For the DNA-binding analysis of the bHLH34 protein, full-length cDNA fragments were amplified using the following primers: for bHLH34, forward 5′-GGGGACAAGTTTGTACAAAAAAGCAGGCTTCATGTATCCATCAATCGAAGACGA-3′ and reverse 5′-GGGGACCACTTTGTACAAGAAAGCTGGGTCAGCAACAGGAGGAAGATTTTTGA-3′. The DNA products were introduced into the pDONR/ZEO vector (Invitrogen, Carlsbad, CA, United States) for DNA sequence analysis. The DNA fragment was then cloned into a pET300/NT-DEST vector by the Gateway system according to the manufacturer’s instruction (Invitrogen). The recombinant protein was expressed in the Escherichia coli strain BL21 (DE3) codon⁺ (Stratagene, La Jolla, CA, United States), and purified by ion exchange chromatography using Ni-charged His-bind resin (Novagen, Darmstadt, Germany).

Four DNA fragments derived from the 5′-upstream region of AtPGR [see Figure 1A; P1, between nucleotide (nt) positions -999 and -875 relative to the AtPGR start codon site; P2, -874 to -735 nt; P3, -734 to -595 nt; P4, -594 to -456 nt] were synthesized by PCR [initial denaturation at 94°C for 2 min, amplification for 30 cycles (denaturation at 94°C for 15 s, annealing 58°C for 15 s, and extension at 72°C for 20 s) and final extension at 72°C for 3 min] with 8 primers (Supplementary Table S1). Six DNA binding elements (DBEs), including DBE1 (-997 to -985 nt), DBE2 (-962 to -950 nt), DBE3 (-843 to -831 nt), DBE4 (-643 to -622 nt), DBE5 (-617 to -605 nt), and DBE6 (-500 to -488 nt), and various mutated DBE1 oligonucleotides were prepared using the two complementary sequences were synthesized as shown in Supplementary Table S1. The double-stranded complementary oligonucleotides used as probes were radiolabeled with γ-³²P-ATP and T4 polynucleotide kinase (New England BioLabs).

DNA–protein binding reactions were carried out at 25°C for 20 min in 20 μL volume containing 100 ng of each purified His-bHLH34 fusion protein, 1 μg of poly(dI-dC), 0.3 pmol of DNA fragments end-labeled with γ-³²P-ATP, and DNA-binding buffer (10 mM Tris-HCl, pH 7.5, 2.5% glycerol, 50 mM KCl, 1 mM EDTA, and 1 mM DTT). The reaction mixture was subjected to 5% non-denaturing PAGE in TBE (44.5 mM Tris-hydroxymethyl aminomethane, 44.5 mM boric acid, and 1 mM EDTA) at 100 V for 2 h at 4°C. The gel with DNA–protein complexes was dried and X-ray film was exposed to this gel.

Arabidopsis seedlings were grown in a growth room under intense light (110 μmol m^-2 sec^-1) at 22°C, 60% relative humidity, and 16 h/8 h light/dark conditions. The bHLH104 T-DNA insertion line SALK_005802 (bhlh104) was acquired from the Arabidopsis T-DNA insertion collection of the Salk Institute (Alonso et al., 2003). To select plants homozygous for the T-DNA insertion, the gene-specific primers 5′-CTCCAGAAAGCGGTGAGTTTTG-3′ and 5′-GAAGAATCACAAGTTTCTGGAG-3′ (forward and reverse, respectively) were utilized for the bhlh104 mutant. Plants yielding no PCR products with the gene-specific primers were subsequently tested for the presence of the T-DNA insertion using the gene-specific forward primer in combination with the T-DNA left border specific primer 5′-GCGTGGACCGCTGCACCT-3′.

The AtPGR promoter 999 (P999)-GUS construct was generated as described previously (Chung et al., 2016), and the overexpression of bHLH34 in the P999-GUS transgenic lines was achieved as described in Supplementary Materials and Methods.

For Glc treatment, 14-day-old Arabidopsis seedlings were submerged in sterilized water containing 6% Glc and sampled at 0, 6, 12, and 36 h under continuous light condition (110 μmol m^-2 sec^-1 light intensity, 24 h light condition) at 22°C, and 60% relative humidity conditions. For ABA or salt stress, 14-day-old Arabidopsis seedlings were submerged in the sterilized water containing 100 μM ABA or 150 mM NaCl and sampled at 0, 3, 6, and 12 h. For osmotic stress, 14-day-old Arabidopsis seedlings were submerged in the sterilized water containing 400 mM mannitol and sampled at 0, 6, 12, and 36 h. In each case, the collected seedlings were promptly frozen in liquid nitrogen and stored at -80°C.

Total RNA was isolated from the frozen samples using the Plant RNeasy Extraction Kit (Qiagen, Valencia, CA, United States). To remove residual genomic DNA from the preparation, total RNA samples were treated with RNase-free DNase I in accordance with the manufacturer’s instructions (Qiagen). The concentration of RNA was exactly quantified via spectrophotometric measurements, and 3 μg of total RNA was separated on 1.2% formaldehyde agarose gels to verify its concentration and monitor its integrity. qPCR was performed on a Rotor-Gene 6000 quantitative PCR apparatus (Corbett Research, Mortlake, NSW, Australia), and each result was analyzed using the RG6000 1.7 software (Corbett Research). Total RNA samples were isolated from the variously treated 14-day-old seedlings using an RNeasy Plant Mini Kit (Qiagen). qPCR was conducted using the SensiMix One-Step Kit (Quantance, London, United Kingdom). Actin 1 (ACT1) served as an internal control, and the quantitative analyses were performed by the Delta Delta C_T method (Livak and Schmittgen, 2001). Each sample was subjected to three independent experiments. RT-PCR was employed to measure the levels of bHLH34 expression in transgenic plants. The amount of RNA used in RT-PCR reactions was 300 ng. After 28 PCR cycles of amplification, 20 μL of each RT-PCR product was loaded onto a 1.2% (w/v) agarose gel in order to visualize the amplified DNA. The primers used for qPCR and RT-PCR reactions are shown in Supplementary Table S2.

For the Glc cotyledon greening test, seeds were sown on the Murashige and Skoog (MS) medium (Murashige and Skoog, 1962) supplemented with 5% or 6% Glc, and grown in a growth chamber under intense light (110 μmol m^-2 sec^-1) at 22°C, and 16 h/8 h light/dark conditions. Cotyledon greening was defined when a cotyledon was fully expanded and turned green. The green cotyledon was counted in triplicate (50 seeds/experiment).

For the ABA cotyledon greening or germination rate tests, seeds were sown on the MS medium supplemented with 0 or 1 μM ABA, and allowed to grow in the growth chamber under the same conditions as the Glc cotyledon greening assay. Germination was defined as an obvious protrusion of the root radicle through the seed coat. The germination rate of each line was measured after 1–8 days. Experiments were carried out in triplicate for each line (50 seeds each). Cotyledon greening of each seedling was measured at 14 days. The experiments were conducted in triplicate for each line (50 seeds each).

For the salt stress test, seeds were sown on the MS medium supplemented with 150 mM NaCl, grown in the growth chamber under the same conditions as the Glc cotyledon greening assay, and analyzed for the percentage of surviving seedlings after 3 weeks. The experiments were conducted in triplicate for each line (50 seeds each).

Chlorophyll (Chl) content of leaves was determined by the spectrophotometric method (Lichtenthaler, 1987). Leaf powder was used to extract Chl pigments in 96% ethanol. Absorbance was analyzed at 648.6 and 664.2 nm. Chl fluorescence measurement was performed using a pulse-modulated fluorometer (Junior-PAM, Heinz Walz, Effeltrich, Germany). Photosystem (PS) II quantum yield (F_v/F_m) was calculated from the F₀ and F_m values as described in the manufacturer’s instructions (Junior-PAM).

Arabidopsis Genome Initiative numbers for the sequences used in this study are as follows: bHLH34 (At3g23210), AtPGR (At5g19930), bHLH104 (At4g14410), bHLH105 (At5g54680), bHLH115 (At1g51070), bHLH28 (At5g46830), bHLH47 (At3g47640), bHLH11 (At4g36060), bHLH121 (At3g19860), AtHXK1 (At4g29130), RAB18 (At5g66400), RD29A (At5g52310), GIN6 (At2g40220), AtAPR2 (At1g62180), ABO3 (At1g66600), ABI1 (At4g26080), AtOZF2 (At4g29190), RD29B (At5g52300).

Recently, we demonstrated that the promoter sequences between nucleotide (nt) positions -999 and -456 of the AtPGR gene modulate transcription expression of the gene during Glc treatment (Chung et al., 2016). To identify transcription factors involved in the regulation of AtPGR, we performed yeast one-hybrid screening of a cDNA library of Arabidopsis (Chung et al., 2016). The yeast one-hybrid method involved in the X-galactosidase (Gal) filter assay was used to isolate transcription factor candidates that bind to P999-binding site (between positions -999 and -456 nt) in the upstream region of AtPGR (Figure 1A and Supplementary Figure S1A). Sequence analysis of the clone isolated from the P999 binding site (-999 to -456 nt) revealed that the gene encoded a basic helix-loop-helix (bHLH) family protein (At3g23210), which have been designated as a bHLH34 in Arabidopsis (Supplementary Figure S1B). bHLH34 cDNA is 963 bp long and encodes a protein of 320 amino acid residues with a calculated molecular weight of 35.6 kDa, which shows 30–56% identity and 52–67% similarity with Arabidopsis proteins bHLH115, bHLH105, bHLH104, bHLH28, bHLH47, bHLH11, and bHLH121. A phylogenetic tree representing the distance groups of sequences was built using a cluster algorithm (Supplementary Figure S1C). bHLH34 harbors a conserved single DNA-binding domain within a basic helix-loop-helix motif in its central region that is 88, 76, and 69% identical to the corresponding region of Arabidopsis bHLH104, bHLH115, and bHLH105 proteins, respectively (Supplementary Figure S1D). Recently, Li et al. (2016) demonstrated that bHLH34 and bHLH104 modulate iron homeostasis in Arabidopsis. Protein interaction assays indicate that both heterodimers and homodimers can form among bHLH34, bHLH104, and bHLH105. The bHLH105, also known as IAA-LEUCINE RESISTANT 3 (ILR3), modulate metal homeostasis, which influences IAA-conjugate hydrolysis (Rampey et al., 2006).

We attempted to express and purify the bHLH34 protein in E. coli to undertake electrophoretic mobility shift assay (EMSA) with the promoter in the region -999 to -456 of AtPGR. An EMSA was carried out to test whether the candidate protein bHLH34 can interact with each of the four DNA fragments (P1: -999 to -875 nt, P2: -874 to -735 nt, P3: -734 to -595 nt, and P4: -594 to -456; Figure 1A). Purified bHLH34 protein bound to all four fragments: P1, P2, P3, and P4 (Figure 1B).

To further investigate bHLH34-binding sites, we aligned the sequences of the P1-P4 fragments, and found that there were conserved cis-elements (5′-GAGA-3′) at six regions in the alignment (Figure 1C). The six DNA binding elements (DBEs), named as DBE1, DBE2, DBE3, DBE4, DBE5, or DBE6, were employed as a probe in an EMSA (Figure 1C). When we used the maltose-binding protein (MBP) as a negative control, there was no DNA–protein complex in the binding assay with DBE1 and DBE2 (Figure 1D, lane 3). Purified bHLH34 protein bound strongly to all six oligonucleotides DBE1–DBE6 (Figure 1D, lane 2). These results indicated that bHLH34 is a DNA-binding protein. To investigate that bHLH34 recognizes these oligonucleotides, excess unlabeled DBE1 was added. An addition of a molar excess of the unlabeled DBE1 oligonucleotide probe gradually diminished the shifted bands of bHLH34 protein in a dose-dependent manner (Figure 1E). These findings indicate that bHLH34 recognizes the six DBE fragments containing the consensus GAGA sequence.

To verify the specificity of binding between bHLH34 and the GAGA element, an EMSA was performed using various mutated DBE1 oligonucleotide fragments (Figure 1F). Mutations introduced in the GAGA sequence abrogated the binding of bHLH34 (Figure 1F, lanes m1, m2, m3, m4, and m5). From these results, we concluded that bHLH34 has sequence-specific binding affinity for the GAGA element in vitro.

On the other hand, consensus DNA sequence recognized by the bHLH proteins is known well as the E-box (5′-CANNTG-3′) (Robinson et al., 2000) and, single copy E-box (between positions -788 and -783 nt) exists in P999 site of the AtPGR gene. To determine whether bHLH34 protein could recognize the E-box, we conducted the EMSA assay with probe containing the E-box sequence (5′-CACTTG–3′; Supplementary Table S1), showing that bHLH34 could bind to E-box sequence (Figure 1G). This result indicate that bHLH34 binds to not only GAGA cis-element but also E-box sequence.

To determine the transcriptional activity of AtPGR by the bHLH34, we examined the AtPGR expression level in P999-GUS transgenic plants, when the bHLH34 was constitutively expressed under the control of Cauliflower Mosaic Virus (CaMV) 35S promoter (OX1-1/P999-GUS, OX2-5/P999-GUS). Two independent lines showing high levels of bHLH34 transgene expression (Figure 2A) were selected for transcription activity analyses. As shown in Figures 2B,C, under normal conditions, OX1-1/P999-GUS and OX2-5/P999-GUS lines increased GUS expression to approximately double that of P999-GUS alone (1.73- and 1.66-fold, respectively). When 6% Glc was applied to seedlings for 12 h, GUS activity in the P999-GUS transgenic plant was elevated in the leaves and roots (Figure 2B). In addition, a significant increase in GUS activity levels (1.44- to 1.47-fold) was observed in P999-GUS transgenic lines overexpressing bHLH34 constructs in comparison with the P999-GUS transgenic plant (Figures 2B,C). Furthermore, to investigate the level and tissue specificities of the AtPGR expression, we measured GUS activity and the transcript level of AtPGR on root and leaf tissues of the P999-GUS transgenic plants treated with or without Glc. As shown in Supplementary Figure S2, GUS and AtPGR expressions analyses indicated that the P999-GUS transgenic lines were expressed more strongly in leaves than in roots under the same condition. Constitutive CaMV35S promoter (35S pro)-GUS served as a positive control for analysis of GUS activity (Figures 2B,D). Taken together, these data indicate that bHLH34 is a functional transcription factor that regulates the expression of the AtPGR gene. Moreover, it appears that bHLH34 functions as a transcriptional activator.

To examine the tissue expression pattern of the bHLH34 and AtPGR, we analyzed the expression patterns of the bHLH34 and AtPGR in various organs and at different developmental stages of Arabidopsis seedlings by quantitative real-time PCR (qPCR). The maximal expression levels of the bHLH34 and AtPGR were detected in leaves, while moderate expression levels were detected in roots and flowers; low expression levels were detected in stem (Figure 3A). The expression patterns of bHLH34 and AtPGR during leaf development from 1 week after germination (WAG) to 5 WAG are also examined (Figure 3B). The expression patterns of bHLH34 and AtPGR were similar at the 1 and 2 WAG stages. In contrast, different expression patterns were observed for bHLH34 and AtPGR at the 3, 4 and 5 WAG stages, indicating that the expression level of the AtPGR is likely to be involved other unidentified regulatory factors or be independent of bHLH34 at different developmental stage.

To determine the in vivo functions of bHLH34, the accumulation of bHLH34 mRNA was assessed in Arabidopsis using qPCR during Glc, ABA, salt, and mannitol treatment. bHLH34 was induced in Arabidopsis seedlings after Glc treatment, and the bHLH34 transcript reached its maximum amount at 12 h followed by a slight decline (Figure 3C). In contrast, only slight upregulation of bHLH34 was observed after 3 h of ABA treatment (Figure 3D). In particular, transcript level of bHLH34 reached a peak within 3 and 6 h after the salt and mannitol treatment, respectively, and decreased for the remainder of the salt and mannitol experiments (Figures 3E,F). The stress-inducible A. thaliana Hexokinase 1 (AtHXK1), Responsive to ABA 18 (RAB18), and Responsive to Desiccation 29A (RD29A) genes served as references for the abiotic stress treatment (Figures 3C–F). These results indicate that bHLH34 is modulated by Glc, ABA, salinity, and osmotic stress.

To investigate bHLH34 function in vivo, bHLH34 was overexpressed in Arabidopsis transgenic plants under the control of the 35S promoter. Twelve homozygous lines (T₃ generation) were obtained, and two independent lines (OX2-1 and OX3-5) showing high levels of transgene expression (Supplementary Figure S3A) were selected for phenotypic characterization. To further evaluate the functional consequences of the loss of bHLH34, bhlh34 RNA interference (RNAi) lines were generated using the exon 1 to exon 2 cDNA sequence. bHLH34 expression was assessed by qPCR in two independent bhlh34 RNAi (ri2-2 and ri5-1) lines. bHLH34 expression was found to be knocked down in the RNAi lines (Supplementary Figure S3A). In addition, to determine whether the bHLH34 RNAi construct influences on the expression levels of the bHLH34 homologs, bHLH104, bHLH105 and bHLH115, in bhlh34 RNAi lines, we performed qPCR assay. The assay revealed that expression of the bhlh34 RNAi construct do not inhibit the expression bHLH104, bHLH105 and bHLH115, indicating that bhlh34 RNAi construct is highly specific to its own target transcripts (Supplementary Figures S3B–D). No morphological differences were observed among WT, bHLH34-overexpressing and bhlh34 RNAi plants, when grown on the Murashige and Skoog (MS) medium containing 1% sucrose (Murashige and Skoog, 1962) (Supplementary Figures S3E–G).

Chung et al. (2011) demonstrated that AtPGR is required for the modulation of Arabidopsis insensitivity to Glc exposure. To determine whether bHLH34 is also related to the Glc response, Glc sensitivity was evaluated by measuring the rate of cotyledon greening. WT, bHLH34-overexpressing (OX2-1, OX3-5), and bhlh34 (ri2-2, ri5-1) seeds were grown on the MS medium containing 5% Glc for 10 days. The cotyledon greening rate of WT seedlings was 33%, and this rate was approximately 25.6% (range 24.3–26.9%) for bhlh34 RNAi lines (Figures 4A,B). In contrast, approximately 75% (range 67.6–82.4%) of the bHLH34-overexpressing cotyledons expanded and turned green (Figures 4A,B). To assess the effects of bHLH34 expression on the ABA response, the seeds of the WT, bhlh34, and bHLH34-overexpressing plants were germinated on the MS medium containing 1% sucrose and 1 μM ABA. The germination percentage of bhlh34 lines was much more affected than that of WT and bHLH34-overexpressing plants by treatment with 1 μM ABA. The germination rate of bHLH34-overexpressing plants was higher than that of the WT (Figure 4C). These results show that the bHLH34 is required for ABA-modulated seed germination in Arabidopsis.

The ABA-induced effect was also evaluated by measuring the cotyledon greening rate. The relative reduction in cotyledon greening of the bhlh34 lines in response to ABA was more profound than that of the WT and bHLH34-overexpressing plants at 14 days after germination. In the WT, the cotyledon greening efficiency was 56.7%, whereas the cotyledon greening efficiency of bHLH34-overexpressing plants was 78.9–87.4% during treatment with ABA. In contrast, the cotyledon greening efficiency of bhlh34 RNAi lines was 11.7–20.6% (Figures 4D,E). These results indicated that the bhlh34 RNAi lines were more likely to be sensitive to Glc and ABA than the WT plant was. On the other hand, the bHLH34-overexpressing plants were less sensitive to exogenous Glc and ABA than were the WT and bhlh34 RNAi plants.

To characterize the effects of salt stress on the bHLH34-overexpression plants, we evaluated the response to treatment with 150 mM NaCl. The survival rate of the WT was slightly above 35% at 3 weeks after germination. Fewer than 30% of plants of the bhlh34 RNAi line (ri2-2) remained alive, and 27% of the bhlh34 leaves (ri5-1) turned green, as compared with 56–60% in OX2-1 and OX3-5 (Figures 4F,G). These results are consistent with the idea that bHLH34 is necessary for the regulation of developmental growth under abiotic stress.

As shown in Figures 2B,C, the induction of GUS activity caused by bHLH34 upon exposure to exogenous Glc may indicate that AtPGR participates in the Glc response through a bHLH34-mediated signaling pathway. To quantify AtPGR expression in the bHLH34 transgenic plants during Glc treatment, a second set of experiments using qPCR was conducted. Figure 5A shows that AtPGR mRNA expression in bHLH34-overexpressing transgenic plants without Glc treatment appears to be higher than that in WT and bhlh34 RNAi lines. By contrast, the transcript abundance of Glc-inducible AtPGR showed greater reduction in bhlh34 RNAi (ri2-2 and ri5-1) plants as compared with WT after Glc treatment. As expected, AtPGR expression was more strongly induced by Glc treatment in the bHLH34-overexpressing transgenic (OX2-1 and OX3-5) lines than in the WT and bhlh34 RNAi plants. Basically, these results are consistent with those obtained in the GUS activity assay after Glc treatment (Figure 2).

To investigate the transcriptional regulatory roles of bHLH34, we screened genes involved in Glc, ABA, and salt stress response and harboring the promoter including GAGA cis-element and E-box sequences. Evaluated were six genes, Glucose-Insensitive 6 (GIN6), Arabidopsis thaliana Adenosine 5′-Phosphosulfate Reductase 2 (AtAPR2), Abscisic acid Insensitive 1 (ABI1), ABA Overly Sensitive 3 (ABO3), Responsive to Desiccation 29B (RD29B), and A. thaliana Oxidation-related Zinc Finger 2 (AtOZF2), which are induced by various abiotic stresses (Leung et al., 1997; Arenas-Huertero et al., 2000; Ren et al., 2010; Huang et al., 2012; Chung et al., 2015).

As shown in Figures 5B–F, Glc- or ABA-induced expression of GIN6 or ABI1 was significantly reduced in bHLH34-overexpressing lines (OX2-1, OX3-5), in comparison with WT and bhlh34 RNAi (ri2-2, ri5-1) plants (Figures 5B,D). However, the induction of transcript levels of AtAPR2 and ABO3 were more pronounced in bHLH34-overexpressing lines than in WT and bhlh34 RNAi plants during Glc or ABA conditions (Figures 5C,E). The induction of transcript levels of dehydration-inducible RD29B and AtOZF2 genes were more pronounced in bHLH34-overexpressing lines than in WT and bhlh34 RNAi lines, but less pronounced in the bhlh34 RNAi lines than in WT following salt treatment (Figures 5F,G). These results indicated that bHLH34 may act positively or negatively depending on the types of abiotic stress-related genes harboring the promoter including GAGA cis-element and E-box sequences.

bHLH34 and bHLH104 sequences are highly homologous, with a significant 56% identity at the amino acid level (Supplementary Figure S4). Furthermore, bHLH34 and bHLH104 harbor a conserved DNA-binding domain within the basic helix-loop-helix motif in their central region that is 88% identical between these proteins (Supplementary Figure S1D). To determine whether the bHLH104 (At4g14410) gene is related to Glc, we obtained the At4g14410-tagged T-DNA insertion mutant SALK_005802. The absence of bHLH104 transcript was verified using qPCR (Figure 6B). The mutant was designated as bhlh104.

To investigate the genetic relation between bHLH34 and bHLH104, a bHLH34 RNAi construct was transformed into the bhlh104 mutant. Ten homozygous lines (T₃ generation) were obtained, and bHLH34 expression was assessed by qPCR in two independent bhlh34/bhlh104 lines (D2-1 and D3-1). bHLH34 expression was knocked down successfully in the double mutant lines (Figure 6A), and the effects of Glc on seed germination or cotyledon greening efficiency in the WT, bhlh34 RNAi (ri2-2), bhlh104, and bhlh34/bhlh104 (D2-1 and D3-1) lines were assessed. Under normal condition, the germination percentage and cotyledon greening rate were similar between WT, bhlh34, and bhlh104 mutant on the MS medium (Figures 6C–E). In contrast, the germination percentage or cotyledon greening rate of bhlh34/bhlh104 lines was much more affected than that of WT, bhlh34, and bhlh104 at 2 days after germination (Figure 6C). In addition, 6-day-old bhlh34/bhlh104 lines displayed paler green leaves (Figures 6D,E), more strongly decreased chlorophyll content, and lower photosynthetic efficiency (F_v/F_m) in comparison with the WT, bhlh34, and bhlh104 mutant, implying that the bhlh34/bhlh104 double mutant had a chlorotic leaf phenotype (Supplementary Figure S5). These results suggest that both bHLH34 and bHLH104 are most likely involved in seed germination and leaf chlorophyll synthesis in Arabidopsis.

To characterize the correlation of bHLH34 and bHLH104 functions in the Glc response, seeds of WT, bhlh34 RNAi (ri2-2),bhlh104, and bhlh34/bhlh104 lines (D2-1 and D3-1) were germinated on the MS medium containing Glc. At 6% Glc, 28.3% of WT and 18.1% of bhlh34 seeds germinated at 7 days, and this rate increased to 91.7% and approximately 81.7% (range 78.3–85%) for bhlh104 and bhlh34/bhlh104 seeds, respectively (Figure 6F). When the bhlh34/bhlh104 lines were allowed to grow for 14 days prior to the assessment of cotyledon greening rates in response to Glc, 23.3% of WT and 21.2% of bhlh34 leaves expanded and turned green, as compared to 73.3% in the bhlh104 mutant. Furthermore, approximately 62.7% (range 60–63.3%) of two bhlh34/bhlh104 lines survived at 14 days (Figures 6G,H). These results show that bhlh104 and bhlh34/bhlh104 mutants are more likely to be insensitive to Glc than the WT and bhlh34 RNAi. This finding indicated that the bhlh104 mutant is epistatic to the bhlh34 under high Glc condition, and bHLH104 may modulate the sensitivity of seed germination and cotyledon greening phenotype more tightly than bHLH34 does in the Glc response as a negative regulator.