent the
465 elicitation of immunity even in the absence of invading pathogens.
466

16
467 Discussion
468
469 In this study, we investigated the underlying mechanism of the NMD-mediated
470 growth–defense trade-off at low temperature. We demonstrated that NMD regulates the
471 transcript levels of four NLR genes in a temperature-dependent manner to suppress defense
472 responses in the absence of pathogens, which is likely important for the growth–defense
473 trade-off.
474 Approximately 20% of endogenous Arabidopsis transcripts containing NMD-triggering
475 signatures, such as uORFs, long 3 UTRs, and introns in 3 UTRs, are targeted by NMD
′ ′

476 (Drechsel et al., 2013; Kurihara et al., 2009). Consistent with the global nature of NMD-
477 mediated regulation, many genes were differentially expressed in the upf1 and upf3 single
478 mutants, with both mutations exerting synergistic effects on the overall DEGs in upf1 upf3 at
479 all temperatures (Figure 2), which is in agreement with previous reports (Drechsel et al.,
480 2013; Nasim et al., 2017). GO analysis of upregulated genes in upf1 upf3 specifically at 16°C
481 showed significant enrichment of terms related to SA biosynthesis, SA signaling, and plant
482 defense (Figure 2F), suggesting that the constitutive immune response occurred in the double
483 mutant only at 16°C. The upf1 upf3 mutant accumulated dramatically higher levels of SA
484 only at 16°C (Figure 3A), leading to a strong defense response, as represented by the ~29-
485 fold upregulation of PR1 (Figure 3E). This result is consistent with the finding that NMD-
486 impaired plants grown at 22–23°C accumulate higher levels of SA than the wild type (Jeong
487 et al., 2011). Increased levels of SA are often accompanied by a ROS burst; the interplay of
488 both factors leads to the HR, which is characterized by PCD at the site of infection
489 (Mammarella et al., 2015; Neuenschwander et al., 1995). Consistent with this, histological
490 analysis of cell death and ROS showed that almost all leaf cells of upf1 upf3 plants were dead
491 and had strong ROS bursts (Figure 3). In addition, we found a temperature-dependent
492 increase in the transcript levels of upstream factors and downstream targets of SA
493 biosynthesis in upf1 upf3 (Figure 4). Plants with CHX-mediated inhibition of NMD also
494 showed a similar temperature-dependent gene expression pattern (Figure 6), suggesting a
495 negative correlation between temperature and the expression of these genes.
496 ABT treatment reduces overall SA levels by inhibiting the PAL pathway of SA
497 biosynthesis (Wang et al., 2012; Yang et al., 2018). ABT treatment at least partially rescued
498 the HR of upf1 upf3 at 16°C (Figure 4), suggesting that SA indeed underlies the HR. The
499 upf1 upf3 plants reverted to wild-type-like growth patterns in the dark at 16°C (Figure 5),
500 which is consistent with the classical finding that light influences a plant’s susceptibility to

17
501 pathogen infection (Helms and McIntyre, 1967) and that the accumulation of SA through the
502 PAL pathway and the subsequent activation of plant defense responses require light (Zeier et
503 al., 2004). As light is essential for the induction and amplification of SA-mediated defense
504 responses that result in the HR in upf1 upf3 at 16°C, further studies will be needed to
505 determine how light signaling is incorporated into the growth–defense trade-off mediated by
506 temperature-dependent NMD.
507 Impaired NMD leads to activation of effector-triggered immunity in Arabidopsis
508 (Riehs-Kearnan et al., 2012), which relies on NLR proteins. To ensure the balance between
509 growth and immunity under normal conditions, the expression of NLRs must be tightly
510 regulated at multiple levels to prevent the inappropriate activation of immune responses
511 (Staiger et al., 2013). NMD functions in the posttranscriptional regulation of NLRs
512 (Gloggnitzer et al., 2014). Global expression analysis of NLR genes showed that at least 12
513 genes produce alternatively spliced variants and 15 NLRs contain introns in their 5 or 3
′ ′

514 UTRs (Tan et al., 2007), suggesting that typical NMD signatures are present in a subset of R
515 gene transcripts.
516 In this study, we revealed an additional layer in the temperature-dependent regulation
517 of NLRs, i.e., posttranscriptional regulation by NMD. Although temperature-dependent
518 modulation of defense responses is well known in plants (Hammoudi et al., 2018; Hua,
519 2013), such that elevated temperatures compromise disease resistance (Huot et al., 2017), the
520 underlying molecular mechanism remains elusive. Evidence suggests that NLR genes might
521 function in temperature sensing and transduction, as a mutation in SUPPRESSOR OF NPR1,
522 CONSTITUTIVE1 (SNC1) led to temperature-independent resistance and growth inhibition
523 (Zhu et al., 2010). However, our current understanding of the temperature-dependent
524 elicitation of defense responses is limited to the differential cellular localization of NLR
525 genes. For instance, at a lower temperature (22°C), NLRs are predominantly localized to the
526 nucleus and hence can induce the defense response; by contrast, at an elevated temperature
527 (28°C), the nuclear localization of NLR proteins is reduced, leading to compromised immune
528 status (Zhu et al., 2010). In this study, we identified 38 NLR genes that were upregulated
529 only at low temperature in upf1 upf3 (Figure 7A). Among these, four NLR genes
530 (AT1G72740, ADR1-LIKE2, AT5G41550, and AT1G72910) showed increased transcript
531 accumulation upon NMD inhibition (CHX treatment) in a temperature-dependent manner and
532 were upregulated in the upf1 upf3 mutant only at 16°C (Figure 7). In addition, the mRNAs of
533 these four NLR genes were more stable in upf1 upf3 than in the wild type, revealing that their
534 expression is under NMD regulation.

18
535 Based on our results, we propose a working model for the temperature-dependent HR
536 of upf1 upf3 (Figure 7H). According to this model, a subset of NLR genes (group 1) is
537 regulated by NMD in a temperature-dependent manner. The accumulation of these NLR
538 transcripts in NMD-defective mutants ultimately induces the expression key defense
539 regulators PAD4, EDS1, PBS3, CBP60g, and SARD1 and SA biosynthesis genes (i.e., ICS1
540 and PAL) in the presence of light. Elevated levels of these transcripts lead to the
541 accumulation of high levels of SA only at lower temperatures, which subsequently induces
542 strong defense responses. Moreover, strong ROS bursts in the upf1 upf3 mutant facilitate this
543 SA-induced defense response, resulting in PCD. SA also contributes to the upregulation of its
544 upstream NLR genes via a positive feedback mechanism to amplify the defense response
545 (Xiao et al., 2003). By contrast, in wild-type plants, functional NMD clears out the transcripts
546 of these NLR genes in the absence of invading pathogens, thereby maintaining the delicate
547 balance between immunity and growth/development, ultimately contributing to plant fitness.
548 In conclusion, our findings uncover a novel layer of regulation to explain the
549 relationship among temperature, defense, and NMD. These findings shed light on the role of
550 NMD in the trade-off between temperature and defense under changing ambient
551 temperatures, which could ultimately help ensure food security in the context of global
552 warming. Further studies on NMD-mediated temperature-dependent regulation of group 1
553 NLR genes should increase our understanding of the mechanisms used by plants to maintain
554 the balance between growth and defense responses to optimize plant fitness.
555
556

19
557 Methods
558
559 Plant materials and growth
560 The upf1-5 (SALK_112922), upf3-1 (SALK_025175), and upf1-5 upf3-1 Arabidopsis
561 mutants were previously described (Arciga-Reyes et al., 2006; Hori and Watanabe, 2005;
562 Nasim et al., 2017). For phenotyping, RNA-seq analysis, and qPCR analysis at different
563 temperatures, 7-day-old seedlings (27°C), 8-day-old seedlings (23°C), and 12-day-old
564 seedlings (16°C) grown under standard LD conditions (16:8 h light:dark) were used, as these
565 plants were at identical developmental stages (Boyes et al., 2001). For dark-grown plants,
566 seeds were germinated on MS plates covered with aluminum foil. 1-aminobenzotriazole
567 (Sigma) was used to inhibit SA biosynthesis.
568
569 RNA-seq
570 Approximately 100 seedlings were harvested at Zeitgeber Time 16 (ZT16) and pooled for
571 RNA extraction using Plant RNA Purification Reagent (Invitrogen). For RNA-seq, an
572 Illumina TruSeq Stranded Total RNA Sample Prep kit was used for library preparation
573 according to the manufacturer’s instructions. The libraries were sequenced with an Illumina
574 HiSeq 2000 sequencer. The raw sequencing data are available at NCBI under accession
575 number GSE134538.
576

577 Analysis of public transcriptome data

578 In addition to the RNA-seq data for the upf1 and upf3 single mutants and the upf1 upf3
579 double mutant at 23°C (GSE87851) (Nasim et al., 2017), datasets from wild-type seedlings
580 treated with exogenous SA (GSE110702) (Ding et al., 2018) and CHX (GSE41432)
581 (Drechsel et al., 2013) were used. To compare the transcriptional changes in plants
582 challenged with different types of pathogens, we used datasets from Pst DC3000-infected
583 (GSE88798) (Mine et al., 2018), Botrytis cinerea-challenged (GSE66290) (Liu et al., 2015),
584 and TCV-inoculated plants (GSE85067) (Wu et al., 2016).
585

586 Bioinformatics analysis

20
587 Public transcriptome data were downloaded from NCBI and checked for quality with FastQC
588 (http://www.bioinformatics.babraham.ac.uk/projects/fastqc). The good-quality RNA-seq
589 datasets were subjected to adapter trimming using Sickle (https://github.com/najoshi/sickle)
590 and mapped against the Arabidopsis reference genome TAIR10 using TopHat 2
591 (https://ccb.jhu.edu/software/tophat/index.shtml) with default parameters. The mapped reads
592 were assembled, and normalized transcript levels were determined using Cufflinks version
593 2.2.1 (Trapnell et al., 2012) and visualized using Integrated Genome Browser (IGB)
594 (http://bioviz.org/). Cuffdiff 2 was used to identify DEGs (Trapnell et al., 2012). The output
595 of Cuffdiff was further analyzed and visualized using the R package CummeRbund (Goff et
596 al., 2012). Heatmaps were generated using Papillon (https://github.com/domenico-
597 somma/Papillon). DEGs were defined as genes with fold-change (FC) values of two or more
598 and Fragments Per Kilobase of transcript per Million (FPKM) values of two or more. GO
599 analysis was performed with BINGO (Maere et al., 2005) and PLAZA (Proost et al., 2014).
600
601 Expression analysis
602 qRT-PCR was used to validate the RNA-seq data from the upf mutants. Total RNA was
603 extracted from seedlings at the identical developmental stage at ZT16 at different
604 temperatures. Plant RNA purification reagent (Invitrogen) was used for RNA extraction. The
605 RNA (~2 g) was reverse transcribed into cDNA using MMLV enzyme (ELPIS Biotech).
μ

606 Quantitative expression analysis was performed using 2X A-Star Real Time PCR Master Mix
607 (BioFACT). Two reference genes, PP2AA3 (AT1G13320) and a SAND family gene
608 (AT2G28390), were used to normalize the data (Hong et al., 2010). All qPCR experiments
609 were conducted in three biological replicates, each with three technical replicates. The
610 statistical significance of differences in gene expression levels among samples was assessed
611 using one-way ANOVA with a 0.05 level of significance (95% confidence interval). Details
612 about the primers used in this study are shown in Supplemental Table 1.
613
614 Quantification of phytohormone and chlorophyll contents
615 To determine endogenous phytohormone levels of the upf mutants at different temperatures,
616 gas chromatography (GC) with a mass spectrometry detector (Agilent GC 6890N/MS 5973N)
617 was performed in SIM mode. The calculations were done based on selected internal standards
618 (d4-SA, d7-PAA, d6-ABA, d5-IAA, dihydroxy-JA, d2-GA1, d2-GA4, and d2-GA20). The
619 phytohormones were isolated from 300 mg whole plant samples grown on MS medium for 3
620 weeks (Nehela et al., 2016). Prior to GC analysis, the samples were methylated and

21
621 trimethylsilylated as described (Hijaz and Killiny, 2014). All measurements were done with
622 three biological replicates. Chlorophyll quantification was performed by measuring
623 absorbance at 665, 649, and 470 nm as described (Frye et al., 2001).
624
625 Analysis of cell death
626 Rosette leaves from ~3- to 4-week-old plants were submerged in trypan blue solution
627 (Sigma) for 3 h (van Wees, 2008). To detect superoxide ion (O2−) and hydrogen peroxide
628 (H2O2), 1 mg/ml NBT (pH 7.5) and 1 mg/ml DAB (pH 5.5) (Thordal-Christensen et al.,
629 1997) were used, respectively.
630
631 Determination of mRNA degradation kinetics
632 Cordycepin (150 g/ml) was used to inhibit transcription in 3-week-old wild-type and upf1
μ

633 upf3 plants (Johnson et al., 2000). Leaves were harvested after 0, 1, 2, 3, and 4 h of treatment,
634 and the transcript levels of the genes of interest were determined via qPCR. The half-lives of
635 the mRNAs were calculated by nonlinear least-square regression using GraphPad Prism 7.0.
636
637 Inhibition of translation
638 To inhibit protein synthesis using CHX, a potent NMD inhibitor, wild-type plants grown at
639 23°C for 6 to 7 days were infiltrated with 20 µM CHX and transferred to 16°C, 23°C, and


640 27°C. Samples were harvested 5 h after the temperature shift and analyzed (Hori and
641 Watanabe, 2005).
642
643

22
644 Funding
645 This work was supported by a National Research Foundation (NRF) of Korea grant
646 funded by the Korean government (NRF-2017R1A2B3009624 to JHA) and Samsung Science
647 and Technology Foundation under Project Number SSTF-BA1602-12.
648
649 Author contributions
650 Z.N., M.F., and J.H.A. designed the research; Z.N. performed the experiments; M.F.
651 performed the initial RNA-seq analyses; K.G. performed phytohormone quantification; H.S.,
652 S.J., and G.Y. helped with the experimental work. Z.N., M.F., and J.H.A. wrote the paper.
653 H.S. helped edit the paper; J.H.A. supervised the research.
654
655 Acknowledgments
656 We thank Prof. Park, Ohkmae Kim (Korea University) for kindly providing research
657 materials. We also thank Young-Ja Kim for her technical assistance. The authors declare no
658 conflict of interest.
659

23
660 References
661 Alcázar, R., and Parker, J.E. (2011). The impact of temperature on balancing immune
662 responsiveness and growth in Arabidopsis. Trends in plant science 16:666-675.
663 Arciga‐Reyes, L., Wootton, L., Kieffer, M., and Davies, B. (2006). UPF1 is required for
664 nonsense‐mediated mRNA decay (NMD) and RNAi in Arabidopsis. The plant journal
665 47:480-489.
666 Bari, R., and Jones, J.D. (2009). Role of plant hormones in plant defence responses. Plant
667 molecular biology 69:473-488.
668 Bernoux, M., Burdett, H., Williams, S.J., Zhang, X., Chen, C., Newell, K., Lawrence, G.J.,
669 Kobe, B., Ellis, J.G., and Anderson, P.A. (2016). Comparative analysis of the flax
670 immune receptors L6 and L7 suggests an equilibrium-based switch activation model.
671 The Plant Cell 28:146-159.
672 Boyes, D.C., Zayed, A.M., Ascenzi, R., McCaskill, A.J., Hoffman, N.E., Davis, K.R., and
673 Görlach, J. (2001). Growth stage–based phenotypic analysis of Arabidopsis: a model
674 for high throughput functional genomics in plants. The Plant Cell 13:1499-1510.
675 Carter, M.S., Doskow, J., Morris, P., Li, S., Nhim, R.P., Sandstedt, S., and Wilkinson, M.F.
676 (1995). A regulatory mechanism that detects premature nonsense codons in T-cell
677 receptor transcripts in vivo is reversed by protein synthesis inhibitors in vitro. Journal
678 of Biological Chemistry 270:28995-29003.
679 Chen, Z., Zheng, Z., Huang, J., Lai, Z., and Fan, B. (2009). Biosynthesis of salicylic acid in
680 plants. Plant signaling & behavior 4:493-496.
681 Cochrane, F.C., Davin, L.B., and Lewis, N.G. (2004). The Arabidopsis phenylalanine
682 ammonia lyase gene family: kinetic characterization of the four PAL isoforms.
683 Phytochemistry 65:1557-1564.
684 Cui, H., Tsuda, K., and Parker, J.E. (2015). Effector-triggered immunity: from pathogen
685 perception to robust defense. Annual review of plant biology 66:487-511.
686 D'Maris Amick Dempsey, A.C., Vlot, M.C.W., and Daniel, F.K. (2011). Salicylic acid
687 biosynthesis and metabolism. The Arabidopsis book/American Society of Plant
688 Biologists 9.
689 Ding, Y., Sun, T., Ao, K., Peng, Y., Zhang, Y., Li, X., and Zhang, Y. (2018). Opposite roles
690 of salicylic acid receptors NPR1 and NPR3/NPR4 in transcriptional regulation of
691 plant immunity. Cell 173:1454-1467. e1415.
692 Dodds, P.N., and Rathjen, J.P. (2010). Plant immunity: towards an integrated view of plant–
693 pathogen interactions. Nature Reviews Genetics 11:539.
694 Drechsel, G., Kahles, A., Kesarwani, A.K., Stauffer, E., Behr, J., Drewe, P., Rätsch, G., and
695 Wachter, A. (2013). Nonsense-mediated decay of alternative precursor mRNA
696 splicing variants is a major determinant of the Arabidopsis steady state transcriptome.
697 The Plant Cell 25:3726-3742.
698 Eulgem, T., and Somssich, I.E. (2007). Networks of WRKY transcription factors in defense
699 signaling. Current opinion in plant biology 10:366-371.
700 Frye, C.A., Tang, D., and Innes, R.W. (2001). Negative regulation of defense responses in
701 plants by a conserved MAPKK kinase. Proceedings of the National Academy of
702 Sciences 98:373-378.
703 Fu, Z.Q., and Dong, X. (2013). Systemic acquired resistance: turning local infection into
704 global defense. Annual review of plant biology 64:839-863.
705 Gangappa, S.N., Berriri, S., and Kumar, S.V. (2017). PIF4 coordinates thermosensory growth
706 and immunity in Arabidopsis. Current Biology 27:243-249.
707 Garcion, C., Lohmann, A., Lamodière, E., Catinot, J., Buchala, A., Doermann, P., and
708 Métraux, J.-P. (2008). Characterization and biological function of the

24
709 ISOCHORISMATE SYNTHASE2 gene of Arabidopsis. Plant physiology 147:1279-
710 1287.
711 Gloggnitzer, J., Akimcheva, S., Srinivasan, A., Kusenda, B., Riehs, N., Stampfl, H., Bautor,
712 J., Dekrout, B., Jonak, C., and Jiménez-Gómez, J.M. (2014). Nonsense-mediated
713 mRNA decay modulates immune receptor levels to regulate plant antibacterial
714 defense. Cell host & microbe 16:376-390.
715 Goff, L.A., Trapnell, C., and Kelley, D. (2012). CummeRbund: visualization and exploration
716 of Cufflinks high-throughput sequencing data. R package version 2.
717 Hammoudi, V., Fokkens, L., Beerens, B., Vlachakis, G., Chatterjee, S., Arroyo-Mateos, M.,
718 Wackers, P.F., Jonker, M.J., and van den Burg, H.A. (2018). The Arabidopsis SUMO
719 E3 ligase SIZ1 mediates the temperature dependent trade-off between plant immunity
720 and growth. PLoS genetics 14:e1007157.
721 Helms, K., and McIntyre, G. (1967). Light-induced susceptibility of Phaseolus vulgaris L. to
722 tobacco mosaic virus infection I. Effects of light intensity, temperature, and the length
723 of the preinoculation dark period. Virology 31:191-196.
724 Hijaz, F., and Killiny, N. (2014). Collection and chemical composition of phloem sap from
725 Citrus sinensis L. Osbeck (sweet orange). PloS one 9:e101830.
726 Hong, S.M., Bahn, S.C., Lyu, A., Jung, H.S., and Ahn, J.H. (2010). Identification and testing
727 of superior reference genes for a starting pool of transcript normalization in
728 Arabidopsis. Plant and Cell Physiology 51:1694-1706.
729 Hori, K., and Watanabe, Y. (2005). UPF3 suppresses aberrant spliced mRNA in Arabidopsis.
730 The Plant Journal 43:530-540.
731 Hua, J. (2013). Modulation of plant immunity by light, circadian rhythm, and temperature.
732 Current opinion in plant biology 16:406-413.
733 Huang, X., Li, Y., Zhang, X., Zuo, J., and Yang, S. (2010). The Arabidopsis LSD1 gene
734 plays an important role in the regulation of low temperature‐dependent cell death.
735 New Phytologist 187:301-312.
736 Huot, B., Castroverde, C.D.M., Velásquez, A.C., Hubbard, E., Pulman, J.A., Yao, J., Childs,
737 K.L., Tsuda, K., Montgomery, B.L., and He, S.Y. (2017). Dual impact of elevated
738 temperature on plant defence and bacterial virulence in Arabidopsis. Nature
739 communications 8:1808.
740 Isken, O., and Maquat, L.E. (2008). The multiple lives of NMD factors: balancing roles in
741 gene and genome regulation. Nature reviews Genetics 9:699.
742 Jeong, H.-J., Kim, Y.J., Kim, S.H., Kim, Y.-H., Lee, I.-J., Kim, Y.K., and Shin, J.S. (2011).
743 Nonsense-mediated mRNA decay factors, UPF1 and UPF3, contribute to plant
744 defense. Plant and cell physiology 52:2147-2156.
745 Johnson, M.A., Pérez-Amador, M.A., Lidder, P., and Green, P.J. (2000). Mutants of
746 Arabidopsis defective in a sequence-specific mRNA degradation pathway.
747 Proceedings of the National Academy of Sciences 97:13991-13996.
748 Jones, J.D., and Dangl, J.L. (2006). The plant immune system. nature 444:323.
749 Kerényi, Z., Mérai, Z., Hiripi, L., Benkovics, A., Gyula, P., Lacomme, C., Barta, E., Nagy, F.,
750 and Silhavy, D. (2008). Inter‐kingdom conservation of mechanism of
751 nonsense‐mediated mRNA decay. The EMBO journal 27:1585-1595.
752 Kurihara, Y., Matsui, A., Hanada, K., Kawashima, M., Ishida, J., Morosawa, T., Tanaka, M.,
753 Kaminuma, E., Mochizuki, Y., and Matsushima, A. (2009). Genome-wide
754 suppression of aberrant mRNA-like noncoding RNAs by NMD in Arabidopsis.
755 Proceedings of the National Academy of Sciences 106:2453-2458.
756 Lee, D.H., Choi, H.W., and Hwang, B.K. (2011). The pepper E3 ubiquitin ligase RING1
757 gene, CaRING1, is required for cell death and the salicylic acid-dependent defense
758 response. Plant Physiology 156.

25
759 Lee, H.-J., Park, Y.-J., Seo, P.J., Kim, J.-H., Sim, H.-J., Kim, S.-G., and Park, C.-M. (2015).
760 Systemic immunity requires SnRK2. 8-mediated nuclear import of NPR1 in
761 Arabidopsis. The Plant Cell 27:3425-3438.
762 Leon, J., Shulaev, V., Yalpani, N., Lawton, M.A., and Raskin, I. (1995). Benzoic acid 2-
763 hydroxylase, a soluble oxygenase from tobacco, catalyzes salicylic acid biosynthesis.
764 Proceedings of the National Academy of Sciences 92:10413-10417.
765 Liu, S., Kracher, B., Ziegler, J., Birkenbihl, R.P., and Somssich, I.E. (2015). Negative
766 regulation of ABA signaling by WRKY33 is critical for Arabidopsis immunity
767 towards Botrytis cinerea 2100. Elife 4:e07295.
768 Maekawa, T., Kufer, T.A., and Schulze-Lefert, P. (2011). NLR functions in plant and animal
769 immune systems: so far and yet so close. Nature immunology 12:817.
770 Maere, S., Heymans, K., and Kuiper, M. (2005). BiNGO: a Cytoscape plugin to assess
771 overrepresentation of gene ontology categories in biological networks. Bioinformatics
772 21:3448-3449.
773 Mammarella, N.D., Cheng, Z., Fu, Z.Q., Daudi, A., Bolwell, G.P., Dong, X., and Ausubel,
774 F.M. (2015). Apoplastic peroxidases are required for salicylic acid-mediated defense
775 against Pseudomonas syringae. Phytochemistry 112:110-121.
776 Mauch-Mani, B., and Slusarenko, A.J. (1996). Production of salicylic acid precursors is a
777 major function of phenylalanine ammonia-lyase in the resistance of Arabidopsis to
778 Peronospora parasitica. The Plant Cell 8:203-212.
779 Mine, A., Seyfferth, C., Kracher, B., Berens, M.L., Becker, D., and Tsuda, K. (2018). The
780 defense phytohormone signaling network enables rapid, high-amplitude
781 transcriptional reprogramming during effector-triggered immunity. The Plant Cell
782 30:1199-1219.
783 Nasim, Z., Fahim, M., and Ahn, J.H. (2017). Possible role of MADS AFFECTING
784 FLOWERING 3 and B-BOX DOMAIN PROTEIN 19 in flowering time regulation of
785 Arabidopsis mutants with defects in nonsense-mediated mRNA decay. Frontiers in
786 plant science 8:191.
787 Nehela, Y., Hijaz, F., Elzaawely, A.A., El-Zahaby, H.M., and Killiny, N. (2016).
788 Phytohormone profiling of the sweet orange (Citrus sinensis (L.) Osbeck) leaves and
789 roots using GC–MS-based method. Journal of plant physiology 199:12-17.
790 Neuenschwander, U., Vernooij, B., Friedrich, L., Uknes, S., Kessmann, H., and Ryals, J.
791 (1995). Is hydrogen peroxide a second messenger of salicylic acid in systemic
792 acquired resistance? The Plant Journal 8:227-233.
793 Pieterse, C.M., Van der Does, D., Zamioudis, C., Leon-Reyes, A., and Van Wees, S.C.
794 (2012). Hormonal modulation of plant immunity. Annual review of cell and
795 developmental biology 28:489-521.
796 Proost, S., Van Bel, M., Vaneechoutte, D., Van de Peer, Y., Inzé, D., Mueller-Roeber, B., and
797 Vandepoele, K. (2014). PLAZA 3.0: an access point for plant comparative genomics.
798 Nucleic acids research 43:D974-D981.
799 Riehs-Kearnan, N., Gloggnitzer, J., Dekrout, B., Jonak, C., and Riha, K. (2012). Aberrant
800 growth and lethality of Arabidopsis deficient in nonsense-mediated RNA decay
801 factors is caused by autoimmune-like response. Nucleic acids research 40:5615-5624.
802 Sano, S., Aoyama, M., Nakai, K., Shimotani, K., Yamasaki, K., Sato, M.H., Tojo, D.,
803 Suwastika, I.N., Nomura, H., and Shiina, T. (2014). Light-dependent expression of
804 flg22-induced defense genes in Arabidopsis. Frontiers in plant science 5:531.
805 Schweingruber, C., Rufener, S.C., Zünd, D., Yamashita, A., and Mühlemann, O. (2013).
806 Nonsense-mediated mRNA decay—mechanisms of substrate mRNA recognition and
807 degradation in mammalian cells. Biochimica et Biophysica Acta (BBA)-Gene
808 Regulatory Mechanisms 1829:612-623.

26
809 Staiger, D., Korneli, C., Lummer, M., and Navarro, L. (2013). Emerging role for RNA‐based
810 regulation in plant immunity. New Phytologist 197:394-404.
811 Tada, Y., Spoel, S.H., Pajerowska-Mukhtar, K., Mou, Z., Song, J., Wang, C., Zuo, J., and
812 Dong, X. (2008). Plant immunity requires conformational charges of NPR1 via S-
813 nitrosylation and thioredoxins. Science 321:952-956.
814 Tan, X., Meyers, B.C., Kozik, A., West, M.A., Morgante, M., St Clair, D.A., Bent, A.F., and
815 Michelmore, R.W. (2007). Global expression analysis of nucleotide binding site-
816 leucine rich repeat-encoding and related genes in Arabidopsis. BMC plant biology
817 7:56.
818 Thordal‐Christensen, H., Zhang, Z., Wei, Y., and Collinge, D.B. (1997). Subcellular
819 localization of H2O2 in plants. H2O2 accumulation in papillae and hypersensitive
820 response during the barley—powdery mildew interaction. The Plant Journal 11:1187-
821 1194.
822 Trapnell, C., Roberts, A., Goff, L., Pertea, G., Kim, D., Kelley, D.R., Pimentel, H., Salzberg,
823 S.L., Rinn, J.L., and Pachter, L. (2012). Differential gene and transcript expression
824 analysis of RNA-seq experiments with TopHat and Cufflinks. Nature protocols 7:562.
825 Uknes, S., Mauch-Mani, B., Moyer, M., Potter, S., Williams, S., Dincher, S., Chandler, D.,
826 Slusarenko, A., Ward, E., and Ryals, J. (1992). Acquired resistance in Arabidopsis.
827 The Plant Cell 4:645-656.
828 van Wees, S. (2008). Phenotypic analysis of Arabidopsis mutants: trypan blue stain for fungi,
829 oomycetes, and dead plant cells. Cold Spring Harbor Protocols 2008:pdb. prot4982.
830 Wang, Y., Hu, J., Qin, G., Cui, H., and Wang, Q. (2012). Salicylic acid analogues with
831 biological activity may induce chilling tolerance of maize (Zea mays) seeds. Botany
832 90:845-855.
833 Ward, E.R., Uknes, S.J., Williams, S.C., Dincher, S.S., Wiederhold, D.L., Alexander, D.C.,
834 Ahl-Goy, P., Métraux, J.-P., and Ryals, J.A. (1991). Coordinate gene activity in
835 response to agents that induce systemic acquired resistance. The Plant Cell 3:1085-
836 1094.
837 Wildermuth, M.C., Dewdney, J., Wu, G., and Ausubel, F.M. (2001). Isochorismate synthase
838 is required to synthesize salicylic acid for plant defence. Nature 414:562.
839 Wu, C., Li, X., Guo, S., and Wong, S.-M. (2016). Analyses of RNA-Seq and sRNA-Seq data
840 reveal a complex network of anti-viral defense in TCV-infected Arabidopsis thaliana.
841 Scientific reports 6:36007.
842 Xiao, S., Brown, S., Patrick, E., Brearley, C., and Turner, J.G. (2003). Enhanced transcription
843 of the Arabidopsis disease resistance genes RPW8. 1 and RPW8. 2 via a salicylic
844 acid–dependent amplification circuit is required for hypersensitive cell death. The
845 Plant Cell 15:33-45.
846 Yang, T., Zhu, L.-s., Meng, Y., Lv, R., Zhou, Z., Zhu, L., Lin, H.-h., and Xi, D.-h. (2018).
847 Alpha-momorcharin enhances Tobacco mosaic virus resistance in tobaccoNN by
848 manipulating jasmonic acid-salicylic acid crosstalk. Journal of plant physiology
849 223:116-126.
850 Zeier, J., Pink, B., Mueller, M.J., and Berger, S. (2004). Light conditions influence specific
851 defence responses in incompatible plant–pathogen interactions: uncoupling systemic
852 resistance from salicylic acid and PR-1 accumulation. Planta 219:673-683.
853 Zhu, Y., Qian, W., and Hua, J. (2010). Temperature modulates plant defense responses
854 through NB-LRR proteins. PLoS pathogens 6:e1000844.
855
856

27
857 Figure legends
858
859 Figure 1. NMD-deficient mutants show growth arrest at 16°C.

860 (A and B) Phenotypes of upf and Col-0 plants grown in soil (A) and on MS plates (B) at
861 16°C, 23°C, and 27°C. (C) Comparison of the phenotypes of 4-week-old Col-0 and upf1 upf3
862 plants grown at 16°C. A close-up view of their shoot apical regions is shown on the right
863 (scale bar for these photos = 0.1 cm). Note that upf1 upf3 and Col-0 seeds were sown at the
864 same time. (D) Growth phenotypes of upf1 upf3 plants of the same age grown at 16°C and
865 shifted to 27°C. The photographs were taken 2 weeks after the temperature shift. Right panel
866 shows a close-up view of the apical region of upf1 upf3. (E) Fresh weights of Col-0 and upf1
867 upf3 plants with or without temperature shift to 27°C. Each circle represents individual data
868 points and the bar represents their mean value. (F) Relative expression levels of SMG7 in upf
869 and Col-0 plants at 16°C, 23°C, and 27°C. Each circle represents the average of three
870 technical replicates for one biological replicate (>10 plants) and the bar represents the mean
871 value for the three biological replicates. *, significant at p < 0.05; **, significant at p < 0.01,
872 ***: p ≤ 0.001; ns: not significant, based on one-way ANOVA with a 0.05 level of
873 significance (95% confidence interval). Scale bar = 1 cm.

874
875 Figure 2. GO analysis of DEGs in the upf mutants.

876 (A–C) Venn diagram showing the DEGs in upf1, upf3, and upf1 upf3 plants at 16°C (A),
877 23°C (B), and 27°C (C). (D) The number of alternative splicing events in upf plants at 16°C,
878 23°C, and 27°C, as determined by RNA-seq. (E) Graphs representing Pearson correlation
879 coefficient values for each splicing event for the upf mutants. (F and G) Top 20 GO terms
880 among upregulated (F) and downregulated (G) genes in upf1 upf3 at 16°C. Y-axis represents
881 GO terms; the primary X-axis represents log2 enrichment, whereas the secondary X-axis
882 shows the negative log of p values. GO terms related to defense are shown in red, whereas
883 GO terms related to growth/development are shown in blue. (H) Gradual increase in the
884 number of defense-related GO terms for upregulated genes and growth/development-related
885 GO terms for downregulated genes in upf1 upf3 at lower temperatures. (I) Expression levels
886 of marker genes of defense and growth/development in upf1 upf3 at 16°C, 23°C, and 27°C.
887 The expression levels are represented as log fold-change values.

888

28
889 Figure 3. A strong immune response underlies the phenotype of upf1 upf3 plants 16°C.

890 (A) Quantification of salicylic acid (SA), phenylacetic acid (PAA), indoleacetic acid (IAA),
891 jasmonic acid (JA), abscisic acid (ABA), and gibberellic acid (GA3) levels in the upf mutants
892 at different temperatures. (B) Phenotypes of SA-hyperaccumulating lsd1-2, SA-deficient
893 sid2, and SA signaling-defective npr1-2 plants at different temperatures. Note that upf1 upf3
894 phenocopied the lsd1-2 mutant at 16°C. (C) Increase in PR1 mRNA levels in upf1 upf3 at
895 16°C. (D) Trypan blue staining of the leaves of upf mutants at different temperatures. Note
896 that only upf1 upf3 and lsd1-2 leaves showed dark staining. (E) NBT staining (left panel) to
897 detect O2− and DAB staining (right panel) to detect H2O2 at different temperatures. (F)
898 Increase in RING1 mRNA levels in upf1 upf3 at 16°C. (A, C, F) Each circle represents the
899 average of three technical replicates for one biological replicate (>10 plants) and the bar
900 represents the mean value for the three biological replicates. *, significant at p < 0.05; **,
901 significant at p < 0.01, ***: p ≤ 0.001; ns: not significant, based on one-way ANOVA with a
902 0.05 level of significance (95% confidence interval).

903

904 Figure 4. SA modulates the temperature-dependent HR of upf1 upf3 plants at 16°C.

905 (A) mRNA levels of key factors involved in the IC and PAL pathways in SA biosynthesis at
906 16°C, 23°C, and 27°C. (B) mRNA levels of PBS3, PAD4, EDS1, SARD1, and CBP60g,
907 important upstream factors of SA biosynthesis, at 16°C, 23°C, and 27°C. (C) Relative mRNA
908 levels of WRKY38, 46, 54, 62, and 70, known downstream targets of the SA pathway, in upf
909 mutants at 16°C, 23°C, and 27°C. (D) Venn diagram of genes that are upregulated in both
910 upf1 upf3 and SA-treated plants at 16°C and 27°C. Enriched biological process (BP) GO
911 terms for the upregulated genes in upf1 upf3 at 16°C and 27°C are shown next to the Venn
912 diagram. (E) Phenotypic rescue of upf1 upf3 grown at 16°C by ABT treatment. Note that
913 normal flowers (arrows) formed in ABT-treated upf1 upf3 plants. Scale bar = 1 cm. (F)
914 Levels of chlorophyll a and b (microgram/gram fresh weight) in upf1 upf3 plants grown in
915 the presence or absence of ABT. (G) Decrease in PR1, SARD1, and CBP60g transcript levels
916 in ABT-treated upf1 upf3 plants. (A, B, C, F, G) Each circle represents the average of three
917 technical replicates for one biological replicate (>10 plants) and the bar represents the mean
918 value for the three biological replicates. *, significant at p < 0.05; **, significant at p < 0.01,
919 ***: p ≤ 0.001; ns: not significant, based on one-way ANOVA with a 0.05 level of
920 significance (95% confidence interval).

29
921

922 Figure 5. The HR of upf1 upf3 plants at 16°C is light dependent.

923 (A) Phenotypes of wild-type Col-0 and upf1 upf3 plants grown in the dark and in the light
924 (long days, 16:8 h light:dark cycle, left panel) at 16°C and 27°C. (B and C) Relative
925 expression levels of PAL1, PAL2, and PAL3 (B) and PR1 (C) in Col-0 and upf1 upf3 plants
926 grown in the light and dark at 16°C. (D) Expression profiles of ICS1, PBS3, CBP60g,
927 SARD1, PAD4, and EDS1 in Col-0 and upf1 upf3 plants grown in the light and dark at 16°C.
928 (B, C, D) Each circle represents the average of three technical replicates for one biological
929 replicate (>10 plants) and the bar represents the mean value for the three biological replicates.
930 *, significant at p < 0.05; **, significant at p < 0.01, ***: p ≤ 0.001; ns: not significant, based
931 on one-way ANOVA with a 0.05 level of significance (95% confidence interval).

932
933
934 Figure 6. SA pathway genes are under NMD regulation in a temperature-dependent
935 manner.
936 (A) Expression of CBP60g, ICS1, EDS1, PAD4, PBS3, and SARD1 in CHX-treated seedlings
937 at different temperatures. (B) Expression profile of RING1 under NMD-inhibited conditions
938 at different temperatures. Seedlings were grown at 23°C and transferred to each temperature
939 for CHX treatment. For each gene, the upper panel represents expression in the form of read
940 coverage for that particular gene in CHX- or mock-treated wild-type plants from public
941 transcriptome data (GSE41432). Each circle represents the average of three technical
942 replicates for one biological replicate (>10 plants) and the bar represents the mean value for
943 the three biological replicates. *, significant at p < 0.05; **, significant at p < 0.01, ***: p ≤

944 0.001; ns: not significant, based on one-way ANOVA with a 0.05 level of significance (95%
945 confidence interval).

946

947 Figure 7. Identification of NLR genes regulated by NMD in a temperature-dependent

948 manner.
949 (A) Venn diagram highlighting the 38 NLR genes that are upregulated in upf1 upf3 plants at
950 16°C. (B) Classification of the 38 NLR genes into seven groups based on their expression
951 patterns in CHX-treated wild-type plants at 16°C, 23°C, and 27°C. (C) Expression levels of
952 group 1 NLR genes determined by RT-PCR. Reads coverage graphs from the public

30
953 transcriptome data (GSE41432) upon CHX treatment are shown on the right. (D) qPCR
954 confirmation of the group 1 NLR genes in CHX-induced NMD-inhibited wild-type plants.
955 Seedlings were grown at 23°C and transferred to each temperature for CHX treatment. (E)
956 Expression levels of group 1 NLR genes in the upf mutants at 16°C, 23°C, and 27°C. (F)
957 mRNA degradation kinetics of group 1 NLR genes in wild-type and upf1 upf3 plants. The
958 half-lives of each gene are listed for both genotypes. SMG7 was used as a positive control.
959 (G) mRNA levels of four NLR genes in WT Col-0 and upf1 upf3 plants grown under long-
960 day conditions (LD) in the dark at 16°C. (H) Proposed model for temperature-dependent
961 trade-off between defense and growth in NMD-deficient mutants. At low temperature, when
962 NMD is defective, plant defense is stimulated and growth is suppressed in the light. Under
963 such conditions, the four group 1 NLR gene products induce the expression of defense-
964 related genes. NMD-compromised plants subsequently exhibit high SA levels, which evoke
965 plant defense. The interplay of high SA levels with a strong ROS burst leads to a strong
966 defense response, ultimately leading to plant death. D: defense; G: growth. (D, E, F, G) Each
967 circle represents the average of three technical replicates for one biological replicate (>10
968 plants) and the bar represents the mean value for the three biological replicates. *, significant
969 at p < 0.05; **, significant at p < 0.01, ***: p ≤ 0.001; ns: not significant, based on one-way
970 ANOVA with a 0.05 level of significance (95% confidence interval).

971

31
A B D Col-0 upf1 upf3
16°C 23°C 27°C
upf1
Col-0 upf1 upf3 16°C
upf3

Col-0 16°C
16°C
16>27°C

upf1
upf3 16>27°C

E 400
*
300
Fresh weight
23°C (mg/plant)
***
C 200

100
Col-0 0
16°C 16>27°C 16°C 16>27°C

F Col-0 upf1 upf3

Relative expression levels

1.0 SMG7

0.5
27°C

0.0
23 C
27 C
16 C
23 C
27 C
16 C
23 C
27 C
16 C
23 C
27 C
°C
°
°
°
°
°
°
°
°
°
°
°
16

upf1
upf3 Col-0 upf1 upf3 upf1
upf3

Figure 1
A 16°C B 23°C C 27°C
Upregulated Downregulated Upregulated Downregulated Upregulated Downregulated
upf1 upf3 upf1 upf3 upf1 upf3 upf1 upf3 upf1 upf3 upf1 upf3
908 459 495 809 191 606 532
1010 774 150 151 160
438 564 636
722 362 619
647 327 699 240
579 1218 804 1133 745
545 1697 547 345 386
952 688 447 543

1715 2412 2625 1233 1007

2876

upf1 upf3 upf1 upf3 upf1 upf3 upf1 upf3 upf1 upf3 upf1 upf3

D Exon skipping Intron retention F G

Upregulated at 16°C Downregulated at 16°C
1000 8000 log2(enrichment) log2(enrichment)
Number of events

0.0 0.5 1.0 1.5 2.0 0.0 0.5 1.0 1.5

800 6000 single-organism cellular process single-organism process
600 single-organism process skin development
4000 immune response epidermis development
400 innate immune response single-organism cellular process
200 2000 systemic acquired resistance epidermal cell differentiation
salicylic acid biosynthetic process epithelial cell differentiation
0 0 defense response incompatible interaction epithelium development
(°C) response to salicylic acid single-organism metabolic process
salicylic acid metabolic process tissue development
Alternative acceptor Alternative donor benzene-containing compound metabolic trichoblast differentiation
process
immune system process root epidermal cell differentiation

5000 1100 single-organism metabolic process organ development

Number of events

cellular response to salicylic acid stimulus single-organism biosynthetic

4000 1000 process
salicylic acid mediated signaling pathway system development
900 regulation of defense response metal ion transport
3000
800 regulation of response to stress ion transport
2000 organic cyclic compound metabolic process cellular carbohydrate metabolic
700 process
salicylic acid mediated signaling pathway carbohydrate metabolic process
1000 600 organic acid metabolic process transition metal ion transport
0 500 cellular aromatic compound metabolic root development
process
(°C)
0 10 20 30 40 0 10 20 30 40
-0
f1

f3

-0
f1

f3
up pf1
up pf1
up

up

up

up
ol

ol

-log (p-value) -log (p-value)

f3
f3

u
C

C
u

E Exon skipping Intron retention H I 16°C 23°C 27°C

1.0 1.0 Growth / 3
Defense development Others
Pearson correlation

0.9 0.9 20
PR1
15 2
log10(fold change)

0.8 10 Up
0.8 10
Frequency

PBS3
Alternative acceptor Alternative donor 5 PR5
1
1.0 1.0 0 0 PAD4

0.9 0.8 5
10 0
0.6 10 Down
0.8 EXP8
0.4 15 20 XTR7
0.7 0.3 SAUR19
temperature -1
up-0
upf1
u p f3

up-0
upf1
u p f3
u p f1

u p f1
ol

ol
f3

f3
C

Figure 2
 ng/FW (g) ng/FW (g)

F
A

C
Relative transcript level C

0
500
Relative transcript level

1000
1500
2000
500

0
1000
1500
2000
o
C

0
1
2
3
4
5
0.0
0.5
1.0
1.5
C u l-0

ns
ol ol up upf1
- -

*
f1 pf

ns
up 0

ns
up 0 ***

ns
16°C
up 3

*
**
u p u f1 u p u f1 C f3
o

***

***
f1 pf f1 pf
u l-0

16°C
u 3

16°C
u 3
C pf3 C pf3 up upf1 ns

ns
f1 pf3

JA
ol
SA

ol

23°C
up
u -0 up-0 C f3

ns
o

**

ns
*
up upf1 u p u f1
f1 pf f1 pf u l-0

23°C

23°C
up upf1

RING1
up 3
ns

up 3

ns
C f3 f1 pf

PR1
27°C
C f3 ol up 3
ol f3
u -0 u -0 C

**
up upf1 up upf1 o

ns
f1 pf f1 pf upl-0

***
u p u f1

27°C

27°C
up 3 up 3 f1 pf
ns

f3 ns

16°C
f3 up 3
C f3
o
u l-0

D
up upf1
f1 pf3
ns

ns

23°C
up
ABA
PAA

C f3

16°C
o
u l-0
up upf1
ns

ns

27°C f1 pf

23°C
up 3
f3

C
20k
40k
60k
80k

o
u l-0
27°C up upf1
ns
ns

f1 pf
16°C

up 3
C f3

sid2
lsd1
upf3
upf1
upf3
upf1
o
Col-0 u l-0
up upf1
ns
ns

f1 pf3
23°C

E
GA3
IAA

up
C f3
o

Figure 3
u l-0
up upf1
ns
ns

16°C

f1 pf
27°C

up 3
f3
B

23°C
Col-0
23°C
16°C

NBT staining
Col-0

27°C
upf3
upf3
upf1

upf1 lsd1

16°C
lsd1

sid2

23°C
sid2

DAB staining
27°C

27°C
Col-0

upf3
upf1
upf3
upf1
npr1

Col-0
upf3
upf1 lsd1
sid2
 A B C
2.0 2.5 2.5 2.0 *** 4.0 2.5 WRKY46
ICS1 PBS3 *** EDS1 PAD4 WRKY38 ***
***
*** 2.0 2.0 *** 2.0
1.5 1.5 * 3.0
ns ns
1.5 ns 1.5 ns ns 1.5
** *** ns
ns
ns
1.0
** * 1.0 2.0
**
1.0 ns 1.0 1.0
ns * *
0.5 0.5 1.0 * ns
ns
0.5 0.5 * 0.5
*
ns ns
Relative transcript level

0.0 0.0 0.0 0.0 0.0 0.0

2.0
PAL1
0.8
PAL2 2.0 *** SARD1
2.0
CBP60g 2.0 *** WRKY54
1.5 WRKY62
*** *** ***
1.5 0.6 1.5 1.5 *** 1.5
* *** 1.0
ns * * *
1.0 ns 0.4 ns 1.0 1.0
1.0 ns ns
**
**
ns ns ns
ns
* ns ns
0.5
0.5 0.2 0.5 0.5 * ns
0.5 * ns * ns

0.0 0.0 0.0 0.0 0.0 0.0

1.5 2.0
up 0
u p u f1
f1 pf3
C f3
up 0
u p u f1
f1 pf3
C f3
up 0
u p u f1
f1 pf3
f3

up 0
u p u f1
f1 pf3
C f3
up 0
u p u f1
f1 pf3
C f3
up 0
u p u f1
f1 pf3
f3

up 0
u p u f1
f1 pf3
C f3
up 0
u p u f1
f1 pf3
C f3
up 0
u p u f1
f1 pf3
f3

up 0
u p u f1
f1 pf3
C f3
up 0
u p u f1
f1 pf3
C f3
up 0
u p u f1
f1 pf3
f3
PAL3 WRKY70
-

-
***
up

up

up

up

up

up

up

up

up

up

up

up
ol

ol

ol

ol

ol

ol

ol

ol

ol

ol

ol

ol
C

C
***
1.5
1.0 16°C 23°C 27°C 16°C 23°C 27°C 16°C 23°C 27°C 16°C 23°C 27°C
ns
ns * ns *
1.0 ***
0.5 ns*
ns ns

E
0.5
ABT treatment no ABT
0.0 0.0
up 0
u p u f1
f1 pf3
C f3
up 0
u p u f1
f1 pf3
C f3
up 0
u p u f1
f1 pf3
f3

up 0
u p u f1
f1 pf3
C f3
up 0
u p u f1
f1 pf3
C f3
up 0
u p u f1
f1 pf3
f3
-

-
up

up

up

up

up

up
ol

ol

ol

ol

ol

ol
C

C
16°C 23°C 27°C 16°C 23°C 27°C

Upregulated Upregulated
D upf1 upf3
at 16°C
upf1 upf3
at 27°C
F G
Chlorophyll a Chlorophyll b
biological biological
1.5 PR1 SARD1 CBP60g

Relative transcript level

response to stress process death process 250 ** 100 *
µg/g fresh weight

*
cell death 2608 939 1601 200 80 * *
signal transduction
60 1.0
cellular 233 150
response to biotic 249
process 685 cellular
stimulus
metabolic 100 40
metabolic
1912
process process 0.5
process 50
protein 20
modification
protein process protein protein 0 0 0.0
metabolic modification metabolic
SA-treated WT process process ABT - + - + ABT - + - + - +
process
5.00E-2 5.00E-7
(GSE110702)

Figure 4
A B
Dark conditions Long-day conditions 2.0 PAL1 ***
16°C 16°C 1.5 ns

1.0
0.5
0.0
2.0 ***
PAL2

Fold change
1.5
**
1.0
27°C 27°C 0.5
0.0
2.0 PAL3 ***
1.5
**
1.0
0.5
Col-0 upf1 upf3 Col-0 upf1 upf3 0.0
LD dark LD dark
Col-0 upf1 upf3

C 400 PR1
D
*** *** ICS1 *** ***
1.00 CBP60g
Fold change

300 EDS1
0.75 *
200 *
0.50 **
100 ns
0.25
Relative expression levels

0.0
LD dark LD dark 0.00
LD dark LD dark LD dark LD dark LD dark LD dark
Col-0 upf1 upf3
Col-0 upf1 upf3 Col-0 upf1 upf3 Col-0 upf1 upf3

*** *** ***

1.00 PAD4 PBS3 SARD1
0.75
0.50
*
0.25 ns ns

0.00
LD dark LD dark LD dark LD dark LD dark LD dark
Col-0 upf1 upf3 Col-0 upf1 upf3 Col-0 upf1 upf3

Figure 5
 A GSE41432
Mock
2.5k
CBP60g CHX 600
ICS1 900
EDS1 B 1.5k
RING1
0 0 0 0
2.5k 600 900 1.5k
0 0 0 0
0.3kb 0.3kb
Relative expression levels

Relative expression levels

2.0 2.0 2.0 2.0 ***
***
1.5 1.5 *** 1.5 *** 1.5
1.0 *** 1.0 1.0 1.0
ns ***
*** **
0.5 0.5 ns
0.5 ns 0.5 **
0.0 0.0 0.0 0.0
Mock CHX Mock CHX Mock CHX
PAD4 PBS3 SARD1
400 1.5k 3k
16°C 23°C 27°C
0 0 0
400 1.5k 3k
0 0 0
0.3kb 23°C
***
Relative expression levels

2.0 2.0 2.0

***
1.5 *** 1.5 1.5
1.0 ** 1.0 1.0
ns
0.5 ns
0.5 ns 0.5 ns
ns

0.0 0.0 0.0

Mock CHX Mock CHX Mock CHX Mock CHX Mock CHX Mock CHX Mock CHX Mock CHX Mock CHX

16°C 23°C 27°C 16°C 23°C 27°C 16°C 23°C 27°C

23°C 23°C 23°C

Figure 6