Phytochrome B enhances seed germination tolerance to high temperature by reducing S‐nitrosylation of HFR1

Abstract

Light and ambient high temperature (HT) have opposite effects on seed germination. Light induces seed germination through activating the photoreceptor phytochrome B (phyB), resulting in the stabilization of the transcription factor HFR1, which in turn sequesters the suppressor PIF1. HT suppresses seed germination and triggers protein S‐nitrosylation. Here, we find that HT suppresses seed germination by inducing the S‐nitrosylation of HFR1 at C164, resulting in its degradation, the release of PIF1, and the activation of PIF1‐targeted SOMNUS (SOM) expression to alter gibberellin (GA) and abscisic acid (ABA) metabolism. Active phyB (phyB^Y276H) antagonizes HFR1 S‐nitrosylation and degradation by increasing S‐nitrosoglutathione reductase (GSNOR) activity. In line with this, substituting cysteine‐164 of HFR1 with serine (HFR1^C164S) abolishes the S‐nitrosylation of HFR1 and decreases the HT‐induced degradation of HFR1. Taken together, our study suggests that HT and phyB antagonistically modulate the S‐nitrosylation level of HFR1 to coordinate seed germination, and provides the possibility to enhance seed thermotolerance through gene‐editing of HFR1.

High temperature and photoreceptor phytochrome B antagonistically modulate the S‐nitrosylation level of HFR1 to coordinate seed germination.

Introduction

Light acts as an essential environmental factor to regulate plant growth through leaf photosynthesis, and modulates seed germination through photoreceptors including phytochromes, cryptochromes, phototropins, and UVR8 (Deng & Quail, ¹⁹⁹⁹; Yu et al, ²⁰¹⁰; Rizzini et al, ²⁰¹¹; Viczian et al, ²⁰¹⁷; Liang et al, ²⁰¹⁹). Among these, phytochromes are the major photoreceptors that accelerate seed germination. Photochromes exist in two distinct photoconvertible forms: the biologically active Pfr (far‐red‐light absorbing) and inactive Pr (red‐light absorbing) forms. The active Pfr can be converted back to the inactive Pr form through a slow nonphoto‐induced reaction or markedly faster through irradiation of FR (far‐red) light, and this property allows adjustment of phytochrome activity through an R/FR‐dependent switch (Rockwell et al, ²⁰⁰⁶; Li et al, ²⁰¹¹; Viczian et al, ²⁰¹⁷). Arabidopsis has five phytochromes, namely, phytochrome A (phyA) through phyE. Due to the high abundance of phyB and the absence of phyA in dry seeds, phyB acts as the main phytochrome that promotes R/FR‐mediated seed germination, and further studies have shown that phyB mediates light‐initiated seed germination partially by destabilizing its partner protein PIF1(Sharrock & Quail, ¹⁹⁸⁹; Sharrock & Clack, ²⁰⁰²; Chen & Chory, ²⁰¹¹). PIF1 belongs to the family of basic helix–loop–helix (bHLH) proteins and directly targets SOMNUS (SOM) as a negative regulator for seed germination, to control GA and ABA biosynthesis and responses, and finally suppress seed germination after far‐red light irradiation (Oh et al, ^{2004, 2007}; Kim et al, ²⁰⁰⁸; Gabriele et al, ²⁰¹⁰). Upon exposure to light treatment, phyB translocates from the cytoplasm to the nucleus, where it interacts with PIF1 to induce the latter's degradation via the ubiquitin‐26S proteasome (Shen et al, ²⁰⁰⁵; Castillon et al, ²⁰⁰⁷; Leivar & Quail, ²⁰¹¹). In addition to PIF1, there exists other bHLH protein, such as LONG HYPOCOTYL IN FAR‐RED1 (HFR1), which also mediates R/FR‐induced seed germination. For example, HFR1 accumulates under light but is degraded by the E3 ubiquitin ligase CONSTITUTIVE PHOTOMORPHOGENSIS 1 (COP1) in the dark (Fairchild et al, ²⁰⁰⁰; Jang et al, ²⁰⁰⁵; Yang et al, ²⁰⁰⁵). HFR1 antagonizes PIF1 transcriptional activity by forming a heterodimer complex to prevent the binding of PIF1 to its target DNA. As the feed‐back mechanism, DET1 and COP10 interact with PIF1 and sustains the stability of PIF1 by promoting HFR1 degradation; thus, the PHYB‐DET1‐HFR1‐PIF1 modules provide a comprehensive molecular framework for light‐irradiated seed germination (Shi et al, ^{2013, 2015}), but further investigation is required to obtain more details.

In addition to light, temperature is another critical environmental signal for modulating the developmental processes of plants, including germination, growth, flowering, disease resistance, and hormonal responses (Howarth & Ougham, ¹⁹⁹³; Wigge, ²⁰¹³; Quint et al, ²⁰¹⁶; Vu et al, ²⁰¹⁹). Suppressing seed germination at an unsuitable high temperature (HT) environment is also an adaption strategy for Arabidopsis to establish vegetative and reproductive growth in an appropriate environment (McClung & Davis, ²⁰¹⁰). HT stress induces the expressions of ABA synthesis‐related genes, such as ZEP and NCEDs, and suppresses the expressions of GA biosynthesis‐related genes, such as GA3ox1 and GA3ox2, and these effects subsequently upregulate ABA levels and downregulate GA levels to inhibit seed germination, such as in the response of imbibed seeds to FR irradiation (Toh et al, ²⁰⁰⁸). The photoreceptor phyB can perceive the temperature shift through its temperature‐dependent reversion from its active Pfr state to inactive Pr state. In fact, increases in temperature reduce the abundance of the biologically active Pfr‐Pfr dimer pool of phyB as well as the abundance of phyB nuclear bodies (Jung et al, ²⁰¹⁶; Legris et al, ²⁰¹⁶). However, the interplay between HT and light signals on seed germination and the underlying mechanisms remain poorly understood.

Nitric oxide (NO) is a small molecular signal that plays a critical role in many different physiological processes in plants, including flowering, defense responses, leaf senescence, stomatal movements, hormone signaling, wounding, and responses to different abiotic stresses (Domingos et al, ²⁰¹⁵; Fancy et al, ²⁰¹⁷). NO and its free‐radical derivatives can react with reduced glutathione (GSH) to form S‐nitrosoglutathione (GSNO), which serves as the major cellular reservoir of NO. High accumulation of GSNO can transfer the NO group to cellular protein thiols to form longer lived S‐nitrosothiols (SNOs). This process is called protein S‐nitrosylation (also often referred to as S‐nitrosation), and often affects the stability of a protein, or alters the function or activity of a protein (Fancy et al, ²⁰¹⁷). In Arabidopsis, S‐nitrosoglutathione reductase (GSNOR) sustains the relative low protein S‐nitrosylation level via scavenging SNO accumulation, and the GSNOR null allele shows a high SNO level, along with defects in thermotolerance and plant development (Lee et al, ²⁰⁰⁸). NO also promotes seed germination through the S‐nitrosylation of ABI5 (Albertos et al, ²⁰¹⁵), but whether or how HT regulates light‐initiated seed germination through protein S‐nitrosylation remains less understood. In this study, we first found that HT compromises light‐induced seed germination, accompanied by a high accumulation of SNOs, while phyB enhances seed germination tolerance to HT stress by sustaining high GSNOR activity to reduce SNO accumulation. Furthermore, we discovered that HT induces the degradation of HFR1 through S‐nitrosylation at Cys‐164 of HFR1, resulting in the release of PIF1 and subsequently activating PIF1‐targeted SOM expression to alter GA/ABA metabolism. The transgenic lines overexpressing phyB or GSNOR1 efficiently scavenge HT‐induced SNOs and therefore reduce HT‐induced S‐nitrosylation of HFR1 to stabilize HFR1, ultimately enhancing seed germination under HT stress. Replacement of cysteine 164 in HFR1 with serine (HFR1^C164S) abolishes HT‐induced S‐nitrosylation of HFR1 and enhances the resistance of this protein to degradation, resulting in increased seed germination under HT stress. Overall, our data reveal the function of Cys‐164 in protein S‐nitrosylation modification of HFR1 and its stability, and propose a critical mechanism by which HT and light coordinate the stability of HFR1 through protein S‐nitrosylation, subsequently fine‐tuning SOM expression and downstream GA/ABA metabolism, in order to ensure seeds germinate under favorable environmental condition.

Results

The photoreceptor phyB enhances the seed germination tolerance to HT

To investigate the crosstalk role of light and HT signals during seed germination, we first evaluated the seed germination under light or HT condition using a phyB‐dependent germination protocol. Under phyB‐on condition to activate phyB (Pfr form), both the wild‐type Col seeds and the phyB‐GFP/phyB seeds (overexpressing the phyB‐GFP fusion under the control of the constitutive 35S promoter in a phyB‐9 mutant background) showed a high level of germination, but the germination rate of Col seeds was slower than that of phyB‐GFP/phyB seeds after 72 h light irradiation (Fig 1A; Appendix Fig S1A and B). HT stress clearly reduced the germination percentage of Col seeds, but the seed germination of the phyB‐GFP/phyB line was relatively higher than that of Col seeds. As the control, the phyB mutant (phyB‐9) showed a much lower germination percentage before or after HT stress. Under phyB‐off condition to inactivate phyB (Pr form), the Col and phyB mutants did not germinate, regardless of the presence of HT stress. However, approximately 12.4% of phyB‐GFP seeds still germinated after far‐red irradiation (Fig 1B). It is possible that there was still residual phyB in this line present as the activate form (Pfr form). These results are consistent with previous studies on lettuce (Borthwick et al, ¹⁹⁵²; Frankland & Smith, ¹⁹⁶⁷; Takaki & Zaia, ¹⁹⁸⁴; Huo et al, ²⁰¹³; Lkhamkhuu et al, ²⁰²⁰), and suggest active phyB enhances seed germination tolerance to HT stress.

Figure 1. The reciprocal effects of phyB and HT on seed germination and dormancy.

• A, B
  Germination percentage of Col, phyB‐9 mutant, the crossed line of phyB‐GFP/phyB, phyB ^Y276H ‐GFP/phyB and phyB ^D307A ‐GFP/phyB in response to light and HT treatment. The top diagram indicates the details for light irradiation and temperature treatments used in this experiment. The postharvested seed was treated as the phyB‐on (A) or phyB‐off protocol for 3 days, and the seed germination percentage was calculated. The experiments were repeated three times, and the values are means ± SD of three biological replicates. Bars labeled with different letters in ‐HT group (normal condition, 22°C) or HT (32°C) are significantly different at P < 0.05 (Turkey's test).

Photoreceptor phyB is reported to integrate light and HT signals through a temperature‐dependent switch from the active Pfr form to the inactive Pr form (Jung et al, ²⁰¹⁶; Legris et al, ²⁰¹⁶). Mutation of Tyr276 in the phyB protein to His (phyB^Y276H) can lock phyB in the active state, thus mimicking the constitutively active version of phyB. It forms the subnuclear photobody structure in a light‐independent pattern even under darkness, whereas the mutation of Asp307 to Ala (phyB^D307A) fails to photoconvert from Pr to Pfr, with very lower seeds germination compared to wild‐type seeds after light irradiation (Su & Lagarias, ²⁰⁰⁷; Hu et al, ²⁰⁰⁹; Zhang et al, ²⁰¹³). To validate the function of phyB signals in enhancing seed germination under HT, we then generated constructs of phyB ^Y276H ‐GFP and phyB ^D307A ‐GFP under the control of 35S, and introduced them into the phyB‐9 mutant background, to generate phyB ^Y276H ‐GFP/phyB and phyB ^D307A ‐GFP/phyB to mimic the constitutive active and inactive phyB signals in planta, respectively. Western analysis showed that all of these transgenic lines presented similar immunoblotting signals using anti‐GFP antibody or anti‐phyB antibody (Appendix Fig S2A–C), suggesting these site mutations did not affect their ectopic expression. A short hypocotyl phenotype of phyB ^Y276H ‐GFP/phyB, but longer hypocotyl for phyB ^D307A ‐GFP/phyB, was observed under red‐light condition (Appendix Fig S2D), which is consistent with previous results (Su & Lagarias, ²⁰⁰⁷; Hu et al, ²⁰⁰⁹; Zhang et al, ²⁰¹³). We then checked the germination percentage of these lines under normal condition at 22°C or at a HT of 32°C. As shown in Fig 1, both the phyB ^D307A ‐GFP/phyB and phyB lines showed lower seed germination before or after HT stress compared with the wild‐type Col line. However, the seed germination of the phyB ^Y276H ‐GFP/phyB line was obviously higher than that of phyB‐GFP/phyB and wild‐type Col line either under phyB‐on or under phyB‐off condition. These data suggest that HT repressed light‐irradiated seed germination by influencing the conformation of Pfr/Pr, possibly blocking the downstream phyB signal.

HT and phyB oppositely control SNO levels by targeting GSNOR

HT induces the accumulation of SNOs in Arabidopsis seedlings, leading to an increased sensitivity to heat stress (Lee et al, ²⁰⁰⁸). It is possible that HT also accumulates more SNOs in seeds leading to damage and loss of viability. Here, we measured the SNO content in germinated seeds before and after HT. As shown in Fig 2A, compared with the control seeds under phyB‐on condition, HT treatment gradually induced the accumulation of SNOs. GSNOR is an enzyme in plants that removes GSNO to prevent the accumulation of SNOs. GSNOR activity increased during the early phase of HT and then gradually decreased after 72 h of HT stress (Fig 2B), suggesting that a long‐term HT treatment lowers the activity of GSNOR allowing GSNO concentrations to increase.

Figure 2. HT induced the SNOs accumulation and GSNOR activity in the imbibed seeds.

• A, B
  Time course of HT‐induced accumulation of SNOs and activity of GSNOR in wild‐type Col seed. The imbibed Col seeds were treated under the phyB‐on protocol. At the indicated time points, the seeds, which were maintained in the dark under normal condition (‐HT, 22°C) or HT (32°C), were collected for measurement of the SNO content (A) and GSNOR activity (B). The experiments were repeated three times, and the values are means ± SD of three biological replicates.
• C, D
  The different accumulation of SNOs and germination percentage of Col, GSNOR1‐ox and the mutant of hot5 and nox1 before or after HT stress. All of these imbibed seeds were treated as per the phyB‐on protocol with or without HT stress, and the content of SNOs and the seed germination percentage were calculated after 3 days of treatment. The experiments were repeated three times, and the values are means ± SD of three biological replicates. **P < 0.05 indicates a significant difference using a student's test.

There exists only one GSNOR1 encoding GSNOR in the Arabidopsis genome (Lee et al, ²⁰⁰⁸; Chen et al, ²⁰²⁰). Both hot5 (the GSNOR1‐null allele) and nox1 (endogenous NO over‐accumulation allele) seeds contain high endogenous SNOs levels and have a loss of thermotolerance to HT stress at the seedling stage (Lee et al, ²⁰⁰⁸). Consistent with this, we found both hot5 and nox1 mutants contained higher SNOs content, and presented lower germination after HT stress (Fig 2C and D). On the contrary, the GSNOR1‐overexpressing line (overexpressing GSNOR1‐HA fusion under the control of 35S, 35S:GSNOR1, abbreviated as GSNOR1‐ox, Appendix Fig S3A) showed higher GSNOR enzyme activity, thus efficiently reducing the overaccumulation of SNOs. In contrast to Col, the GSNOR1‐ox line showed higher seed germination under HT stress (Fig 2D). These data suggest that GSNOR activity scavenges HT‐induced endogenous SNOs to enhance seed germination tolerance to HT.

Next, we investigated the effect of phyB on SNO content and GSNOR activity after HT stress. HT did not efficiently accumulate SNOs in the germinated seeds of phyB ^Y276H ‐GFP/phyB (as the active phyB mimic), but induced more SNOs in phyB ^D307A ‐GFP/phyB (as the inactive phyB mimic), compared with phyB‐GFP/phyB and Col under phyB‐on condition (Fig 3A). The GSNOR activity in the germinated seeds of phyB ^Y276H ‐GFP/phyB was also higher than that in phyB‐GFP/phyB and phyB ^D307A ‐GFP/phyB under HT stress (Fig 3B). These results hint that active phyB (Pfr) enhances GSNOR activity to scavenge SNO accumulation in response to HT stress. It is possible that phyB affects the transcriptional level of GSNOR1 or the stability of the protein during HT stress, consequently regulating its overall activity. To confirm this, we first detected the transcription GSNOR1 in a different phyB genetic background and did not find the transcriptional level different among Col, phyB, phyB‐GFP/phyB, phyB ^Y276H ‐GFP/phyB or phyB ^D307A ‐GFP/phyB lines, before or after HT (Appendix Fig S3B). However, HT induced the degradation of GSNOR1 in the phyB ^D307A ‐GFP/phyB, but the GSNOR1 protein was stable in the phyB‐GFP/phyB or phyB ^Y276H ‐GFP/phyB lines (Appendix Fig S3C), hinting that active phyB possibly stabilized GSNOR1. Consistently, using CO‐IP showed that GSNOR1‐HA could be specially immunoprecipitated by phyB‐GFP (Appendix Fig S3D), suggesting that there was an interaction between GSNOR and phyB in planta. We also performed genetic analysis by introducing phyB ^Y276H ‐GFP into the hot5 mutant background to generate phyB ^Y276H ‐GFP/hot5, which showed as low germination as hot5 under HT stress. Similarly, the crossed line of double transgenic phyB ^D307A ‐GFP/GSNOR1‐ox showed as high germination as GSNOR1‐ox after HT stress (Appendix Fig S3E). These data correlate with our results above and suggest that active phyB (Pfr form) stabilized GSNOR1 activity to enhance seed germination, probably by scavenging HT‐induced SNOs.

Figure 3. HT and phyB signal antagonistically regulate endogenous SNOs content, GSNOR activity, and seed germination.

• A, B
  Effects of HT on endogenous SNO levels (A) and GSNOR activity (B) in the imbibed seeds of Col, phyB‐9 mutant, phyB‐GFP/phyB, phyB ^Y276H ‐GFP/phyB, and phyB ^D307A ‐GFP/phyB under phyB‐on condition with HT (32°C) or without HT (22°C) for 3 days, and the endogenous SNO content (A) and GSNOR activity was measured. The experiments were repeated three times, and the values are means ± SD of three biological replicates. Bars labeled with different letters are significantly different at P < 0.05 (Turkey's test).
• C
  PIF1 is required for phyB‐dependent seed germination under HT stress. The seeds of Col, pif1, PIF1‐MYC, phyB‐GFP/PIF1‐MYC, and phyB ^Y276H ‐GFP/ PIF1‐MYC were treated as per the phyB‐on protocol with HT (32°C) or without HT (22°C) for 3 days, and the seed germination percentage was recorded. The experiments were repeated three times, and the values are means ± SD of three biological replicates. Bars labeled with different letters are significantly different at P < 0.05 (Turkey's test).

HT and phyB affect the S‐nitrosylation level of HFR1 regulating its degradation

It is known that red light‐activated phyB targets PIF1 to induce PIF1 degradation controlling seed germination (Oh et al, ²⁰⁰⁴; Yang et al, ²⁰²⁰), while a warm environment reduces the biologically active Pfr‐Pfr dimer pool and nuclear phyB photobodies (Legris et al, ²⁰¹⁶). We hypothesize that HT inactivates phyB, thus blocking PIF1 degradation to suppress seed germination. To test such a possibility, we compared the seed germination of the pif1 null mutant and the transgenic line overexpressing PIF1 (PIF1‐MYC), overexpressing the PIF1‐MYC fusion under the control of the 35S promoter (Appendix Fig S4A). In contrast to Col seeds, the pif1 mutant showed higher germination while PIF1‐MYC seeds showed lower germination when subjected to HT stress (Fig 3C), suggesting that PIF1 negatively regulates seed germination under HT stress. We then introduced phyB‐GFP or phyB ^Y276H ‐GFP into the PIF1‐MYC line (Appendix Fig S4B), and found that both phyB‐GFP/PIF1‐MYC and phyB ^Y276H ‐GFP/PIF1‐MYC lines showed statistically higher seed germination than that of PIF1‐MYC under phyB‐on condition after HT stress. Even phyB ^Y276H ‐GFP/PIF1‐MYC still sustained higher seed germination (Fig 3C), suggesting that constitutively activating the nuclear phyB signal in phyB ^Y276H ‐GFP influences PIF1 activity. However, we found that the PIF1 transcriptional level and the PIF1 protein abundance did not dramatically drop down in the imbibed seeds of phyB ^Y276H ‐GFP after HT stress (Appendix Fig S4C and D), hinting that the higher seed germination tolerance to HT in the phyB ^Y276H ‐GFP mimic is possibly irrelevant to the transcriptional level of PIF1 or its protein abundance, and there exists another component that restrains PIF1 activity.

Apart from phyB, the PIF1 partner HFR1 can sequestrate PIF1 to accelerate seed germination after red‐light irradiation (Shi et al, ²⁰¹³). Thus, we speculated whether the phyB signal suppresses PIF1 activity through HFR1. To confirm such a possibility, we first tested the effect of HT on the expression of HFR1, and found HT did not significantly change the transcriptional level of HFR1 in the imbibed seeds (Appendix Fig S5A). We then generated a transgenic line overexpressing a HFR1‐Flag fusion under the control of the 35S promoter (35S:HFR1‐Flag, abbreviated as HFR1‐Flag, Appendix Fig S5B). In contrast to the stable HFR1 under the normal condition of 22°C (Appendix Fig S6A), HT triggered the statistical degradation of HFR1‐Flag (Fig 4A). These data suggest that HT‐induced degradation of HFR1 releases the activity of PIF1 to suppress seed germination. Meanwhile, treatment with a proteasome inhibitor, MG132, suppressed HT‐induced degradation of HFR1 under HT stress (Appendix Fig S6B), indicating that the HT‐induced degradation of HFR1 depends on the ubiquitination pathway.

Figure 4. HT and phyB signal mediates protein S‐nitrosylation modification of HFR1 and its degradation. The experiments were performed the three biological repeats with similar results, one of them was represented.

• A, B
  Time‐course effect of HT on the S‐nitrosylation and protein degradation of HFR1. The transgenic HFR1‐Flag (A) or HFR1 ^C164S ‐Flag (B) were incubated with 100 μM cycloheximide (CHX) to block de novo protein synthesis and treated as per the phyB‐on condition plus HT stress at 32°C, and the sample was collected at indicated time for HFR1‐Flag or HFR1^C164S‐Flag protein abundance analysis, the anti‐Flag was used to monitor HFR1‐Flag or HFR1^C164S‐Flag abundance. The anti‐RPT5 was used as the loading control (upper panel). As for the S‐nitrosylation analysis, the transgenic seed of HFR1‐Flag or HFR1 ^C164S ‐Flag was incubated with 100 μM of cycloheximide (CHX) and 10 μM of MG132 and treated as phyB‐on protocol, and the total protein from imbibed seeds was extracted, and the formation of S‐nitrosylated HFR1 was monitored using the biotin‐switch method and detected by anti‐biotin antibody. Anti‐RPT5 was used as the loading control (bottom panel).
• C, D
  Cys‐164 in HFR1 meditates the protein S‐nitrosylation of HFR1 and its protein degradation through GSNOR1. The transgenic seeds of HFR1‐Flag, HFR1 ^C164S ‐Flag, HFR1‐Flag/GSNOR1‐ox, and HFR1‐Flag/hot5 were treated as described above for 12 h, and the S‐nitrosylation modification level (C) or the protein degradation rate of HFR1(or HFR1^C164S, D) was measured using anti‐Flag and anti‐biotin antibodies, respectively. Anti‐RPT5 antibody was used as the loading control.
• E, F
  Effect of phyB signal on S‐nitrosylation modification or degradation level of HFR1. The crossed line of HFR1‐Flag/phyB, HFR1‐Flag/phyB‐GFP/phyB, HFR1‐Flag/phyB ^Y276H ‐GFP/phyB, and HFR1‐Flag/phyB ^D307A ‐GFP/phyB was treated as described above for 12 h, and the S‐nitrosylation modification level (D) or the protein degradation rate of HFR1(or HFR1^C164S, F) was measured by anti‐Flag and anti‐biotin antibody, respectively. Anti‐RPT5 antibody was used as the loading control.

As S‐nitrosylation of ABI5 triggers its degradation to accelerate seed germination (Albertos et al, ²⁰¹⁵), we wondered whether HT induces the degradation of HFR1 through S‐nitrosylation. To validate such a possibility, we first checked the S‐nitrosylation level of HFR1 in response to NO or HT. We prepared and purified recombined HFR1‐His protein from E. coli and found that treatment with the NO donor GSNO indeed induced the S‐nitrosylation of HFR1‐His in a dose‐dependent manner, and adding the reductant DTT treatment reduced GSNO‐induced S‐nitrosylated HFR1 (Appendix Fig S6C). Here, we also introduced HFR1‐Flag into the gsnor1 null mutant (hot5) to obtain HFR1‐Flag/hot5, or into the GSNOR1‐ox line to obtain the HFR1‐Flag/GSNOR1‐ox line (Appendix Fig S6D). We observed that HT treatment indeed induced S‐nitrosylation of HFR1 in the HFR1‐Flag line (Fig 4C) The S‐nitrosylated protein abundance of HFR1‐FLAG in the imbibition seed of HFR1‐FLAG/GSNOR1‐ox was lower, but relatively higher in the HFR1‐FLAG/hot5 line when subjected to HT stress (Fig 4C). Thus, these data suggest that GSNOR activity modulates HT‐induced S‐nitrosylation of HFR1. Correspondingly, the degree of protein degradation of HFR1‐Flag in the crossed HFR1‐Flag/hot5 was more obvious than that in HFR1‐Flag, whereas the degree of degradation of HFR1‐Flag was retarded in HFR1‐Flag/GSNOR1‐ox (Fig 4D). These findings suggest that GSNOR1‐mediated S‐nitrosylation of HFR1 determines its degradation under HT stress.

As described above, phyB affected the activity of GSNOR activity and SNO levels after HT. We then evaluated the effect of phyB signals on S‐nitrosylation levels of HFR1 and its degradation under HT stress by crossing HFR1‐Flag/phyB with phyB‐GFP, the active mimic of phyB ^Y276H ‐GFP and inactive mimic of phyB ^D307A ‐GFP. In contrast to the HFR1‐Flag/phyB line in which strong accumulation of S‐nitrosylated HFR1‐Flag was seen, along with very low total HFR1‐Flag after HT exposure, introducing phyB‐GFP into HFR1‐Flag/phyB line (HFR1‐Flag/phyB‐GFP/phyB) efficiently reduced the accumulation of S‐nitrosylated HFR1 and increased the abundance of HFR1‐Flag, although this was not the case when introducing phyB ^D307A ‐GFP into the HFR1‐Flag/phyB line (HFR1‐Flag/phyB ^D307A ‐GFP/phyB). However, expressing the active mimic of phyB ^Y276H ‐GFP completely suppressed the production of S‐nitrosylated HFR1‐Flag, as well as blocking the degradation of HFR1‐Flag in the HFR1‐Flag/phyB ^Y276H ‐GFP/phyB line (Fig 4E and F). We also crossed phyB ^Y276H ‐GFP/phyB with HFR1‐Flag or HFR1‐Flag/hot5 to generate HFR1‐Flag/phyB ^Y276H ‐GFP/phyB and HFR1‐Flag/phyB ^Y276H ‐GFP/phyB/hot5. Unlike HFR1‐Flag/phyB ^Y276H ‐GFP/phyB, HT still induced the degradation of HFR1‐Flag, accompanied with high S‐nitrosylated HFR1 level in the line of HFR1‐Flag/phyB ^Y276H ‐GFP/phyB/hot5 (Appendix Fig S6E and F). These results suggest that active phyB suppresses the S‐nitrosylation status of HFR1 to stabilize HFR1 protein in response to HT stress, and that such an effect requires GSNOR1.

Stabilizing HFR1 hijacks PIF1 activity to promote seed germination

To evaluate the role of HFR1 in controlling seed germination under HT stress, we compared the different seed germination of the hfr1 null mutant (hfr1‐4 and hfr1‐201) and the transgenic HFR1‐Flag line (Appendix Fig S7A–C). In agreement with previous studies, the hfr1‐4 and hfr1‐201 mutants showed lower germination under phyB‐on condition (Shi et al, ²⁰¹³). Here, we found that HT stress strongly inhibited the seed germination of Col and hfr1 mutants, but about 23.3% of the HFR1‐Flag seeds still germinated (Fig 5A), suggesting that HFR1 acts as the positive regulator to enhance thermotolerance of seed germination. As the balance of ABA and GA determines the status of seeds germination or dormancy, we thus measured the endogenous ABA and GA content in seeds of Col, the hfr1 mutant and HFR1‐Flag. HT reduced the bioactive GA4 content in Col and hfr1 mutant seeds, but less so in HFR1‐Flag seeds (Fig 5B). Conversely, HT increased the content of ABA in the Col and hfr1 mutants, but the increased ABA content in the HFR1‐Flag seeds was lower (Fig 5C). These data suggest that HFR1 enhances seed germination under HT through altering GA/ABA metabolism.

Figure 5. HFR1 modulates seed germination under HT through PIF1 and its target SOM.

• A
  The different seed germination percentage among Col, hfr1 mutant and transgenic line overexpressing native HFR1 and its variant HFR1 ^C164S . The Col, hfr1‐4, hfr1‐201, HFR1‐Flag, and HFR1 ^C164S seed was treated as per the phyB‐on protocol with or without HT (+HT or ‐HT) for 3 days, and the seed germination percentage was recorded. The experiments were repeated three times, and the values are means ± SD of three biological replicates. Bars labeled with different letters are significantly different at P < 0.05 (Turkey's test).
• B, C
  The endogenous GA4 and ABA content in the seeds as above with or without HT treatment. All of these seeds as above were treated as per the phyB‐on protocol without HT (HT) or HT (+HT) for 24 h, and the endogenous ABA (B) and GA₄ (C) were measured. The experiments were repeated three times, and the values are means ± SD of three biological replicates. Bars labeled with different letters are significantly different at P < 0.05 (Turkey's test).
• D
  RT‐qPCR analysis of the SOM transcriptional level in the imbibed seeds of Col, hfr1, HFR1‐Flag, PIF1‐MYC, PIF1‐MYC/HFR1‐Flag, and PIF1‐MYC/HFR1 ^C164S ‐Flag treated as phyB‐on condition with or without HT stress for 24 h, PP2A was used for normalization. Values are shown as means ± SD of three biological replicates. Bars labeled with different letters are significantly different at P < 0.05 (Turkey's test).
• E
  The seed germination percentage of Col, PIF1‐MYC, PIF1‐MYC/HFR1‐Flag, PIF1‐MYC/HFR1 ^C164S ‐Flag, HFR1‐Flag/hot5, and HFR1 ^C164S /hot5 were treated as per the phyB‐on protocol with or without HT treatment for 3 days, and the seed germination percentage was calculated. The experiments were repeated three times, and the values are means ± SD of three biological replicates. Bars labeled with different letters are significantly different at P < 0.05 (Turkey's test).

Given that SOM is a main locus altering the balance of GA/ABA to suppress seed germination under HT, and that PIF1 targets SOM and activates its expression (Kim et al, ²⁰⁰⁸; Park et al, ²⁰¹¹), it is possible that HFR1 sequestrates PIF1 to attenuate its binding to the SOM promoter, thus decreasing SOM expression. To verify this, we measured the effects of HFR1 on SOM expression under HT stress. RT‐qPCR analysis showed a relatively higher level of transcripts of SOM in the hfr1‐4 mutant, but lower in HFR1‐Flag line, when compared with the Col seeds after 12 h of HT stress. Overexpressing HFR1 reduced PIF1‐induced SOM expression in the imbibed seed of the PIF1‐MYC/HFR1‐MYC line (Fig 5D). Meanwhile, overexpressing HFR1‐Flag in the PIF1‐MYC line (HFR1‐Flag/PIF1‐MYC) reduced the inhibitory effect of PIF1‐MYC on seed germination under HT stress (Fig 5E). These data indicate that HFR1 antagonizes PIF1 to suppress SOM expression during seed germination under HT. Furthermore, we utilized the transient protoplast system to determine the antagonistic mechanism between HFR1 and PIF1 in regulating SOM expression. As shown in Fig 6A and B, using pSOM:LUC as the reporter, transiently expressing PIF1‐MYC obviously activated the expression of LUC after 12 h of HT treatment. However, coexpressing HFR1‐Flag with PIF1‐MYC suppressed PIF1‐induced SOM expression. In agreement with this, our genetic analysis showed that the double mutants hfr1/som or PIF1‐MYC/som showed as high germination as the single som mutant under HT stress (Appendix Fig S8). These data support the notion that HFR1 counteracts PIF1‐activated SOM expression under HT, which is consistent with the relatively higher germination phenotype of HFR1‐Flag or PIF1‐MYC/HFR1‐Flag lines under HT stress (Fig 5E). During this process, active phyB (Pfr) was also required for the inhibitory effect of HFR1 on SOM expression, as coexpressing HFR1 and PIF1 effectors in the protoplast from phyB‐9 mutant did not so efficiently suppress SOM expression as compared to those protoplasts from the wild‐type Col line (Fig 6B).

Figure 6. HFR1 and PIF1 competitively regulate SOM expression in response to HT stress.

• A, B
  HFR1 and PIF1 antagonistically regulate SOM expression by transient protoplast transformation analysis. The SOMpro:LUC were co‐expressed with PIF1, native or mutated HFR1 effectors in the protoplast from wild‐type or phyB mutant under phyB‐on condition and HT for 12 h (B), and the firefly luciferase and renilla luciferase (LUC/REN) ratio represents SOMpro:LUC activity relative to the internal control (35Spro:REN). Data are means ± SD of three biological replicates. Bars labeled with different letters are significantly different at P < 0.05 (Turkey's test). Schematic diagram of the effectors and reporters containing different truncated SOM promoter are shown (A).
• C, D
  HFR1 and phyB affect the association of PIF1 to the SOM promoter by CHIP‐qPCR analysis. The diagram at the top shows the promoter sequences of the SOM gene. The blue box represents a G‐box element (CACGTG) and the green box represents an E‐box element (CANNTG). Underlying black lines covering G‐ or E‐box indicate various promoter fragments used for PCR. The primer sequences used for ChIP‐PCR are shown in Appendix Table S1. The graph indicates the percentage of DNA fragments that coimmunoprecipitated with PIF1‐myc (closed bars) relative to the input DNAs. The E1 fragment within the exon was used as the internal control. To test the effect of HFR1 on the binding activity of PIF1(C), the imbibed seed of Col, PIF1‐Myc, the crossed PIF1‐MYC/HFR1‐Flag, and PIF1‐MYC/HFR1 ^C164S ‐Flag treated as per the phyB‐on protocol under HT for 12 h were used for CHIP‐qPCR analysis. To test the effect of phyB on the binding activity of PIF1 (D), the imbibed seed of Col, PIF1‐MYC, PIF1‐MYC/phyB, PIF1‐MYC/phyB‐GFP, and PIF1‐MYC/phyB ^Y276H ‐GFP was used. Values are shown as means ± SD of triple experiments. Asterisks indicate significant difference by student's t test (**P < 0.01).

As HFR1 interacts with PIF1 as a heterodimer to prevent the PIF1 from binding to DNA, and PIF1 can bind the promoter region of SOM to initiate SOM expression (Kim et al, ²⁰⁰⁸), we wondered whether HFR1 influences the binding of PIF1 to the promoter region of SOM under HT. As shown in Fig 6C, our CHIP‐qPCR analysis showed that the P3, P4 and P5 fragments, mainly the P3 fragment, within the promoter region of SOM were enriched in the transgenic imbibed PIF1‐MYC seed after 12 h of HT treatment, but coexpressing HFR1‐Flag significantly reduced the enrichment of PIF1‐MYC to these fragments in the cotransgenic seeds of PIF1‐Myc/HFR1‐Flag. Meanwhile, active phyB (Pfr) also suppressed the binding of PIF1 to the SOM promoter under HT stress, as more F3 fragment was enriched in the imbibed seeds of PIF1‐MYC/phyB line than that in PIF1‐MYC/phyB‐GFP or PIF1‐MYC/ phyB ^Y276H ‐GFP lines (Fig 6D).

S‐nitrosylation of HFR1 at Cys‐164 mediates the degradation of HFR1 under HT

Bioinformatic analysis revealed that HFR1 only contains one Cys residue (Cys‐164), which could be a potential target site for S‐nitrosylation (Appendix Fig S9A). To decipher the function of this site, Cys164 in HFR1 was replaced with Ser to generate the mutated HFR1 ^C164S construct. We expressed the recombinant native His‐HFR1 and mutated His‐HFR1^C164S proteins and purified them from E. coli. The biotin switch analysis was subsequently used to check S‐nitrosylation levels of native or mutated HFR1^C164S. Compared with native HFR1 which could be S‐nitrosylated by the NO donor GSNO, mutation of Cys‐164 to Ser completely abolished the S‐nitrosylation of HFR1^C164S in vitro (Appendix Fig S9B). We also generated a transgenic line overexpressing the HFR1 ^C164S ‐Flag fusion driven by the 35S promoter (termed as HFR1 ^C164S ‐Flag) Western analysis using an anti‐Flag antibody showed a strong immunoblotting band, indicating strong expression of HFR1^C164S in these transgenic lines (Appendix Fig S9C). Compared with the rapid degradation of HFR1 in the HFR1‐Flag line, HT‐induced degradation of HFR1^C164S in the HFR1 ^C164S ‐Flag line was remarkably slower (Fig 4B). HT‐induced S‐nitrosylation modification also disappeared (Fig 4C). These data support the notion that HT triggers the S‐nitrosylation modification of HFR1 at Cys‐164, subsequently inducing its degradation.

Overexpressing the degradation‐resistance version of HFR1^C164S enhances seed germination thermotolerance

To explore the biological function of Cys‐164 of HFR1, we compared the seed germination of HFR1‐Flag and HFR1 ^C164S ‐Flag lines after HT stress. Overexpressing HFR1 ^C164S ‐Flag enhanced the seed germination tolerance to HT stress in contrast to the transgenic HFR1‐Flag line or the non‐transgenic Col line (Fig 5A). We also crossed HFR1‐Flag or HFR1 ^C164S ‐Flag with the hot5 mutant and found that HFR1‐Flag/hot5 showed very lower germination, as also seen with the hot5 mutant, after HT stress. However, overexpressing protein‐degradation‐resistance version of HFR1 ^C164S ‐Flag under the hot5 background (HFR1 ^C164S ‐Flag/hot5) obviously enhanced seed germination after HT stress (Fig 5E), although the SNO levels in the germinated seed of HFR1‐Flag appeared to be as high as that of HFR1 ^C164S ‐Flag/hot5 under HT condition (Appendix Fig S10). These data suggest that Cys‐164 contributes to HT‐induced S‐nitrosylation of HFR1 for its degradation, whereas mutation of Cys‐164 abolishes SNO‐induced S‐nitrosylation of HFR1, thus stabilizing HFR1 to enhance seed germination tolerance to HT. In agreement with this, our CHIP‐PCR and transient protoplast transformation experiment showed that overexpressing HFR1 ^C164S ‐Flag, rather than overexpressing HFR1‐Flag, was more efficient at suppressing the binding of PIF1‐Myc to the promoter region of SOM (as shown by CHIP analysis in Fig 6D), thus suppressing SOM expression (as shown by transient protoplast analysis in Fig 6B) to increase GA biosynthesis and suppress ABA biosynthesis (Fig 5B and C). These data confirm that the degradation‐resistant HFR1^C164S shows stronger competition with PIF1 to repress SOM expression, ultimately enhancing seed germination tolerance to HT stress.

Discussion

The photoreceptor phyB integrates light and temperature signals for seed germination via GSNOR

Accumulated evidence demonstrates that the photoreceptor phyB and its downstream PIFs play a central role in integrating HT and light signals (Leivar & Quail, ²⁰¹¹; Balcerowicz, ²⁰²⁰). Due to the presence of phyB and the lack of phyA in dry seeds, phyB‐mediated signaling mainly promotes germination in imbibed seeds in the early stages (Shinomura et al, ¹⁹⁹⁴). In this study, we investigated the reciprocal effects of HT and phyB on seed germination, and found phyB signaling enhanced seed germination tolerance to HT stress, as the transgenic line overexpressing phyB showed a high seed germination percentage, and the phyB mutant showed very lower germination, either under phyB‐on or phyB‐off condition (Fig 1A and B). Thus, these data are in agreement with the previous opinion that phyB functions as the thermosensor (Jung et al, ²⁰¹⁶; Legris et al, ²⁰¹⁶), and suggest that HT integrates phyB signals to coordinate seed germination. It has been reported that ambient HT reduces the abundance of the active Pfr form of phyB along with the abundance of the associated nuclear bodies, thus attenuating the degradation of its downstream target PIF4, resulting in the long hypocotyl phenotype called as thermomorphogenesis (Jung et al, ²⁰¹⁶; Legris et al, ²⁰¹⁶). Expressing the thermally stable variants of phyB repressed the HT‐induced reduction of nuclear body size as well as presenting little thermoresponsive elongation growth (Legris et al, ²⁰¹⁶). In this study, we also compared the different seed germination of the thermostable variant (phyB ^Y276H ‐GFP) to mimic light‐independent active phyB status, and the thermolabile variant (phyB ^D307A ‐GFP), to mimic inactive phyB status. In response to ambient HT stress, we found ecotypic expressing the thermostable phyB ^Y276H ‐GFP enhances seed germination tolerance for HT, and vice versa for the thermolabile phyB ^D307A ‐GFP (Fig 1A and B). These findings further confirm the previous conclusion that phyB acts as the thermosensor to perceive ambient HT (Jung et al, ²⁰¹⁶; Legris et al, ²⁰¹⁶), and extends the function of phyB in modulating seed germination thermotolerance, in addition to promoting hypocotyl elongation in response to ambient HT. Certainly, excepting phyB, other factors such as PIF7 RNA hairpin and ELF3 have also been suggested as thermosensors to perceive the ambient high stress during thermomorphogenesis (Chung et al, ²⁰²⁰; Fiorucci et al, ²⁰²⁰; Jung et al, ²⁰²⁰). Therefore, their possible roles in modulating HT‐induced seed thermoinhibition needs to be investigated in future.

A previous study found that the GSNOR1‐null mutant hot5 showed an increased SNO content and defects in thermotolerance (Lee et al, ²⁰⁰⁸). In this study, we found that the hot5 mutant showed a lower germination rate after HT stress than the wild‐type Col seeds. The nox1 mutant with a high level of SNOs also showed lower seed germination after HT stress (Fig 2). Conversely, the overexpression of GSNOR1 increased the seed germination with lower SNO production under HT stress, suggesting GSNOR activity positively controls seed germination under HT stress through efficiently scavenging HT‐induced SNO over‐accumulation. The main function of GSNOR in plants is to metabolize GSNO and thereby avoid protein S‐nitrosylation to alter protein stability and functions (Lee et al, ²⁰⁰⁸). Here, we found that HT repressed the activity of GSNOR and increased the SNO levels in imbibed seeds, which correlated to the lower germination viability of the hot5 mutant under HT. On the contrary, expressing the thermostable variant of phyB ^Y276H ‐GFP obviously reduced HT‐induced SNO production and sustained a high level of GSNOR activity, resulting into their high seed germination under HT. Consistently, overexpressing the thermolabile variant of phyB ^D307A ‐GFP showed relative lower GSNOR activity under HT (Fig 3A and B), suggesting that thermostability of phyB activity enhances the activity of GSNOR, subsequently scavenging HT‐induced SNO generation, thus avoiding the downstream protein S‐nitrosylation and sustaining normal protein function for seed germination in response to HT stress. Supporting this conclusion, our genetic analysis also revealed that overexpressing thermostable phyB ^Y276H ‐GFP in the hot5 mutant background showed as lower germination as that of the hot5 mutant, whereas co‐overexpressing GSNOR1 with the phyB ^D307A ‐GFP (phyB ^D307A ‐GFP/GSNOR1) background obviously enhances seed germination under HT (Appendix Fig S3E). CO‐IP result also demonstrated the interaction of phyB and GSNOR1 in planta (Appendix Fig S3D). These data strongly support the notion that the effect of phyB on enhancing seed germination thermotolerance to HT dependents on GSNOR activity. Because Cys‐10 of GSNOR1 is targeted by GSNOR itself for S‐nitrosylation under hypoxia stress (Zhan et al, ²⁰¹⁸), inducing its conformational change for ATG8‐mediated selective autophagy of GSNOR1, it is possible that HT‐induced ROS/RNS might trigger S‐nitrosylation of GSNOR1 for its degradation, consequently removing GSNOR activity during HT stress. However, active phyB interacts with GSNOR1(Appendix Fig S3D), thus protecting GSNOR1 from degradation. Therefore, the stabilized GSNOR1 potentially further scavenges HT‐induced SNOs and thus sustaining a suitable cellular redox status under HT stress. As the activity of phyB is also regulated by posttranslational modifications, such as phosphorylation and SUMOation (Sadanandom et al, ²⁰¹⁵), the next step would be to explore whether GSNOR‐mediated protein S‐nitrosylation, or indeed oxidation by ROS, also controls the phyB activity under HT stress.

S‐nitrosylation of HFR1 coordinates seed germination or dormancy in response to HT

Our results above showed that active phyB enhances the seed germination tolerance to HT stress, suggesting that phyB signaling triggers its downstream components controlling seed germination. PIF1, as a typical bHLH transcriptional factor, is targeted by phyB for its degradation, thus initiating seed germination after light irradiation (Shen et al, ²⁰⁰⁵; Castillon et al, ²⁰⁰⁷; Yang et al, ²⁰²⁰). Here, we found that the phyB ^Y276H ‐GFP line, as the constitutive active phyB mimic, can efficiently enhance seed germination under HT, even under the PIF1‐MYC background (Fig 3C). The combined physiological, genetic, and transient transformation experiments revealed that the HFR1‐PIF1 heterodimer targets SOM to activated its expression under HT, ultimately altering ABA/GA metabolism for seed thermoinhibition (Figs 5 and 6). These data suggest the novel function of PIF1 to be downstream of phyB to suppress seed germination in response to HT stress, and not only to act as a negative factor to suppress light‐induced seed germination (Kim et al, ²⁰⁰⁸; Shi et al, ²⁰¹³; Jiang et al, ²⁰¹⁶; Majee et al, ²⁰¹⁸). However, light did not obviously trigger PIF1 degradation under HT condition (Appendix Fig S4); possibly HT suppressed light‐dependent photoconversion of the inactive Pr form to the active Pfr form of phyB.

Given that PIF1 associates with seed thermoinhibition, and active phyB enhances seed germination under HT, but that HT does not affect the PIF1 protein abundance, we propose the existence of other factors rather than phyB to restrain PIF1 activity for seed germination under HT stress. Thus, we turned our attention to HFR1, which was previously reported to sequestrate PIF1 activity (Shi et al, ^{2013, 2015}). Here, we indeed found the hfr1 null mutant showed lower germination with high SOM expression and altered GA/ABA levels under HT stress, while overexpression of HFR1 showed high seed germination with lower SOM expression (Fig 5A–C). Further biochemical and genetic analysis also showed that HFR1 counteracts PIF1 to modulate SOM expression (Fig 5D and E). Therefore, our data suggest that the PIF1/HFR1 heterodimer is also involved in regulating seed germination in response to HT stress, not just participating into light‐induced seed germination. Here, we also noticed HT induced the degradation of HFR1, which possibly depends on the proteasome‐mediated ubiquitin pathway (Fig 4). This might explain how HT induced the transactivity of PIF1 on SOM expression through degrading HFR1 to release PIF1. NO has been reported to trigger S‐nitrosylation of the ABI5 protein to induce its degradation, thus promoting seed germination (Albertos et al, ²⁰¹⁵). Similarly, we also found HT‐induced SNOs trigger the S‐nitrosylation of HFR1 at Cys‐164. Replacing Cys‐164 with Ser in HFR1 abolished its S‐nitrosylation during HT stress, which also attenuated HT‐induced HFR1 degradation (Fig 4). This suggests that S‐nitrosylation of HFR1 is a necessary process for its degradation under HT stress. Possibly S‐nitrosylated HFR1 enhances its interaction with a E3 ligase, such as COP1, to trigger proteasome‐dependent degradation of HFR1 (Duek et al, ²⁰⁰⁴; Jang et al, ²⁰⁰⁵; Yang et al, ²⁰⁰⁵). Recently, evidence has also demonstrated that a HFR1 variant from Cardamine hirsuta is more stable due to its low binding affinity to COP1, therefore enhancing HFR1 total activity but attenuating PIFs activity during the shadow response (Paulisic et al, ²⁰²¹). Other components, such as COP10‐DET1‐DDB1(CDD)‐based E3 ligase, are also possibilities to facilitate the degradation of HFR1 in response to HT (Shi et al, ²⁰¹⁵). Consistent with this, S‐nitrosylation of GSNOR1 is necessary for autophagy‐mediated degradation of itself during hypoxic stress (Zhan et al, ²⁰¹⁸). Similarly, S‐nitrosylated ABI5 also showed high affinity to the E3 ligase, such as KEG, for its degradation during seed germination (Albertos et al, ²⁰¹⁵). It is possible that S‐nitrosylated HFR1 recruited more E3 ligase, possibly COP1 or the DDB1‐CUL4 complex, for its degradation. Here, we found that the degradation‐resistance HFR1 variant (HFR1^C164S) efficiently sequestrated PIF activity, therefore suppressing PIF1‐induced SOM expression for higher seed germination compared with the native HFR1 under HT stress (Fig 4). The CHIP analysis also revealed that overexpressing HFR1^C164S attenuated the binding ability of PIF1 to the SOM promoter, thus reducing PIF1‐mediated SOM expression (Fig 6C), which may explain why the degradation‐resistance HFR1 variant enhances seed germination tolerance to HT stress. These data suggest the multifaceted function of phyB in enhancing seed germination thermotolerance. Meanwhile, previous study showed that NO causes the S‐nitrosylation of ABI5, leading to the degradation of ABI5 during seed germination by recruiting CUL4 and KEG E3 ligase complexes (Albertos et al, ²⁰¹⁵). However, HT induced the accumulation of ABI5 (Ren et al, ²⁰²²), and ABI5 can bind to the promoter of SOM to activate SOM expression (Lim et al, ²⁰¹³). Simultaneously, HT induced the overaccumulation of NO and its metabolic SNO derivatives. It seems that NO signal triggers the S‐nitrosylation of ABI5 for its degradation, with such a process possibly being blocked under HT condition that induced extreme accumulation of SNO, probably inactivating KEG activity. Unlike under normal condition that ABI5 recruits KEG E3 ligase for its degradation, HT induced the degradation of HFR1 through recruiting COP1 E3 ligase. It appears that S‐nitrosylation of different protein substrates recruits different E3 ligases for substrate protein degradation in response to changeable environment temperatures.

In summary, we reported the molecular mechanisms by which the photoreceptor phyB perceives the environment temperature to ensure seed germination at favorable condition. Based on our data, we propose a working model to illustrate the function of phyB in sensing the ambient environmental temperature to coordinate seed germination or dormancy. As shown in Fig 7, under normal condition, light‐activated phyB (Pfr form) shifts into the nucleus to stabilize HFR1, thus triggering the degradation of PIF1 to block PIF1‐dependent SOM expression, ultimately increasing GA content and decreasing ABA levels to initiate seed germination. However, ambient HT impairs the active phyB signal to reduce phyB‐mediated PIF1 degradation, simultaneously suppressing GSNOR activity to induce the overaccumulation of SNOs, which triggers S‐nitrosylation of HFR1 for its degradation and disrupts the PIF1‐HFR1 heterodimer to release PIF1 activity. As a result, the accumulated PIF1 activates SOM expression to suppress seed germination. During this process, mimicking the active phyB status can sustain high activity of GSNOR to efficiently scavenge HT‐induced SNO levels, thus antagonizing HT‐induced S‐nitrosylation of HFR1 to stabilize HFR1, which then detains PIF1 activity to suppress SOM expression, ultimately enhancing seed germination. Furthermore, we identify Cys‐164 of HFR1 as the sole site for HT‐induced S‐nitrosylation responsible for its degradation. Substituting Cys‐164 with Ser abolishes HT‐induced S‐nitrosylation of HFR1 and its degradation The stabilized variant of HFR1 efficiently binds to PIF1 to repress its transactivation activity on SOM expression, ultimately enhancing seed germination under HT stress. Overall, our study highlights the critical role of phyB in coordinating light and temperature signals to fine‐tune seed germination, and expands our understanding on protein S‐nitrosylation in controlling seed germination or plant fitness in response to a harsh environment. Our findings also propose the new route to improve seed germination tolerance to HT via gene‐engineering modification of HFR1.

Figure 7. Proposed model to illustrate the role of HFR1 in coordinate HT and phyB signal for seed germination or dormancy in response to HT.

Under normal condition, the active phyB (Pfr form) induces the degradation of PIF1 in the nucleus, HFR1also sequestrates PIF1 to restrain its transactivation ability to SOM expression, thus the lower level of SOM sustains the higher GA and lower ABA to promote seed germination. However, ambient HT impairs the active phyB form (Pr form) to attenuate PIF1 degradation. HT also induces more SNOs, which trigger the S‐nitrosylation of HFR1 at Cys‐164, which induced the degradation of HFR1, thus releasing PIF1 to activate SOM expression, thereby the higher level of SOM induced more ABA and less GA accumulation to induced seed dormancy. However, on substituting Cys‐164 with Ser HT‐induced S‐nitrosylation of HFR1 (named as HFR1^C164S) was abolished. The degradation‐resistance version of HFR1 still sequestrates PIF1 and antagonizes PIF1‐dependent SOM expression, thus enhancing seed germination tolerance to HT.

Materials and Methods

Plant materials and treatment

The Columbia ecotype of Arabidopsis was used in this study except when indicated. PhyB mutant seeds were obtained from Arabidopsis Biological Resource Center, and hfr1‐4 and hfr1‐201 seeds were kind gifts from Professor Lin Li (Zhang et al, ²⁰¹⁹). phyB‐GFP seeds were kindly provided by Professor Jorge J. Casal (Legris et al, ²⁰¹⁷), and nox1 and hot5 seeds were obtained from Professors Zhenming Pei (He et al, ²⁰⁰⁴) and Elizabeth Vierling (Lee et al, ²⁰⁰⁸). GSNOR1‐ox seeds were obtained from Professor Gary J. Loake (Feechan et al, ²⁰⁰⁵). Plants were grown under a long‐day photoperiod WL at 22°C. For the HT treatment, surface‐sterilized seeds were plated on 1/2 Murashige‐Skoog (PhytoTechnology Laboratories) agar plates containing 0.1% sucrose and 1% phytoagar (pH 5.7) and incubated at 32°C with different light irradiation condition.

Germination assay

The seed germination assay was performed as previously described (Li et al, ²⁰²⁰; Yang et al, ²⁰²⁰). For the phyB‐dependent germination assay, the seeds were surface‐sterilized and plated on 1/2 MS agar plates containing 0.1% sucrose and 1% phytoagar (pH 5.7). The plates were exposed to weak white light for 1 h and then irradiated with far‐red light (3.2 μmol/m²/s) for 5 min to inactivate phyB (indicated as the phyB‐off condition) or exposed to far‐red light (3.2 μmol/m²/s) for 5 min and then to red light (13.12 μmol/m²/s) for 5 min to activate phyB (indicated as the phyB ‐on condition). After the light treatment, all of the seeds were incubated in the dark for 3 days, and the germination frequency (protrusion of the radical) was then assayed. Growth chambers with 740‐nm LEDs (far‐red light) were used for the light treatment. At least 100 seeds were used for each experimental point, and three biological replicates were used for the statistical analyses. For ambient HT treatment, the seed was irradiated with far‐red light or red‐light as phyB‐on or phyB‐off protocol, and the then immediately placed into the chamber at 32°C for 3 days under darkness, and the seed germination was quantified; the seed at 32°C under darkness was used as the control experiments (Li et al, ²⁰²⁰).

Site‐directed mutagenesis

Site‐directed mutagenesis of HFR1 was performed using QuikChange II Site‐Directed Mutagenesis Kits (Stratagene Corp.). The pET28a‐HFR1 or pRI101‐phyB‐GFP plasmid was used as the template, and the primer information is listed in Appendix Table S1. The mutations were confirmed by sequencing.

Generation of transgenic Arabidopsis plants

The HFR ^C164S fragment was cloned in the pRI101‐6Flag vector using in‐fusion technology (Clontech). The primer information was listed in Appendix Table S1. The generated constructs were used for transformation of the GV3101 agrobacterium strain. For plant transformation, Arabidopsis plants (Col‐0) were transformed using the floral dip method. Seeds were harvested and plated on selection medium to identify T1 transgenic plants. Approximately 100 of the T2 seeds were plated on selection medium in MS agar plates, and transgenic lines with a 3:1 (resistant/sensitive) segregation ratio were selected. T3 progenies homozygous for resistance to the selection medium were used in further studies.

Analyses of GSNOR activity and SNO content

GSNOR activity was determined using a previously reported method (Lee et al, ²⁰⁰⁸). In brief, the imbibed seeds were collected, ground with liquid nitrogen in 10 mM phosphate buffer (pH 7.2) and centrifuged at 10,000 g and 4°C for 10 min. The supernatant was incubated at 25°C with 0.4 mmol of NADH and 1 mmol of GSNO. GSNOR activity was determined by monitoring the consumption of NADH at 340 nm. The total SNO content in the seeds was quantified using a previously described method (Rusterucci et al, ²⁰⁰⁷). In brief, the seed extracts were prepared in 0.1 mol/L of phosphate (pH 7.2), 0.1 mol/L EDTA, and 0.1 mol/L EGTA and then incubated for 10 min in solution A (1% sulfanilamide in 0.5 mol/L HCl) or solution B (solution A plus 0.2% HgCl₂) to obtain the diazonium salt. The formation of the azo dye was obtained by reacting the incubated solutions with an equal volume of solution C [0.02% N‐(1‐naphthyl) ethylenediamine dihydrochloride in 0.5 mol/L HCl] for 10 min. The absorbance was measured at 550 nm. The SNO content in tissue extracts was quantified by determining the difference in absorbance between the reaction with solution B and that with solution A. The values obtained were compared with those of a standard curve constructed using GSNO.

ABA and GA content analysis

The measurement of ABA was performed as described previously, with small modifications (Jiang et al, ²⁰¹⁶; Li et al, ²⁰²⁰). Briefly, 100 mg of imbibed seeds was homogenized in 2 ml of precooled methanol:isopropanol (20:80 [v/v]) solution containing 0.2% (v/v) formic acid using a TissueLyser (JX‐24) with zirconia beads for 3 min at 30 Hz. ABA was extracted at −20°C overnight. The supernatant was collected after 14,000 g centrifugation at 4°C for 15 min and dried with a flow of nitrogen. The residue was dissolved with 100 μl of cold methanol solution containing internal standard d6‐ABA (CDN Isotopes). Quantification of ABA was performed with an ultra‐performance liquid chromatography–tandem mass spectrometry system consisting of an ultra‐performance liquid chromatography system (Waters) and a triple quadruple tandem mass spectrometer (AB Sciex). For GA measurements, about 200 mg of imbibed seeds were weighed and finely ground in liquid nitrogen. Internal standards of 2 ng g^{‐1 2}H₂‐GA₄ were added to the samples, followed by extraction with 1 ml solvent (methanol/H₂O, 80/20, v/v) at 4°C for 12 h. The supernatants were sequentially passed through the preconditioned tandem solid‐phase extraction cartridges containing C18 adsorbent (50 mg) and strong anion exchange adsorbent (200 mg). The strong anion exchange cartridge was then rinsed with 2 ml of 20% methanol (v/v), and the targeted acidic phytohormones were eluted with 3 ml acetonitrile and 1% formic acid (v/v). The eluent was evaporated under a mild liquid nitrogen stream at 35°C and re‐dissolved in 100 μl H₂O. The solution was acidified with 10 μl formic acid and extracted twice with 1 ml ether. The combined ether phase was dried under nitrogen gas and reconstituted in 100 μl acetonitrile, and then 10 μl triethylamine (20 mmol/ml) and 10 μl 3‐bromoactonyltrimethylammonium bromide (20 μmol/ml) were added. The reaction solution was vortexed at 35°C for 30 min and then evaporated under nitrogen gas. The samples were dissolved in 200 ml 10% acetonitrile (v/v) and subjected to nano‐liquid electrospray ionization (ESI) quadrupole time‐of‐flight (Q‐TOF) mass spectrometry (MS) for analysis. Three independent biological replicates were performed for each sample.

RNA extraction and quantitative RT‐PCR

After the different light or chemical treatments, total RNA from the seeds was extracted using the TRIzol Reagent (Invitrogen) according to the manufacturer's recommended protocol (Li et al, ²⁰²⁰). After DNAase I treatment, the first‐strand cDNAs were synthesized using 2 μg of total RNA in a volume of 20 μl using M‐MuLV reverse transcriptase (Fermentas) with oligo(dT)18 primer. The relative transcript level of each gene was determined by real‐time PCR using the SYBR green I master mix with a Roche LightCycler 96 PCR instrument (Roche), as described previously. At least different biological replicates were used to confirm the gene expression patterns, and PP2A was used as an internal gene expression control. The primers used in this study are listed in Appendix Table S1.

Western analysis

The total protein from imbibed seeds was extracted for western analysis with extraction buffer (50 mM Tris–HCl, pH 7.5, 75 mM NaCl, 15 mM EGTA,15 mM MgCl_2, 1 mM DTT, 0.1% Tween 20, 1 mM NaF, 0.2 M NaV, 2 mM Na‐pyrophosphate, 60 mM β‐glycerol phosphate and 1× proteases inhibitor mix, Roche) and then centrifuged for 10 min at 15,800 g and 4°C. The protein concentration was determined using the Bio‐Rad Protein Assay (Bio‐Rad) based on the Bradford method. Fifty micrograms of total protein were loaded per well in SDS‐acrylamide/bisacrylamide gel for electrophoresis using Tris–glycine–SDS buffer. The proteins were electrophoretically transferred to an Immobilon‐P polyvinylidene difluoride membrane (Millipore) using the Trans‐Blot Turbo instrument (Bio‐Rad). The membranes were blocked in Tris‐buffered saline‐0.1% Tween 20 containing 5% blocking agent and probed with antibodies diluted in blocking buffer. Anti‐Flag (Sigma, 1:3,000), anti‐His (Proteintech, anti‐GFP Clontech, 1:3,000), anti‐HA (Invitrogen, 1:3,000), anti‐RPT5 (Youke Biotech, 1:5,000), anti‐GSNOR1(AS09647, Agrisera,1:2,000), anti‐phyB (AS214565, Agrisera, 1:1,000), ECL‐peroxidase‐labeled anti‐rabbit (Sigma, 1:10,000) and anti‐mouse (Sigma, 1:10,000) antibodies were used in the western analyses. Detection was performed using an ECL Advance Western Blotting Detection Kit (Amersham), and the chemiluminescence was detected using a Tianlen 5600 ECL system (Shanghai, China). The band signal was quantified with ImageJ software.

Chromatin immunoprecipitation (CHIP) analysis

Chromatin affinity purification was conducted according to previous described method (Li et al, ²⁰²⁰). In brief, the germinated seeds were crosslinked in fixation buffer (10 mM Tris–HCl, pH 8.0, 0.4 M Suc, 1 mM EDTA, 1 mM PMSF, 0.25% [v/v] Triton X‐100, and 1% [v/v] formaldehyde) under a vacuum for 20–30 min, and the fixation procedure was stopped by adding 0.125 M Gly for 10 min. After washing three times with 50 ml of cold ddH₂O, the germinated seeds were ground to a fine powder in liquid nitrogen, and the nuclei were isolated and purified using the extract buffer (10 mM Tris–HCl, pH 8.0, 0.25 M Suc, 10 mM MgCl₂, 1% [v/v] Triton X‐100, 5 mM EDTA, 5 mM β‐mercaptoethanol, 1 mM PMSF, 5 μg/ml leupeptin, 1 μg/ml aprotinin, 5 μg/ml antipain, 1 μg/ml pepstatin, and 50 μM MG‐132). The nuclei were then sheared to an average length of 300 to 500 bp by sonication and then immunoprecipitated with anti‐FLAG M2 gel (catalog no. A2220; Sigma‐Aldrich), or anti‐MYC antibody (catalog no 20168, Invitrogen). The protein‐DNA complexes were released by incubation with 300 μl of ChIP elution buffer (20 mM Tris–HCl, pH 7.5, 5 mM EDTA, 50 mM NaCl, 1% [w/v] SDS, and 50 μg/ml proteinase K) for 1 h at 65°C. Immunoprecipitated DNA was purified using a PCR Purification Kit (New England Biolabs), and qPCR was conducted to measure the amounts of target gene fragment on a Light Cycler 480 Real‐Time PCR machine (Roche) using SYBR Green PCR Master Mix. TUB2 was used to normalize the qPCR results in each ChIP sample. Each biological replicate contained three technical replicates. All primers are listed in Appendix Table S1.

In vivo and in vitro S‐nitrosylation assays

We used the biotin switch method, which converts –SNO into biotinylated groups to detect protein S‐nitrosylated modification (Jaffrey & Snyder, ²⁰⁰¹; Forrester et al, ²⁰⁰⁹; Albertos et al, ²⁰¹⁵). For in vitro S‐nitrosylation, purified HFR1 or HFR1^C164S recombinant protein was exchanged for HEN buffer containing 25 mM HEPES, pH7.7, and 1 mM EDTA with a dialyzer, and then about 30 μg protein was incubated with the NO donor GSNO (200 μM, Calbiochem) in a reaction volume of 100 μl, and the reaction was run at room temperature under dark condition for 60 min with regular vortexing. After GSNO incubation, the sample was precipitated with cold acetone, and resuspended in 300 μl of blocking buffer (250 mM HEPES, pH7.7, 4 mM EDTA, 1 mM neocuproine, 2.5% SDS and 0.1% S‐methylmethane thiosulfonate). As the negative control, treatment with the reducing agent (DTT, 20 mM; Sigma) after GSNO incubation was carried out for 1 h under the same condition to check reversibility of the modification. After reaction at 50°C for 25 min with frequently vortex mixing, the sample was then precipitated with cold acetone and dissolved in 48 μl of HENS buffer (250 mM of HEPES, pH7.7, 4 mM of EDTA, 1 mM of neocuproine, 1% SDS), and then was added 6 μl of 20 mM sodium ascorbate and 6 μl of 4 mM biotin‐HPDP for ‐SNO conversion and labelling. The reaction was run for 1 h at room temperature. All the above steps were carried out in the dark. Aliquots of the sample were separated by SDS‐PAGE without boiling, and then analyzed by immunoblotting using an anti‐biotin antibody (Cell Signaling Technology, Cat# 7075). In order to normalize the equivalent protein abundance, a small fraction of purified protein (e.g.10 μl) without biotin switch treatment was used as the “input” protein for immunoblotting with anti‐His antibody (Santa Cruz Biotechnology, Cat # SC‐8036).

As for detecting in vivo S‐nitrosylation of HFR1 or HFR1^C164S, the imbibed seed was pretreated with 100 μM of cycloheximide to block de novo protein synthesis and 50 μM of MG132 to block the protein degradation. After HT treatment, the seed sample was homogenized in extraction buffer (50 mM of Tris–HCl, pH 7.5, 75 mM of NaCl, 15 mM of EGTA, 15 mM of MgCl₂, 0.1% Tween 20, of 1 mM NaF, 0.2 M of NaV, 2 mM of sodium pyrophosphate, 50 μM of MG132, 100 μM of cycloheximide, and 60 mM of beta‐glycerol phosphate) containing a complete protease inhibitor cocktail (Roche). After precipitating with cold acetone, the protein was dissolved in the HEN buffer (250 mM of HEPES, pH 7.7, 1 mM of EDTA) containing 0.1 mM of neocuproine, 50 μM of MG132, 100 μM of cycloheximide, and protease inhibitor cocktail (Sigma‐Aldrich, Cat # P9599), approximately 200 mg total protein in 100–200 μl of HEN buffer was incubated with equal volume of blocking buffer (250mM HEPES, pH7.7, 1mMEDTA, 0.1 mM neocuproine, 2.5% [w/v] SDS, and 25 mM of S‐methylmethane thiosulfonate) at 50°C for 1 h. After the block, proteins were precipitated with cold acetone, washed with 70% (v/v) acetone three times, and dissolved in 170 μl of HEN buffer supplemented with 1% (w/v) SDS. The sample was labeled for 1 h at room temperature by adding 10 μl of 1 M sodium ascorbate and 20 μl of 4 mM biotin‐HPDP, followed by precipitation and washed with acetone. The pellet was dissolved in 400 μl of 250 mM HEPES, pH 7.7, 1 mM EDTA, and 1% (w/v) SDS and neutralized with 800 ml of neutralization buffer (25 mM of HEPES, pH 7.7, 100 mM of NaCl, 1 mM of EDTA, and 0.5% [v/v] Triton X‐100). The sample was then mixed with 30 μl of streptavidin beads (Thermo Scientific, Cat #29202) and incubated at 4°C overnight. The beads were washed four times with washing buffer (25 mM of HEPES, pH 7.7, 600 mM of NaCl, 1 mM of EDTA, and 0.5% [v/v] Triton X‐100). The proteins were then eluted by boiling in 50 μl of HEN buffer containing 1% 2‐mercaptoethanol. The samples were analyzed by SDS–PAGE and immunoblotting with anti‐Flag to detect the S‐nitrosylated HFR1 or HFR1^C164S. An equally small fraction of extracted protein without biotin‐switch treatment was used as the input to normalize against the total protein amount.

Protoplast transient transformation assay

To assay the transient transcription activity of the LUC reporter, a 2.4 kb promoter fragment of SOM, or 2‐kb promoter fragment of CYP707A2, was amplified from genomic DNA and inserted into pGreenII 0800‐LUC to generate the P _RGA ‐LUC reporter construct as previous method (Li et al, ²⁰²⁰). The coding sequences of PIF1 and HFR were amplified and inserted into the pGreenII 62‐SK vector under the control of the 35S promoter. All of the primers used for these constructs are listed in Appendix Table S1. Arabidopsis mesophyll protoplasts were prepared and transiently transfected as previously described (Yoo et al, ²⁰⁰⁷). The relative REN activity was used as an internal control, and the LUC/REN ratios were calculated (Li et al, ²⁰²⁰).

Author contributions

Songbei Ying: Resources; validation; investigation; methodology. Wenjun Yang: Resources; validation; investigation; methodology. Ping Li: Resources; validation; project administration. Yulan Hu: Resources; validation; methodology. Shiyan Lu: Resources; validation; investigation; methodology. Yun Zhou: Conceptualization; formal analysis; funding acquisition. Jinling Huang: Conceptualization; writing – original draft; writing – review and editing. John T Hancock: Writing – review and editing. Xiangyang Hu: Conceptualization; supervision; funding acquisition; validation; investigation; writing – original draft; project administration; writing – review and editing.

Disclosure and competing interests statement

The authors declare that they have no conflict of interest.

Supporting information

Appendix

EMBO reports (2022) 23: e54371

Data availability

No primary datasets have been generated and deposited.