Characterization of Maize Phytochrome-Interacting Factors in Light Signaling and Photomorphogenesis

Phytochrome-interacting factors regulate light signaling and photomorphogenesis in maize.

Abstract

Increasing planting density has been an effective means of increasing maize (Zea mays ssp. mays) yield per unit of land area over the past few decades. However, high-density planting will cause a reduction in the ratio of red to far-red incident light, which could trigger the shade avoidance syndrome and reduce yield. The molecular mechanisms regulating the shade avoidance syndrome are well established in Arabidopsis (Arabidopsis thaliana) but poorly understood in maize. Here, we conducted an initial functional characterization of the maize Phytochrome-Interacting Factor (PIF) gene family in regulating light signaling and photomorphogenesis. The maize genome contains seven distinct PIF genes, which could be grouped into three subfamilies: ZmPIF3s, ZmPIF4s, and ZmPIF5s. Similar to the Arabidopsis PIFs, all ZmPIF proteins are exclusively localized to the nucleus and most of them can form nuclear bodies upon light irradiation. We show that all of the ZmPIF proteins could interact with ZmphyB. Heterologous expression of each ZmPIF member could partially or fully rescue the phenotype of the Arabidopsis pifq mutant, and some of these proteins conferred enhanced shade avoidance syndrome in Arabidopsis. Interestingly, all ZmPIF proteins expressed in Arabidopsis are much more stable than their Arabidopsis counterparts upon exposure to red light. Moreover, the Zmpif3, Zmpif4, and Zmpif5 knockout mutants generated via CRISPR/Cas9 technology all showed severely suppressed mesocotyl elongation in dark-grown seedlings and were less responsive to simulated shade treatment. Taken together, our results reveal both conserved and distinct molecular properties of ZmPIFs in regulating light signaling and photomorphogenesis in maize.

Maize (Zea mays ssp. mays) has become the highest producing crop globally (FAO statistics, http://www.fao.org/faostat). Average maize yields in the United States increased over sevenfold from the 1930s to 2010 (Mansfield and Mumm, 2014). This increase can be largely attributed to the synergistic improvement of genetic gain and management practices (Tollenaar and Lee, 2002). In particular, increasing planting density has been a key factor for the increased maize yields in the central Corn Belt in the United States (from 30,000 plants ha⁻¹ in the 1930s to approximately 70,000 plants ha⁻¹ in 2010; Mansfield and Mumm, 2014). A series of studies by Duvick and others have shown that newer maize hybrids are better adapted to high-density planting than older ones, mainly due to a series of morphological changes associated with grain production efficiency, including reduced leaf angle and tassel branch number, and increased tolerance to various biotic and abiotic stresses associated with high-density planting, such as drought, shading, pathogens, and insects (Duvick, 2005a, 2005b; Lauer et al., 2012; Mansfield and Mumm, 2014). Optimal plant architecture, such as reduced plant and ear height and increased culm strength, could contribute to lodging resistance; elevated leaf angle and smaller tassel sizes allow plants to intercept sunlight more efficiently at high planting densities; the reduced anthesis-silking interval could help to synchronize male and female flowering, thus improving kernel setting (Duvick, 2005a, 2005b; Lauer et al., 2012; Mansfield and Mumm, 2014).

Numerous studies have established that phytochromes (phys), the major photoreceptors of red and far-red light signals in plants, play critical roles in regulating plant growth and development throughout their life cycle, from seed germination, seedling deetiolation, and vegetative growth to flowering and seed setting (Franklin and Quail, 2010). Especially, a role of phys in conferring the developmental plasticity in response to the changing light environments has been well documented in a number of plant species (Gururani et al., 2015). For example, in response to anticipation of shading (reduced red to far-red light ratios [R:FR]) or actual shading (reduced R:FR plus reduced photoactive irradiation) at high planting densities, plants trigger a series of adaptive responses, including the promotion of hypocotyl (or stem) and petiole elongation, reduced branching, reorientation of the growth direction of leaf or branch, and early flowering, collectively known as the shade avoidance syndrome (SAS; Franklin, 2008). Despite the fact that these responses are believed to increase the chance of individual success, they are detrimental to yield production of crops at high planting densities. Thus, understanding the molecular mechanism governing SAS in plants, particularly in crops, will be essential to guide the breeding of shade-tolerant cultivars suitable for high-density planting.

A consensus pathway of phy regulating SAS has been recently established in the model dicot plant species Arabidopsis (Arabidopsis thaliana). Arabidopsis has five phys, phyA to phyE. phyA is the major photoreceptor for perceiving far-red light, whereas phyB to phyE are the major photoreceptors for sensing red light, with phyB playing a dominant role (Franklin and Quail, 2010). Phys are synthesized in cytoplasm in the inactive red light-absorbing (Pr) form, which are changed to the active far-red light-absorbing (Pfr) form by red light irradiation. The activated phys are translocated to the nucleus to form nuclear bodies (Van Buskirk et al., 2012; Klose et al., 2015). In the nucleus, phys interact directly with a set of basic helix-loop-helix (bHLH) transcription factors called PHYTOCHROME INTERACTING FACTORs (PIFs). In Arabidopsis, the AtphyB single mutant displays various traits reminiscent of a constitutive shade avoidance response, such as elongated hypocotyls and petioles, acceleration of flowering, and higher apical dominance under high R:FR, indicating that it is the major photoreceptor for canopy shade. Recent studies showed that under high R:FR conditions, active phyB can promote the degradation of PIF proteins (PIF3, PIF4, PIF5, and probably PIF1) or phosphorylation of PIF7, whereas low R:FR signals reduce the levels of active phyB, thus promoting stabilization of PIF proteins (PIF3, PIF4, and PIF5) or dephosphorylation of PIF7, which in turn stimulates changes in the downstream transcriptional network (such as up-regulation of auxin biosynthetic genes) and induces growth responses to shade (Lorrain et al., 2008; Leivar et al., 2012; Mansfield and Mumm, 2014; Mizuno et al., 2015; Xie et al., 2017).

There is a paucity of evidence to support a role of phys in regulating SAS in maize. The maize genome contains six phy genes: PHYA1, PHYA2, PHYB1, PHYB2, PHYC1, and PHYC2 (Sheehan et al., 2004). Similar to the phyB mutant in Arabidopsis, the maize phyB1 phyB2 double mutant also displays constitutive shade avoidance responses, such as increased plant height, elongated internodes, and tendency to lodge (Kebrom et al., 2006, 2010; Sheehan et al., 2007). Similarly, the maize elongated mesocotyl1 mutant, which carries a lesion in the ZmHY2 gene encoding phytochromobilin synthase, also shows pronounced elongation of the mesocotyl and fails to deetiolate under red or far-red light conditions (Sawers et al., 2002). Recent studies also showed that overexpression of ZmPhyA1 causes increased plant and ear height in maize (Yu et al., 2018) and that phy-mediated signaling is involved in the suppression of axillary bud outgrowth in response to canopy shade in maize, as in Arabidopsis and rice (Oryza sativa; Kebrom et al., 2006, 2010; Whipple et al., 2011).

Given the central roles of PIF proteins in integrating multiple signaling pathways (light, temperature, hormones, biotic and abiotic stresses, etc.), thus optimizing plant growth and development tailored to its environments (Paik et al., 2017), there is strong interest to exploit them for biotechnological applications to produce improved crops that can be better adapted to adverse environmental conditions. A number of previous studies have attempted to identify the maize PIF genes and conducted functional analyses of some of the members (Kumar et al., 2016; Shi et al., 2018; Gao et al., 2019). For example, Gao et al. (2019) identified six maize PIF genes (named ZmPIF1–ZmPIF6) and showed that ZmPIF1 and ZmPIF3 might regulate response to salt or drought stresses in rice, while Shi et al. (2018) showed that overexpression of ZmPIF4 confers a constitutive shade avoidance response in Arabidopsis. Despite the progress made, systematic characterization of the ZmPIF gene family in regulating phy-mediated light signaling and SAS remains forthcoming.

Here, we cloned seven potential ZmPIF genes and systematically analyzed their molecular properties (expression patterns, protein localization and stability, and interaction with phys). We show that all seven ZmPIFs possess conserved function with their Arabidopsis counterparts in rescuing the phenotypes of the Arabidopsis pifq mutant. In addition, we performed an initial characterization of these ZmPIF members in light signaling and photomorphogenesis in maize. Our results establish a foundation for future dissection of ZmPIFs in regulating plant architecture and SAS in maize and might provide potential targets for genetic improvement of maize for adapting to high-density planting.

RESULTS

Cloning and Phylogenetic Analysis of the ZmPIF Genes

In order to identify all genes potentially encoding PIF proteins in maize, we first performed BLAST analysis using the bHLH domain of Arabidopsis PIF3 (AtPIF3) in the data set (Gramene gene model set 5b+ for Refgen_v3-translations) in MaizeGDB (https://www.maizegdb.org/). A total of 200 hits with an E-value cutoff of 1e-4 were kept, and self-BLAST was performed on the resulting sequence list to eliminate all the redundant sequences. In total, 123 proteins were kept, and the corresponding sequences were downloaded. Each sequence was manually checked to determine the presence of the complete bHLH domain and the conserved active phyB-binding (APB) domain. Finally, seven potential PIF homologs were identified in the maize B73 genome (Supplemental Table S1), and the DNA fragments were amplified and cloned into the pGADT7 vector (Clontech) for sequencing. We successfully cloned the full-length cDNAs for all seven putative ZmPIF genes via reverse transcription (RT) PCR using a lab-owned B73 inbred line, indicating that these genes are all actively transcribed. Sequence analysis showed that the cloned cDNA sequences of five ZmPIFs were 100% identical to the predicted cDNA sequences deposited in the AGP B73v3 and/or AGP B73v4 version. However, the sequences of the cloned cDNAs of GRMZM2G165042 and GRMZM2G016756 differed from the predicted cDNA sequences deposited in the AGP B73v3 and/or AGP B73v4 database. To examine whether the differences might be caused by alternative splicing, as previously reported for the Arabidopsis PIF6 gene (Penfield et al., 2010), we randomly picked 20 cDNA clones for sequencing analysis. Indeed, we found that both GRMZM2G165042 and GRMZM2G016756 had two alternative splicing variants (named isoform α and isoform β, respectively), but no clone matching the predicted cDNA sequences was found (Supplemental Fig. S1). This result suggests that the predicted cDNA for GRMZM2G165042 and GRMZM2G016756 might be inaccurate.

Phylogenetic analysis revealed that these seven ZmPIFs could be grouped into three clades and were renamed according to their respective clades (Fig. 1A, left). GRMZM2G062541 was named ZmPIF3.3, as it was grouped into the same clade with the previous reported proteins ZmPIF3.1 (GRMZM2G115960) and ZmPIF3.2 (GRMZM2G387528), and these three proteins share homology with AtPIF3 (Fig. 1A). The previously reported ZmPIF4 (GRMZM5G865967) and ZmPIF5 (GRMZM2G165042; Shi et al., 2018) were renamed ZmPIF4.1 and ZmPIF4.2, respectively, as they were clustered together in the phylogenetic tree (Fig. 1A), and GRMZM5G865967 has higher sequence similarity to AtPIF4 (88% identity) than to AtPIF5 (41% identity). Similarly, GRMZM2G065374 and GRMZM2G016756 were named ZmPIF5.1 and ZmPIF5.2, respectively, as they were grouped into the same clade and GRMZM2G016756 has higher sequence similarity to AtPIF5 (61% identity) than to AtPIF4 (47% identity; Fig. 1A). In addition, analysis of ZmPIF protein sequence motifs using MEME (Bailey et al., 2009) showed that all seven ZmPIF proteins have a highly conserved bHLH domain and APB motif, while ZmPIF3s (ZmPIF3.1, ZmPIF3.2, and ZmPIF3.3) also have an active phyA-binding motif at their N termini, similar to AtPIF1 and AtPIF3 (Fig. 1A, right).

Figure 1.

Characterization of the ZmPIF gene family. A, Phylogenetic analysis and motif comparisons of the ZmPIF and AtPIF proteins. The phylogenetic tree was constructed based on their full-length amino acid sequences using the maximum likelihood method (left). The presence of active phyA-binding (APA), APB, and bHLH motifs is depicted as boxes and shown on the right. Bar = 100 amino acid residues. B, Expression analysis of ZmPIFs in various tissues of maize seedlings. Different tissues of three-leaf stage seedlings of the maize inbred line B73 were harvested and then used to perform RT-quantitative PCR (qPCR) analysis. The mRNA level of maize Tubulin5 was used as a reference. Data are means and sd of three independent biological replicates. Asterisks indicate significant differences compared with leaf tissue using Student’s t test (**, P < 0.001). C, Coleoptile; L, leaf; M, mesocotyl; R, root.

Expression Profiles of ZmPIF Genes and Protein Subcellular Localization

In order to better understand the function of each ZmPIF gene, their differential tissue expression profiles were investigated in 10-d-old maize seedlings. Using transcriptomic data from maize B73 (Sekhon et al., 2011), an expression heat map was constructed for all seven ZmPIFs in different tissues from various developmental stages (Supplemental Fig. S2). Our RT-qPCR assay showed that the expression patterns of these seven ZmPIF genes were highly similar, all with a relatively high expression level in leaf and lower expression levels in root, mesocotyl, and coleoptile (Fig. 1B), which was in line with the transcriptomic data of maize B73 (Supplemental Fig. S2).

To examine whether the transcript levels of ZmPIFs are regulated by light, 7-d-old dark-grown maize B73 seedlings were transferred to white light (WL) for various times or retained in darkness as controls. Notably, the expression levels of all ZmPIF genes fluctuated during the time course examined, but in general, ZmPIF3.1, ZmPIF3.2, and ZmPIF3.3 showed a similar profile, being rapidly down-regulated by light treatment (Supplemental Fig. S3, top). ZmPIF4s and ZmPIF5s shared a similar profile, being up-regulated by light following prolonged light exposure (2–3 h; Supplemental Fig. S3, middle and bottom). These results suggest that the homologous genes in the same clade may more likely perform similar functions.

In order to detect the localization of these ZmPIF proteins, we expressed GFP fusion of each ZmPIF protein in Nicotiana benthamiana leaf cells. Resembling the typical localization of AtPIFs, all seven ZmPIFs localized to the nucleus, and four of them (ZmPIF3.1, ZmPIF3.2, ZmPIF3.3, and ZmPIF4.2) formed obvious nuclear bodies, whereas ZmPIF4.1, ZmPIF5.1, and ZmPIF5.2 showed uniform nuclear localization without distinct nuclear bodies in N. benthamiana leaf cells (Fig. 2).

Figure 2.

ZmPIF proteins localize to the nucleus. The expression constructs Pro-35S:ZmPIFs-GFP were individually cotransformed with a nuclear protein marker construct (Pro35S:mRFP-AHL22; Xiao et al., 2009) into N. benthamiana leaves. The N. benthamiana leaves were incubated for 48 h in the dark after transformation and then transferred to WL for 3 to 6 h prior to imaging using a confocal microscope. Bars = 20 μm.

ZmPIF Proteins Physically Interact with ZmphyB

Since all seven ZmPIF proteins possess a putative APB motif, which is necessary for interacting with phyB, we performed a yeast two-hybrid (Y2H) assay to test this possibility. Surprisingly, we found that ZmPIFs interacted with ZmphyB1 under all conditions examined (darkness or light treatment, presence or absence of phycocyanobilin [PCB]; Fig. 3; Supplemental Fig. S4). A similar interaction was observed between AtPIF3 and AtphyB (Fig. 3; Supplemental Fig. S4). However, no visible interactions were observed between ZmPIF proteins and full-length ZmphyB2 in yeast under the same conditions (Fig. 3). Western-blot analysis showed that the ZmphyB2 protein is normally expressed in yeast (Supplemental Fig. S5), suggesting that ZmphyB1 and ZmphyB2 may have distinct affinities for interacting with ZmPIFs.

Figure 3.

ZmPIF-ZmphyB interaction analyzed by Y2H assay. The Y2H assay shows that ZmPIFs interact with ZmphyB1, but not with ZmphyB2, in yeast. Full-length ZmphyB1 and ZmphyB2 were fused with the DNA-binding domain (BD) as the baits. Each full-length ZmPIF member was fused with the activation domain (AD) as the prey. The interaction between AtPIF3 and AtphyB was used as a positive control. Empty vectors were used as negative controls. Yeast cells (AH109) coexpressing the indicated combinations of constructs were grown on nonselective (synthetic dextrose [SD]-T-W) or selective (SD-T-W-H) medium with 1 mm 3-aminotriazole (3AT) in the presence (+PCB) or absence (−PCB) of 25 μm phycocyanobilin (PCB) under continuous red light (Rc [R]; 4 μmol photons m⁻² s⁻¹) or far-red light (FR; 3 μmol photons m⁻² s⁻¹) or in darkness (D) for 3 d.

We next performed luciferase complementation imaging (LCI) and bimolecular fluorescence complementation (BiFC) assays to test the interaction between ZmPIFs and ZmphyBs. The LCI assay showed that all combinations of ZmPIFs with either ZmphyB1 or ZmphyB2 could reconstitute the functional LUC activity in plant cells (Fig. 4, A and B), but the negative controls could not reconstitute the LUC activity in plant cells (Supplemental Fig. S6). The interactions between ZmPIFs with ZmphyB1 and ZmphyB2 were also confirmed using BiFC assay in the nucleus of N. benthamiana leaf cells (Fig. 4, C and D; Supplemental Fig. S7). Additionally, Y2H and LCI assays showed that all maize PIF proteins could also interact with AtphyB (Supplemental Fig. S8), indicating that the interaction between phyB and PIF is conserved between maize and Arabidopsis.

Figure 4.

ZmPIFs interact with ZmphyB1 and ZmphyB2 in plant cells. A and B, LCI assay showing the ZmPIFs-ZmphyB interaction in N. benthamiana leaf cells. Each ZmPIF member was fused to cLUC. The full-length ZmphyB1 or ZmphyB2 was fused to nLUC. The N. benthamiana leaves were infiltrated with the indicated combinations, incubated for 48 h in the dark, and then transferred to WL for 3 to 6 h prior to photographing using an in vivo imaging system. Empty vectors were used as negative controls (Supplemental Fig. S6). C and D, BiFC assay showing the ZmPIFs-ZmphyB interaction in N. benthamiana leaf cells. Full-length ZmPIF and ZmphyB proteins were fused to the split N-terminal (nYFP [yellow fluorescent protein]) or C-terminal (cYFP) fragments of YFP, respectively. The N. benthamiana leaves infiltrated with the combinations were adapted in darkness for 48 h and then exposed to light for 3 to 6 h before examination with a confocal microscope. Red fluorescence protein (RFP) served as the internal control. The empty vectors were used as negative controls (Supplemental Fig. S7). Bars = 20 μm.

ZmPIFs Complement the Arabidopsis pifq Mutant Phenotype

To further examine whether ZmPIFs could functionally complement the Arabidopsis pifq mutant phenotype, we overexpressed each ZmPIF member (ZmPIF-OE) in the pifq mutant background. The transcript levels of ZmPIFs in the transgenic lines were determined using RT-qPCR to verify successful transformation (Supplemental Fig. S9). Since the Arabidopsis pifq mutant showed a constitutive photomorphogenic-like phenotype with shortened hypocotyls and open apical hooks in darkness (Leivar et al., 2008b), the ZmPIF-OE transgenic lines were checked for hypocotyl length in dark-grown seedlings. The phenotypic analysis showed that most of the dark-grown ZmPIF-OE seedlings exhibited significantly elongated hypocotyls and partially closed apical hooks compared with pifq, and this effect was most robust in the ZmPIF4.1-OE lines (Fig. 5, A–F). Interestingly, dark-grown ZmPIF5-OE transgenic seedlings exhibited an exaggerated apical hook phenotype, reminiscent of the Arabidopsis PIF5-overexpressing lines (Fig. 5C). We therefore investigated whether ZmPIF5 is involved in the ethylene signaling pathway regulating hook opening, as is the case for AtPIF5 (Khanna et al., 2007). As expected, treatment with ACC aggravated the triple response phenotype of the ZmPIF5-OE seedlings (Supplemental Fig. S10, middle). By contrast, treatment with AgNO₃ abolished the triple response phenotype observed for the ZmPIF5-OE seedlings (Supplemental Fig. S10, bottom). These observations provide additional support for the functional conservation between AtPIF5 and ZmPIF5.

Figure 5.

ZmPIFs rescue the phenotype of Arabidopsis pifq mutant seedlings. A to C, The phenotype of dark-grown Arabidopsis pifq mutant seedlings is partially complemented by ZmPIF overexpression. Columbia-0 (Col-0) and pifq mutant as well as pifq mutant seedlings expressing Pro-35S:ZmPIFs-GFP were grown in darkness for 4 d. Bars = 1 mm. D to F, Quantification of hypocotyl length of the seedlings shown in A to C. Data represent means from at least 20 plants, and the error bars are sd. Asterisks indicate significant differences compared with pifq mutant plants using Student’s t test (**, P < 0.001). G and H, ZmPIF overexpression restores the expression of PIF-dependent genes in the Arabidopsis pifq mutants. The expression of marker genes including XTH15 and IAA19 was analyzed by RT-qPCR in 4-d dark-grown seedlings of Col-0, pifq, and pifq expressing Pro-35S:ZmPIFs-GFP. ACT2 was used as the internal control. Data represent means and sd of three independent biological replicates. Asterisks indicate significant differences compared with pifq plants using Student’s t test (**, P < 0.001).

To further confirm the regulation of ZmPIFs on hypocotyl elongation-related genes, we measured the transcript levels of XTH15 and IAA19, two representative genes involved in dark-induced hypocotyl elongation (Zhang et al., 2013a). Consistent with the morphological phenotype, the expression levels of both XTH15 and IAA19 were dramatically elevated in the ZmPIFs overexpression lines (Fig. 5, G and H). Taken together, these data clearly demonstrate that these ZmPIFs possess a conserved function with AtPIFs in repressing photomorphogenesis.

ZmPIF Proteins Are Slowly Degraded in Arabidopsis Seedlings in Response to Red Light

In Arabidopsis, light-activated phys accumulate in the nucleus, where they directly interact with PIFs, leading to light-induced phosphorylation and degradation of PIF proteins (PIF3, PIF4, PIF5, and probably PIF1 as well; Ni et al., 2013). Therefore, we examined whether these ZmPIF proteins have similar degradation kinetics to the AtPIFs in etiolated Arabidopsis seedlings irradiated with Rc. As shown in Figure 6A, all ZmPIFs:GFP proteins accumulated in the nucleus and did not form nuclear bodies before Rc treatment. Interestingly, fluorescent nuclear bodies were detected for ZmPIF3.1:GFP, ZmPIF3.2:GFP, ZmPIF3.3:GFP, and ZmPIF4.2:GFP fusion proteins after a short high Rc exposure (R 1min) and remained visible even after a low Rc exposure for 1 h (R 1h), a behavior reminiscent of that of AtPIF7 (Leivar et al., 2008a; Fig. 6A). However, the ZmPIF4.1:GFP, ZmPIF5.1:GFP, and ZmPIF5.2:GFP fusion proteins did not form nuclear bodies upon Rc treatment and appeared much more stable and similar to their accumulation in the nucleus in the dark (Fig. 6A). Consistent with this, western-blot analysis showed that degradation of ZmPIFs was much slower (up to 6 or 24 h in the light; Fig. 6B) than degradation of the Arabidopsis PIF proteins. These results suggest that these ZmPIF proteins exhibit distinct properties in light-induced nuclear body formation and are more stable than the Arabidopsis PIF proteins in response to red light.

Figure 6.

Red light induces slow degradation of ZmPIF proteins in Arabidopsis. A, Epifluorescence imaging of GFP fluorescence in hypocotyl cell nuclei of Arabidopsis transgenic seedlings expressing ZmPIF-GFP as indicated. Seedlings were grown in the dark for 4 d and then maintained in darkness (D), exposed to high Rc of 198 μmol photons m⁻² s⁻¹ for 1 min (R 1min), or exposed to low Rc of 4 μmol photons m⁻² s⁻¹ for 1 h (R 1h). Samples were fixed in 4% (v/v) paraformaldehyde and examined using a fluorescence microscope. Bar = 10 μm. B, Immunoblot analysis of ZmPIF protein stability in response to red light. Seedlings of transgenic lines expressing ZmPIFs-GFP fusion proteins were grown for 4 d in the dark (D) or exposed to Rc (4 μmol photons m⁻² s⁻¹) for 1, 3, 6, or 24 h. ZmPIF fused to a GFP tag was detected in total protein extracts by immunoblot using anti-GFP antibodies. Protein extracts from pifq were included as a control. The detection of Actin using anti-Actin antibodies is shown as a loading control.

Knockout of ZmPIFs by CRISPR/Cas9 Causes Short Mesocotyls in Maize under Darkness

To investigate the in vivo function of these ZmPIF genes in regulating maize growth and development, we knocked out the endogenous ZmPIF genes in the maize wild-type inbred line ZC01 using Clustered Regularly Interspaced Short Palindromic Repeats/Cas9 (CRISPR/Cas9) technology. To overcome functional redundancy, homologous ZmPIF genes of the same subgroup were constructed in the same CRISPR/Cas9 vector system for multiple target site cleavage (Supplemental Fig. S11). We successfully obtained Zmpif3.1/Zmpif3.2/Zmpif3.3 triple knockout, Zmpif4.1/Zmpif4.2 double knockout, and Zmpif5.1 single knockout mutants. Results of sequencing analysis of these ZmPIF knockout mutant lines are shown in Supplemental Figure S12. All the homozygous ZmPIFs maize mutants exhibit shorter mesocotyls than the wild type in etiolated maize seedlings (Fig. 7, A and B), which were similar to the short-hypocotyl phenotype of the Arabidopsis pif mutant. Remarkably, the mesocotyl was hardly detectable in the etiolated seedlings of Zmpif3s triple knockout mutants (Fig. 7A). Tissue section analysis showed that the mesocotyl cells of the Zmpifs lines were significantly shorter than those of wild-type plants under dark conditions (Fig. 7, C and D). Several cell elongation-related genes were also detected to be down-regulated in the Zmpifs lines as compared with wild-type plants (Fig. 7E). These data suggest that ZmPIFs function as repressors of photomorphogenesis and positively regulate mesocotyl growth in dark-grown maize seedlings through promoting cell elongation.

Figure 7.

Knockout mutants of ZmPIFs show inhibited mesocotyl elongation in etiolated seedlings. A, Phenotypes of the etiolated wild-type (WT) and Zmpif3s, Zmpif4s, and Zmpif5.1 knockout mutant seedlings grown for 7 d after germination in darkness. Red arrowheads indicate the coleoptilar nodes between the mesocotyl and coleoptile. The seedlings were digitally extracted for comparison. Bar = 2 cm. B, Mesocotyl lengths of the etiolated seedlings represented in A. The values are means ± sd (n ≥ 10). Asterisks indicate significant differences compared with wild-type plants using Student’s t test (**, P < 0.001). C, Methylene Blue-stained longitudinal sections of etiolated wild-type and Zmpif3s, Zmpif4s, and Zmpif5.1 knockout mutant mesocotyls. Bars = 50 μm. D, Cell lengths of the mesocotyl of the etiolated seedlings represented in C. The values are means ± sd (n ≥ 30). Asterisks indicate significant differences compared with wild-type plants using Student’s t test (**, P < 0.001). E, Expression analysis of three cell expansion-related genes (ZmEXPA3, ZmEXPB4, and ZmEXPB6) in the mesocotyls of wild-type and Zmpif3s, Zmpif4s, and Zmpif5.1 mutant seedlings grown for 2 d after germination in darkness. The 2-d-old whole seedlings of Zmpif3s mutants were harvested for expression analysis instead of mesocotyls, as the mesocotyls of Zmpif3s mutants were too short to be harvested. The transcript levels were first normalized to Tubulin5. Data are means and sd of three independent biological replicates. Asterisks indicate significant differences compared with wild-type plants using Student’s t test (**, P < 0.001).

ZmPIFs Act as Positive Regulators in Response to Shade

Light quality is a crucial environmental factor that influences hypocotyl elongation. To investigate whether ZmPIFs are involved in regulating hypocotyl growth in Arabidopsis under low R:FR conditions, we analyzed the hypocotyl responses of Col-0, pifq, and ZmPIF-OE plants to simulated shade conditions. Hypocotyl measurements showed that the Arabidopsis transgenic plants overexpressing ZmPIF3.1, ZmPIF3.2, ZmPIF3.3, and ZmPIF4.2 appeared to be similar to pifq under simulated shade (WL+FR); however, the ZmPIF4.1-OE, ZmPIF5.1-OE, and ZmPIF5.2-OE plants had significantly longer hypocotyls than pifq under shade (Fig. 8, A and B). These results suggest that ZmPIF4.1, ZmPIF5.1, and ZmPIF5.2 may have stronger effects on promoting hypocotyl growth than ZmPIF3s and ZmPIF4.2 in response to shade in Arabidopsis. In addition, we noted that the ZmPIF4.1-OE transgenic lines displayed a constitutive shade avoidance response, resulting in early flowering and exaggerated petiole elongation compared with Col-0 and the pifq mutant under long-day conditions (Supplemental Fig. S13). This phenotype was reminiscent of the AtPIF4 overexpression plants (Kumar et al., 2012).

Figure 8.

ZmPIFs act as positive regulators in response to shade. A, Heterologous expression of ZmPIFs confers increased responses to simulated shade in Arabidopsis seedlings. Visible phenotypes of Col-0, pifq mutant, and ZmPIFs-OE seedlings (in the pifq background) were grown in constant WL (28 μmol photons m⁻² s⁻¹) for 2 d from germination and then retained in WL (high R:FR = 6.48) or transferred to WL+FR (low R:FR = 0.145) for an additional 5 d. The seedlings were digitally extracted for comparison. Bars = 2 mm. B, Quantification of hypocotyl lengths for Col-0, pifq mutant, and ZmPIFs-OE seedlings in the pifq background grown as in A. Data represent mean values and sd from at least 20 seedlings. Letters indicate significant differences by two-way ANOVA. C, Knockout of ZmPIFs by CRISPR/Cas9 reduces sensitivity to the simulated shade treatment in maize seedlings. Seedlings of the wild type (WT) and Zmpif mutant lines were grown for 3 d under continuous WL (35 μmol photons m⁻²s⁻¹) and then transferred to light/dark cycles (10 h of light/14 h of dark) for an additional 7 d. The light conditions were WL either supplemented with far-red light (low R:FR of 0.121) or not (high R:FR of 7.7). Red arrowheads indicate the lengths of the first leaf sheaths. The seedlings were digitally extracted for comparison. Bar = 2 cm. D, Quantification of first leaf lengths of the seedlings shown in C. Data represent means ± sd (n ≥ 6). Letters indicate significant differences by two-way ANOVA.

To further investigate the potential roles of ZmPIFs in this process in maize, we conducted a similar simulated shade treatment with the maize ZmPIFs knockout mutants. As expected, the wild-type maize seedlings exposed to such simulated shade conditions displayed longer first leaf sheaths characteristic of the shade avoidance response, but this response was attenuated in the Zmpifs mutants (Fig. 8, C and D). Interestingly, the extent of this reduction in responsiveness was more obvious in the Zmpif3s triple knockout mutant than in the Zmpif4s double knockout mutant and the Zmpif5.1 single knockout mutant (Fig. 8, C and D), indicating that ZmPIF3s are the dominant positive regulators of early shade response in maize.

DISCUSSION

Phys are the predominant photoreceptors that sense changes in light quality in an ambient environment to regulate plant growth and development. Several Arabidopsis PIF proteins, including PIF3, PIF4, PIF5, and PIF7, have been shown to play a particularly prominent role in shade avoidance by directly activating the expression of targeting genes, including the auxin synthesis genes (TAA1 and YUC), a cell wall-associated gene (XTH15), and transcription factors such as ATHB-2, PIF3-LIKE1 (PIL1), and LONG HYPOCOTYL IN FAR-RED1 (Zhang et al., 2013b). We recently showed that, in response to simulated shade, accumulation of AtPIF proteins increases and they directly repress the expression of MIR156s, thus releasing the downstream SPL gene family of transcription factors to regulate various aspects of shade avoidance responses (Xie et al., 2017). In this study, we conducted an initial characterization of the molecular properties and functions of ZmPIFs in regulating light signaling and photomorphogenesis in maize. As previously reported (Kumar et al., 2016), we identified seven potential ZmPIFs in maize, which share the conserved bHLH domain and APB motif. However, only members of the ZmPIF3s subgroup (ZmPIF3.1, ZmPIF3.2, and ZmPIF3.3) also have a phyA-binding motif, similar to AtPIF1 and AtPIF3 (Fig. 1A, right), suggesting that ZmPIF3s might be closer to AtPIF1 and AtPIF3 on an evolutionary scale. We found that all ZmPIF genes are constitutively expressed in most tissues examined, with relatively higher expression in leaves, an expression pattern also similar to that of AtPIFs (Jeong and Choi, 2013).

We collected several lines of evidence to show that these ZmPIF proteins share several similar molecular properties with the Arabidopsis PIFs. First, we showed that, like the AtPIFs, all ZmPIF proteins are capable of physically interacting with the full-length ZmphyB1 and AtphyB in both yeast and plant cells (Figs. 3 and 4; Supplemental Fig. S8). Although no interaction was detected between ZmPIFs and the full-length ZmphyB2 in yeast cells (Fig. 3), interaction between them was detected in plant cells using both LCI and BiFC assays (Fig. 4). Surprisingly, in contrast to the earlier report showing that the AtphyB (NT)-AtPIF3 interaction is red light inducible, we detected interaction between ZmPhyB1 and AtphyB (full length) with ZmPIFs under both darkness and light conditions in our Y2H assay (Fig. 3; Supplemental Fig. S8A). We speculate that this might be due to different lengths of phyB constructs (N-terminal fragment versus full length) or different configurations of phyB fusions with the GAL4 DNA-binding domain (GBD) used in these assays (phyB-GBD versus GBD-phyB; Ni et al., 1999). Second, we showed that all of these ZmPIFs are targeted to the nucleus like the Arabidopsis PIFs, and most of them (ZmPIF3.1, ZmPIF3.2, ZmPIF3.3, and ZmPIF4.2) are able to form nuclear bodies upon exposure to light (Figs. 2 and 6A). Third, we showed that all of these ZmPIFs can complement the mutant phenotypes of the pifq mutant to different degrees (Fig. 5) and that some of the ZmPIF overexpression transgenic plants (ZmPIF4.1-OE, ZmPIF5.1-OE, and ZmPIF5.2-OE) display an enhanced shade avoidance response phenotype in response to simulated shade (Fig. 8, A and B). It is particularly interesting that the ZmPIF4.1 overexpression plants displayed a strong constitutive SAS phenotype, similar to its Arabidopsis counterpart AtPIF4 (Supplemental Fig. S13A), which has been shown to be a strong promoting factor of elongated growth in response to shade or elevated temperature (Lorrain et al., 2008; Kumar et al., 2012; Sun et al., 2012). Also noteworthy, dark-grown ZmPIF5-OE transgenic seedlings exhibited an exaggerated apical hook phenotype similar to the Arabidopsis PIF5-overexpressing lines (Fig. 5C; Supplemental Fig. S10). Fourth, using the CRISPR/Cas9 technology, we created a set of ZmPIFs knockout mutants (Zmpif3s triple mutant, Zmpif4s double mutant, and Zmpif5.1 single mutant) and, as expected, we found that these knockouts all displayed reduced mesocotyl phenotypes in dark-grown seedlings, with the strongest phenotype being observed with the Zmpif3s triple knockouts (Fig. 7). We further showed that these knockouts all exhibited a compromised elongation response to simulated shade treatment (Fig. 8C). Together, these data suggest that ZmPIFs play a conserved function with the Arabidopsis PIFs in regulating seedling photomorphogenesis and the shade avoidance response.

Our results also provided evidence suggesting that these ZmPIFs may differ in several aspects from the AtPIFs. First, it has been shown that, upon brief light exposure, all AtPIFs are able to form nuclear bodies where they interact with activated phys (Bauer et al., 2004; Al-Sady et al., 2006), but we found that only the ZmPIF3.1:GFP, ZmPIF3.2:GFP, ZmPIF3.3:GFP, and ZmPIF4.2:GFP fusion proteins, but not others, formed visible nuclear bodies in Arabidopsis seedlings after a short exposure to high intensity of red light (Fig. 6A, R 1min), hinting that other ZmPIFs (ZmPIF4.1, ZmPIF5.1, and ZmPIF5.2) may have distinct interacting properties with Atphys. Second, previous studies showed that AtPIFs (PIF1, PIF3, and PIF5) have a relatively short half-life (less than 5 min) when exposed to red light (Shen et al., 2007, 2008). However, we found that all the ZmPIFs:GFP fusion proteins are much more stable than their Arabidopsis counterparts and they remained visible even after prolonged exposure to low Rc (1, 3, 6, and 24 h), a behavior reminiscent of that of AtPIF7 (Leivar et al., 2008a; Fig. 6B). There are several possible explanations for the observed slow degradation of ZmPIFs in Arabidopsis. One possibility is that, as heterologously expressed proteins, these ZmPIFs might not be efficiently recognized by the E3 ubiquitin ligases responsible for targeted degradation of AtPIFs, or it may reflect the intrinsic property of ZmPIFs being more stable proteins. In this regard, it is worth indicating that light-induced phosphorylation of AtPIF3 at multiple sites is necessary for the recruitment of the Light-Response Bric-a-Brack E3 ubiquitin ligases to target both PIF3 and phyB for degradation in vivo (Ni et al., 2013, 2014). Another possibility is that light-induced phosphorylation of ZmPIFs might be impaired in Arabidopsis, leading to their slow degradation. More detailed studies are required to address these possibilities.

Besides participating in the phy pathway, recent studies demonstrated that Arabidopsis PIFs serve a broader function, as a signaling hub that integrates environmental signals with multiple phytohormone biosynthetic or signaling pathways (Leivar and Quail, 2011; Pham et al., 2018). Consistent with this notion, recent studies showed that OsPIL1/OsPIL13 (a PIF4 homolog) plays a role in regulating the drought stress response by reducing internode elongation in rice (Todaka et al., 2012). Similarly, it was shown in an earlier study that overexpression of ZmPIF1 and ZmPIF3 (renamed ZmPIF3.1 and ZmPIF3.2 in this study) in rice also enhances drought tolerance (Gao et al., 2015, 2018). Furthermore, it was shown that ZmPIF4 (renamed ZmPIF4.1 in this study) physically interacts with the Arabidopsis DELLA protein REPRESSOR OF GA1-3 (RGA), indicating a potential interaction between ZmPIF4 and a GA signaling pathway on plant growth (Shi et al., 2018). We suspect that these ZmPIFs are likely to play important roles in regulating diverse aspects of plant growth and development, as well as in response to a wide range of biotic and abiotic stresses, through mediating signaling cross talk and integration with various hormone signaling pathways. Considering that there may be functional redundancy among ZmPIF members, further studies using additional single and various combinations of higher order mutants may help to answer this question.

Given that the phyB-PIF signaling module is highly conserved in plant architecture regulation and stress responses in plants, it was considered to be a preferred candidate target for crop improvement (Sawers et al., 2005; Wang and Wang, 2015). Previous attempts to improve crop shade tolerance by overexpressing phyB in several crops have demonstrated promise, despite some undesirable side effects (Gururani et al., 2015; Carriedo et al., 2016). It will be worthwhile to test the effect of overexpressing ZmphyB in future studies to attenuate SAS in maize. Alternatively, manipulating the downstream factors of phy (such as ZmPIFs) may also help to attenuate the effect of SAS in maize. In this study, we found that knocking out ZmPIFs in maize could substantially reduce the shade avoidance in maize seedlings. Next, we will examine whether these adult Zmpifs plants and their single mutants also exhibit attenuated SAS. Additionally, our previous study revealed a direct functional link between the phyB-PIF module and SPL factors in mediating shade avoidance responses in Arabidopsis (Xie et al., 2017) and speculated that this link may also operate in maize and other cereal crops (Wei et al., 2018). These downstream SPL factors might represent valuable targets for optimizing plant architecture for high-density planting. With the rapid development of genome/functional genome research and gene-editing technology in crops, we believe that more effective strategies will surface to modify SAS when a better understanding of light signaling mechanisms is achieved in crops.

MATERIALS AND METHODS

Plant Materials and Growth Conditions

The Arabidopsis (Arabidopsis thaliana) mutant used in this study was the pifq quadruple mutant (pif1, pif3, pif4, pif5; Leivar et al., 2008b), and the Arabidopsis ecotype Col-0 was used as the wild-type control. All the Arabidopsis transgenic plants used in this study were generated in the pifq background. The AtPIF5-OE (35S:PIF5-HA) transgenic line was kindly provided by Dr. Jiaqiang Sun (Nozue et al., 2007). Surface-sterilized seeds were sown on one-half-strength Murashige and Skoog medium (1% [w/v] Suc and 0.8% [w/v] agar, pH 5.8) and stratified for 3 d at 4°C. Adult plants were grown in soil under long-day conditions (16 h of light/8 h of dark) at 22°C.

Nicotiana benthamiana seeds were directly sown into the soil and grown in the greenhouse for about 1 month. Maize (Zea mays ssp. mays) inbred line seedlings were grown in growth chambers under WL conditions (14 h of light/10 h of dark, light intensity of 200 μmol photons m⁻² s⁻¹) at 28°C. Various tissues of 10-d-old maize B73 seedlings were used for RNA extraction to detect the tissue expression patterns of ZmPIFs. For detection of the mRNA expression of ZmPIFs in response to light, B73 seedlings were grown in darkness for 7 d and then transferred to WL for various times or retained in darkness as controls.

Total RNA Extraction and RT-qPCR Assays

Total RNA was isolated with Trizol reagent (Invitrogen), and reverse transcription reactions were performed using the manufacturer’s instructions of the first-strand cDNA synthesis kit (Tiangen Biotech). RT-qPCR was performed using the SYBR Green SuperReal PreMix Plus (Tiangen Biotech) on an Applied Biosystems 7500 real-time PCR detection system. The expression levels of Tubulin5 and ACT2 were used as internal controls for RT-qPCR in maize and Arabidopsis, respectively. The expression levels of genes were calculated using the relative 2^–ΔΔCt method (Saleh et al., 2008). The sequences of the primers are listed in Supplemental Table S2.

Phylogenetic Analysis

The full-length amino acid sequences of PIF in maize and Arabidopsis were aligned using the ClustalW program with default parameters in the alignment window of MEGA6 software (Tamura et al., 2013). A phylogenetic tree was constructed using the PIF sequences of maize and Arabidopsis using a maximum likelihood method, the Jones-Taylor-Thornton model, and partial deletion parameters.

Subcellular Localization Analysis

Coding sequences of each ZmPIF gene without the terminator were cloned into the pCambia1305 binary vector through the XbaI site with the CaMV35S promoter and a GFP tag to generate the Pro-35S:ZmPIFs-GFP vectors (primers are shown in Supplemental Table S2), and all binary vectors were introduced into Agrobacterium tumefaciens strain GV3101. For subcellular localization analysis of ZmPIFs, the A. tumefaciens GV3101 cells harboring Pro-35S:ZmPIFs-GFP and a nuclear protein marker construct (Pro-35S:mRFP-AHL22; Xiao et al., 2009) were coinjected into N. benthamiana leaf epidermal cells. GFP and RFP signals in transient transformed plants were observed using confocal microscopy (Zeiss LSM710) after 2 to 3 d.

Y2H Assay

The pGADT7 vector containing the GAL4 activation domain (GAD) and the pGBKT7 vector containing the GBD were obtained from Clontech. Full-length coding regions of ZmphyB1, ZmphyB2, and AtphyB were cloned into the pGBKT7 vector at the NdeI and EcoRI sites to generate GAL4-BD-ZmphyB1, GAL4-BD-ZmphyB2, and GAL4-BD-AtphyB, respectively. For GAL4-AD-PIFs, the coding region of each ZmPIF member was cloned into the pGADT7 vector at the EcoRI site. The BD- and AD-fused plasmids were cotransformed into the yeast strain AH109 according to the manufacturer’s instructions (Clontech Yeast Protocols Handbook). The phyB-ZmPIF interaction assay with PCB was performed as described by Shimizu-Sato et al. (2002), with minor modification. Briefly, the positive yeast colonies were first selected on SD-Leu-Trp medium and then plated on SD-Leu-Trp-His medium with or without 25 μm PCB (Scientific Frontier) under continuous far-red light (3 μmol photons m⁻² s⁻¹) or red light (4 μmol photons m⁻² s⁻¹) or in darkness for 3 d. The LacZ activity assay was performed to quantify protein-protein interactions according to the manufacturer’s instructions, using o-nitrophenyl β-d-galactopyranoside as the substrate. The sequences of the primers are listed in Supplemental Table S2.

LCI Assays

The vectors for the LCI assay (pCAMBIA1300-nLUC and pCAMBIA1300-cLUC) were described previously in Chen et al. (2008). The full-length coding regions of ZmphyB1, ZmphyB2, and AtphyB were fused to the pCAMBIA1300-nLUC vector at the SalI/KpnI sites to generate ZmphyB1-nLUC and ZmphyB2-nLUC, respectively. The full-length coding regions of ZmPIFs were fused to the pCAMBIA1300-cLUC vector at the SalI/KpnI sites to generate the corresponding ZmPIFs-cLUC vectors. LCI assays were performed as described by Chen et al. (2008). Briefly, both the nLUC- and cLUC-fused proteins were transformed into the A. tumefaciens strain EHA105 and infiltrated into N. benthamiana leaves with the indicated combinations. The negative controls (the empty vectors and split-LUC constructs) were supplemented with an equal amount of the control strain harboring the 35S:GUS reporter gene to confirm successful transfection. Samples were incubated in darkness for 48 h after the infiltration and then transferred to WL for 3 to 6 h. The luciferase activity was photographed using the NightShade LB985 Plant Imaging System (Berthold Technologies) after spraying with 20 mg mL⁻¹ potassium luciferin (Gold Biotech). These experiments were independently repeated at least three times. The sequences of the primers are listed in Supplemental Table S2.

BiFC Assays

The vectors for BiFC assays (p2YN, p2YC, pSPYNE-35S, and pSPYCE-35S) were described previously (Walter et al., 2004; Yang et al., 2007). Full-length coding sequences of ZmPIFs were fused in-frame with the N terminus of YFP. Full-length coding sequences of ZmphyB1 and ZmphyB2 were fused in-frame with the C terminus of YFP. ZmPIF3s and ZmPIF4s were inserted to p2YN vector at the PacI/SpeI sites, and ZmPIF5.1 and ZmPIF5.2 were inserted to pSPYNE-35S vector at the SalI site. ZmphyB1 and ZmphyB2 were inserted to p2YC and pSPYCE-35S vectors at the SalI site. The N. benthamiana leaves were coinfiltrated with A. tumefaciens EHA105 cells carrying the indicated plasmid pairs. The N. benthamiana plants were grown in the dark for 48 h after infiltration and then exposed to light for 3 to 6 h before analysis by confocal microscopy (Zeiss LSM710). The primers used in the BiFC assays are listed in Supplemental Table S2.

Arabidopsis Transformation and Phenotypic Analysis

For Arabidopsis transformation, the A. tumefaciens GV3101 cells harboring various Pro-35S:ZmPIFs-GFP constructs were separately transformed into the Arabidopsis pifq mutant plants using the floral dip method (Clough and Bent, 1998) to generate various ZmPIFs-OE/pifq lines. More than 20 independent lines of each transformation were selected with hygromycin, and two independent transgenic lines were verified by western-blot analysis for further studies.

For dark-grown seedlings, seeds were irradiated with continuous WL for 6 h at 22°C to stimulate germination and then placed at 22°C for 4 d under dark conditions before phenotypic analysis of seedlings. For analyzing the responses to simulated shade conditions, seedlings were grown in WL (28 μmol photons m⁻² s⁻¹, R:FR of 6.48) for 2 d at 22°C and then kept in WL or transferred to WL with supplemental continuous far-red light (WL+FR, R:FR of 0.145) for an additional 5 d before measurements were taken. Hypocotyl lengths were measured from at least 20 seedlings using the ImageJ software. Data were analyzed using Student’s t test (Statistical Analysis System package, version 8.01; SAS Institute).

Epifluorescence Microscopy

For fluorescence microscopy analyses, seedlings were grown in the dark for 4 d and then maintained in darkness, given a 1-min high Rc pulse (R 1min), or exposed to a low Rc for 60 min of 4 μmol photons m⁻² s⁻¹ (R 1h). The Rc was provided by the Percival LED-41HL2 growth chamber. Etiolated seedlings exposed to Rc over time as indicated were collected under dim-green safelight and fixed in 4% (v/v) paraformaldehyde butter as described by Zuo et al. (2012). Then the whole seedling was immersed with one drop of 4′,6-diamino-phenylindole (Vector Laboratories; catalog no. HT-1200) and examined with a Zeiss Axio Observer microscope. Representative cells were recorded by photography with a digital Zeiss camera system.

Protein Extraction and Immunoblots

For western-blot analysis, Arabidopsis seedlings were grown in the dark for 5 d and transferred to Rc (4 µmol photons m^–2 s^–1) for various durations. Two hundred milligrams of seedling powder was homogenized in a hot extraction buffer as described by Bauer et al. (2004), and heated for 5 min at 95°C. The supernatant was used for further experiments. Aliquots from each sample containing equal amounts of protein were subjected to PAGE as described by Al-Sady et al. (2006). The GFP tag was detected by western-blot assay using an anti-GFP antibody (Abmart). As a loading control, Actin was detected using an anti-Actin antibody (Abmart).

Generation and Mutation Analysis of CRISPR/Cas9 Knockout Lines in Maize

To generate ZmPIFs knockout transgenic lines, the expression vector of the dual-single guide RNA (sgRNA)/Cas9-mediated targeted deletion was constructed as described by Zhao et al. (2016) with minor modifications. Briefly, a series of target sites located on all the exons of each gene were analyzed using the SnapGene Viewer 3.2 software according to the reported criteria of 5′-GG-(N)18-NGG-3′. All the target sites were additionally BLASTed for specificity in the MaizeGDB database. Finally, two sgRNAs that specifically target sequences of individual ZmPIF genes were selected (Supplemental Table S3). The different forward primers ZmPIFs-sgR-F1 and ZmPIFs-sgR-F2 paired with the same reverse primer sgR-R were used to assemble sgRNA1 and sgRNA2 fragments through overlap-PCR, respectively. These sgRNA fragments driven by the maize ubiquitin U6-1 promoter were cloned into the CPB vector (Zhao et al., 2016) using the HindIII restriction site and an In-fusion HD Cloning Kit (TaKaRa). To overcome functional redundancy, the homologous ZmPIF genes of the same subgroup were combined in a single CRISPR/Cas9 vector system (Supplemental Fig. S11). Each construct was confirmed by PCR and sequencing analysis. All the constructs were introduced into the strain EHA105 and transformed into the immature embryo of a recipient maize inbred line ZC01 using the conventional A. tumefaciens-mediated approach.

The genotypes of selected knockout transgenic lines were confirmed by DNA sequencing using specific primers listed in Supplemental Table S2. The target region of each ZmPIF gene was amplified from ZC01 and the transgenic lines, and then PCR products were either directly sequenced or cloned into the pCR TA clone vector (Transgene) and sequenced. The mutated sequences of each ZmPIF gene in transgenic lines were revealed by alignment of sequences between ZC01 and the transgenic lines.

Phenotypic Analysis of Maize Zmpifs Knockout Mutants and Simulated Shade Treatment

All the homozygous Zmpifs maize mutants and the corresponding wild-type inbred maize line (ZC01) were uniformly buried at a 2-cm depth in plastic pots (8 × 8 cm) filled with soil. The same planting density of four kernels per plastic pot was used for all experiments. To measure the mesocotyls, the seedlings were grown in constant darkness at 28°C for 10 d. For simulated shade treatment, maize seeds were germinated and grown in a growth chamber (Percival LED-41HL2) at 28°C for 3 d under continuous WL (35 μmol photons m⁻² s⁻¹ photosynthetically active radiation; R:FR of 7.7) and then transferred to light/dark cycles (10 h of WL/14 h of dark) with WL+FR (low R:FR of 0.121) or with WL (high R:FR of 7.7) for an additional 7 d before measurement.

Mesocotyl Semithin Sections

To measure the length of the mesocotyl cells, the mesocotyl middle portion of 10-d-old maize etiolated seedlings was cut into 1 × 1 × 0.5-mm pieces. The specimens were treated according to the method of Kong et al. (2016). The specimens were cut to sections of 1 μm thickness on a microtome (Leica RM2155), and the sections were stained with 0.5% (w/v) Toluidine Blue and observed with a Leica DMLB microscope. The epidermal cells on the central region of the mesocotyl were observed using a Nikon Eclipse 80i upright microscope on the bright-field setting. Two fields were observed for each mesocotyl, and ∼20 to 30 cells per field were measured under 10× magnification. The average length of measured cells from three mesocotyls was used to represent cell length for each genotype.

Statistical Analysis

To determine the significant differences among the various genotypes treated with or without far-red light, the method of Brady et al. (2015) for two-way ANOVA with interaction was used by performing the aov function in the stats package in R version 3.5.0. The Tukey’s honestly significant difference method was used for all pairwise comparisons, with P values corrected for multiple comparisons to control against type I errors. Student’s t test was adopted to evaluate the significant differences in hypocotyl length of the ZmPIF-OE transgenic Arabidopsis plants in dark conditions and in mesocotyl length of the etiolated Zmpifs knockout maize seedlings.

Accession Numbers

Sequences of the maize genes analyzed in this work are available at http://ensembl.gramene.org/genome_browser/index.html/: ZmPIF3.1 (GRMZM2G115960), ZmPIF3.2 (GRMZM2G387528), ZmPIF3.3 (GRMZM2G062541), ZmPIF4.1 (GRMZM5G865967), ZmPIF4.2 (GRMZM2G165042), ZmPIF5.1 (GRMZM2G065374), ZmPIF5.2 (GRMZM2G016756), ZmphyB1 (GRMZM2G124532), ZmphyB2 (GRMZM2G092174), ZmEXPB4 (GRMZM2G154178), ZmEXPB6 (GRMZM2G176595), ZmEXPA3 (GRMZM2G074585), and Tubulin5 (GRMZM2G099167). Sequences of the Arabidopsis genes analyzed in this work are available at TAIR under the following accession numbers: AtPIF1 (AT2G20180), AtPIF3 (AT1G09530), AtPIF4 (AT2G43010), AtPIF5 (AT3G59060), AtPIF6 (AT3G62090), AtPIF7 (AT5G61270), AtPIF8 (AT4G00050), AtphyB (AT2G18790), XTH15 (AT4G14130), IAA19 (AT3G15540), and ACT2 (AT3G18780).

Supplemental Data

The following supplemental materials are available.

• Supplemental Figure S1. Diagram showing the annotated and actual exon-intron structure (α- and β-isoforms) of GRMZM2G165042 and GRMZM2G016756.

• Supplemental Figure S2. A heat map illustrating the expression levels of the seven ZmPIF genes in different tissues from various developmental stages.

• Supplemental Figure S3. Light regulation of the ZmPIF genes.

• Supplemental Figure S4. Quantitative Y2H liquid assay.

• Supplemental Figure S5. Western-blot analysis of ZmphyB2 protein in yeast cells.

• Supplemental Figure S6. GUS histochemistry staining showing proper transformation of negative controls in the LCI assay.

• Supplemental Figure S7. Negative controls for the BiFC assay.

• Supplemental Figure S8. ZmPIFs physically interact with AtphyB.

• Supplemental Figure S9. Expression analysis of the ZmPIFs transgenes in the selected Arabidopsis transgenic lines.

• Supplemental Figure S10. Effects of ZmPIF5 overexpression on ethylene-related responses.

• Supplemental Figure S11. Vector construction for genome editing of the ZmPIF genes.

• Supplemental Figure S12. Sequence analysis of the target sites in the Zmpifs knockout mutant lines.

• Supplemental Figure S13. Adult phenotypes of the Arabidopsis ZmPIFs-OE plants.

• Supplemental Table S1. Characteristics of the ZmPIF gene family.

• Supplemental Table S2. Primers used in this study.

• Supplemental Table S3. The targeted genes and two sgRNA target sequences with PAM.