Phytochrome B Induces Intron Retention and Translational Inhibition of PHYTOCHROME-INTERACTING FACTOR3^{1  [OPEN]}

Abstract

The phytochrome B (phyB) photoreceptor stimulates light responses in plants in part by inactivating repressors of light responses, such as PHYTOCHROME-INTERACTING FACTOR3 (PIF3). Activated phyB inhibits PIF3 by rapid protein degradation and decreased transcription. PIF3 protein degradation is mediated by EIN3-BINDING F-BOX PROTEIN (EBF) and LIGHT-RESPONSE BTB (LRB) E3 ligases, the latter of which simultaneously targets phyB for degradation. In this study, we show that PIF3 levels are additionally regulated by alternative splicing and protein translation in Arabidopsis (Arabidopsis thaliana). Overaccumulation of photo-activated phyB, which occurs in the mutant defective for LRB genes under continuous red light, induces a specific alternative splicing of PIF3 that results in retention of an intron in the 5′ untranslated region of PIF3 mRNA. In turn, the upstream open reading frames contained within this intron inhibit PIF3 protein synthesis. The phyB-dependent alternative splicing of PIF3 is diurnally regulated under the short-day light cycle. We hypothesize that this reversible regulatory mechanism may be utilized to fine tune the level of PIF3 protein in light-grown plants and may contribute to the oscillation of PIF3 protein abundance under the short-day environment.

Phytochromes are important photoreceptors regulating growth and development throughout a plant’s lifecycle. Among the five members of the phytochrome family in Arabidopsis (Arabidopsis thaliana), phytochrome B (phyB) is relatively abundant in light and plays a major role in controlling hypocotyl growth under prolonged continuous red light (Rc; Li et al., 2011). PHYTOCHROME-INTERACTING FACTORs (PIFs) are basic helix-loop-helix transcription factors that repress light responses (Leivar et al., 2008; Shin et al., 2009; Pham et al., 2018). PIF3, the first characterized member of the PIF family (Ni et al., 1998), highly accumulates in dark-grown seedlings to maintain etiolation, and levels of PIF3 rapidly decline upon light exposure, allowing establishment of photomorphogenesis in seedlings (Bauer et al., 2004; Zhang et al., 2013). During de-etiolation, photoactivated phyB directly binds to PIF3 and induces PIF3 phosphorylation, which is necessary for the 26S proteasome-mediated degradation of PIF3 (Al-Sady et al., 2006; Ni et al., 2013). In addition, phyB down-regulates PIF3 transcription by about 3-fold during de-etiolation and interferes with PIF3 DNA binding ability (Shi et al., 2016; Park et al., 2018). Inhibition of PIF3 at the level of translation has not been reported, although light signals generally promote translation activity globally (Paik et al., 2012; Chen et al., 2018).

After seedling establishment, normal light-grown plants maintain lower steady-state expression of PIFs, which continue to modulate plant development in response to light and temperature (Leivar and Monte, 2014; Pham et al., 2018). For example, PIF3 plays prominent roles in promoting dark-period elongation growth under short-day conditions (Soy et al., 2012, 2016) and in freezing tolerance (Jiang et al., 2017).

Studies on phyB-induced rapid protein degradation of PIF3 have revealed that PIF3 can be degraded through at least two pathways. During de-etiolation, F-box proteins EIN3-BINDING F BOX PROTEIN1 (EBF1) and EBF2 mediate PIF3 degradation via SCF^EBFs (for Skp1, Cullins, F-box) ubiquitin E3 ligases to facilitate photomorphogenesis of plants (Dong et al., 2017). Plants deficient in EBFs exhibit reduced light sensitivity and inefficient de-etiolation. Under high-fluence red light, PIF3 can also be degraded via LIGHT-RESPONSE BTBs (LRBs), the ubiquitin E3 ligases CRL3^LRBs (for Cullin3-RING ligase) that target phyB for degradation to attenuate light responses (Ni et al., 2014). Plants deficient in LRBs show hypersensitivity to light in a phyB-dependent manner (Christians et al., 2012; Ni et al., 2014; Dong et al., 2017).

To study the functional dynamics between EBFs and LRBs in the regulation of PIF3, we eliminated both types of PIF3 E3 ligases and generated an ein3 ebf lrb mutant line. Using this mutant line, we uncovered an unexpected mechanism by which phyB inactivates PIF3: i.e. inhibition of PIF3 protein translation via alternative splicing (AS).

RESULTS AND DISCUSSION

To knock out both ubiquitin E3 families of PIF3, we generated the ein3 ebf lrb hextuple mutant by crossing the viable ein3 ebf1 ebf2 (ein3 ebf) mutant with lrb1 lrb2 lrb3 (lrb; Fig. 1). Using reverse transcription (RT)-PCR, these mutants were further confirmed to lack the corresponding gene transcripts (Supplemental Fig. S1). The resulting phenotypes under Rc showed that the ein3 ebf mutant exhibited longer hypocotyls compared to ein3, while lrb showed shorter hypocotyls compared to ecotype Columbia of Arabidopsis (Col; Fig. 1, A and B), consistent with previous observations (Christians et al., 2012; Ni et al., 2014; Dong et al., 2017). The ein3 ebf lrb mutant almost phenocopied lrb in having a shortened hypocotyl under Rc, suggesting that lrb strongly suppressed ein3 ebf (Fig. 1, A and B). The hyper-photosensitive phenotype of ein3 ebf lrb seemed at odds with the anticipated accumulation of PIF3 in high levels and prompted us to examine PIF3 protein levels in these mutants.

One would expect that by eliminating both E3 ubiquitin ligase families that regulate PIF3 abundance, PIF3 protein would be stabilized in the ein3 ebf lrb mutant. However, immunoblotting showed that PIF3 protein accumulated only in ein3 ebf but not in lrb or in ein3 ebf lrb mutants (Fig. 1, C and D). PIF3 protein levels were much higher in ein3 ebf than in ein3 ebf lrb, while PIF3 transcript levels were comparable between these two mutant lines (Fig. 1, D and E). In addition, PIF3 activities matched with PIF3 protein levels in these mutants, based on the expression of the PIF3 target genes XTR7 and XTH33 (Fig. 1, F and G). In the dark, relative levels of PIF3 protein corresponded with their respective mRNA levels among the mutants tested (Supplemental Fig. S2), suggesting that the regulation of PIF3 protein levels by LRBs and EBFs is light dependent. Taken together, the lrb mutations appeared to have suppressed the PIF3 overaccumulation of ein3 ebf by decreasing PIF3 protein expression without affecting its transcript levels, and one possible explanation for this is that lrb potentially inhibits translation of PIF3 under Rc. As lrb mutations primarily cause phyB accumulation and light hypersensitivity (Christians et al., 2012; Ni et al., 2014), we hypothesized that phyB may negatively regulate PIF3 translation.

Light causes widespread AS in plant transcriptomes (Shikata et al., 2014; Hartmann et al., 2016; Xin et al., 2017; Godoy Herz et al., 2019). Although it has not been experimentally demonstrated that PIF3 undergoes light-regulated AS, based on the information from Araport (Arabidopsis Information Portal; www.araport.org), transcripts with AS patterns at an intron of ∼500 bp in the PIF3 5′ untranslated region (UTR) have been found in the database (Fig. 2A). We refer to this intron as Intron_AS in this study and monitored its splicing level by RT-quantitative PCR (qPCR). We found that levels of the Intron_AS-retained form were significantly increased relative to total PIF3 mRNA in Col under Rc as compared to dark (Dc; Fig. 2B), suggesting that Intron_AS retention in PIF3 mRNA is a light-dependent event. Furthermore, mutation of phyB essentially abolished the Rc-induced increase of Intron_AS retention (Fig. 2B), indicating that phyB is critical in red light-responsive Intron_AS retention in PIF3 mRNA. Moreover, lrb mutations caused higher Intron_AS retention under Rc (Supplemental Fig. S3), which is consistent with higher levels of phyB accumulation in lrb (Christians et al., 2012; Ni et al., 2014).

To better visualize the retention of Intron_AS, we performed gel electrophoresis of the PCR products using a primer pair that would produce a longer product when Intron_AS is retained and a shorter product when Intron_AS is spliced (Fig. 2C). The ACTIN2 (ACT2) primer pair produced only the short-sized band (the spliced form) from all complementary DNA (cDNA) samples, while the same primer pair produced a long-sized band (the unspliced form) from genomic DNA (Fig. 2C), suggesting that there was no intron retention in ACT2 mRNA in those samples. By contrast, the PIF3 primer sets produced both the intron-containing and intron-spliced bands from the cDNA samples. The Intron_AS-retained form was strongly enriched under Rc compared to Dc in Col, but the light-induced intron retention was nearly absent in phyB-9. This result shows again that Intron_AS retention in PIF3 mRNA occurred under red light and that this event was dependent on phyB (Fig. 2C). As an additional control, we examined other introns in PIF3 mRNA and found no evidence of retention of any other introns (Supplemental Fig. S4). These data show that photoactivated phyB can induce an AS event that results in the retention of Intron_AS in PIF3 mRNA.

Extensive studies have shown that the 5′-UTR can influence its mRNA translation (Araujo et al., 2012; von Arnim et al., 2014). Since Intron_AS resides within the 5′ UTR of PIF3 mRNA, we next asked whether the retention of Intron_AS may affect PIF3 translation. We generated T7 promoter-driven constructs in which either normally spliced PIF3 5′ UTR (5U) or Intron_AS-retained 5′ UTR (5IU) was inserted upstream of the PIF3 coding sequence. In a mammalian cell-based in vitro translation (IVT) assay using either DNA or RNA as template, 5U produced higher levels of PIF3 proteins than did 5IU (Fig. 3A). Clearly, retaining the Intron_AS sequence in PIF3 5′ UTR inhibited protein translation even in a heterologous IVT system.

To test the effect of Intron_AS in planta, we generated transgenic plants in a pif3-3 mutant background that carried the following transcription cassettes under the 35S promoter: 5U PIF3 or 5IU PIF3, in which the normally spliced 5′ UTR or the Intron_AS-containing 5′ UTR, respectively, was placed upstream of the PIF3 coding region (Fig. 3B). Three independent 5U PIF3 and 5IU PIF3 transgenic lines with similar levels of total PIF3 mRNA transcript were selected, and Intron_AS-containing PIF3 transcript was confirmed to exist only in 5IU PIF3 transgenic lines (Supplemental Fig. S5). It appeared that in the context of this artificial Intron_AS-containing PIF3 transgene, Intron_AS could be spliced, but very inefficiently, as most of the transcript retained the intron (Supplemental Fig. S5C). In darkness, when PIF3 protein was stable, immunoblotting showed that PIF3 protein was expressed in 5U PIF3 lines but was almost undetectable in 5IU PIF3 lines (Fig. 3B). Quantification showed that retention of Intron_AS in PIF3 5′ UTR caused an ∼80% reduction in PIF3 protein levels (Fig. 3B). These results indicate that forced retention of Intron_AS upstream of the PIF3 coding sequence inhibits protein production. As a result, the 5IU PIF3 transgene was unable to restore the phenotype of pif3-3 in hypocotyl elongation under Rc, in contrast to the rescue seen with 5U PIF3 (Fig. 3, C and D). We deduce from these results that the retention of Intron_AS in the 5′ UTR of PIF3 mRNA can inhibit PIF3 protein expression both in vitro and in vivo, consequently affecting PIF3-regulated hypocotyl elongation.

uORF-mediated translation inhibition regulates gene expression involved in light response as well as other developmental and metabolic processes in plants (von Arnim et al., 2014; Kurihara et al., 2018). The PIF3 Intron_AS sequence contains multiple AUG start codons and putative uORFs (Supplemental Fig. S6). The longest uORF, which is 129 nucleotides in length and located just upstream of the PIF3 main ORF, is considered the most likely candidate uORF to inhibit downstream PIF3 translation, even though the start codon context is suboptimal (Kozak, 2001). To test this idea, we carried out a dual-luciferase protoplast transient assay where the 5′ UTR testing effector was inserted upstream of the firefly luciferase (LUC) reporter with a 35S promoter-driven Renilla luciferase (RLUC) as the internal control (Fig. 3E). The testing effectors included PIF3 5U, 5IU, or Intron_AS with the mutated form of the uORF start codon 5I_ASU (5IUm). In another set, we explicitly tested the effect of the uORF without other Intron_AS sequences: either the uORF with an ATG start codon, or mutated uORF with a CTG start codon (uorf). Relative LUC activity in 5IU and uORF was significantly lower than that in 5IUm and uorf, respectively (Fig. 3E). These results indicate that the 129-nt uORF of Intron_AS can inhibit translation of the downstream ORF.

We noticed that the inhibitory effect of Intron_AS inclusion was not as strong in this transient assay (Fig. 3E, 5U versus 5IU) compared to the in planta assay (Fig. 3B), which was probably due to the difference in the assay system. In addition, the dramatic inhibition of protein expression seen in the experiments of Figure 3, A and B, could be caused by the combination of the uORFs in Intron_AS, as compared to the Figure 3E experiment, in which the 129-nt uORF was tested exclusively. Regardless, our data support the conclusion that retention of Intron_AS, which contains uORFs, causes a blockage in PIF3 expression at the level of translation.

It is well established that phyB activation leads to a sharp decline in PIF3 abundance during de-etiolation (Bauer et al., 2004; Al-Sady et al., 2006; Dong et al., 2017). phyB-stimulated PIF3 protein degradation accounts for the predominant portion of PIF3 down-regulation (Al-Sady et al., 2006), while decreased transcription of PIF3 also occurs (Shi et al., 2016). Here, our data show that phyB can additionally inhibit PIF3 protein translation via retention of Intron_AS under prolonged red light. However, questions remain as to how important the translation control is relative to the control of protein stability and transcription, and under what natural circumstances the plants would utilize this type of regulatory mechanism. To this end, we examined the kinetics of Intron_AS retention rate during de-etiolation. When dark-grown seedlings were irradiated with red light, PIF3 protein level rapidly and sharply declined within minutes and dropped >10 fold within hours after light exposure, while the increase of Intron_AS retention was more evident in prolonged light-grown plants than during the 24 h of de-etiolation (Supplemental Fig. S7). This result suggests that phyB-induced AS is unlikely to play a major role in de-etiolation but may be a feature associated with light-grown plants. As most light-dependent AS profiling studies are performed under the de-etiolation condition (Shikata et al., 2014; Hartmann etal., 2016; Kurihara et al., 2018), this may be a reason why PIF3 has not been reported to undergo light-induced AS.

In light-grown plants, PIF3 protein abundance oscillates under the short-day diurnal cycle to promote predawn hypocotyl growth (Soy et al., 2012, 2016). Interestingly, unlike PIF4 and PIF5, whose transcription is strongly diurnally regulated, transcriptional regulation of PIF3 is minimal under this condition, while PIF3 protein accumulates in the Dc period and declines during the day (Soy et al., 2012). We examined PIF3 Intron_AS retention in plants grown in a short-day light cycle. The Intron_AS retention rate of PIF3 oscillated diurnally, rising postdawn and dropping around dusk (Fig. 4A). This pattern would correspond with reduced PIF3 translation during the day and normal protein synthesis during the night, which is consistent with peak accumulation of PIF3 protein at predawn (Soy et al., 2012). We thus suggest that regulation of PIF3 protein synthesis, in coordination with PIF3 degradation, contributes to diurnal oscillation of PIF3 protein levels under short-day conditions.

In summary, we have identified an additional regulatory mechanism of phyB on PIF3 and revealed that light signals can inactivate PIF3 at the level of translation, apart from regulating its transcription, DNA binding, and protein stability (Fig. 4B). In particular, our findings explain why the lrb E3 mutations can suppress the ebf E3 mutants phenotypically and at the level of PIF3 protein accumulation. This is because lrb-induced phyB accumulation leads to a block of PIF3 protein synthesis, which would alleviate the problems caused by the deficient PIF3 protein degradation in ebf (Fig. 1). Light signals can regulate uORF inclusion in the 5′ UTR of several genes as a result of global change of transcription start sites (Kurihara et al., 2018). In PIF3, alternative transcription initiation may also occur (Fig. 2A), but it is not related to the inclusion of the uORFs. Regarding the extent of translational regulation relative to transcription and protein degradation of PIF3, we suggest that during de-etiolation, the massive irreversible PIF3 protein degradation probably accounts for most of the rapid and dramatic decline of PIF3 levels, while the AS-coupled translational regulation is minor, if it occurs at all. In normal light-grown plants, PIF3 expresses at a reduced level, and a reversible control mechanism, as revealed in this study, may operate to fine-tune PIF3 levels, likely in conjunction with other PIF3 regulatory mechanisms. This idea is supported by our data showing that retention of Intron_AS oscillates in the short-day diurnal cycle. It is possible that Intron_AS-coupled translational regulation of PIF3 may play important roles in other light-modulated physiological processes in plants.

MATERIALS AND METHODS

Plant Materials and Growth Condition

Arabidopsis (Arabidopsis thaliana) mutants in the Col ecotype were grown under long-day conditions (16 h light/8h dark) at 22°C. The parental lines of the higher-order mutant, ein3 ebf1 ebf2 (ein3 ebf) and lrb1 lrb2 lrb3 (lrb), were previously described in An et al. (2010), and Ni et al. (2014), respectively. The pif3 mutant was described in Dong et al. (2017).

Seeds were surface sterilized using 30% (v/v) bleach and sowed on one-half strength Murashige and Skoog plates (2.2 g/L Murashige and Skoog Basal Salt, 0.5 g/L MES, and 8 g/L agar, pH 5.7). The seeds were then stratified at 4°C for 2 d, placed under white light for 3 h to induce germination, and either put into Dc or 10 μmol m⁻² s⁻¹ Rc for 3–4 d followed by phenotypic observation, RT-qPCR analysis, and immunoblot analysis.

Generation of Transgenic Plants

For generation of 35S: 5U PIF3 and 35S: 5IU PIF3 plants, the 5U PIF3 coding sequence was amplified directly from cDNA, while the 5IU PIF3 coding sequence was generated by overlapping PCR of 5IU, which was amplified from genomic DNA, and the PIF3 coding sequence, which was amplified from cDNA. The resulting 5U PIF3 and 5IU PIF3 coding sequences were inserted into pjim19 (Kan) using XhoI/SacI restriction enzymes. The vectors were then transformed into the pif3-3 background by the floral dip method using Agrobacterium tumefaciens strain GV3101. Primers used are listed in Supplemental Table S1.

RT-qPCR

Total RNA was extracted using Qiagen RNeasy Plant Mini kit according to the instructions. Reverse transcription was performed using SuperScript III reverse transcriptase (Invitrogen), oligo(dT) 12-18 (Invitrogen), and 1 µg total RNA as template. The resulting cDNA samples were diluted 10-fold with nuclease-free water, and 2 µL was used for qPCR assay in a CFX96 real-time system (Bio-rad). ACT2 was used as internal control. Primers used for RT-qPCR are listed in Supplemental Table S1.

Immunoblot

Total protein was extracted from whole seedlings using denaturing buffer (8 m urea, 0.1 m NaH₂PO₄, and 0.1 m Tris-HCl, pH 8.0). Protein concentration was determined by the Bio-rad protein assay. The same amount of total protein was subjected to SDS-PAGE, transferred to polyvinylidene difluoride film, blocked by 5% (w/v) milk in 1× Tris-buffered saline plus 0.1% (v/v) Tween 20, and incubated with anti-PIF3 or anti-RPN6 at 4°C overnight. After washing, film was further incubated with horseradish peroxidase-conjugated secondary antibody for 1 h at room temperature and exposed to x-ray film. The x-ray film was then developed and fixed in a dark room.

IVT Assay

For DNA as template, either the 5U PIF3 or the 5I1U PIF3 coding sequence was inserted into pT7CFE1-Myc (Pierce) vector using NdeI/SacI, yielding pT7-5U PIF3 and pT7-5I1U PIF3, respectively. One microgram DNA was used as template in a 25 μL IVT system following the instructions of the 1-step Human Coupled IVT kit–DNA (Pierce). The resulting IVT product was mixed with 100 μL loading buffer, boiled, and subjected to immunoblot analysis using anti-PIF3 antibody. Ponceau S staining was used as loading control. Primers used for construct cloning are listed in Supplemental Table S1.

For RNA as template, the pT7-5U PIF3 and pT7-5I1U PIF3 plasmids were linearized by Not I digestion and purified using the QIAquick PCR purification kit. One microgram linearized plasmid was used as template for in vitro transcription using the HiScribe T7 High Yield RNA Synthesis Kit (New England Biolabs). Five micrograms RNA was used as template for IVT, and immunoblotting was conducted in the same way as with DNA as template.

Dual Luciferase Assay

The 35S promoter was cloned into pGreenii 0800-LUC using Hind III/BamH I, yielding pGreenii 35S-LUC. Then the 5U, 5IU, 5IUm, uORF, and uorf of PIF3 were cloned into pGreenii 35S-LUC using BamH I/Spe I. Primers used are listed in Supplemental Table S1.

Protoplasts were isolated based on a protocol previously described (Yoo et al., 2007). The leaves of Col adult plants grown under long-day conditions for 3–4 weeks were used for protoplast isolation. Ten-microgram plasmids were used for transfection. After incubation in Dc for 12 h, luciferase activity was measured following the instructions of the Dual-Luciferase Reporter Assay System (Promega). Each plasmid was measured with four biological replicates and at least two technical repeats.

Accession Numbers

Sequence data can be found in the Arabidopsis Genome Initiative database under the following accession numbers: AT2G18790 (phyB), AT1g09530 (PIF3), AT2G46260 (LRB1), AT3G61600 (LRB2), AT4G01160 (LRB3), AT3G20770 (EIN3), AT2G25490 (EBF1), AT5G25350 (EBF2), AT4G14130 (XTR7), AT1G10550 (XTH33), AT3G18780 (ACT2), and AT1G29150 (RPN6).

Supplemental Data

The following supplemental materials are available.

• Supplemental Figure S1. RT-PCR confirmation of the ebf lrb higher-order mutant.

• Supplemental Figure S2. LRB mutations do not affect PIF3 protein level in the dark.

• Supplemental Figure S3. LRB mutations induce PIF3 Intron_AS retention under red light.

• Supplemental Figure S4. Neither light nor phyB regulates the retention of other introns in PIF3 mRNA.

• Supplemental Figure S5. Expression level of PIF3 transcripts in the 35S:5U PIF3/pif3-3 and 35S:5IU PIF3/pif3-3 transgenic lines.

• Supplemental Figure S6. Sequences of Intron_AS and the 129-nt uORF of PIF3 mRNA.

• Supplemental Figure S7. Kinetics of PIF3 protein degradation and Intron_AS retention during de-etiolation.

• Supplemental Table S1. Primers used in this study.

ACKNOWLEDGMENTS

We thank Peter Quail and Dr. Weimin Ni for providing lrb1 lrb2 lrb3 mutant seeds. We also thank Hongwei Guo for providing ein3 ebf1 ebf2 mutant seeds.

LITERATURE CITED

Author notes