Petunia PHYTOCHROME INTERACTING FACTOR 4/5 transcriptionally activates key regulators of floral scent

Floral scent emission of petunia flowers is regulated by light conditions, circadian rhythms, ambient temperature and the phytohormones GA and ethylene, but the mechanisms underlying sensitivity to these factors remain obscure. PHYTOCHROME INTERACTING FACTORs (PIFs) have been well studied as components of the regulatory machinery for numerous physiological processes. Acting redundantly, they serve as transmitters of light, circadian, metabolic, thermal and hormonal signals. Here we identified and characterized the phylogenetics of petunia PIF family members (PhPIFs). PhPIF4/5 was revealed as a positive regulator of floral scent: TRV-based transient suppression of PhPIF4/5 in petunia petals reduced emission of volatiles, whereas transient overexpression increased scent emission. The mechanism of PhPIF4/5-mediated regulation of volatile production includes activation of the expression of genes encoding biosynthetic enzymes and a key positive regulator of the pathway, EMISSION OF BENZENOIDS II (EOBII). The PIF-binding motif on the EOBII promoter (G-box) was shown to be needed for this activation. As PhPIF4/5 homologues are sensors of dawn and expression of EOBII also peaks at dawn, the prior is proposed to be part of the diurnal control of the volatile biosynthetic machinery. PhPIF4/5 was also found to transcriptionally activate PhDELLAs; a similar positive effect of PIFs on DELLA expression was further confirmed in Arabidopsis seedlings. The PhPIF4/5–PhDELLAs feedback is proposed to fine-tune GA signaling for regulation of floral scent production. Supplementary Information The online version contains supplementary material available at 10.1007/s11103-024-01455-8.

Floral VOC production is controlled at different levels.Transcriptional activation of biosynthetic genes includes MYB-transcription factors: ODORANT1 (ODO1) and EMISSION OF BENZENOIDS (EOB) I, and EOBII.The latter acts upstream of EOBI and ODO1 (Verdonk et al. 2005;Spitzer-Rimon et al. 2010, 2012;Colquhoun et al. 2011).GRAS protein PHENYLPROPANOID EMISSION-REGULATING SCARECROW-LIKE (PES) has also been shown to enhance scent via transcriptional activation of scent-related genes (Shor et al. 2023a).The recently identified repressor of scent UNIQUE PLANT PHENYLPRO-PANOID REGULATOR (UPPER) has been proposed to act at the post-transcriptional level (Shor et al. 2023b).In addition, VOC biosynthesis is determined by Phe availability which, in turn, is controlled by feedback of both its level and the levels of synthesized VOCs (Lynch and Dudareva 2020;Liao et al. 2021).VOC emission is also influenced by the genes facilitating their release into the environment (Cna'ani et al. 2015b;Adebesin et al. 2017;Liao et al. 2021Liao et al. , 2023)).
Volatile production in Petunia x hybrida is diurnally governed and is increased at night to attract its nocturnal pollinators (Fenske et al. 2018).Diurnal control of physiological processes usually involves both circadian and light signaling (Dodd et al. 2005;Nozue et al. 2007;Inoue et al. 2017).Scent emission is regulated by the circadian clock components LATE ELONGATED HYPOCOTYL and GIGAN-TEA (Fenske et al. 2015;Brandoli et al. 2020).The effect of light on scent production has not been studied in depth.Emission of the major VOCs by petunia flowers is higher under constant light than constant darkness (Fenske et al. 2015;Shor et al. 2023a).Moreover, light quality affects VOC biosynthesis.Far-red light activates floral scent via the light photoreceptor PHYTOCHROME A (PHYA) (Shor et al. 2023a).Red and blue light spectral ranges also impact the regulation of VOC levels (Colquhoun et al. 2013); however, the molecular mechanisms underlying this regulation remain obscure.
Physiological responses to far-red and red light are regulated by the basic helix-loop-helix (bHLH) transcription factors termed PHYTOCHROME INTERACTING FAC-TORs (PIFs), which physically interact with photoreceptors (Pedmale et al. 2016;Pham et al. 2018).PIFs have been well-studied in Arabidopsis (Leivar and Quail 2011) and characterized in the model plants tomato (Solanum lycopersicum), rice (Oryza sativa) and maize (Zea mays L.) (Nakamura et al. 2007;Kumar et al. 2016;Rosado et al. 2016).No petunia PIFs have been identified.In Arabidopsis, the PIF family has seven members: PIF1 and PIF3-8 (Leivar and Quail 2011).Members of the best-studied quartet-PIF1, PIF3, PIF4 and PIF5-act redundantly in the regulation of numerous physiological processes, among them, hypocotyl growth, leaf senescence and circadian rhythms (Leivar et al. 2008;Shor et al. 2017).PIFs not only transmit light cues, they also act as integrators of environmental and endogenous (phytohormone signaling, plant metabolic status) signals (Paik et al. 2017;Pham et al. 2018;Shor et al. 2018).PIFs intersect with the auxin, ethylene, gibberellin (GA), brassinosteroid, and jasmonic acid pathways to coordinate plant growth and development (De Lucas and Prat 2014;Leivar and Monte 2014).PIF4 (and to a lesser extent, PIF5) has been identified as a convergence point for the transduction of various signals (Quint et al. 2016;Zhao and Bao 2021).Moreover, PIF4 binds to a much higher number of DNA targets than PIF3 or PIF5 (Zhang et al. 2013), supporting the implication of PIF4's control of a broader spectrum of biological processes.PIF functioning is mediated by DEL-LAs-repressors of GA signaling.DELLA-induced protein degradation has been shown for the PIF quartet members (De Lucas et al. 2008;Feng et al. 2008;Li et al. 2016).DELLAs also physically interact with PIF4 and PIF3, preventing their binding to target promoters (Feng et al. 2008).Interestingly, petunia DELLAs (PhDELLAs) are involved in the regulation of floral volatile production-they activate the expression of scent-related genes, whereas GA, known to promote the degradation of DELLA proteins (Sun 2010), suppresses scent (Ravid et al. 2017).
In view of the involvement of PIF family members in the coordination of numerous signals, including those shown to impact petunia scent (GA, light, temperature) (Cna'ani et al. 2015a;Ravid et al. 2017;Shor et al. 2023a), their involvement-and specifically that of PIF4 homologues-in the regulation of volatile production seems plausible.Here we identified petunia PIF family members and revealed that PhPIF4/5 positively regulates floral scent production by activating the expression of VOC-biosynthesis genes.The mechanism of this regulation includes promoter activation of the dawn-expressed positive regulator of scent, EOBII, suggesting PhPIF4/5's involvement in initiating the scentproduction regulatory system at daybreak; and upregulation of PhDELLAs, which suggests participation of PhPIF4/5 in GA-signal modulation of the VOC-biosynthesis machinery.

Plant material and growth conditions
Petunia × hybrida line Mitchell diploid (W115) plants were grown in a glasshouse under 25 °C day/20°C night temperatures with a 16 h light/8 h dark photoperiod.For analysis of PhPIF4/5 expression under far-red light and in the dark, flowers were detached from the plants and placed in a growth chamber (Percival) at 22 °C, under far-red/dark (FR, FR: 730 ± 10 nm, 20 µmol m − 2 s − 2 ) lighting conditions with 16 h light/8 h dark photoperiod, or in constant darkness (D).Light was provided by light-emitting diodes (LED30-HL1).The Arabidopsis thaliana (ecotype Columbia[Col]-0) PIFoverexpressing (PIF-OX) transgenic lines and pifQ (pif-1pif3pif4pif5) mutants were described previously (Shor et al. 2017(Shor et al. , 2018)).Arabidopsis seeds were imbibed and coldtreated at 4 °C for 4 days, then sown on Petri dishes with Murashige and Skoog (MS) medium (Duchefa Biochemie, Netherlands) with 2% (w/v) sucrose.Plants were entrained in 14 h light:10 h dark with 100 µmol m − 2 s − 1 white light (supplied by Philips fluorescent lights TLD 18 W/840) at 23 °C for 8 or 10 days before being transferred for 2 days to continuous light (LL) or DD, respectively.The 10-day-old (for LL conditions) or 12-day-old (for DD conditions) seedlings were sampled for RNA extraction.

Transient suppression of PhPIF4/5 using TRV vectors
Localized transient suppression of PhPIF4/5 was performed using tobacco rattle virus (TRV) as described previously (Shor et al. 2023a).To generate pTRV2-PhPIF4/5, 145 bp of PhPIF4/5 was amplified from cDNA using primers 5 and inserted into pTRV2.As a control, pTRV2 carrying a fragment of CHALCONE SYNTHASE (pTRV2-CHS), shown previously not to affect floral VOC production, was used (Spitzer et al. 2007).pTRV2-PhPIF4/5 or pTRV2-CHS was introduced into Agrobacterium tumefaciens strain AGLO and mixed with agrobacteria carrying pTRV1 in a 1:1 ratio (in inoculation solution containing 200 µM acetosyringone and 10 mM MgCl 2 ) prior to inoculation of flower petals at anthesis.Agroinfiltrated petal regions of 2 day postanthesis (dpa) or 3 dpa flowers were used for the experiments.

Transient overexpression of PhPIF4/5
For overexpression of PhPIF4/5, the CDS was amplified from petunia cDNA using primers 5

'-A T G A A C C C T T G T C T T C C T G A A-3', 5'-C T A A A A A T G T T T A T G G G C T-3'
and cloned into a binary vector under a 35 S promoter.Agrobacterium tumefaciens strain AGLO was transformed with pCGN1559-35Spro:PhPIF4/5 (PhPIF4/5-OX) or pDGB3α2-35Spro:DsRED (DsRED-OX) used as a control, and then injected into petals at anthesis.Agroinfiltrated petal regions of 2 dpa flowers were used for the experiments.

Collection of emitted VOCs
Emitted floral scent compounds were collected for 24 h by localized headspace from the agroinfiltrated petal regions of flowers at 2 dpa (Skaliter et al. 2021).Glass tubes containing 100 mg Porapak Type Q polymer held in place with a plug of silanized glass wool were used as columns.Volatiles were eluted by 1.35 mL hexane + 0.45 mL acetone, and 2 µg isobutylbenzene was added to each sample as an internal standard, followed by GC-MS (Shor et al. 2023a).The experiments were performed in 3-4 biological repeats with similar results.

Gene expression analysis
RNA was extracted from agroinfiltrated petal regions using the Tri-Reagent kit (Sigma-Aldrich) and treated with RNase-free DNase I (Thermo Fisher Scientific) prior to cDNA synthesis using ImProm-II (Promega) reverse transcriptase and oligo(dT) primers.Two-step real-time quantitative PCR (qRT-PCR) was performed on a CFX Opus 384 Real-Time PCR System (Bio-Rad) using 2X qPCRBIO SyGreen Blue Mix Hi-ROX (PCR Biosystems).Raw transcript level data were normalized to EF1α.For Arabidopsis samples, tubulin was used as the reference gene.Quantification calculations were carried out using the 2 −ΔΔCT formula as described (Nozue et al. 2007).The primers are shown in Supplementary Table S1.The experiments were performed in 2-3 biological repeats with similar results.

Promoter activation assay
To test for activation of the EOBII promoter by PhPIF4/5, DsRED CDS was cloned under the 1345-bp promoter region of EOBII (EOBIIpro) or under a mutated EOBII promoter lacking the G-box motif CACGTG (EOBIImpro) into pCGN1559-35Spro:PhPIF4/5 or pCGN1559 without PhPIF4/5 (control).Agrobacterium tumefaciens strain AGLO was transformed with each of the obtained plasmids: pCGN1559-35Spro:PhPIF4/5-EOBIIpro:DsRED, pCGN1559-35Spro:PhPIF4/5-EOBIImpro:DsRED, pCGN1559-EOBIIpro:DsRED.Each of these bacteria was co-infiltrated into petals of petunia flowers at anthesis or into leaves together with agrobacteria carrying pART27-35Spro:YFP.Inoculated tissues were analyzed ca.72 h after agroinfiltration.Transcriptional activation of EOBII promoter was evaluated in inoculated petal regions by DsRED fluorescence level or by DsRED mRNA level.YFP was used as a normalization factor.For imaging of DsRED and YFP fluorescence, a fluorescence binocular microscope (FLOUIII; Leica) was used, and UV light with DsRED and YFP filters was applied.The levels of YFP and DsRED

Petunia PIFs
PIFs of S. lycopersicum (Rosado et al. 2016) which, like petunia, belongs to the Solanaceae family, were used to identify putative petunia PIFs.Using TBLASTN against the CDSs predicted from the Petunia axillaris genome (https:// solgenomics.net),the top hits were found and presence of the bHLH domain in these proteins was verified by PROSITE tool (https://prosite.expasy.org).This search resulted in seven putative petunia PIFs (PhPIFs).Neighbor-joining phylogenetic tree, based on multiple protein sequence alignment by ClustalW, was generated for the PhPIFs and those previously characterized in model plants A. thaliana and S. lycopersicum (Fig. 1).In petunia, similar to PIFs in tomato, the homologues of the Arabidopsis central PIF quartet (homology with AtPIFs was confirmed by reciprocal BLASTp) were: PhPIF1a and PhPIF1b, which were most closely related to AtPIF1; PhPIF3 which was homologous to AtPIF3; and the protein termed PhPIF4/5 that clustered with signal were measured by ImageJ software (Mean Gray Value).mRNA levels of DsRED and YFP were analyzed by qRT-PCR.The experiments were performed three times with similar results.
Fig. 1 Phylogeny of petunia PIFs (PhPIFs).Neighbor-joining tree based on degree of sequence similarity between Petunia Axillaris (Peaxi), Solanum lycopersicum (Slyc) and Arabidopsis thaliana (AT) PIF proteins, built using the p-distance method with partial deletion option applied.Scale length represents number of substitutions per site.Numbers on the branches indicate percentage of replicate trees in which the associated taxa clustered together in the bootstrap test (1000 replicates) profiling of PhPIF4/5 by qRT-PCR.This analysis revealed an increase in transcript level with bud development to a peak in 2.5 cm buds, then a gradual decrease (Fig. 2A).During the daytime, PhPIF4/5 mRNA levels peaked in the afternoon in mature flowers (Fig. 2B), similar to the orthologs in tomato (SlPIF4) and Arabidopsis (PIF4) which are highly expressed in the middle of the day with a peak at around ZT8 (Soy et al. 2014;Rosado et al. 2016).
In Arabidopsis, expression of PIF4 and PIF5 is induced by red and far-red light (Huq and Quail 2002;Oh et al. 2020).To test whether the biological properties of PhPIF4/5 are similar to those of the Arabidopsis homologues, we examined PhPIF4/5 expression in response to far-red lighting.Petunia flowers at anthesis were placed under far-red/dark (FR) conditions or in constant darkness (D), and mRNA levels of PhPIF4/5 were evaluated 2 days later.Expression of PhPIF4/5 was significantly higher under FR vs. D conditions (Fig. 2C), demonstrating a far-red-sensitive expression pattern analogous to that in Arabidopsis.

PhPIF4/5 positively affects petunia floral scent production
To examine the involvement of PhPIF4/5 in the regulation of floral scent production, PhPIF4/5 was locally and transiently suppressed in petunia petals by virus-induced gene AtPIF4 and AtPIF5.Pairwise protein sequence alignment between PhPIF4/5 and AtPIF4 or AtPIF5 revealed similar sequence identity (32%), supporting PhPIF4/5's relation to both of these Arabidopsis PIFs.
Next, we evaluated accumulation of PIF transcripts in petunia petals using previously generated RNA-seq data (Shor et al. 2023b).PhPIF1b and PhPIF7b were not expressed in petals of mature flowers, but their transcripts could be found in the leaf transcriptome (Villarino et al. 2014) available at https://solgenomics.net/, as transcripts comp20642 and comp33265, respectively.All other petunia PIFs-PhPIF1a, PhPIF3, PhPIF4/5, PhPIF7a and PhPIF8-were expressed in the corolla.Characterization of temporal changes in PIF mRNA levels, based on RNA-seq data, revealed that in 1 dpa flowers, PhPIF3 and PhPIF7a transcripts are higher accumulated in the morning, PhPIF8 is higher in the evening.Expression levels of PhPIF1a and PhPIF4/5 did not differ at these time points (Supplementary Fig. S1).
According to the Arabidopsis model, PIF4 acts as a hub, integrating numerous environmental and developmental signals, including those regulating floral scent (Quint et al. 2016;Paik et al. 2017).Based on this we reasoned that PhPIF4/5 may also play a role in regulation of VOC production.Focusing on PhPIF4/5 for the further analysis, we performed detailed developmental and diurnal expression Fig. 2 Transcript levels of PhPIF4/5 in petunia petals.(A) Levels of PhPIF4/5 (at 15.00 h) in developing flowers, length of buds is indicated.(B) PhPIF4/5 expression during daytime in flowers at anthesis.Plants were grown under 16-h light/8-h dark photoperiod (lights on 06.00 h/ lights off 22.00 h).For statistical analysis in (A) and (B), one-way ANOVA with post-hoc Tukey HSD test was applied (P ≤ 0.05).(C) Effect of far-red light on PhPIF4/5 expression.Flowers at anthesis were exposed to far-red/ dark conditions (16-h light/8-h dark photoperiod) or to constant dark for 2 days.Samples were collected from flowers in darkness (D) or under far-red light (FR) at 11.00 h.Significance of differences between treatments was calculated using Student's t-test, *P ≤ 0.05.EF1α was used as an internal reference gene.Data are means ± SEM, n = 3-4 compounds-methyl benzoate, benzaldehyde, phenylethyl alcohol, benzyl alcohol-decreased in response to PhPIF4/5 suppression (Fig. 3A).Consequently, total VOC emission from infected areas was 2 times lower in TRV-PhPIF4/5 petals than in controls (Fig. 3B).
To further confirm the involvement of PhPIF4/5 in scent regulation, PhPIF4/5 was locally and transiently overexpressed in petals by inoculation with agrobacteria silencing (VIGS) using TRV (Shor et al. 2023a).pTRV2 plasmids with a fragment of PhPIF4/5 (TRV-PhPIF4/5) or without it (control) were agroinfiltrated into the petals.A ca. 5-fold reduction in PhPIF4/5 mRNA accumulation in the inoculated areas was confirmed by qRT-PCR (Supplementary Fig. S2A).Two days post-infiltration, the levels of emitted VOCs were evaluated by localized headspace followed by GC-MS.The levels of the most abundant scent in PIF-overexpressing seedlings were evaluated under constant light conditions (LL) at different times of the daybefore subjective dawn (ZT-2), at dawn (ZT0) and in the middle of the day (ZT5) (Fig. 5A).RGA1 was not affected by overexpression of any of the PIF quartet members.In contrast, RGA2 expression was responsive to PIF3 and PIF4: it increased at dawn and before dawn in PIF4-and PIF3-overexpressing seedlings, respectively, as compared to wild-type (WT) plants.Due to the functional redundancy among PIFs, the effect of individual PIFs' suppression on DELLA expression was not evaluated.Instead, the effect of PIFs on RGA2 expression was additionaly confirmed in a quadruple pif mutant (pifQ) at different time points in constant darkness (DD) (Fig. 5B).RGA2 mRNA levels were lower in pifQ than in the WT during the subjective night (ZT-5, ZT-3) and at the beginning of the subjective day (ZT2).These results suggest that PIFs positively affect RGA2 expression independently of light/dark conditions, indicating that the PIFs' activity in this regulation is probably not related to their role in light-signal transduction.Activation of DELLAs by PIF quartet members in different plants and under different lighting conditions (LL and DD) further supports the impact of PIFs in DELLA-regulated scent-related processes.

PhPIF4/5 transcriptionally activates EOBII
Arabidopsis PIFs 4 and 5, homologues of PhPIF4/5, are accumulated at the end of night (Leivar and Monte 2014) and implement their regulatory effect on their target genes' expression at around dawn (Seaton et al. 2018).Among the known regulators of petunia floral VOC production, EOBII and EOBI are expressed at dawn/in the early morning (Spitzer-Rimon et al. 2010, 2012).To evaluate whether morning activation of EOBI and/or EOBII expression is related to PhPIF4/5 activity, we examined the presence of PIF-binding motifs-G-box (CACGTG) and/or PBE-box (CACATG) (Hornitschek et al. 2012;Zhang et al. 2013)in the promoter sequences (2000 bp upstream of ATG) of EOBI and EOBII.EOBI promoter did not contain any of these motifs, whereas EOBII promoter contained G-box at -850 bp.To examine the ability of PhPIF4/5 to activate the EOBII promoter, we agroinfiltrated petals with the coding sequence of DsRED fluorescent protein under the EOBII promoter (EOBIIpro:DsRED) together with 35Spro:PhPIF4/5 (EOBIIpro + PhPIF4/5) or without it (EOBIIpro, control), and with 35Spro:YFP used for normalization.Activity of the EOBII promoter was evaluated by relative DsRED/YFP fluorescence signal, detected by image analysis.Transient localized overexpression of PhPIF4/5 (EOBIIpro + PhPIF4/5) caused a 2-fold increase in the DsRED/YFP ratio, compared to the control (EOBIIpro) carrying PhPIF4/5 under the 35 S promoter (PhPIF4/5-OX) or DsRED under 35 S promoter (DsRED-OX), used as a control.Agroinfiltration with PhPIF4/5-OX led to a three orders of magnitude increase in PhPIF4/5 transcript levels compared to the control (Supplementary Fig. S2B).Localized headspace analysis of the infiltrated regions revealed that emission of all of the main scent compounds was significantly higher in PhPIF4/5 -overexpressing tissues than in controls.Total emission from PhPIF4/5-OX petals was ca.3.5 times higher than from peals, inoculated with DsRED-OX (Fig. 3C, D).Taken together, these results indicated that PhPIF4/5 positively regulates floral VOC emission.

PhPIF4/5 activates expression of scent-related genes
To deeply characterize the positive effect of PhPIF4/5 on scent production, transcript levels of the known regulators of scent and of the genes involved in VOC biosynthesis were analyzed in inoculated petal areas.The samples for analysis were collected in the morning as PIF4 and PIF5 often act as dawn-active inducers of gene expression (Seaton et al. 2018).A qRT-PCR analysis revealed that TRV-based PhPIF4/5 silencing causes a significant reduction in mRNA levels of scent-related genes (Fig. 4A).The genes encoding biosynthesis enzymes of the phenylpropanoid pathway (PAAS, PAL2, BSMT, IGS), shikimate pathway (EPSPS, ADT1), and activators of VOC biosynthesis (EOBI, EOBII, ODO1) showed lower expression in TRV-PhPIF4/5 than in the TRV control.Unlike all other tested VOC-biosynthesis genes, BPBT was not affected in TRV-PhPIF4/5-inoculated petals.Suppression of PhPIF4/5 also did not affect positive regulators of scent PhDELLAs.
In line with PhPIF4/5 silencing, transient overexpression of PhPIF4/5 caused a 2-to 4-fold increase in transcript accumulation of regulators EOBI, EOBII and ODO1.The mRNA levels of all tested biosynthesis genes were also significantly higher in PhPIF4/5-OX petals than in DsRED-OX controls (Fig. 4B).Transcript levels of both PhDELLA1 and PhDELLA2 were significantly increased in response to PhPIF4/5 overexpression, indicating the involvement of PhDELLAs in PhPIF4/5-mediated activation of VOC production.These data suggest that PhPIF4/5 enhances scent production via activation of the expression of positive regulators of scent production and VOC-biosynthesis genes.
To further confirm the previously unreported effect of PIFs on DELLA expression, and to test the generality of the PIFs' influence on DELLAs, we used Arabidopsis (Col-0) transgenic plants overexpressing PIF quartet members PIF1, PIF3, PIF4 or PIF5 (Shor et al. 2018).Transcript levels of REPRESSOR OF GA 1 (RGA1) and RGA2, homologous to PhDELLA1 and PhDELLA2 (Shor et al. 2023a), Overexpression of PhPIF4/5 in petals together with DsRED expressed under the mutated EOBII promoter (EOBIImpro + PhPIF4/5) or under the native EOBII promoter (EOBI-Ipro + PhPIF4/5) revealed that this mutation abolishes the promoter's activation by PhPIF4/5 (Fig. 6B).These data suggest that PhPIF4/5 activates the EOBII promoter and that the presence of the G-box motif affects this activation.
(Fig. 6A), indicating activation of the EOBII promoter by PhPIF4/5.The ratio between DsRED and YFP was evaluated at the transcript level by qRT-PCR as well, considering that this method provides more accurate results, as DsRED accumulation in cells can be reflected by different dynamics in promoter activity and protein degradation.These experiments also revealed a strong increase in DsRED mRNA level as a result of 35Spro:PhPIF4/5 expression (Fig. 6B).Moreover, EOBII promoter was activated by transient overexpression of PhPIF4/5 in petunia leaves as well (Supplementary Fig. S3A, B).To validate the importance of G-box in PIF4/5-mediated activation of EOBII, we generated a mutated EOBII promoter that lacks this motif (EOBIImpro).As a control, binary vector without PhPIF4/5 (EOBIIpro) was used.For normalization, petals were co-infiltrated with a vector carrying 35 S:YFP.(A) Relative DsRED levels, estimated by fluorescence in petal tissues expressing DsRED under the native EOBII promoter.Tissues were imaged by fluorescence microscope and the DsRED/YFP ratio was calculated (n = 6).Significance of differences between treatments was calculated using Student's t-test, *P ≤ 0.05.(B) Effect of PhPIF4/5 on expression levels of DsRED measured by qRT-PCR in petal tissues expressing DsRED under the native or mutated EOBII promoter.Samples were collected at 12.00 h (n = 4).Expression of DsRED was normalized to YFP transcript levels and then to the maximum for all samples in the experiment.For statistical analysis, one-way ANOVA with post-hoc Tukey HSD test was applied (P ≤ 0.05).Data are means ± SEM.Plants were entrained in 14 h light/10 h dark before being transferred to free running conditions.(A) Independent PIF-overexpressing (PIF-OX) transgenic and wild-type (WT) control lines were sampled in constant light (LL).(B) pifQ (pif1pif3pif4pif5) mutant and WT control plants sampled in constant darkness (DD).Samples were collected at the indicated times for qRT-PCR.n = 3. Tubulin (tub) was used as an internal reference gene.Expression of the targets was normalized to that in the WT at each time point (ZT).Data are means ± SEM.Significance of differences between treatments was calculated using Student's t-test, *P ≤ 0.05 PIF5 homologues in the regulation of phenylpropanoid biosynthesis in flowers has not been investigated.
The petunia PIF family is represented by seven members (PhPIFs) which are similar to those in tomato (Rosado et al. 2016) (Fig. 1).Five of them, including PhPIF4/5, are expressed in the corolla.The name PhPIF4/5 was given based on its similarity to the two Arabidopsis proteins, PIF4 and PIF5.The presence of one protein that is similar to both of these Arabidopsis PIFs has also been observed in other eudicots, and duplication in the PIF4 clade has been found specifically in several eudicots of the family Brassicaceae (Rosado et al. 2016).
Here we revealed the involvement of PhPIF4/5 in specialized metabolism during advanced stages of flower development, i.e., floral scent production.Both VOC-biosynthesis genes and positive regulators of the pathway were significantly activated by PhPIF4/5, leading to increased levels of the emitted volatiles.Moreover, PhPIF4/5 was shown to act upstream of the central activator of scent, EOBII.Whereas overexpression of PhPIF4/5 increased transcript levels of all of the tested scent-related genes, suppression of PhPIF4/5 did not affect BPBT or PhDELLAs (Fig. 4).This might be due to the redundancy among PIF family members, similar to that described for other PIF-regulated physiological processes (Leivar et al. 2012;Shor et al. 2017).The fact that PIF4, PIF5 and PIF3 often regulate the same target genes, as revealed by CHIP-seq (Oh et al. 2012;Zhang et al. 2013), highlights the overlapping functions of these PIFs and their tendency toward redundance.Interestingly, the pattern of BPBT expression also differed from that of other VOC-biosynthesis genes in experiments evaluating scent production in response to far-red light and to suppression of PHYA (Shor et al. 2023a).In those experiments, the genes encoding the main regulators of VOC biosynthesis, including EOBII, were upregulated in response to the tested conditions.However, BPBT, known to be activated by EOBII under normal conditions (Spitzer-Rimon et al. 2010), was not affected.This suggests the presence of an additional mechanism governing the control of BPBT expression, which is opposite to that mediated by EOBII.
PIFs 4 and 5 are dawn-active transcriptional factors (Seaton et al. 2018) as well as the major positive regulator of scent, EOBII.Activation of EOBII promoter, which contains a PIF-binding G-box motif, by PhPIF4/5 and the importance of G-box for this activation (Fig. 6) suggest that PhPIF4/5 is required for peak EOBII transcript accumulation at dawn.This subsequently initiates the cascade of the events-including activation of ODO1 and expression of VOC-biosynthesis genes-resulting in the emission of volatiles in the evening.Similar PIF-mediated induction of a morning expression peak has been shown for the core circadian clock component CIRCADIAN CLOCK

Discussion
The scent-production machinery in petunia is sensitive to light, temperature, the circadian clock and the phytohormones GA and ethylene (Underwood et al. 2005;Cna'ani et al. 2015a;Fenske et al. 2015;Ravid et al. 2017;Shor et al. 2023a).PIFs are known to be involved in the transduction and integration of these environmental and developmental signals (Leivar and Monte 2014;Paik et al. 2017).Interestingly, anthocyanin pigmentation, which originates from the same phenylpropanoid pathway as petunia floral VOCs and is shown to be coregulated with them (Zvi et al. 2008;Cna'ani et al. 2015b), is sensitive to PIFs (Liu et al. 2015;Seaton et al. 2018).Numerous processes are redundantly regulated by PIFs; e.g., overlapping functions of PIF quartet members have been shown for photomorphogenesis, the transition to scotomorphogenesis and circadian clock entrainment (Shin et al. 2009;Shor et al. 2017).PIF1 and PIF3 negatively regulate chlorophyll biosynthesis and chloroplast development (Shin et al. 2009;Stephenson et al. 2009).PIF3 and PIF4 integrate GA and light signals to modulate hypocotyl growth (De Lucas et al. 2008;Feng et al. 2008).The latter has been specifically shown to incorporate additional signaling pathways (Choi and Oh 2016).For example, in the control of hypocotyl elongation, PIF4 integrates sensitivity to light, temperature, brassinosteroid signals and, together with the circadian clock, provides time-gating of auxin biosynthesis for growth initiation (Koini et al. 2009;Oh et al. 2012;Zhu et al. 2016;Pereyra et al. 2022).Functions of PIF4 and PIF5 largely overlap, and they often work collaboratively in the same pathways, among them: transmission of output signals from the circadian oscillator, regulation of ethylene signaling and biosynthesis during leaf senescence, shade avoidance and anthocyanin biosynthesis (Nozue et al. 2007;Lorrain et al. 2008;Sakuraba et al. 2014;Liu et al. 2015;Liebsch and Keech 2016).Yet, the mechanisms of PIF4 and PIF5 regulation of the same process can differ.For example, in inhibiting light-induced anthocyanin production, PIF5 suppresses late anthocyanin-biosynthesis genes by direct interaction with their promoters, whereas PIF4 directly represses expression of the positive regulator of anthocyanin biosynthesis PRODUCTION OF ANTHO-CYANIN PIGMENT 1 (Liu et al. 2015(Liu et al. , 2021)).Moreover, PIF4 and PIF5 regulation of the phenylpropanoid pathway, i.e., anthocyanin production, in Arabidopsis seedlings is specific to growth conditions.Whereas under standard conditions, pif4pif5 mutants display a photoperiod-dependent low-anthocyanin phenotype, indicating their positive effect on anthocyanin biosynthesis (Seaton et al. 2018), lightinduced anthocyanin accumulation is inhibited by PIF4 and PIF5 (Liu et al. 2015(Liu et al. , 2021)).Although it has been extensively studied in vegetative organs, the role of PIF4 and al. 2008;Feng et al. 2008;Li et al. 2016), while PIF4 activates DELLA expression.In the petunia corolla, the levels of GA-which acts as a repressor of floral scent production (Ravid et al. 2017)-decrease during flower development in parallel with increasing sensitivity to GA and initiation of scent emission (Patrick et al. 2021;Shor et al. 2023b).Considering the GA-DELLA-PIF4/5 interactions in petunia, demonstrated previously (Ravid et al. 2017) and in the current work, PhDELLA-PhPIF4/5 fine-tuning may be a part of the GA-sensitivity-adjustment mechanism in petals.
Taken together, we propose that the PhPIF4/5-mediated positive effect on VOC production includes transcriptional activation of both EOBII and PhDELLAs (Fig. 7).Even though PhDELLAs have been shown to increase EOBII expression (Ravid et al. 2017), PhPIF4/5's effect on EOBII is not only exerted through PhDELLAs, because a reduction in scent was observed in flowers with suppressed PhPIF4/5, where PhDELLAs were not affected.Interestingly, VOCbiosynthesis genes can be activated by PhPIF4/5 in EOBIIand PhDELLAs-independent manner.For example, PAAS is not affected by EOBII (Spitzer-Rimon et al. 2010) and positively regulated by PhDELLAs (Ravid et al. 2017), however, PAAS level decreased in response to PhPIF4/5 suppression, which did not cause reduction in PhDEL-LAs' levels.Thus, PhPIF4/5 may link partially independent routes leading to scent production, allowing for precise regulation of scent-related genes.Considering the role of PhPIF4/5 in the regulation of scent, and the signaling-hub function of PIFs in the adjustment of numerous physiological processes (Leivar and Quail 2011;Choi and Oh 2016), an evaluation of PhPIF4/5's participation in the integration of external and internal signals for the production of floral volatiles is warranted.
ASSOCIATED 1 (CCA1) (Shor et al. 2017).Genes with a diurnal expression pattern similar to that of, e.g., EOBII and CCA1, are often controlled by both light-sensing regulators and a circadian oscillator (Seaton et al. 2018).Therefore, it can be suggested that EOBII is also under circadian control.Among the core components of the circadian oscillator, only the PSEUDO-RESPONSE REGULATOR (PRR) transcription factors bind specifically to G-box on the target promoters (Liu et al., 2016).Hence, it would be reasonable to evaluate the involvement of PRRs in the regulation of EOBII expression and floral scent production.
PhPIF4/5, similar to PhDELLAs-the transmitters of GA signaling-positively regulate scent and expression of VOC-biosynthesis and regulatory genes (Ravid et al. 2017).However, it is well established that at the protein level, DELLAs negatively affect PIF activity (De Lucas et al. 2008;Feng et al. 2008;Li et al. 2016).Here we provide evidence for the positive transcriptional/post-transcriptional effect of PIFs on DELLAs: of petunia PhPIF4/5 on the expression of PhDELLA1 and PhDELLA2 and of Arabidopsis PIF4 and PIF3 on the expression of RGA2 (Figs. 4 and  5).Till now, transcriptional regulation of DELLAs by PIFs had only been demonstrated for Arabidopsis PIF1 (PIL5) which positively affects RGA1 and RGA2 expression in germinating seeds (Oh et al. 2007).The findings presented here expand our understanding of the mechanism of PIF4-DELLA fine-tuning and of PIF-mediated modulation of GA signaling (Supplementary Fig. S4), a model that includes the following feedback loops: PIF4, together with PIF5, is involved in the control of GA biosynthesis (Filo et al. 2015), and the latter negatively controls the abundance of DELLA proteins, leading them to degradation (Sun 2010); DELLAs induce protein degradation of PIFs (De Lucas et

Fig. 5
Fig. 5 PIF4 activates expression of DELLA in Arabidopsis seedlings.Plants were entrained in 14 h light/10 h dark before being transferred to free running conditions.(A) Independent PIF-overexpressing (PIF-OX) transgenic and wild-type (WT) control lines were sampled in constant light (LL).(B) pifQ (pif1pif3pif4pif5) mutant and WT control plants sampled in constant darkness (DD).Samples were collected at

Fig. 7
Fig. 7 A model showing PhPIF4/5 input in petunia floral scent emission.Petunia PHYTOCHROME INTER-ACTING FACTOR 4/5 (PhPIF4/5) positively regulates VOC production and expression of scent-related genes.This includes transcriptional activation of the master regulator of scent EMISSION OF BENZENOIDS II (EOBII) and upregulation of PhDELLA transcripts