PEAPOD repressors modulate and coordinate developmental responses to light intensity in Arabidopsis

Summary

• Higher plants adapt to different light intensities by altering hypocotyl elongation, stomatal density, seed size, and flowering time. Despite the importance of this developmental plasticity, knowledge of the underlying genetic and molecular mechanisms modulating and coordinating responses to light intensity remains incomplete. Here, I report that in Arabidopsis the PEAPOD (PPD) repressors PPD1 and PPD2 prevent exaggerated responses to light intensity.

• Genetic and transcriptome analyses, of a ppd deletion mutant and a PPD1 overexpression genotype, were used to identify how PPD repressors modulate the light signalling network.

• A ppd1/ppd2 deletion mutant has elongated hypocotyls, elevated stomatal density, enlarged seed, and delayed flowering, whereas overexpression of PPD1 results in the reverse. Transcription of both PPD1 and PPD2, upregulated in low light and downregulated in higher light, is activated by PHYTOCHROME INTERACTING FACTOR 4. I found PPDs modulate light signalling by negative regulation of SUPPRESSOR OF phyA‐105 (SPA1) transcription. Whereas PPDs coordinate many of the responses to light intensity – hypocotyl elongation, flowering time, and stomatal density – by repression/de‐repression of SPA1, PPD regulation of seed size occurs independent of SPA1.

• In conclusion PPD repressors modulate and coordinate developmental responses to light intensity by altering light signal transduction.

Introduction

Higher plants alter many aspects of their growth and development in response to changes in ambient light intensity. For example, Arabidopsis thaliana adapts to higher light intensities by inhibiting hypocotyl elongation, producing large cotyledons, leaves, and seeds, delaying flowering, and increasing stomatal density (Li et al., ²⁰⁰⁶; Jiao et al., ²⁰⁰⁷; Casson et al., ²⁰⁰⁹; Franklin & Quail, ²⁰¹⁰). When plants are grown in lower light levels there is a reversal of all these traits (Li et al., ²⁰⁰⁶; Jiao et al., ²⁰⁰⁷; Casson et al., ²⁰⁰⁹; Franklin & Quail, ²⁰¹⁰). Although these adaptations have a significant effect on plant growth and reproductive success, we know little about the molecular mechanism(s) modulating and coordinating responses to light intensity.

Plants use multiple photoreceptors including the red/far‐red‐sensing phytochromes (PHYs) and the blue/UV‐A‐sensing cryptochromes (CRYs) to perceive light. These photoreceptors act in multiple ways to regulate a complex light signalling network. In the dark, PHYTOCHROME INTERACTING FACTORs (PIFs), a set of basic helix–loop–helix transcription factors, promote the etiolation of seedlings by activating genes that increase cell elongation. Upon red light activation, phyB is translocated to the nucleus and phosphorylates PIF proteins, which are then ubiquitinated and degraded by the 26S proteasome (Park et al., ²⁰⁰⁴; Al‐Sady et al., ²⁰⁰⁶; Shen et al., ²⁰⁰⁷). In addition, during de‐etiolation, phyB downregulates PIF transcription, inhibits PIF translation, and blocks PIF protein binding to its targets (Park et al., ^{2012, 2018}; Shi et al., ²⁰¹⁶; Dong et al., ²⁰²⁰). cry1 and cry2 also bind to PIFs and disrupt their capacity to activate gene expression (Ma et al., ²⁰¹⁶; Pedmale et al., ²⁰¹⁶). Although this photoreceptor‐mediated reduction in PIF function in the light is required for seedling de‐etiolation, PIFs also have an essential role promoting plant developmental responses to changes in light intensity and high temperature (Leivar & Monte, 2014).

A complex of CONSTITUTIVE PHOTOMORPHOGENIC 1 (COP1), a RING‐finger E3 ubiquitin ligase, and SUPPRESSOR OF phyA‐105 (SPA) proteins is another major hub of the light signalling network in plants (Huang et al., ²⁰¹⁴). The SPA proteins (SPA1–4) act to enhance COP1 activity (Hoecker & Quail, 2001; Laubinger et al., ²⁰⁰⁴; Yang & Wang, ²⁰⁰⁶). In darkness, the COP1/SPA complex interacts with, and targets for degradation, a set of positive photomorphogenic transcription factors, including ELONGATED HYPOCOTYL 5 (HY5) and LONG HYPOCOTYL IN FAR‐RED (HFR1) (Osterlund et al., ²⁰⁰⁰; Saijo et al., ²⁰⁰³; Yang et al., ^2005a,b). Though essential for etiolated growth in darkness, the COP/SPA complex also plays a significant role in modulating plant responses to changes in light intensity. In the light interaction of cry1 and phyB with COP1 and SPA1 inactivates the COP1/SPA1 complex, leading to increased levels of HY5 and HFR1 that limit hypocotyl elongation (Wang et al., ²⁰⁰¹; Yang et al., ²⁰⁰¹; Laubinger et al., ²⁰⁰⁶; Hornitschek et al., ²⁰⁰⁹; Lian et al., ²⁰¹¹; Liu et al., ²⁰¹¹; Ranjan et al., ²⁰¹¹; Lu et al., ²⁰¹⁵). HY5 and HFR1 restrict PIF‐directed gene expression (Shi et al., ²⁰¹³; Toledo‐Ortiz et al., ²⁰¹⁴). The COP1/SPA complex has an important role in optimizing growth and development by preventing exaggerated photomorphogenesis in seedlings grown in higher light (Wang et al., ²⁰⁰¹; Yang et al., ^2005b). Among the SPA gene family, SPA1 has the most prominent role as a repressor of excessive photomorphogenesis (Laubinger et al., ²⁰⁰⁴; Fittinghoff et al., ²⁰⁰⁶; Balcerowicz et al., ²⁰¹¹; Lu et al., ²⁰¹⁵). Recently, Burko et al. (2020) reported that HY5 acts as a direct activator of SPA1 expression, thereby providing a light‐regulated feedback mechanism to prevent hyperphotomorphogenesis of seedlings grown in high light.

In this study, I show that Arabidopsis PEAPOD (PPD) genes PPD1 and PPD2 act to prevent exaggerated developmental responses to light intensity. Previous reports indicated the PPD genes regulate leaf and seed growth by limiting cell proliferation (White, 2006; Karidas et al., ²⁰¹⁵; Baekelandt et al., ²⁰¹⁸; Liu et al., ²⁰²⁰). PPD proteins belong to the TIFY superfamily, which also contains the ZIM, JAZ, and TIFY8 family proteins (Bai et al., ²⁰¹¹). A large unique and highly conserved N‐terminal PPD domain (White, 2006) distinguishes PPD proteins from other TIFY proteins. PPD proteins interact with an assortment of different proteins, with these interactions contributing to the control of cell proliferation during development. PPD‐interacting proteins include the KINASE‐INDUCIBLE DOMAIN INTERACTING (KIX) domain proteins KIX8 and KIX9 (Gonzales et al., ²⁰¹⁵) and NOVEL INTERACTOR OF JAZ (NINJA), which act as adaptors for the co‐suppressor TOPLESS (TPL) (Baekelandt et al., ²⁰¹⁸), LIKE HETEROCHROMATIN PROTEIN 1 (LHP1), which is involved in maintaining gene silencing by chromatin modification (Zhu et al., ²⁰¹⁹), and the MYC transcription factors MYC3 and MYC4 (Liu et al., ²⁰²⁰). Counteracting PPD/KIX repressor action, the F‐box protein STERILE APETALA promotes cell proliferation by targeting PPD1, PPD2, KIX8, and KIX9 for degradation (Wang et al., ²⁰¹⁶; Li et al., ²⁰¹⁸).

Here, I report that seedlings of a ppd deletion mutant have pronounced long hypocotyls, indicating a possible role for PPDs in light signalling. A detailed examination of the ppd mutant hypocotyl phenotype, together with transcriptome profiling and genetic analysis, was undertaken to determine how PPDs modulate photomorphogenesis. This analysis identified mechanisms for the regulation of PPD1 and PPD2 expression by light intensity, the feedback regulation of light signal transmission by PPDs, and the PPD‐mediated coordination of adaptive responses to light intensity.

Materials and Methods

Plant materials and growth conditions

Arabidopsis Δppd mutant and transgenic Δppd::PPD1 complemented (PPD1C‐1, PPD1C‐2) and Δppd::PPD1 overexpression (PPD1‐OE26) genotypes in a Landsberg erecta (Ler) ecotype background were described previously (White, 2006). The Δppd mutant and transgenic Δppd::PPD1‐OE26 genotypes were backcrossed with wild‐type Columbia‐0 (Col‐0) for four generations to obtain Δppd and PPD1‐OE in a Col‐0 ecotype background. Transfer DNA (T‐DNA) insertion mutant lines ppd1‐1 (SALK_149924C), ppd1‐2 (SALK_057237C), both in At4g14713, and ppd2‐1 (SALK_142698C) in At4g14720 were from the Arabidopsis Biological Resource Center (ABRC). Mutants cry1 (CS6955), phyB‐5 (CS6213), cop1‐6 (CS69041), hy5 (SALK_096651C), spa1‐7 (CS869043), pif4‐2 (CS66043), and transgenic overexpression lines 35SPHYB (CS8037) and pif4‐2::35SPIF4‐MYC (CS69173), also obtained from ABRC, were crossed with Δppd or PPD1‐OE26 in the appropriate Ler or Col‐0 ecotype and double mutants selected from F₂ populations. A combination of phenotyping, genotyping by PCR with gene‐specific primers (Supporting Information Table S1) and progeny testing identified double mutant homozygous lines. The kix8;kix9 double mutant previously described by Gonzalez et al. (2015) with T‐DNA insertion alleles in kix8 (GAB_422H04) and kix9 (SAIL_1168_G09) was supplied by Dirk Inzé.

For hypocotyl measurements, seeds were stratified at 4°C for 3 d in the dark and then germinated and seedlings grown in soil mix in a controlled‐environment growth cabinet at 22°C, 65% relative humidity, with a long day 16 h : 8 h, light : dark cycle. Unless stated otherwise, photosynthetic photon flux density (PPFD) was 180 μmol m⁻² s⁻¹ over the waveband 400–700 nm, from a combination of Sylvania GRO‐LUX F36W/GRO‐T8 and Phillips TLD 58W/840 fluorescent tubes. Continuous red or blue light treatments were provided from LED lamps at 10 μmol m⁻² s⁻¹. Light intensity was measured with a quantum photometer (model Li‐189: Li‐Cor, Santa Clara, CA, USA).

For whole‐genome transcript analysis, seeds were surface sterilized with 70% ethanol, 0.01% Triton X‐100 for 10 min, followed by 100% ethanol for 5 min, and then air dried on sterile filter paper and transferred to media plates containing ½ Murashige & Skoog (½MS) salts and vitamins (Duchefa Biochemie, Haarlem, the Netherlands), 1% sucrose, and 0.8% agar. Plates were incubated for 3 d at 4°C in the dark and grown at 24°C under a 14 h : 10 h, light : dark daily cycle. White light was provided by fluorescent tubes (Philips TLD 58W/840) at a PPFD of 100 μmol m⁻² s⁻¹.

Plasmid construction and plant transformation

GeneArt AG (Regensburg, Germany) synthesized a gene cassette designed to overexpress a C‐terminal MYC epitope tagged version of the Arabidopsis PPD1 protein in transgenic plants. This cassette included the PPD1 protein coding sequence with stop codon removed, a UBQ10 intron sequence inserted where PPD1 intron1 is normally positioned (between aa 20 and 21), and addition of a C‐terminal tag sequence encoding a linking spacer (GGGS), 3× MYC epitopes (GEQKLSEEDLN), followed by two stop codons (TAG and TAA). The cassette was cloned into pENTR221, using Gateway^® technology, and then subcloned into the Gateway adapted plant binary vector pRSh1 to produce p35SPPD1MYC. In this vector, the PPD1MYC cassette is under the control of a cauliflower mosaic virus 35S promoter. Arabidopsis Δppd mutant Ler ecotype plants grown in soil were transformed with Agrobacterium tumefaciens strain GV3101 containing p35SPPD1MYC using a modified floral dipping method (Clough & Bent, 1998). Transgenic T₁ plants were selected using Basta^®, then checked for the presence of the PPD1 transgene by PCR analysis. Segregation analysis of T₂ seedlings using phenotypic markers identified 12 independent, stable, homozygous Δppd::35SPPD1MYC transgenic lines. All lines had the small leaf characteristic of PPD1 transgene overexpression (White, 2006). Data are presented for 35SPPD1MYC‐2, 3, and 6, seedlings.

Hypocotyl measurements

At 6 d post‐germination (dpg), seedlings were sampled 1 h after the initiation of the light cycle and photographed with a Canon EDS 1300D digital camera. Fifteen seedlings were sampled of each genotype and experiments repeated at least twice (n > 30). Hypocotyl lengths were measured using NIH ImageJ software (Schindelin et al., ²⁰¹²).

To determine hypocotyl cell number and length, 6 dpg seedlings sampled 1 h into the light cycle and mounted in water were photographed using a Zeiss Axiophot microscope (Auckland, New Zealand), Leica DFC320 digital camera (Hobsonville, New Zealand), and Leica image software. The lengths of each of the individual hypocotyl cells in a protruding epidermal cell file were measured using ImageJ software. Cell numbers and lengths were measured for eight hypocotyls of each genotype.

Stomatal density analysis

To determine hypocotyl stomatal density, seedlings sampled 6 dpg were cleared in 85% lactic acid, for 1 min at 101°C, 15 psi, mounted in the same solution on microscope slides, and the total number of stomata on each hypocotyl counted with an Olympus BX50 microscope (Auckland, New Zealand) using interference contrast (n = 40).

Measurement of flowering time

Individual plants were grown in separate pots under shorter day conditions of 12 h : 12 h, light : dark. Flowering time was determined as the number of days post‐germination taken for inflorescences to attain 0.5 cm in height and the number of leaves in the rosette at flowering recorded. The experiment was replicated twice with six plants per genotype in each experiment (n = 12).

RNA‐sequencing transcriptome analysis

Three biological replicates per genotype of 10‐d‐old seedlings grown on ½MS medium were harvested 6 h into the light phase of the 14 h : 10 h, light : dark cycle and immediately frozen in liquid nitrogen (N₂). Total RNA was isolated using an RNeasy Plant Kit (74904; Qiagen) and genomic DNA contamination eliminated by on‐column DNase I digestion using an RNase‐Free DNase Set (79254; Qiagen). Oxford Gene Technology carried out the RNA‐sequencing (RNA‐seq) library preparation, high‐throughput sequencing, and data analysis. RNA quality of the samples was determined using a 2100 Bioanalyzer (Agilent, Santa Clara, CA, USA), and messenger RNA libraries prepared using an Illumina TruSeq RNA Sample Prep Kit v.2. Sample sequencing was performed on the Illumina HiSeq2000 platform using TruSeq v.3 chemistry. Sequencing was paired‐ended and over 100 cycles. Library peak insert size was c. 120 bp, and average paired‐end reads were 55 073 058 per sample. Reads were mapped to the TAIR10 Arabidopsis genome assembly using Bowtie v.2.02 (Langmead & Salzberg, 2012), and splice junctions were identified using TopHat v.2.09 (Kim et al., ²⁰¹³). Cufflinks v.2.1.1 (Trapnell et al., ²⁰¹²) was used to perform transcript assembly, abundance estimations, and differential expression between wild‐type Col‐0 and Δppd Col‐0 samples. Genes with false discovery rate (FDR) < 0.05 were considered differentially expressed. A list of Δppd differentially expressed genes (DEGs), with log₂ fold changes and P values, is given in Table S2.

Gene expression analysis using real‐time quantitative PCR

For quantitative real‐time reverse transcription (qRT)‐PCR analysis, total RNA was isolated from 6 dpg seedlings grown under a 16 h : 8 h, light : dark cycle. Seedlings were immediately frozen in liquid N₂ and RNA isolated using an RNeasy Plant kit (74 904; Qiagen). Any possible DNA contamination was removed by DNase I digestion. First‐strand complementary DNA was synthesized from 1 µg total RNA, using both random hexamer and oligo(dT)₁₈ primers and iScript Advanced Reverse Transcriptase (1725037; Bio‐Rad) according to the manufacturer's protocol. Quantitative PCR (qPCR) analyses were undertaken using a LightCycler 480 real‐time PCR instrument (Roche), LightCycler 480 SYBR Green Master reagents v.12 (Roche), and gene‐specific oligonucleotide primers designed using Primer3 software (Untergasser et al., ²⁰¹²) (Table S1). The LightCycler 480 instrument software (LightCycler) was used to calculate transcript abundance, and relative expression of a gene was determined from the ratio of test samples to reference samples. ACTIN2 and TUBULIN were the reference genes used to normalize samples. Each experiment used three biological replicates and at least four technical replicates.

Gene Ontology terms analysis

Gene set enrichment analysis was carried out using agriGO (Tian et al., ²⁰¹⁷), where enrichment of Gene Ontology (GO) terms in the Δppd mutant differential RNA‐seq dataset was determined relative to GO term frequency in the Arabidopsis genome (TAIR10 2017 assembly). GO terms with FDR < 0.05 were regarded as significantly enriched.

Western immunoblots

Immunoblots were performed on proteins extracted from 30 seedlings grown for 6 dpg with a 16 h : 8 h, light : dark cycle, PPFD 180 μmol m⁻² s⁻¹. Seedlings were homogenized in liquid N₂, resuspended in 100 µl of 2× loading buffer (18 ml of β‐mercaptoethanol was added to 1 ml of 2× NuPAGE lithium dodecyl sulphate sample buffer; NP0008; Thermo Fisher Scientific, Waltham, MA, USA), heated at 95°C for 5 min, and then centrifuged at 23°C, 17 000  g for 10 min to pellet debris. Protein concentration was measured with the Bio‐Rad RC DC protein assay. Samples adjusted to 30 µg protein were heated at 95°C for 5 min and separated on a 12% Mini‐Protean TGX gel (Bio‐Rad), Tris‐glycine buffer (25 mM Tris, 192 mM glycine, 0.1% sodium dodecyl sulphate, pH 8.3). After transfer to polyvinylidene difluoride membrane, overnight at 4°C, blots were blocked with 5% low‐fat skim milk powder and 1% polyvinylpyrrolidone for 1 h at room temperature (RT) with agitation in 1× TBS‐T (20 mM Tris‐base, 137 mM sodium chloride, 0.1% Tween 20, pH 7.6) freshly prepared from stocks of pH‐adjusted 10× Tris‐buffered saline and Tween 20. Blocking solution was discarded and the membrane incubated 1 h at RT in 1 : 5000 dilution of α HY5 (N2), R1245‐2, rabbit polyclonal antibody (Abiocode, Aqoura Hills, CA, USA) in 10 ml of TBS‐T containing 5% low‐fat skim milk powder. After discarding the primary antibody solution, the membrane was rinsed and washed three times for 15 min in 20 ml 1× TBS‐T. The membrane was then incubated with 1 : 5000 dilution of goat anti‐rabbit immunoglobulin G (IgG)‐horseradish peroxidase (HRP) (1706515; Bio‐Rad) in 10 ml 1× TBS‐T containing 5% low‐fat skim milk powder for 1 h. The blot was rinsed and washed in 20 ml 1× TBS‐T three times for 15 min. Detection was carried out with ECL SuperSignal West Dura HRP substrate (37071; Thermo Fisher Scientific) using an Azure C600 imager (Azure Biosystems, Dublin, CA, USA). Actin levels were detected by reprobing membranes with 1 : 70 000 dilution of mouse monoclonal antibodies against α‐ACTIN (A0480; Sigma), followed by 1 : 70 000 dilution of goat anti‐mouse IgG‐HRP (1706516, Bio‐Rad). PageRuler™ prestained protein ladder (26617; Thermo Fisher Scientific) was used to determine that HY5 and α‐ACTIN bands were of expected molecular weights. The experiment was performed three times. HY5 signal intensity for each genotype was determined using ImageJ software and normalized using the α‐ACTIN loading control.

Chromatin immunoprecipitation–quantitative PCR assay

For chromatin immunoprecipitation (ChIP)–qPCR assays, seedlings of Ler Δppd::35SpPPD1‐c‐Myc S‐6 were grown for 18 d with a 16 h : 8 h, light : dark cycle. An Abcam ChIP kit–plants (ab117137) was used for the analysis, following the protocol provided. Anti c‐Myc monoclonal antibodies (ab56; Abcam) were used to detect PPD1/DNA, whereas antibodies IgG and anti‐H3K9 me2 supplied with the kit were used as reference and positive controls, respectively. Briefly, 1 g of shoot tissue was infiltrated with 20 ml of 1.0% formaldehyde for 10 min, and then the DNA protein cross‐linking was quenched by adding 1.25 ml of 2.0 M glycine for 5 min. After removal of formaldehyde, tissue was reduced to a fine powder in liquid N₂. Chromatin was isolated and sonicated five times for 15 s at 4°C to produce DNA fragments of 500–700 bp in size. Anti c‐Myc, IgG, and anti‐H3K9 me2 antibodies were bound to separate strip wells in the assay plate provided. After removal of input DNA samples, chromatin was transferred to each antibody‐bound strip well. Precipitated DNA was recovered and purified using spin columns. Samples were analysed by qPCR using a LightCycler 480 real‐time PCR instrument (Roche), LightCycler 480 SYBR Green Master reagents v.12 (Roche), and either primers for SPA1 promoter fragments (Table S1), At4g03770, which is associated with demethylated H3K9, or TUBULIN. Enrichment was calculated by normalizing target DNA with ACTIN 2 DNA and then as a percentage of input DNA.

Seed weight measurements

The weight of batches of 100 seeds of each genotype were determined using a Mettler Toledo analytical balance. Results are the averages of seven replicates.

Leaf growth analysis

Plants were grown in soil with a 16 h : 8 h, light : dark cycle for 18 d. Both plants and individual (third, L3) leaves were photographed with a Canon EDS 1300D digital camera. For epidermal cellular observations the L3 leaves of four plants of each genotype were cleared in 100% ethanol overnight. Leaves were cut at the margins to allow flattening and mounted in 85% lactic acid on microscope slides. The abaxial epidermis in the half of the blade nearest to the leaf apex was examined using differential interference optics to determine cellular type composition. A representative example of the abaxial epidermis of each genotype was photographed with a Zeiss Axiophot microscope, Leica DFC320 digital camera, and Leica image software, and cell outlines drawn from a printed image.

Statistical information

Statistical significance was determined using one‐way ANOVA, followed by post hoc Dunnet's or Tukey's analysis. Pairwise statistical significance was determined using a two‐tailed Student's t‐test. The number of replicates and error bars is indicated in each of the figure legends. Statistical analysis was done in the graphing programme SigmaPlot 13.0.

Results

PPD genes modulate hypocotyl elongation

To evaluate the role PPD genes have in light‐regulated development, the hypocotyl phenotype of Arabidopsis seedlings of Δppd, a mutant with deletion of both PPD1 (At4g14714) and PPD2 (At4g14720) (White, 2006), was compared with wild‐type seedlings. Either in the original Ler ecotype or after introgression into a Col‐0 ecotype background, the Δppd mutation resulted in a hyper‐elongated hypocotyl phenotype when seedlings were grown in high light (180 μmol m⁻² s⁻¹) (Fig. 1a,b,d,e). It has been reported (Gonzales et al., ²⁰¹⁵; Liu et al., ²⁰²⁰) that KIX8 or KIX9 proteins associate with PPD1 or PPD2 to form repressor complexes controlling leaf and seed development. I found that a kix8;kix9 double mutant also has a hyperelongate hypocotyl phenotype (Fig. S1). Hence, it appears that the short hypocotyl of wild‐type seedlings grown in the presence of high light requires both PPD1/2 and KIX8/9 proteins. Introduction of a PPD1 transgene (PPD1 promoter and open reading frame) into the Δppd Ler mutant (PPD1C‐1, PPD1C‐2, PPD1‐OE) fully restored wild‐type hypocotyl length (Fig. 1a,b). This complementation of the mutant hypocotyl length occurred whether transgene expression was at wild‐type levels (PPD1C‐1, PPD1C‐2) or greatly elevated, as in the previously described (White, 2006) PPD1‐OE overexpression genotype (Fig. 1c). Complementation of the Δppd mutant hypocotyl phenotype by a PPD1 transgene with wild‐type levels of expression (PPD1C‐1, PPD1C‐2) might suggest that only PPD1, and not PPD2, is required for normal photomorphogenesis. However, because the hypocotyl length of null mutants with T‐DNA inserts in PPD1 (ppd1‐1, ppd1‐2) or PPD2 (ppd2‐1) (Fig. 1d,e) was not significantly different from Col‐0 wild‐type, it appears that either of the two PPD genes can regulate photomorphogenic development.

Fig. 1.

PEAPODs (PPDs) regulate seedling hypocotyl length in Arabidopsis thaliana. (a) Seedlings of wild‐type, Δppd mutant, and mutant complemented with a PPD1 transgene (PPD1C‐1, PPD1C‐2, and PPD1‐OE) all in a Landsberg erecta (Ler) ecotype background. (b) Hypocotyl lengths of n > 30 of each genotype shown in (a). (c) PPD1 expression levels in PPD1C‐1, PPD1C‐2 and PPD1‐OE seedlings, relative to Ler wild‐type, determined by quantitative reverse transcription PCR. (d) Seedlings of wild‐type, Δppd mutant, PPD1‐OE transgenic and transfer DNA insertion mutations in PPD1 (ppd1‐1, ppd1‐2) or PPD2 (ppd2‐1) all in a Col‐0 ecotype background. (e) Hypocotyl lengths of n > 30 of each genotype shown in (d). (f) Comparison of the number of protruding epidermal cells in a file the length of hypocotyls of Col‐0, Δppd, and PPD1‐OE seedlings. (g) Mean lengths of individual protruding epidermal cells in files from apical to basal position on hypocotyls of Col‐0, Δppd, and PPD1‐OE seedlings. Cell numbers and lengths were measured for eight hypocotyls of each genotype. Bar, 2 mm. Error bars represent mean ± SEM. Seedlings sampled 6 d post‐germination were grown in soil at 22°C, with a 16 h : 8 h, white light (180 µmol m⁻² s⁻¹) : dark cycle. The statistical significance in (b, e, f) was determined by one‐way ANOVA, with Dunnet's post hoc analysis (P < 0.001). Equivalent groups are denoted by different letters. Box plots denote a range from the first to the third quartiles, with the median indicated by a narrow horizontal line, the fifth percentile by a thick horizontal line, and the 95^th percentile by a circle. The whiskers below and above the box plot are 1.5 times the interquartile range. The statistical significance in (c) was determined with Student's t‐test, ns, not significant. **, P < 0.001.

Since PPD genes have a known role controlling cell proliferation during leaf and seed development, a comparison was made of the number and mean lengths of protruding epidermal cells over the length of hypocotyls of wild‐type Col‐0, Δppd and PPD1‐OE genotypes (Fig. 1e,f). Because neither the loss of PPDs nor the overexpression of PPD1 altered the number of protruding cells in hypocotyls, the long hypocotyl phenotype of the Δppd mutant appears to be solely due to the observed elongation of hypocotyl cells (Fig. 1f).

To further evaluate the influence of PPD expression on photomorphogenesis, I tested the effect of darkness, continuous red or blue light at 10 µmol m⁻² s⁻¹, or 16 h of 60 µmol m⁻² s⁻¹ white light followed by 8 h darkness on the hypocotyl lengths of Col‐0, Δppd, and PPD1‐OE seedlings (Fig. 2a). Though not significantly different from wild ‐type when grown in darkness, the Δppd mutant had longer hypocotyls and PPD1‐OE shorter hypocotyls when grown with continuous red or blue or low‐intensity white light. Furthermore, additional PPD1 overexpression transgenic lines, obtained by introducing a 35SPPD1MYC cassette into Δppd Ler, also had shorter hypocotyls than wild‐type Ler when grown in low‐intensity light conditions (Fig S2). Therefore, the PPD genes appear to act as positive regulators of photomorphogenesis, limiting hypocotyl elongation in a range of light conditions.

Fig. 2.

PEAPOD (PPD) expression is regulated by light intensity and modulates photomorphogenesis in Arabidopsis thaliana. (a) Effect of darkness (Dc), continuous red (Rc) or blue (Bc) light at 10 µmol⁻² s⁻¹, or 16 h : 8 h, white light (60 µmol⁻² s⁻¹) : darkness (W 60), on the hypocotyl length of wild‐type, Δppd, and PPD1‐OE seedlings, all in Col‐0 ecotype background, 6 d post‐germination. Different letters indicate statistically equivalent groups as assessed by one‐way ANOVA with Tukey's post hoc analysis (P < 0.05). Box plots denote a range from the first to the third quartiles, with the median indicated by a narrow horizontal line, the fifth percentile by a thick horizontal line, and the 95^th percentile by a circle. The whiskers below and above the box plot are 1.5 times the interquartile range. (b) Comparison of PPD1 and PPD2 expression in Col‐0 seedlings grown with high (180 µmol m⁻² s⁻¹) relative to low (20 µmol m⁻² s⁻¹) white light intensity. (c) Expression of PPD1 and PPD2 in mutants phyB, cry1, cop1‐6, hy5, and pif4‐2 or overexpression transgenic PHYBOE and 35SPIF4‐MYC seedlings grown with 180 µmol m⁻² s⁻¹ white light intensity, relative to wild‐type (WT). Error bars represent mean ± SEM. Statistical significance differences in (b, c) from wild‐type were determined by the Student's t‐test, *, P < 0.01; **, P < 0.001. Differences between PPD1 and PPD2 were not significant. Seedlings used for the quantitative reverse transcription PCR expression analysis were sampled 6 d post‐germination and grown in soil at 22°C, with a 16 h : 8 h, light : dark cycle.

PPD expression is regulated by light intensity

An examination of Arabidopsis community RNA‐seq databases (Arabidopsis RNA‐seq Database, The Arabidopsis Information Resource, Genevisible) revealed both PPD1 and PPD2 are likely to be upregulated by dark or shade treatments and downregulated at higher photon irradiances. The regulation of PPD1/2 expression by light intensity was confirmed by qRT‐PCR analysis, comparing expression of wild‐type seedlings grown in high vs low light levels (Fig. 2b) with expression downregulated in high light. Community RNA‐seq data also suggested that PPD2 expression may be regulated by components of the light signalling network. An analysis of PPD1/2 expression in various light‐signalling mutants indicated both genes are upregulated in phyB, cry1, and hy5 mutants and, conversely, downregulated in cop1‐6 and PHYBOE seedlings (Fig. 2c). Hence, expression of the PPD genes is likely to be regulated by a component of the light‐signalling pathway that is in turn modulated by red or blue light intensity. Because phyB, cry1, and hy5 are likely to have elevated levels and cop1‐6 and PHYBOE reduced levels of PIF protein activity, I examined the possibility that expression of PPD1 and PPD2 is regulated by PIFs. Published ChIP data indicate that PIF4 may bind to chromatin 5′ of PPD2 (Oh et al., ²⁰¹²). Significantly, the expression of PPD1 and PPD2 was downregulated in pif4‐2, a loss‐of‐function mutant, and upregulated in pif4‐2::35SPIF4‐MYC, a PIF4 overexpression genotype (Fig. 2c). Together, these results suggest PIF4 acts as an activator of PPD1 and PPD2 transcription. Further investigation is required to discover if this regulation of PPD expression by PIF4 is direct or indirect.

Genetic analysis illustrates PPDs are positive regulators of photomorphogenesis

Genetic analysis was undertaken to examine PPD interactions with the light‐signalling network (Fig. 3). The additive effect on hypocotyl elongation of double mutants phyB;Δppd and cry1;Δppd suggests PPDs act as positive regulators of the light‐signalling network downstream of the red and blue photoreceptors. Since photoactivated phyB and cry1 inhibit multiple regulatory hubs in light signalling, I also determined the effects of PHYBOE;Δppd, hy5;Δppd and pif4‐2;Δppd combinations on hypocotyl length (Fig. 3). Overexpression of PHYB was completely epistatic to the Δppd mutant, preventing hypocotyl elongation. The hy5;Δppd double mutant exhibited an additive hyperelongated hypocotyl phenotype, whereas the pif4‐2;Δppd combination, although significantly longer than pif4‐2, was closer to wild‐type in length. Although these genetic interaction results provided additional evidence that PPDs act to modulate photomorphogenesis, they did not reveal how PPDs intersect with the light‐signalling cascade.

Fig. 3.

Genetic analysis shows that PEAPODs (PPDs) are positive regulators of photomorphogenesis and a component of light signalling in Arabidopsis thaliana. (a) Hypocotyl length of Landsberg erecta (Ler) wild‐type, Δppd Ler, phyB, phyB;Δppd, phyBOE, and phyBOE;Δppd. phyB, phytochrome B.(b) Hypocotyl length of Col‐0 wild‐type, Δppd Col‐0, cry1, cry1;Δppd, hy5, hy5;Δppd, pif4‐2, and pif4‐2;Δppd seedlings. cry1, cryptochrome 1; hy5, elongated hypocotyl 5; pif4, phytochrome interacting factor 4. Seedlings were sampled after 6 d growth in 180 µmol m⁻² s⁻¹ light, 16 h : 8h, light : dark conditions. n > 30 for each genotype. Statistical significance differences were determined by one‐way ANOVA followed by Tukey's post hoc analysis (P < 0.05). Different letters denote equivalent groups. Box plots denote a range from the first to the third quartiles, with the median indicated by a narrow horizontal line, the fifth percentile by a thick horizontal line, and the 95^th percentile by a circle. The whiskers below and above the box plot are 1.5 times the interquartile range.

PPDs modulate light signalling by negative regulation of SPA1 expression

To understand transcriptome changes underlying the Δppd elongated hypocotyl phenotype, RNA isolated from Δppd and wild‐type seedlings was compared by RNA‐seq analysis. A total of 2830 DEGs were expressed in Δppd compared with wild‐type, with 927 (33%) upregulated and 1903 (67%) downregulated (Table S2). Upregulated and downregulated Δppd DEGs were separately analysed for enrichment of GO terms (Fig. 4a,b). GO terms significantly overrepresented in the Δppd mutant upregulated DEGs were those involving epidermal/stomatal development and gene regulation. Since there is extended proliferation of meristemoids (stomatal precursor stem cells) (White, 2006) and increased stomatal density in hypocotyls of the Δppd mutant (Fig. S3), it was not surprising that expression of genes regulating the proliferative steps of stomatal development – HOMEODOMAIN GLABROUS 2 (HDG2, At1g05230), SCREAM 2 (SCRM2, At1g12860), CLAVATA3/ESR‐RELATED 9 (CLE9, At1g26600), EPIDERMAL PATTERNING FACTOR 2 (EPF2, At1g34245), TOO MANY MOUTHS (TMM, At1g80080), ARABIDOPSIS RESPONSE REGULATOR 16 (ARR16, At2g40670), SCREAM (SCRM, At3g26744), AGAMOUS‐LIKE 16 (AGL16, At3g57230), ERECTA‐LIKE 2 (ERL2, At5g07180), SPEECHLESS (SPCH, At5g53210), BREAKING OF ASYMMETRY IN THE STOMATAL LINEAGE (BASL, At5g60880), ERECTA‐LIKE 1 (ERL1, At5g62230) – were all upregulated in the mutant.

Fig. 4.

Comparison of genome‐wide differential gene expression between wild‐type and Δppd mutant revealed PEAPODs (PPDs) regulate the light signalling network in Arabidopsis thaliana (At). A selected subset of Gene Ontology terms significantly enriched in genes identified by RNA‐sequencing analysis is shown: (a) downregulated; (b) upregulated. (c) Venn diagram showing the overlap of downregulated and upregulated Δppd differentially expressed genes (DEGs) with light‐regulated ELONGATED HYPOCOTYL 5 (HY5) direct target genes. (d) Western blot detection of HY5 protein in wild‐type, Δppd, and PPD1‐OE seedlings. HY5 was detected with an anti‐HY5 (N2) antibody, and blots were reprobed with an anti‐α‐ACTIN antibody as loading control. (e) Relative HY5 levels normalized against α‐ACTIN loading controls using ImageJ. Three biological replicates. (f) Quantitative reverse transcription PCR analysis of (SUPPRESSOR OF phyA‐105) SPA1 transcript levels in Δppd, PPD1‐OE, cop1‐6, hy5, hy5;Δppd, pif4‐2, 35SPIF4‐MYC, and kix8;kix9 seedlings, relative to Col‐0 wild‐type. cop1, constitutive photomorphogenic 1; hy5, elongated hypocotyl 5; kix, kinase‐inducible domain interacting; pif4, phytochrome interacting factor 4. (g) Comparison of SPA1 expression in Col‐0 seedlings grown with high (180 µmol m⁻² s⁻¹) relative to low (20 µmol m⁻² s⁻¹) white light intensity. Error bars represent mean ± SEM. Statistical significance differences from wild‐type were determined by Student's t‐test, *, P < 0.01; **, P < 0.001.

GO terms enriched in downregulated DEGs predominantly related to responses to light and light intensity. Most notably, flavonoid biosynthesis genes that are direct targets of HY5 activation (Burko et al., ²⁰²⁰) – FLAVONOL‐7‐O‐RHAMNODYLTRANSFERASE (UGT89C1, At1g06000), PRODUCTION OF ANTHOCYANIN PIGMENT 1 (PAP1, At1g56650), 4‐COUMARATE COA LIGASE 3 (4CL3, At1g65060), PHE AMMONIA LYASE 1 (PAL1, At2g37040), CINNAMYL ALCOHOL DEHYDRASE 4 (CAD, At3g19450), CHORISMATE MUTASE 1 (CM1, At3g29200), FLAVANONE 3‐HYDROXYLASE (F3H, At3g51240), CHALCONE ISOMERASE (CHI, At3g55120), CHALCONE ISOMERASE LIKE (CHIL, At5g05270), FLAVONOL SYNTHASE 1 (FLS1, At5g08640), CHALCONE SYNTHASE (CHS, At5g13930) – were downregulated in the Δppd DEGs dataset. I also identified additional HY5 target genes in the Δppd dataset (Table S2). Of 297 genes reported as high‐confidence HY5 direct targets (Burko et al., ²⁰²⁰), 71 were differentially expressed in Δppd, a 2.34‐fold enrichment compared with the by chance prediction of 30, given 27 655 protein‐coding loci in the Arabidopsis genome (Cheng et al., ²⁰¹⁷). Interestingly, 64 of the 71 HY5 target genes identified as DEGs in the mutant were downregulated (Fig. 4c). This result suggested reduced HY5 function in the Δppd mutant. Although HY5 transcription was not altered (Fig. S4), the amount of HY5 protein was reduced in Δppd and increased in PPD1‐OE compared with wild‐type seedlings (Fig. 4d,e). Therefore, PPDs affect the level of HY5 protein, which is known to be targeted for degradation by the COP1/SPA complex (Saijo et al., ²⁰⁰³).

Because HY5 is an antagonist to PIF activity, I examined the Δppd DEGs dataset for alterations in expression of PIF‐target genes. Increased expression of PIF‐target genes that are positive regulators of hypocotyl growth in shade, TRYPTOPHAN AMINOTRANSFERASE OF ARABIDOPSIS 1 (TAA1, At1g70560), WAG2 (At3g14370), and HOMEOBOX‐LEUCINE ZIPPER PROTEIN 3 (HAT3, At3g60390), suggested elevated PIF activity in the Δppd mutant (Leivar & Monte, 2014; Li et al., ²⁰²¹). Upregulation of TAA1 and HAT3 in the Δppd mutant was confirmed by qRT‐PCR analysis (Fig. S5a). The possibility that PPDs alter PIF activity was supported by downregulation of PHOTOSYSTEM II LIGHT HARVESTING COMPLEX GENE 2.3 (LHCB2.3, At3g27690) and PHOTOSYSTEM II SUBUNIT Q‐2 (PSBQ‐2, At4g05180), PIF‐repressed, photosynthesis‐related genes (Pham et al., ²⁰¹⁸).

It has been reported that light limits hypocotyl elongation by inhibiting brassinosteroid (BR) biosynthesis via a PIF‐mediated mechanism (Wei et al., ²⁰¹⁷; Martínez et al., ²⁰¹⁸). For example, PIF4 mediates the de‐repression of DWF4 (At3g50660), which encodes a rate‐limiting step in BR biosynthesis. The BR response to light signalling is also mediated by BAS1 (At2g26710), a cytochrome P450 involved in BR metabolism (Sandhu et al., ²⁰¹²). Interestingly, DWF4 transcription was upregulated in the Δppd mutant but downregulated in PPD1‐OE, whereas expression of BAS1 was down‐egulated in Δppd and upregulated in PPD1‐OE (Fig. S5b). These results suggest BRs may increase in the absence of PPDs and decrease with PPD1 overexpression, possibly by the influence of PPDs on PIF activity.

Hypocotyl elongation is also regulated by interaction between the light and gibberellin (GA) signalling pathways. Accumulation of GA‐signalling DELLA repressors suppresses PIF activity by disrupting interactions between PIFs and their target gene promoters (De Lucas et al., ²⁰⁰⁸; Feng et al., ²⁰⁰⁸) and by stimulating PIF protein degradation (Li et al., ²⁰¹⁶). Because DELLA proteins are destabilized by the COP1/SPA complex (Blanco‐Touriñán et al., ²⁰²⁰), resulting in increased GA signalling, I examined the Δppd DEGs dataset for examples of GA‐responsive gene expression. Significantly, the GA‐activated GIBBERELLIN‐REGULATED PROTEIN 1 (GASA1, At1g75750) was upregulated. Also, GIBBERELLIN 20‐OXIDASE 1 (GA5, At4g25420), which encodes an enzyme converting inactive GA precursors into active GA and is negatively regulated by GA (Bouquin et al., ²⁰⁰¹), was downregulated in the Δppd mutant. These results suggested an increase in GA signal transmission in the Δppd mutant.

Next, the Δppd DEGs dataset was scanned for any altered expression of light‐signalling genes that could result in reduced levels of HY5 protein, elevated PIF activity, and increased GA signalling. Intriguingly, expression of SPA1 (At2g46340) was elevated in the Δppd mutant (Table S2). The resulting increase in COP1/SPA activity could account for the altered light signalling found in the Δppd mutant. It is also noteworthy that many aspects of the Δppd seedling phenotype – longer hypocotyls than wild‐type when grown in continuous blue or red light, or 16 h low‐intensity white light (Fig. 2a), and reduced expression of CHS (Table S1) – resemble SPA1 overexpression genotypes (Yang & Wang, 2006). SPA1 was subsequently examined as a possible target for PPD‐mediated modulation of photomorphogenesis. qRT‐PCR expression analysis showed SPA1 was upregulated in the mutant but downregulated in PPD1‐OE seedlings (Fig. 4f). Furthermore, SPA1 expression increased in the kix8;kix9 double mutant. Hence, the PPD/KIX repressor complex appears to act as a negative regulator of SPA1 expression. Since it has been reported that HY5 directly activates SPA1 expression (Burko et al., ²⁰²⁰), I determined the effect of hy5 mutation on SPA1 expression. As expected, hy5 mutant seedlings had lower SPA1 transcript levels than wild‐type did. By contrast, SPA1 expression was upregulated in cop1‐6, pif4‐2 and an hy5;Δppd double mutant, but downregulated in PIF4 overexpression (35SPIF4‐MYC) seedlings. It has been reported that a switch from dark to light results in the upregulation of SPA1 transcription (Fittinghoff et al., ²⁰⁰⁶; Balcerowcz et al., ²⁰¹¹). qRT‐PCR analysis comparing SPA1 transcripts in wild‐type seedlings grown in high vs low light levels showed upregulation of expression in high light (Fig. 4g). These results, with SPA1 expression being the opposite of PPD1/2 in a range of light conditions and light‐signalling mutants (Figs 2, 4), are consistent with repression of SPA1 transcription by PPDs.

SPA1 is essential for PPD modulation of photomorphogenesis

To test the hypothesis that PPDs modulate photomorphogenesis by negative regulation of SPA1 expression, hypocotyl lengths of combinations between Δppd and PPD1‐OE, and a recessive mutant allele of spa1 (spa1‐7, a T‐DNA insertion mutant), were determined for seedlings grown in a variety of light conditions (Fig. 5a,b). In either higher (180 µmol m⁻² s⁻¹) or lower white light (60 µmol m⁻² s⁻¹), or continuous red or blue light (10 µmol m⁻² s⁻¹), the spa1 mutation was epistatic to both Δppd and PPD1‐OE. This spa1 epistasis occurred at a wide range of continuous red or blue fluence rates (Fig. 5c,d). Furthermore, hypocotyls of the Δppd mutant were longer, and PPD1‐OE shorter, than wild‐type at all red or blue fluence rates examined.

Fig. 5.

SUPPRESSOR OF phyA‐105 (SPA1) is essential for PEAPOD (PPD) modulation of photomorphogenesis and flowering time in Arabidopsis thaliana. (a) Seedlings of Col‐0, Δppd mutant, PPD1‐OE, spa1, spa1;Δppd, and spa1;PPD1‐OE, after 6 d growth in either 180 µmol m⁻² s⁻¹ (W 180) or 60 µmol m⁻² s⁻¹ (W 60) white light 16 h : 8 h, light : dark or continuous red (Rc) or blue (Bc) light at 10 µmol m⁻² s⁻¹, or continuous dark (Dc). (b) Hypocotyl lengths of n > 30 of each genotype and light condition shown in (a). Fluence dose response curves for seedlings of Col‐0, Δppd mutant, PPD1‐OE, spa1, spa1;Δppd, and spa1;PPD1‐OE, in continuous (c) red or (d) blue light, after 6 d (n = 20). Flowering time for plants grown in short day length of 12 h : 12 h, light : dark was determined as (e) number of leaves at bolt and (f) days to flowering, for n = 12 for each genotype. Statistical significance differences from wild‐type were determined by one‐way ANOVA followed by Tukey's post hoc analysis (P < 0.05). Different letters denote equivalent groups. Box plots denote a range from the first to the third quartiles, with the median indicated by a narrow horizontal line, the fifth percentile by a thick horizontal line, and the 95^th percentile by a circle. The whiskers below and above the box plot are 1.5 times the interquartile range. Error bars in (c) and (d) are ± SEM.

SPA1 regulates photoperiodic flowering time in Arabidopsis plants grown with short days by interacting with and targeting the degradation of CONSTANS (CO), an activator of FLOWERING‐TIME (FT) transcription (Laubinger et al., ²⁰⁰⁶; Ranjan et al., ²⁰¹¹). Consequently, spa1 mutants flower early when grown in short day conditions. When grown in short days, 12 h : 12 h, light : dark, Δppd mutant plants flowered later and had a greater number of leaves at bolt than wild‐type plants did (Fig. 5e,f). By contrast, PPD1‐OE plants began flowering earlier and with fewer leaves than wild‐type plants did. Significantly, spa1 was epistatic for flowering to either Δppd or PPD1‐OE. Additionally, because SPA1 function requires COP1, I also tested genetic combinations between Δppd and PPD1‐OE and cop1‐6. The cop1‐6 mutant was epistatic to both Δppd and PPD1‐OE for hypocotyl length in cop1‐6;Δppd and cop1‐6;PPD1‐OE seedlings (Fig. S6). These genetic interaction results confirm that SPA1 is required for PPD modulation of photomorphogenesis.

PPD repression of SPA1 expression is possibly indirect

Data from published ChIP‐seq assays using affinity tags PPD2‐HBH or PPD2‐GS^yellow expressed in Arabidopsis cell cultures did not identify SPA1 as a possible direct target of PPD2 (Besbrugge et al., ²⁰¹⁸). Furthermore, ChIP–qPCR analysis using transgenic 35SPPD1‐Myc seedlings did not detect binding of PPD1 to 1045 bp of the SPA1 promoter (Fig. S7). Taken together these results suggest repression of SPA1 expression by PPDs is via an indirect mechanism.

PPDs modulate stomatal development via SPA1

Previous reports indicated PPDs directly control leaf flatness by repressing CYCD3 gene expression (Baekelandt et al., ²⁰¹⁸) and seed size by repressing expression of GIF1 (Liu et al., ²⁰²⁰). The regulation of PPD gene expression by PIF4 (Fig. 2) suggests a possible mechanism coordinating the complex set of developmental responses to light intensity, some by the direct action of PPDs and others via PPD regulation of light signal transmission. But which developmental responses to light intensity, other than hypocotyl elongation, are controlled by PPD modulation of light signalling? Comparison of seed weights of Col‐0, Δppd, PPD1‐OE, spa1, spa1:Δppd, and spa1;PPD1‐OE plants showed that SPA1 is not required for PPD regulation of seed development (Fig. S8). Deletion or downregulation of the PPD genes results in mutant plants with a propeller‐like rosette of narrow, enlarged, dome‐shaped leaves. The characteristic leaf phenotype of ppd mutants is due to the combined effects of changes to both the primary and secondary cell cycle arrest fronts during development (White, 2006; Karidas et al., ²⁰¹⁵; Baekelandt, et al., ²⁰¹⁸). Although the position of the primary cell cycle arrest front is not altered in ppd mutants, its exaggerated convex shape results in additional cell proliferation in the centre of the leaf blade. Delayed progression of the secondary cell cycle arrest front results in the extended proliferation of meristemoids, stomatal precursor cells. A previous report (Baekelandt et al., ²⁰¹⁸) showed direct PPD regulation of CYCD3 gene expression controls the primary cell proliferation front but does not influence meristemoid cell proliferation during leaf development. The COP1/SPA complex has previously been implicated in the light‐mediated control of stomatal development (Kang et al., ²⁰⁰⁹; Ranjan et al., ²⁰¹¹). To determine if PPD regulation of SPA1 expression controls stomatal development, I compared the leaf curvature, leaf meristemoid cell proliferation, and hypocotyl stomatal density of Col‐0, Δppd, PPD1‐OE, spa1, spa1:Δppd, and PPD1‐OE:Δppd genotypes (Fig. 6). The intermediate leaf curvature phenotype of spa1:Δppd and the small flat leaves of spa1;PPD1‐OE plants (Fig. 6a,b) suggest SPA1 is unlikely to be involved in PPD regulation of the primary cell proliferation front during leaf development. Further research examining the shape of the primary cell proliferation arrest front in these combination genotypes is required. However, the absence of extended meristemoid cell proliferation in leaves of spa1:Δppd double mutants and the epistasis of spa1 to Δppd for hypocotyl stomatal number (Fig. 6c,d) clearly show that SPA1 acts downstream of PPDs to control stomatal density. Hence, it is likely that PPDs control photomorphogenesis, flowering time, and stomatal density indirectly by regulating SPA1, but seed size and leaf flatness by direct gene repression. Further investigation is required to establish whether the increased stomatal density in Δppd mutant hypocotyls is due to extended proliferation of meristemoid cells or to the recruitment of more protodermal cells to stomatal development.

Fig. 6.

SUPPRESSOR OF phyA‐105 (SPA1) is required for PEAPOD (PPD) regulation of meristemoid epidermal cell proliferation and stomatal density in Arabidopsis thaliana. (a) Photographs of Col‐0, Δppd, PPD1‐OE, spa1, spa1; Δppd, and spa1;PPD1‐OE plants at 18 d. (b) Side view of third leaf (L3) at 18 d. (c) Representative drawings of cells in the abaxial epidermis of the third leaves of spa1, Δppd, and Δppd;spa1 18‐d‐old plants. Stomata are illustrated in yellow and meristemoids in red. (d) Stomatal density on hypocotyls of Col‐0, Δppd, PPD1‐OE, spa1, spa1;Δppd, and spa1;PPD1‐OE seedlings at 6 d post‐germination, n = 40. Statistical significance differences from wild‐type or PPD1‐OE were determined by one‐way ANOVA with Tukey's post hoc test (P < 0.05). Different letters indicate statistically equivalent groups. Box plots denote a range from the first to the third quartiles, with the median indicated by a narrow horizontal line, the fifth percentile by a thick horizontal line, and the 95^th percentile by a circle. The whiskers below and above the box plot are 1.5 times the interquartile range. Plants and seedlings were grown at 22°C, with a 16 h : 8 h, white light (180 µmol m⁻² s⁻¹) : dark cycle.

Discussion

Anchored in the soil and utilizing light for energy, higher plants adapt to different light intensities with coordinated changes in growth and development. In this study, I show that PPD proteins are positive regulators of light signalling in Arabidopsis, acting to prevent excessive hypocotyl elongation when seedlings are grown in either higher or lower light intensity. PPDs also control seed size, the density of stomata, and flowering time (Liu et al., ²⁰²⁰; Figs 5, 6, S8). I found expression of PPD1 and PPD2 is activated by PIF4, with transcription upregulated in lower light and downregulated in higher light (Fig. 2b,c). PPDs modulate the transmission of light signalling by repression/de‐repression of SPA1 expression (Fig. 4f). SPA1 is required for PPD regulation of hypocotyl elongation, stomatal density, and flowering time but is not involved in the PPD control of seed size. Therefore, PPD1/2 expression coordinates developmental responses to light intensity in part indirectly by modulating light signalling (Figs 3, 4, 5, 6), and in part by directly targeting gene expression (Baekelandt et al., ²⁰¹⁸; Liu et al., ²⁰²⁰).

I propose the following working model for the modulation of light signalling by PPDs (Fig. 7). In seedlings grown in lower light intensities, high PIF4 activity results in increased PPD1/2 expression. The resulting higher levels of PPDs repress SPA1 transcription by an indirect mechanism, leading to reduced COP1/SPA activity and greater stability of HY5 and HFR1, thereby preventing exaggerated etiolation. This signalling cascade also antagonizes PIF activity, acting as a negative feedback loop to constrain PPD1/2 expression. In higher light intensities, activated phytochromes and cryptochromes disrupt PIF4 activity, resulting in lower levels of PPD1/2 expression and de‐repression of SPA1 transcription. The resulting increase in COP1/SPA activity reduces the levels of HY5 and HFR1, providing a mechanism to prevent hyper‐photomorphogenesis and optimize plant growth in the presence of higher light. This repression/de‐repression of signalling is appropriate for situations, such as variable light intensity, where a plant’s exaggerated response is a greater risk than a restrained response. Expression of SPA1 is regulated by the concerted action of two light‐intensity‐mediated mechanisms: activation by HY5 (Burko et al., ²⁰²⁰), and the repression/de‐repression by PPDs described here. Concerted molecular switching mechanisms provide more accurate signal transduction than activation or de‐repression alone (Ghusinga et al., ²⁰²¹). In addition, the levels of both HY5 and PPDs are controlled by COP1/SPA‐mediated negative feedback mechanisms. I propose this combination of modulation mechanisms provides a ‘rheostat‐like’ adaptive response, adjusting sensitivity of the signalling network to extremes of light intensity. In high light, the combination of enhanced HY5 activation and PPD de‐repression of SPA1 expression limits light signal transmission. In low light, however, diminished HY5 activation and strong PPD repression, resulting in reduced SPA1 transcription, increases sensitivity to light signalling. Protein cofactors also play an essential role in the HY5 and PPD regulation of SPA1 expression. HY5 lacks a transcription activation domain and requires B‐box‐containing proteins BBX20, 21, 22 as essential cofactors for the control of hypocotyl elongation (Bursch et al., ²⁰²⁰). Similarly, PPDs require KIX8/9 to form a functional repressor complex (Gonzalez et al., ²⁰¹⁵; Liu et al., ²⁰²⁰) to limit hypocotyl elongation (Fig. S1) and restrict SPA1 transcription (Fig. 5). Chromatin remodelling may also contribute to the PPD repression of target gene transcription. PPDs interact with adaptors KIX8/9 or NINJA to recruit TPL and engage histone deacetylases to mediate transcriptional repression of target genes (Gonzalez et al., ²⁰¹⁵; Baekelandt et al., ²⁰¹⁸). Another PPD‐interacting protein, LHP1, is a component of the Polycomb complex that reads and maintains H3K27me3 chromatin silencing (Zhu et al., ²⁰¹⁹).

Fig. 7.

A proposed model for PEAPOD (PPD) repressor modulation of light signal transmission. In low light the concerted action of reduced ELONGATED HYPOCOTYL 5 (HY5) activation and PPD repression limits SUPPRESSOR OF phyA‐105 (SPA1) expression resulting in increased sensitivity of light signalling. In high light, enhanced HY5 activation and PPD de‐repression of SPA1 expression limits light signalling. COP1, CONSTITUTIVE PHOTOMORPHOGENIC 1; cry1, cryptochrome 1; HFR1, LONG HYPOCOTYL IN FAR‐RED; phyB, phytochrome B; PIF4, PHYTOCHROME INTERACTING FACTOR 4. Thick lines indicate strong action, and thin lines indicate weaker effects. Arrows represent activation, and bars represent repression. Solid lines represent direct action, whereas unknown molecular mechanisms are represented by dashed lines. [Correction added after online publication 25 May 2022: the right‐hand panel of the figure has been updated.]

Although hypocotyl lengths of PPD1 overexpression seedlings grown in lower light were shorter, hypocotyls of the same genotype grown in higher light intensity were not significantly different in length from wild‐type (Figs 1, 2, S2). The difference in hypocotyl length between wild‐type and spa1 mutant was also substantially less when seedlings were grown in high light than in lower light (Fig. 5). This may be because, at higher light intensities, elevated levels of activated phyB and crys limit COP1/SPA1 function and PIF activity to such an extent that any decrease in SPA1 expression has no significant effect on hypocotyl elongation. Notably, the reduced SPA1 expression in PPD1‐OE seedlings did influence stomatal development and flowering time. PPD1‐OE seedlings grown in high light had a lower hypocotyl stomatal density and flowered earlier than wild‐type seedlings (Figs 5, 6). These differences in response may reflect which parts of the light signalling network influence a particular trait. For example, the COP1/SPA1 complex controls flowering time in Arabidopsis by directly targeting CO degradation (Laubinger et al., ²⁰⁰⁶; Ranjan et al., ²⁰¹¹), whereas both PIFs and COP1/SPA indirectly determine hypocotyl elongation (Leivar & Monte, 2014).

Because PPDs fine‐tune higher plant developmental responses to light intensity, they could be a target for improving the yield and resilience of crops. PPD orthologues are present in a wide range of higher plants, including some monocots, but they appear to be absent from the Poaceae (grasses) (White, 2006). Despite this limitation, loss of PPD/KIX repressor function has been shown to result in increased biomass and seed size, highly valuable agronomic traits, in a variety of eudicot plants, including soybean (Ge et al., ²⁰¹⁶; Nguyen et al., ²⁰²⁰). However, because PPDs regulate a complex set of traits, simply reducing PPD/KIX function could result in unwanted pleiotropic effects. These may be avoided by using a targeted approach, changing a few PPD‐regulated traits to improve crop yields, while not disrupting most of the response to light intensity. Seed size is a trait that might be ‘uncoupled’ from the PPD‐mediated coordinated response to light intensity (Fig. S8). This and other potential applications of PPDs for crop improvement would benefit from a more comprehensive knowledge of the mechanisms and promoter DNA elements involved in PPD‐mediated repression of gene expression.

Author contributions

DWRW designed and performed the research, analysed data, and wrote the manuscript.

Supporting information

Fig. S1 Hypocotyl length of kix8;kix9 double mutant.

Fig. S2 Overexpression of PPD1 limits hypocotyl elongation in low but not high light intensity.

Fig. S3 Stomatal density.

Fig. S4 HY5 expression.

Fig. S5 Expression of PIF4‐regulated genes.

Fig. S6 COP1 is required for PPD modulation of photomorphogenesis.

Fig. S7 ChIP assay.

Fig. S8 PPDs influence seed weight independently of SPA1.

Table S1 Genotyping and qPCR primers.

Table S2 RNA‐seq dataset.

Please note: Wiley Blackwell are not responsible for the content or functionality of any Supporting Information supplied by the authors. Any queries (other than missing material) should be directed to the New Phytologist Central Office.

Data availability

Data supporting results reported in this papert are available from the author upon request. Data underlying the RNA‐seq analysis can be accessed at the National Center for Biotechnology Information following the link https://www.ncbi.nlm.nih.gov/sra/PRJNA826666.