The Molecular Mechanism of GhbHLH121 in Response to Iron Deficiency in Cotton Seedlings

Iron deficiency caused by high pH of saline–alkali soil is a major source of abiotic stress affecting plant growth. However, the molecular mechanism underlying the iron deficiency response in cotton (Gossypium hirsutum) is poorly understood. In this study, we investigated the impacts of iron deficiency at the cotton seedling stage and elucidated the corresponding molecular regulation network, which centered on a hub gene GhbHLH121. Iron deficiency induced the expression of genes with roles in the response to iron deficiency, especially GhbHLH121. The suppression of GhbHLH121 with virus-induced gene silence technology reduced seedlings’ tolerance to iron deficiency, with low photosynthetic efficiency and severe damage to the structure of the chloroplast. Contrarily, ectopic expression of GhbHLH121 in Arabidopsis enhanced tolerance to iron deficiency. Further analysis of protein/protein interactions revealed that GhbHLH121 can interact with GhbHLH IVc and GhPYE. In addition, GhbHLH121 can directly activate the expression of GhbHLH38, GhFIT, and GhPYE independent of GhbHLH IVc. All told, GhbHLH121 is a positive regulator of the response to iron deficiency in cotton, directly regulating iron uptake as the upstream gene of GhFIT. Our results provide insight into the complex network of the iron deficiency response in cotton.


Introduction
With the worldwide decrease in available arable land, the efficient use of soil has become an important research direction. About 30% of the world's arable land is saline-alkali soil with pH above 8.2 [1][2][3]. Saline soil refers to sodium salt soil mainly containing NaCl, while alkaline soil mainly refers to that containing Na 2 CO 3 and NaHCO 3 [4,5]. Alkaline soil means soil with a high pH value. When soil pH is above 7.4, the solubility of iron hydroxide is decreased to 10 −18 M, at which level it cannot be effectively absorbed and utilized by plants [6]. Iron is an essential microelement for plant growth and development, playing important roles in the electron transport chain (ETC) and in enzymatic reactions that contribute to many physiological metabolic processes such as photosynthesis, respiration, nitrogen fixation, and protein as well as nucleic acid synthesis [7][8][9]. Iron deficiency inhibits plant growth and development, evidenced in leaf chlorosis with decreased chlorophyll content, plant biomass, iron content, and increased iron reductase activity [10][11][12]. Cotton is a global cash crop with higher saline-alkali tolerance than other food crops [1,13]. However, when the alkaline level in the soil is too high, the cotton seedling will still

Iron Deficiency Induces Differential Expression of Genes Related to Iron Deficiency with Consequent Low Photosynthetic Efficiency in Cotton
Fe has a very low solubility in saline-alkali soil with high HCO 3 concentrations, resulting in limited uptake by plant roots [4]. Iron deficiency caused cotton leaf chlorosis, thicker roots, and fewer root hairs ( Figure 1A) [16,17]. In order to study the regulation mechanism governing the cotton seedling stage response to iron deficiency, we first irrigated the laboratory-grown cotton seedlings with 1 / 2 Hoagland nutrient solution either with (+Fe) or without Fe 2+ (−Fe). Under −Fe conditions, for two weeks, cotton seedlings appeared stunted and had chlorotic leaves ( Figure 1B). Leaf chlorophyll content ( Figure 1C), leaf biomass ( Figure 1D), Fe content ( Figure 1E), and net photosynthesis rate ( Figure 1G, Supplemental Figure S1) were significantly lower in the −Fe group than those in the control group, while ferric chelate reductase activity (FCR) was increased ( Figure 1F). The photosynthetic electron transport machinery is composed of two light energy-driven photosystems (PSs), PSI and PSII, along with the cytochrome (Cyt) b 6 f and ATP synthase. In linear electron transport (LET), electrons extracted from water by PSII were transported to PSI through Cytb 6 f and eventually produced NADPH [53,54]. The iron deficiency treatment suppressed expression of genes encoding proteins involved with PSI, PSII, Cytb 6 f, and Ferredoxin (Fd) (Supplemental Figure S2). Iron deficiency also affected the accumulation of reactive oxygen species (ROS) in cotton leaves (Supplemental Figure S3). Staining with nitroblue tetrazolium (NBT) (Supplemental Figure S3A) and diaminobenzidine tetrahydrochloride (DAB) (Supplemental Figure S3B) alike was darker on leaves of −Fe plants than those from +Fe control plants. These results indicate that the content of O 2 -and H 2 O 2 on cotton leaves is increased under −Fe conditions. Examination of superoxide dismutase (SOD), catalase (CAT), and peroxidase (POD) revealed the −Fe group exhibiting higher activities of SOD (Supplemental Figure S3C) and POD (Supplemental Figure S3D) in cotton leaves compared with the control group. CAT content did not differ significantly (Supplemental Figure S3E). All told, cotton grown under iron deficiency conditions showed low photosynthetic efficiency.
In order to study transcriptional regulation in iron-deficient cotton, a preliminary analysis was carried out of the expression of genes that may be involved in the response to iron deficiency. Firstly, the amino acid sequences of genes known to be involved in iron transport and iron signal regulation in Arabidopsis were extracted, and homologous genes in the cotton genome were identified by construction of a phylogenetic tree (Supplemental Figure S4). Cotton seedlings were then grown for two weeks with +Fe or −Fe treatment and the leaves harvested for reverse transcription qRT-PCR analysis. Expression of GhbHLH121, GhPYE (bHLH IVb), and GhbHLH104 (bHLH IVc) was found to be increased in the −Fe group, while GhbHLH115 (bHLH IVc) did not show a difference. Members of the bHLH Ib subfamily which functioned in a FIT-dependent transcriptional regulatory pathway were differentially impacted, GhbHLH38 was elevated significantly in the −Fe group, but another subfamily member, GhbHLH101, did not show a significant difference. GhFIT (bHLH IIIa) and GhBTS, which encoded a member of the protease degradation pathway, were both increased under −Fe conditions. The FIT-dependent iron-uptake genes GhIRT1, GhFRO2, and GhAHA2 also showed significantly higher expression in the −Fe group ( Figure 1H). These results indicate that iron deficiency is a substantial stressor for cotton seedlings and has direct impacts on the transcription of iron uptake-and homeostasis-related genes. Figure 1. Iron deficiency induces differential expression of cotton iron deficiency regulatory genes, resulting in iron deficiency phenotypes. (A) Schematic diagram of cotton seedling growing on normal (left) and saline-alkali (pH > 8.2, right) soil. Iron exists in form of soluble iron under normal soil conditions, which can be absorbed by plants, while in saline-alkali soil, iron mainly exists as insoluble iron, which leads to cotton suffering from iron deficiency stress, noted as chlorosis of cotton leaves. (B) Representative images of the phenotypes of cotton seedlings grown for two weeks under +Fe ( 1 /2 Hoagland nutrient solution) or −Fe ( 1 /2 Hoagland nutrient solution without Fe) conditions. Statistics were determined from three biological replicates, and each experiment contained ten seedlings. Scale bar = 5 cm. (C-G) Histograms of chlorophyll content (C), leaf biomass (D), Fe content (E), ferric chelate reductase activity (FCR) (F), and net photosynthesis rate (G) in cotton seedling leaves. Seedlings were grown for two weeks in +Fe or −Fe conditions. (H) Expression of marker genes of the cotton iron deficiency regulatory genes. Relative expression was determined by qRT-PCR of genes in cotton seedling leaves. Seedlings were the same plants as in (C-G). The regulatory relationship between genes refers to the regulatory network of Arabidopsis and does not mean its Figure 1. Iron deficiency induces differential expression of cotton iron deficiency regulatory genes, resulting in iron deficiency phenotypes. (A) Schematic diagram of cotton seedling growing on normal (left) and saline-alkali (pH > 8.2, right) soil. Iron exists in form of soluble iron under normal soil conditions, which can be absorbed by plants, while in saline-alkali soil, iron mainly exists as insoluble iron, which leads to cotton suffering from iron deficiency stress, noted as chlorosis of cotton leaves. (B) Representative images of the phenotypes of cotton seedlings grown for two weeks under +Fe ( 1 / 2 Hoagland nutrient solution) or −Fe ( 1 / 2 Hoagland nutrient solution without Fe) conditions. Statistics were determined from three biological replicates, and each experiment contained ten seedlings. Scale bar = 5 cm. (C-G) Histograms of chlorophyll content (C), leaf biomass (D), Fe content (E), ferric chelate reductase activity (FCR) (F), and net photosynthesis rate (G) in cotton seedling leaves. Seedlings were grown for two weeks in +Fe or −Fe conditions. (H) Expression of marker genes of the cotton iron deficiency regulatory genes. Relative expression was determined by qRT-PCR of genes in cotton seedling leaves. Seedlings were the same plants as in (C-G). The regulatory relationship between genes refers to the regulatory network of Arabidopsis and does not mean its presence in cotton. Values represent means ± SD of three biological replicates. Significant differences were determined by Student's t-test, ** p < 0.01.

GhbHLH121 Encodes a Transcription Factor Involved in Responding to Iron Deficiency in Root and Shoot Tissues
Iron deficiency induces the expression of GhbHLH121, which encodes a homolog of AtbHLH121 (86.62% amino acid similarity), an upstream regulatory TF for iron homeostasis in Arabidopsis (Supplemental Figure S5). The specific function of GhbHLH121 in the cotton has not been characterized. GhbHLH121 encodes a protein of the bHLH IVb subfamily that features a canonical bHLH domain (Supplemental Figures S4 and S5). The TM-1 reference genome for allotetraploid upland cotton features two copies of GhbHLH121 from the A and D subgenomes which have 97.44% similarity in their DNA sequences (Supplemental Figure S5). Transcriptome data show that GhbHLH121-A T and GhbHLH121-D T are highly expressed in roots and stems, followed by leaves, but expressed at lower levels in petals and sepals of floral organs. The expression patterns of GhbHLH121-AT and GhbHLH121-DT were not significantly different because they were more expressed in roots and leaves than in floral organs (Figure 2A). qRT-PCR further showed that GhbHLH121 expression predominantly occurred in roots and leaves at the seedling stage and continued to increase as leaves grew ( Figure 2B), with minimal levels in floral organs. In addition, the expression of GhbHLH121 in roots and leaves after −Fe treatment was significantly higher than that in the control ( Figure 2C,D). The subcellular localization signals of GhbHLH121-A T and GhbHLH121-D T were distributed in the nucleus and cytoplasm ( Figure 2E). Taken together, these findings indicate that GhbHLH121 is an iron deficiency-induced gene expressed in both roots and leaves. The homeologs of GhbHLH121-A T and GhbHLH121-D T are not different in terms of coding sequence, transcriptional activity, or protein localization. Therefore, the gene name will not be distinguished in subsequent experiments and is written as GhbHLH121 hereafter.

GhbHLH121 Encodes a Transcription Factor Involved in Responding to Iron Deficiency in Root and Shoot Tissues
Iron deficiency induces the expression of GhbHLH121, which encodes a homolog of AtbHLH121 (86.62% amino acid similarity), an upstream regulatory TF for iron homeostasis in Arabidopsis (Supplemental Figure S5). The specific function of GhbHLH121 in the cotton has not been characterized. GhbHLH121 encodes a protein of the bHLH IVb subfamily that features a canonical bHLH domain (Supplemental Figures S4 and S5). The TM-1 reference genome for allotetraploid upland cotton features two copies of GhbHLH121 from the A and D subgenomes which have 97.44% similarity in their DNA sequences (Supplemental Figure S5). Transcriptome data show that GhbHLH121-AT and GhbHLH121-DT are highly expressed in roots and stems, followed by leaves, but expressed at lower levels in petals and sepals of floral organs. The expression patterns of GhbHLH121-AT and GhbHLH121-DT were not significantly different because they were more expressed in roots and leaves than in floral organs ( Figure 2A). qRT-PCR further showed that GhbHLH121 expression predominantly occurred in roots and leaves at the seedling stage and continued to increase as leaves grew ( Figure 2B), with minimal levels in floral organs. In addition, the expression of GhbHLH121 in roots and leaves after −Fe treatment was significantly higher than that in the control ( Figure 2C,D). The subcellular localization signals of GhbHLH121-AT and GhbHLH121-DT were distributed in the nucleus and cytoplasm ( Figure 2E). Taken together, these findings indicate that GhbHLH121 is an iron deficiencyinduced gene expressed in both roots and leaves. The homeologs of GhbHLH121-AT and GhbHLH121-DT are not different in terms of coding sequence, transcriptional activity, or protein localization. Therefore, the gene name will not be distinguished in subsequent experiments and is written as GhbHLH121 hereafter. Values represent means ± SD of three biological replicates. Significant differences were determined by Student's t-test, ** p < 0.01.

Suppression of GhbHLH121 in Cotton Seedlings Reduced Tolerance to Iron Deficiency
In order to analyze the role of GhbHLH121 in regulating iron deficiency in cotton, we performed an Agrobacterium-mediated VIGS assay (virus-induced gene silence) to suppress GhbHLH121 expression in cotton seedlings. The VIGS-treated seedlings were incubated in hydroponic culture with 1 / 2 Hoagland nutrient solution either with (+Fe) or without Fe 2+ (−Fe) for three weeks. Under −Fe conditions, TRV2:00 plants showed the obvious corresponding phenotype, that is, leaf chlorosis ( Figure 3A, Supplemental Figure S6). The expression of GhbHLH121 was confirmed to be suppressed in TRV2:GhbHLH121 plants relative to TRV2:00 plants ( Figure 3B) under both +Fe and −Fe conditions. Overall, under the −Fe conditions, TRV2:00 and TRV2:GhbHLH121 both showed decreased chlorophyll and Fe content, and this resulted in less leaf biomass (Figure 3C,F,G). Under the +Fe conditions, suppression of GhbHLH121 led to slightly but significantly less biomass and Fe content than the control ( Figure 3C,G). In the −Fe conditions, suppression of GhbHLH121 greatly decreased chlorophyll and Fe content, leaf biomass, and FCR relative to TRV2:00 plants ( Figure 3C-G). The concomitant suppression of GhbHLH121 further decreased the net photosynthetic rate ( Figure 3E, Supplemental Figure S7) and also the expression of genes encoding proteins involved with PSI, PSII, Cytb 6 f, and Fd (Supplemental Figure S8). In contrast, SOD activity in TRV2:GhbHLH121 plants was significantly higher than that in TRV2:00 under the −Fe conditions, though there was no significant difference in POD and CAT activities (Supplemental Figure S9). Suppression of GhbHLH121 thus reinforced the damage from iron deficiency, indicating that GhbHLH121 should be a positive regulator of the response to iron deficiency. Values represent means ± SD of three biological replicates. Significant differences were determined by Student's t-test, ** p < 0.01.

Suppression of GhbHLH121 in Cotton Seedlings Reduced Tolerance to Iron Deficiency
In order to analyze the role of GhbHLH121 in regulating iron deficiency in cotton, we performed an Agrobacterium-mediated VIGS assay (virus-induced gene silence) to suppress GhbHLH121 expression in cotton seedlings. The VIGS-treated seedlings were incubated in hydroponic culture with 1 /2 Hoagland nutrient solution either with (+Fe) or without Fe 2+ (−Fe) for three weeks. Under −Fe conditions, TRV2:00 plants showed the obvious corresponding phenotype, that is, leaf chlorosis ( Figure 3A, Supplemental Figure S6). The expression of GhbHLH121 was confirmed to be suppressed in TRV2:GhbHLH121 plants relative to TRV2:00 plants ( Figure 3B) under both +Fe and −Fe conditions. Overall, under the −Fe conditions, TRV2:00 and TRV2:GhbHLH121 both showed decreased chlorophyll and Fe content, and this resulted in less leaf biomass (Figure 3C,F,G). Under the +Fe conditions, suppression of GhbHLH121 led to slightly but significantly less biomass and Fe content than the control ( Figure 3C,G). In the −Fe conditions, suppression of GhbHLH121 greatly decreased chlorophyll and Fe content, leaf biomass, and FCR relative to TRV2:00 plants ( Figure 3C-G). The concomitant suppression of GhbHLH121 further decreased the net photosynthetic rate ( Figure 3E, Supplemental Figure S7) and also the expression of genes encoding proteins involved with PSI, PSII, Cytb6f, and Fd (Supplemental Figure S8). In contrast, SOD activity in TRV2:GhbHLH121 plants was significantly higher than that in TRV2:00 under the −Fe conditions, though there was no significant difference in POD and CAT activities (Supplemental Figure S9). Suppression of GhbHLH121 thus reinforced the damage from iron deficiency, indicating that GhbHLH121 should be a positive regulator of the response to iron deficiency. The last leaves of seedlings from TRV2:00 and the TRV2:GhbHLH121 seedlings. Statistics were determined from three biological replicates, and each experiment contained ten seedlings. Scale bar = 5 cm. (B) Expression of GhbHLH121 in TRV2:00 and the TRV2:GhbHLH121 seedlings in +Fe or −Fe solution. Expression was determined by qPCR using RNA from leaf of (A). (C-G) Histograms of chlorophyll content (C), FCR activity (D), net photosynthesis rate (E), leaf biomass (F), and Fe content (G) in seedlings from (A). Photo: net photosynthesis rate. Relative expression was determined by qRT-PCR in cotton leaves. Values represent means ± SD of three biological replicates. Significant differences were determined by Student's t-test, * p < 0.05, ** p < 0.01. A further transcriptional examination of genes involved in iron uptake and iron signaling was conducted by using the material of the TRV2:GhbHLH121 and TRV2:00 (Table 1). qRT-PCR further showed that suppression of GhbHLH121 did not affect the expression of GhbHLH101 and GhbHLH38 (bHLH Ib), nor that of GhIRT1 and GhFRO2 under the +Fe conditions. How-ever, under −Fe conditions, GhbHLH38, GhIRT1, and GhFRO2 all were significantly lower in TRV2:GhbHLH121 than in TRV2:00 controls, while GhbHLH101 was not differentially expressed. In addition, GhFIT was significantly lower in TRV2:GhbHLH121 than in TRV2:00 under both +Fe and −Fe conditions. Meanwhile, expression of the iron transport gene GhYSL1 decreased under −Fe, but that of GhNAS4, GhZIF1, and GhFRD3 was increased. However, there was no significant difference in the expression of GhZIF1 and GhFRD3 between TRV2:GhbHLH121 and TRV2:00 plants. GhYSL1 and GhNAS4 were decreased in TRV2:GhbHLH121 under −Fe conditions. It has been reported that PYE regulates YSL1, NAS4, and FRD3 to facilitate iron redistribution in plants, while BTS regulates iron homeostasis by affecting the stability of bHLH IVc proteins [40,55]. This study found both GhPYE and GhBTS to be induced under −Fe conditions. Furthermore, in +Fe conditions, GhPYE showed similar expression in TRV2:GhbHLH121 and TRV2:00 plants, but GhBTS was decreased with TRV2:GhbHLH121. Meanwhile, in −Fe conditions, GhPYE and GhBTS exhibited significantly lower expression in TRV2:GhbHLH121 than in TRV2:00. These findings support that GhbHLH121 is actively involved in responding to iron deficiency in cotton through regulating iron uptake and iron signaling under iron deficiency. Table 1.
Differential expression of Fe deficiency-responsive genes in TRV2:00 and the TRV2:GhbHLH121 seedlings in +Fe or −Fe solution. The relative gene expression levels measured by qPCR in leaves of the TRV2:00 and the TRV2:GhbHLH121 seedlings in +Fe or −Fe solution. The expression of GhHistone3 was used to normalize mRNA levels, and the gene expression level in the TRV2:00 under Fe-sufficient conditions was set to 1. Values represent means ± SD of three biological replicates. Significant differences from the corresponding wild type are indicated by an asterisk (* p < 0.05, ** p < 0.01), as determined by Student's t-test. In order to study the effects of iron deficiency stress on the submicroscopic structure of cotton leaves, we observed the leaves of TRV2:GhbHLH121 and TRV2:00 plants treated with +Fe or −Fe by transmission electron microscopy (TEM). In +Fe conditions, both TRV2:GhbHLH121 Plants 2023, 12, 1955 8 of 23 and TRV2:00 exhibited normal chloroplast morphology with the thylakoids well piled up ( Figure 4A). The chloroplasts contained oval starch granules and a few plastid globules (PGs). The mitochondrial was also in a round shape with the inner ridges arranged in an orderly manner ( Figure 4B). The nuclear double membrane and the cytoplasmic membrane structure were clearly observed ( Figure 4C). Under −Fe conditions, however, submicroscopic observation of TRV2:00 leaf cells revealed obvious damage to the morphology of organelles. The internal structure of chloroplasts was disordered with loose arrangements of thylakoids ( Figure 4A). Chloroplast sizes were significantly different with the width decreased ( Figure 4A; Supplemental Figure S10B), aspect ratio increased (Supplemental Figure S10C), and PG number increased (Supplemental Figure S9D). Mitochondrial structure was likewise damaged with the loss of inner ridges ( Figure 4B) and increased aspect ratio (Supplemental Figure S10E). The nucleus overall remained a normal shape ( Figure 4C), but the cell membrane was also damaged or broken ( Figure 4C). Thus, iron deficiency leads to chloroplast thylakoid structural disorder in cotton leaves, and the suppression of GhbHLH121 increases the damages associated with iron deficiency, which confirms it as playing a positive role in iron homeostasis. With suppression of GhbHLH121, the chloroplast structure was damaged, the membrane lipid peroxidation was aggravated, and ROS accumulated; the photosynthetic electron transport system was destroyed, which resulted in the decrease in the photosynthetic rate.

Ectopic Expression of GhbHLH121 in Arabidopsis Confers Enhanced Tolerance to Iron Deficiency
To investigate the role of GhbHLH121 in the regulation of iron homeostasis, transgenic Arabidopsis lines were generated through ectopically expressed GhbHLH121 (GhbHLH121-OE) fused to the cauliflower mosaic virus (CaMV) 35S promoter. Under +Fe conditions, there was no significant difference between the Arabidopsis seedlings of GhbHLH121-OE and WT in terms of growth state, including root length, chlorophyll content, FCR, and Fe content. Under −Fe conditions, WT seedlings exhibited a pronounced iron deficiency phenotype, including inhibited root growth, reduced FCR, and Fe content ( Figure 5A-E). Compared with WT, GhbHLH121-OE seedlings showed amelioration of iron deficiency symptoms, that is, the root lengths were twice that of WT ( Figure 5A,C) and FCR and Fe contents were higher ( Figure 5D,E). These results suggest that the ectopic expression of GhbHLH121 can reduce iron deficiency damages to Arabidopsis.

Ectopic Expression of GhbHLH121 in Arabidopsis Confers Enhanced Tolerance to Iron Deficiency
To investigate the role of GhbHLH121 in the regulation of iron homeostasis, transgenic Arabidopsis lines were generated through ectopically expressed GhbHLH121 (GhbHLH121-OE) fused to the cauliflower mosaic virus (CaMV) 35S promoter. Under +Fe conditions, there was no significant difference between the Arabidopsis seedlings of GhbHLH121-OE and WT in terms of growth state, including root length, chlorophyll content, FCR, and Fe content. Under −Fe conditions, WT seedlings exhibited a pronounced iron deficiency phenotype, including inhibited root growth, reduced FCR, and Fe content ( Figure 5A-E). Compared with WT, GhbHLH121-OE seedlings showed amelioration of iron deficiency symptoms, that is, the root lengths were twice that of WT ( Figure 5A,C) and FCR and Fe contents were higher ( Figure 5D,E). These results suggest that the ectopic expression of GhbHLH121 can reduce iron deficiency damages to Arabidopsis. Furthermore, we analyzed whether the tolerance of GhbHLH121-OE plants to iron deficiency is related to the expression of iron-related genes by qRT-PCR (Table 2, Supplemental Figure S11). In the transgenic lines evaluated here, expression of GhbHLH121 ( Figure 5B, Supplemental Figure S12) was increased under −Fe. In addition, the iron Furthermore, we analyzed whether the tolerance of GhbHLH121-OE plants to iron deficiency is related to the expression of iron-related genes by qRT-PCR (Table 2, Supplemental Figure S11). In the transgenic lines evaluated here, expression of GhbHLH121 ( Figure 5B, Supplemental Figure S12) was increased under −Fe. In addition, the iron uptake genes IRT1 and FRO2 were upregulated together with FIT and bHLHIb (Table 2). Under +Fe conditions, ectopic expression of GhbHLH121 did not affect the expression of FIT, bHLHIb, IRT1, and FRO2. In addition, expression of PYE was significantly higher in GhbHLH121-OE than in WT under −Fe conditions, but the expression of YSL1, NAS4, and FRD3 did not differ significantly. Expression of BTS was higher in GbhHLH121-OE than in WT under −Fe conditions. While bHLH IVc itself is not activated by iron deficiency and is not affected by GhbHLH121 under either −Fe or +Fe conditions, the observed expression patterns suggest that GhbHLH121 may effectively alleviate the symptoms of iron deficiency in Arabidopsis by regulating genes related to iron homeostasis. Table 2. Differential expression of Fe deficiency-responsive genes in WT and GhbHLH121-OE seedlings in +Fe or −Fe solution. The relative gene expression levels measured by qPCR in WT and GhbHLH121-OE seedlings grown for 7 days under +Fe or −Fe conditions. The expression of TUB2 was used to normalize mRNA levels, and the gene expression level in the wild type under Fe-sufficient conditions was set to 1. Values represent means ± SD of three biological replicates. Significant differences from the corresponding wild type are indicated by an asterisk (* p < 0.05, ** p < 0.01), as determined by Student's t-test.

GhbHLH121 Interacts with GhbHLH IVc Genes and GhPYE
In order to study the molecular regulatory network of GhbHLH121 in cotton, GhbHLH121 was inserted into pGBKT7 to construct a bait vector, which was used to screen a library of cotton proteins via the yeast two-hybrid system. A total of 85 genes whose products potentially interact with GhbHLH121 were screened from the yeast library (Supplemental Table S1), including GhbHLH104, GhbHLH115 (bHLH IVc TF), and PYE (bHLH IVb TFs). Yeast two-hybrid assay and Luciferase complementation assays confirmed that GhbHLH121 can interact with GhbHLH104, GhbHLH115, and PYE ( Figure 6A,B). The bimolecular fluorescence complementation technology (BIFC) further confirmed these interactions ( Figure 6D). In addition, YFP fluorescence signals were observed to be localized in the nucleus for three fusion proteins ( Figure 6C). However, Figure 2E shows that the subcellular localization signals of GhbHLH121 were distributed in the nucleus and cytoplasm. Figure 6D shows that heterodimer of GhbHLH121 and GhbHLH104, GhbHLH115, PYE resides in the nucleus. confirmed these interactions ( Figure 6D). In addition, YFP fluorescence signals were observed to be localized in the nucleus for three fusion proteins ( Figure 6C). However, Figure 2E shows that the subcellular localization signals of GhbHLH121 were distributed in the nucleus and cytoplasm. Figure 6D shows that heterodimer of GhbHLH121 and GhbHLH104, GhbHLH115, PYE resides in the nucleus.

GhbHLH121 Directly Activates Transcription of GhbHLH38, GhFIT, and GhPYE Independent of GhbHLH104 and GhbHLH115
bHLH transcription factors regulate the expression of target genes by recognizing and binding E-box DNA motifs [56,57]. Multiple E-boxes are present in the promoters of GhbHLH38, GhFIT, and GhPYE ( Figure 7A; Supplemental Table S2). In order to explore whether GhbHLH121 can bind to these E-box motifs, we generated constructs with the Ebox sites mutated with CA to TC (mE-box) [58]. Yeast one-hybrid assay confirmed

GhbHLH121 Directly Activates Transcription of GhbHLH38, GhFIT, and GhPYE Independent of GhbHLH104 and GhbHLH115
bHLH transcription factors regulate the expression of target genes by recognizing and binding E-box DNA motifs [56,57]. Multiple E-boxes are present in the promoters of GhbHLH38, GhFIT, and GhPYE ( Figure 7A; Supplemental Table S2). In order to explore whether GhbHLH121 can bind to these E-box motifs, we generated constructs with the E-box sites mutated with CA to TC (mE-box) [58]. Yeast one-hybrid assay confirmed GhbHLH121 could bind to E-box motifs on the promoters of GhbHLH38, GhFIT, and GhPYE ( Figure 7B). Mutation of the E-boxes in each promoter decreased this binding efficiency ( Figure 7D). GhbHLH121 could bind to E-box motifs on the promoters of GhbHLH38, GhFIT, and GhPYE ( Figure 7B). Mutation of the E-boxes in each promoter decreased this binding efficiency ( Figure 7D). To further verify whether GhbHLH121 has transcriptional activation effects on GhbHLH38, GhFIT, and GhPYE, the promoters of GhbHLH38, GhFIT, and GhPYE were fused into the LUC vector as reporter genes, for which GhbHLH121 was used as the effector ( Figure 7C). Co-injection of constructs into tobacco leaves revealed that GhbHLH121 can activate the transcription of GhbHLH38, GhFIT, and GhPYE ( Figure 7D,E).
Lei et al. reported AtbHLH121 alone had no notable effect on FIT and bHLH38 promoter activity, while AtbHLH104 and AtbHLH115 each induced significant activation [46]. In addition, bHLH IVc TFs were reported to directly bind the bHLH121 promoter [46]. Therefore, we designed the experiment to explore whether the regulatory networks of cotton and Arabidopsis thaliana are consistent. GhbHLH104 and GhbHLH115 were used as effectors and the GhbHLH121 promoter in the reporter ( Figure 7F). GhbHLH104 and GhbHLH115 are unable to activate the transcription of GhbHLH121 (Figure 7G,H). Thus, GhbHLH121 can activate transcription of GhbHLH38, GhFIT, and GhPYE, and this activation was independent of GhbHLH104 and GhbHLH115.

Effects of Iron Deficiency Stress on Cotton Seedling Development
Soil salinization is a major environmental threat to the agriculture industry worldwide and affects approximately 20% of the world's cultivated land and nearly half of all irrigated land, and the impact is becoming increasingly severe [59]. For example, the high pH in saline-alkali soil causes a lack of absorbable iron, which is harmful to the growth and development of seedling-stage cotton. Iron is required for a wide variety of plant metabolic processes but is particularly crucial for photosynthesis, as it acts as a cofactor in both photosystems [60,61]; consequently, Fe deficiency severely limits photosynthetic efficiency [62]. This study is the first to conduct a comprehensive and in-depth analysis of the photosynthetic efficiency of cotton seedlings under iron deficiency. Iron-deficient cotton seedlings exhibited leaf chlorosis due to reduced leaf Fe content and chlorophyll content, which in turn led to decreased leaf photosynthetic rate (Figure 1). There are two main mechanisms for the reduction in plant photosynthesis caused by abiotic stress: stomatal limitation and non-stomatal limitation [63,64]. Stomatal limitation refers to the decrease in stomatal conductance, which directly affects the photosynthesis [65]. Iron deficiency did not affect transpiration rate stomatal conductance and intercellular CO 2 concentration (Supplemental Figure S1). The non-stomatal limitations mainly include the destruction of chloroplast structure, the decrease in chlorophyll content, the inhibition of Ru-BP carboxylase activity, the decrease in glycolate oxidase activity, the decrease in photosynthetic phosphorylation activity, and the change in reactive oxygen species metabolism [62]. This reduction in photosynthetic rate in iron-deficient cotton leaves might be mainly attributable to the following two causes. First, with regard to the overall impact on transcriptional regulation, iron deficiency conditions inhibited the expression of genes involved in photosystem elements, including PSI, PSII, Cytb 6f , and Fd (Supplemental Figure S2), with consequent weakening of the electron transfer efficiency of the photosynthetic chain. This pattern has previously been reported in Arabidopsis [38], phytoplankton [66,67], and sugar beet (Beta vulgaris L. cv. F58-554H1) [68]. Second, we observed the ultracellular structure of cotton leaf cells to be damaged under iron deficiency, predominantly on chloroplast and mitochondrial ( Figure 4). Therefore, the decrease in photosynthetic rate under iron deficiency stress in cotton seedlings may be mainly caused by non-stomatal limiting factors.

GhbHLH121 Is a Positive Regulator for the Response to Iron Deficiency in Cotton
The response of AtbHLH121 to iron deficiency is a subject of debate (Figure 8) [45][46][47], but here we observed the expression of GhbHLH121 to increase under iron deficiency ( Figure 2). AtbHLH121 was a transcription factor and positive regulator of iron homeostasis in Arabidopsis [45][46][47]. Under −Fe conditions, Atbhlh121 mutants have decreased tolerance to iron deficiency, exhibiting strong inhibition of primary root growth and FCR activity along with low fresh weight and chlorophyll content [45][46][47].

GhbHLH121 is a Positive Regulator for the Response to Iron Deficiency in Cotton
The response of AtbHLH121 to iron deficiency is a subject of debate (Figure 8) [45][46][47], but here we observed the expression of GhbHLH121 to increase under iron deficiency ( Figure 2). AtbHLH121 was a transcription factor and positive regulator of iron homeostasis in Arabidopsis [45][46][47]. Under −Fe conditions, Atbhlh121 mutants have decreased tolerance to iron deficiency, exhibiting strong inhibition of primary root growth and FCR activity along with low fresh weight and chlorophyll content [45][46][47].

Figure 8.
A working model illustrating the roles of GhbHLH121 in response to iron deficiency in cotton. When cotton is subjected to iron deficiency, expression of GhbHLH121 is increased, though GhbHLH104 and GhbHLH115 do not activate its transcription. GhbHLH121 forms a heterodimer with GhbHLH104 or GhbHLH115, but can activate the expression of GhbHLH38, GhFIT, and GhPYE independent of them.
We found that ectopic expression of GhbHLH121 in Arabidopsis can increase the iron content of transgenic Arabidopsis GhbHLH121-OE and improve their tolerance to iron deficiency ( Figure 5). Conversely, suppressing its expression in cotton resulted in the leaves becoming chlorotic, chlorophyll and leaf iron contents being reduced, and the photosynthetic rate decreasing (Figure 3). In addition, iron deficiency in cotton caused the ultracellular structure of leaves to be severely damaged, chloroplast thylakoids more loosely arranged, and mitochondria almost completely hollowed (Figure 4). These results suggest that inhibiting expression of GhbHLH121 decreases cotton tolerance to iron deficiency.
In cotton, GhbHLH121 is an iron deficiency response gene, which activates the expression of downstream genes only when cotton faces iron deficiency. Guerinot et al. propose a model in which phosphorylated AtbHLH121 is allowed to accumulate when plants become Fe-deficient and interacts with subgroup IVc bHLH transcription factors. These When cotton is subjected to iron deficiency, expression of GhbHLH121 is increased, though GhbHLH104 and GhbHLH115 do not activate its transcription. GhbHLH121 forms a heterodimer with GhbHLH104 or GhbHLH115, but can activate the expression of GhbHLH38, GhFIT, and GhPYE independent of them.
We found that ectopic expression of GhbHLH121 in Arabidopsis can increase the iron content of transgenic Arabidopsis GhbHLH121-OE and improve their tolerance to iron deficiency ( Figure 5). Conversely, suppressing its expression in cotton resulted in the leaves becoming chlorotic, chlorophyll and leaf iron contents being reduced, and the photosynthetic rate decreasing (Figure 3). In addition, iron deficiency in cotton caused the ultracellular structure of leaves to be severely damaged, chloroplast thylakoids more loosely arranged, and mitochondria almost completely hollowed (Figure 4). These results suggest that inhibiting expression of GhbHLH121 decreases cotton tolerance to iron deficiency.
In cotton, GhbHLH121 is an iron deficiency response gene, which activates the expression of downstream genes only when cotton faces iron deficiency. Guerinot et al. propose a model in which phosphorylated AtbHLH121 is allowed to accumulate when plants become Fe-deficient and interacts with subgroup IVc bHLH transcription factors. These heterodimers bind to the promoters of subgroup Ib genes and induce their expression. Subgroup Ib transcription factors then form heterodimers with FIT to activate the transcription of FRO2, IRT1, and other FIT-regulated genes. The induction of Fe uptake genes enhances Fe uptake and alleviates iron deficiency [45]. However, whether iron deficiency affects phosphorylation of GhbHLH121 in cotton remains to be seen.

The Transcriptional Regulatory Network of Cotton GhbHLH121
Due to whole-genome duplication and allopolyploidization, the cotton genome contains many more members of the bHLH transcription factor family than does Arabidopsis (Supplemental Figure S4; Supplemental Table S3), and they also exhibit specific transcriptional patterns. In Arabidopsis, the bHLH Ib subgroup has four members whose expression is induced by iron deficiency [69]. In contrast, the cotton genome features seven members and only GhbHLH38 is induced by iron deficiency (Table 1). Moreover, GhbHLH101 is barely expressed in leaves (Supplemental Figure S13). Among iron homeostasis genes, the core gene FIT (bHLH29, bHLH IIIe) has four homologs in the cotton genome, of which only a pair of duplicated homeologs are stably expressed in leaves (Supplemental Figures  S4 and S13; Supplemental Table S3). Of the bHLH IVc subgroup [70] in Arabidopsis, only AtbHLH115 is induced by iron deficiency [71]. Meanwhile, cotton members of this subgroup comprise two homologs of GhbHLH104 and ten homologs of GhbHLH115 (Supplemental Figure S4), with expression of GhbHLH104 being significantly induced by iron deficiency in cotton seedlings (Table 1). There are conflicting reports regarding the transcriptional activity of bHLH12 under iron deficiency in Arabidopsis [45][46][47]. Our study confirmed that GhbHLH121 is induced under iron deficiency in cotton seedlings ( Figure 2C). In general, bHLH transcription factors have been reported to play important roles in the regulation of plant iron homeostasis [50,51], and our results confirmed that cotton also uses members of this family to regulate iron homeostasis. However, the observed transcriptional differences indicate that cotton has formed an iron deficiency regulation network with its own distinct characteristics.
Few studies have reported on the regulation of response to iron deficiency in cotton [52]. In the present work, it was found that GhbHLH121, as an upstream iron deficiencyresponsive protein, forms heterodimers with GhbHLH104, GhbHLH115, and GhPYE ( Figure 5) that act to increase the expression of iron-responsive genes involved in regulation of cotton iron homeostasis ( Table 1). The 35S Pro :GhbHLH121:GFP signal was initially detected in both nucleus and cytoplasm ( Figure 2E), but the presence of GhbHLH104, GhbHLH115, or GhPYE caused it to localize only to the nucleus ( Figure 5C). GhbHLH121 does not have a transmembrane signal peptide, but does co-localize with GhGCN5 and GhNRT in the cytoplasm (Supplemental Figure S14). The post-transcriptional regulation of GhbHLH121 remains to be further explored. All told, this and prior evidence indicate that bHLH121 is a positive iron regulator in both Arabidopsis [45][46][47] and cotton.
While both cotton and Arabidopsis are dicotyledonous plants, we observed some notable differences in their respective iron regulatory networks (Figure 8, Supplemental Figure S15). Under +Fe conditions, inhibiting the expression of GhbHLH121 decreased expression of GhbHLH38, GhFIT, GhPYE, and GhBTS. Meanwhile, under −Fe conditions, the genes GhbHLH38, GhFIT, GhPYE, GhBTS, GhIRT3, GhFRO2, GhAHA2, GhNAS4, and GhYSL1 all exhibited significantly lower expression than in controls, but no significant difference was observed in the expression of GhbHLH101 (Table 1). This suggests that in cotton, GhbHLH101 is not involved in the regulation of iron homeostasis by GhbHLH121, which is different from prior reports in Arabidopsis [45][46][47]. We also found GhbHLH121 to activate transcription of GhFIT, GhbHLH38, and GhPYE ( Figure 7A-E), which is not consistently attested in Arabidopsis [45][46][47]. Gao et al. and Kim et al. found that bHLH121 did not bind the promoter of FIT, but Lei et al. reported bHLH121 to bind the FIT promoter and interact with bHLH IVc TFs, which may activate the binding on the FIT promoter [45][46][47]. However, bHLH121 alone had no notable effect on FIT and bHLH38 promoter activity, while bHLH104 and bHLH115 each induced significant activation [46]. In addition, bHLH IVc TFs were reported to directly bind the bHLH121 promoter [46], which is different to our results. Here we found that GhbHLH121 directly binds to the promoters of GhbHLH38, GhFIT, and GhPYE and activates their transcription independent of GhbHLH104 and GhbHLH115 ( Figure 7A-E). GhbHLH104 and GhbHLH115 were both induced by iron deficiency, but their expression was not associated with GhbHLH121 ( Figure 7F-H and Figure 8).
Ultimately, our work revealed that GhbHLH121 has a positive regulatory role in the iron homeostasis network ( Figure 8). As an upstream iron deficiency-responsive protein, GhbHLH121 forms a heterodimer with GhbHLH104, GhbHLH115, and GhPYE. It also activates the expression of GhbHLH38, GhFIT, and GhPYE independent of GhbHLH104 and GhbHLH115 and regulates the expression of Fe uptake genes (such as GhIRT3, GhAHA2, and GhFRO2) and Fe transport genes (such as GhYSL1 and GhNAS4).

Plant Materials and Growth Conditions
Cotton plants (Gossypium hirsutum, accession Texas Marker-1 [TM-1]) were grown in a growth chamber with a 16/8 h light/dark cycle at 28 • C/25 • C and 70% relative humidity. Cotton seeds were sterilized and germinated by the paper towel method. In brief, cotton seeds were sterilized with 3% H 2 O 2 for 30 min, rinsed 4-5 times with ddH 2 O, and then laid flat on soaked germinating paper. The germinating paper was rolled into a tube and placed in a germinating bag, which was then placed vertically in an incubator for four days. Afterwards, the seedlings were transferred to a hydroponic bucket filled with water for three days and then incubated with 1 / 2 Hoagland nutrient solution with 75 µM Fe 2+ -EDTA (+Fe) or without it (−Fe) for two weeks.
Arabidopsis thaliana ecotype Columbia (Col-0) seedlings were grown in a growth chamber with a 16/8 h light/dark cycle at 22 • C and 70% relative humidity. Briefly, seeds were sterilized and transferred to plates with half-strength Murashige and Skoog (MS) medium. After stratification at 4 • C for two days, the plates were transferred to the growth chamber. Fe-sufficient medium (+Fe) consisted of 1 / 2 MS medium with 1% (w/v) sucrose, 0.8% (w/v) agar, and 75 µM Fe 2+ -EDTA, while Fe-deficient medium (−Fe) consisted of the same ingredients without Fe 2+ -EDTA.

Construction of Materials for Virus-Induced Gene Silencing (VIGS-GhbHLH121) in TM-1
VIGS was performed following a previously described protocol [72]. The 1-300 bp nucleotide sequence of GhbHLH121 was amplified by specific primers (Supplemental Table  S4) from TM-1 cDNA and cloned into a pTRV2 vector digested with EcoRI and BamHI. The resulting construct was transformed into Agrobacterium tumefaciens GV3101 to obtain the TRV2:GhbHLH121 bacterial solution. Cotton seeds were sterilized and germinated by the paper towel method, transferred to a hydroponic bucket filled with water for three days, and finally inoculated with TRV2:GhbHLH121, TRV2:CLA (positive control), or TRV2:00 (negative control) bacterial solution.

Generation of Transgenic Arabidopsis
The full-length coding sequence of GhbHLH121 was amplified by specific primers (Supplemental Table S4) from TM-1 cDNA and fused with GUS to generate 35S:GhbHLH121:GUS vectors (GhbHLH121-OE). The promoter of GhbHLH121 was amplified by specific primers (Supplemental Table S4) from TM-1 DNA and inserted into the pBI121 vector digested with HindIII and XbaI to generate GhbHLH121pro:GUS vectors. The constructs were then transformed into Agrobacterium tumefaciens strain GV3101 and injected into Col-0 Arabidopsis by the floral dip method. Positive transgenic plants were selected on 1 / 2 MS medium containing 50 mg/mL kanamycin. T3 plants were used for phenotype analysis. Primers used for the vector construction are listed in Supplemental Table S4.

Real-Time Quantitative Reverse Transcription PCR (qRT-PCR)
Total RNA was extracted by means of the Biospin Plant Total RNA Extraction Kit (BioFlux, Cluj, Romania). The RNA was then used for the synthesis of first-strand cDNA with the SuperScript III reverse transcriptase kit (Invitrogen, Waltham, MA, USA). Afterwards, qRT-PCR was performed using the Light Cycler FastStart DNA Master SYBR Green I Kit (Roche, Basel, Switzerland) according to the manufacturer's instructions (Bio-Rad, Hercules, CA, USA). Means and standard deviations were calculated from three biological replicates. Primers used for the qRT-PCR are listed in Supplemental Table S5.

Phylogenetic Analysis
For phylogenetic analysis, protein sequences were retrieved from TAIR and COT-TONOMICS. Sequences of bHLH transcription factors involved in Fe homeostasis were aligned using ClustalX 1.5. The phylogenetic tree included multiple plant species and was constructed using MEGA 5 with the NJ method, and the selected model was the Poisson model with 1000 bootstrap replications [73]. We used the ITOL9 web server to beautify the resulting phylogenetic tree. After growing cotton seedlings in +Fe or −Fe solution, the penultimate leaf was placed in the 3 cm leaf chamber of an LI-6400 portable photosynthesizer (LI-6400XT; Li-Cor, Darmstadt, Germany) for evaluation of photosynthetic parameters. Photosynthesis rate, stomatal conductance, transpiration rate, and intercellular CO 2 concentration were determined simultaneously. Conditions in the measurement chamber consisted of a flow rate of 400 mol/s, CO 2 concentration in the sample chamber of 400 mmol/mol, and relative humidity of 60%. Means and standard deviations were calculated from three biological replicates.

Fe Concentration Measurement
To determine Fe concentration, samples (Arabidopsis seedlings or cotton leaves) were harvested and dried at 65 • C for three days. About 100 mg dry weight was mixed with 5 mL of 1 M HNO 3 and 1 mL of H 2 O 2 for 20 min at 220 • C. Each sample was then diluted to 16 mL with ultrapure water and the Fe concentration determined using a Thermo SCIENTIFIC ICP-MS (iCAP6300, Waltham, MA, USA). Means and standard deviations were calculated from three biological replicates.

Ferric Chelate Reductase Assays
To determine ferric chelate reductase activity (FCR), cotton leaves were harvested and placed in assay solution containing 0.1 mM Fe 3+ -EDTA and 0.3 mM ferrozine in the dark for 20 min. An identical assay solution without the addition of plant tissue was used as a blank. After incubation, the absorbance of the purple-colored Fe(II)−ferrozine complex was measured at 562 nm and concentration calculated using a molar extinction coefficient of 28.6 mM −1 cm −1 [74]. Means and standard deviations were determined from three biological replicates.

NBT Staining and DAB Staining
Cotton leaves were put into a tube with NET staining solution (G1023, 100 mL, Solarbio, Beijing, China) or DAB staining solution (G1022, 100 mL, Solarbio, Beijing, China), with the leaves being completely immersed in the solution. Leaves were subjected to 15 min of vacuumization and then incubated overnight at room temperature. The next day, the leaves were decolorized with absolute ethanol for 10 min, 95% for 10 min, and then 70% until completely decolorized.

Measurements of SOD, POD, and CAT Activity
To determine physiological parameters of oxidative defense, 0.1 g cotton penultimate leaf was used in the SOD Quantification Assay Kit (BC0170, Solarbio, Beijing, China), POD Quantification Assay Kit (BC0095, Solarbio, Beijing, China), and CAT Quantification Assay Kit (BC0205, Solarbio, Beijing, China) as described by the manufacturer. Means and standard deviations were calculated from three biological replicates.

Subcellular Localization
To produce the constructs 35S Pro :GhbHLH121:GFP, 35S Pro : GhNRT:GFP, and 35S Pro : GhGCN5:GFP, full-length coding sequences of GhbHLH121, GhNRT, and GhGCN5 were inserted into the pBINGFP4 vector. The constructs were transformed into Agrobacterium tumefaciens strain GV3101 and subsequently injected into three-week-old Nicotiana benthamiana leaves. After cultivation in the dark for two days, fluorescence in the epidermal cells was visualized on a confocal microscope (Zeiss LSM780, Oberkochen, Germany) with the excitation laser at 488 nm. The primers used for vector construction are listed in Supplemental Table S4. For each combination, three biological replicates were performed.

Transmission Electron Microscopy
For TEM analysis of treated plants, cotton penultimate leaves were immersed in a solution containing 2.5% glutaraldehyde. The leaves were fixed with osmic acid for two hours, rinsed with phosphate buffer solution, and then subjected to ethanol gradient dehydration. After embedding with an embedding agent overnight, thin sections were prepared with an ultrathin microtome, stained with uranyl acetate and lead citrate, and examined using a Hitachi H-7650 transmission electron microscope at 80 kV. For the yeast two-hybrid screening, the full-length coding sequence of GhbHLH121 was inserted into pGBKT7 as the bait vector and transformed into yeast strain Y2H to produce the bait strain. Cotton leaf and root cDNA libraries cloned into the pGADT7-Rec vector were used to produce the bait strain with the Make Your Own "Mate & Plate" Library System (Cat. No. 630490, TaKaRa, Dalian, China). The library strain was then mated with the bait strain, and the mating mixture was spread onto SD/-Trp/-Leu/-His medium and incubated at 30 • C for 8-10 days. Positive clones were amplified by Matchmaker Insert Check PCR Mix 2 (Cat. No. 630497, TaKaRa, Dalian, China).
For the yeast two-hybrid assay, the full-length coding sequence of GhbHLH121 was inserted into pGBKT7 as the bait vector, and the full-length coding sequences of GhbHLH104, GhbHLH115, and GhPYE were inserted into pGADT7 to produce the prey vectors. The bait vector and each prey vector were introduced together into Y2HGlod, and the transformed cells were spread onto SD/-Trp/-Leu/-His medium and incubated at 30 • C for 2-3 days. The combination of pGBKT7-53 and pGADT7-T was used as a positive control and the combination of pGBKT7-Lam and pGADT7-T as a negative control (Matchmaker Gold Yeast Two-Hybrid System, Cat. No. 630489, TaKaRa, Dalian, China). Primers used for the vector construction are listed in Supplemental Table S4. For each combination, three biological replicates were performed.

Bimolecular Fluorescence Complementation
The full-length coding sequence of GhbHLH121 was inserted adjacent to the sequence for the N-terminal fragment of YFP, denoted as YNE-GhbHLH121. The full-length coding sequences of GhbHLH104, GhbHLH115, and GhPYE were similarly inserted alongside the sequence for the C-terminal fragment of YFP, and the resulting constructs were denoted 35S Pro :GhbHLH104:YCE, 35S Pro :GhbHLH115:YCE, 35S Pro :GhPYE:YCE, respectively. The constructs were transformed into A. tumefaciens strain GV3101 and bacteria harboring the effectors and reporters were mixed in a 1:1 volume ratio. The obtained mixtures were then injected into the abaxial sides of three-week-old N. benthamiana leaves. After two days of cultivation in the dark, epidermal cells were imaged on a confocal microscope (Zeiss LSM780, Oberkochen, Germany), with the YFP signal elicited by a 514 nm excitation laser and recorded. Primers used for the vector construction are listed in Supplemental Table S4. For each combination, three biological replicates were evaluated.

Luciferase Complementation Imaging Assay
The full-length coding sequence of GhbHLH121 was inserted alongside the sequence encoding the C-terminal fragment of luciferase, producing the construct cLUC-GhbHLH121. The full-length coding sequences of GhbHLH104, GhbHLH115, and GhPYE were likewise inserted alongside that encoding the N-terminal fragment of luciferase, which constructs were denoted as GhbHLH104-nLUC, GhbHLH115-nLUC, and GhPYE-nLUC, respectively. The constructs were transformed into A. tumefaciens strain GV3101 and bacteria harboring the effectors and reporters were mixed in a 1:1 volume ratio. The resulting mixtures were then injected into the abaxial sides of three-week-old N. benthamiana leaves. After two days of cultivation in the dark, Luciferin (100 µM) spray (Sigma, 103404-75-7, St. Louis, MO, USA) was smeared on the backs of the leaves and luciferase signals were obtained by a low-light cooled charge-coupled device imaging apparatus (Tanon 5200, Shanghai, China). The positive controls were FTL9 and FD1 [76]. Primers used for the vector construction are listed in Supplemental Table S4. For each combination, three biological replicates were assessed.

4.9.
Validation of DNA-Protein Interactions 4.9.1. Yeast One-Hybrid Assay E-box-containing fragments of the promoters for GhbHLH38, GhFIT, and GhPYE were inserted into vectors to produce bait vectors and an mE-box containing a mutation of CA to TC [58], which were then digested by BstBI and the linearized plasmids transferred into yeast strain Y1H gold to yield the pBait-AbAi bait strain. Yeasts were screened according to the concentration of AbAi. The full-length coding sequence of GhbHLH121 was inserted into the pGBKT7 vector to construct the prey vector, which was then introduced into the pBait-pAbAi strain and the bacteria spread on SD/-Leu/AbA and incubated at 30 • C for two to three days. The yeast one-hybrid assay was then conducted using the Matchmaker Gold Yeast One-Hybrid Library Screening System kit (Cat. No. 630491, TaKaRa, Dalian, China). For each combination, three biological replicates were performed.

Luciferase Assay
The promoters of GhbHLH38, GhFIT, and GhPYE were inserted into the pGreen II 0800-LUC vector to create reporter constructs. The full-length coding sequence of GhbHLH121 was inserted into the pGreen II 62SK vector as the effector. The constructs were transformed into A. tumefaciens strain GV3101 and bacteria harboring the effectors and reporters were mixed in a 1:1 volume ratio. The resulting mixtures were then injected into the abaxial sides of three-week-old N. benthamiana leaves. After two days of cultivation in the dark, luciferase signals were obtained by observing leaves coated with Luciferin (100 µM) spray (Sigma, 103404-75-7, St. Louis, MO, USA) under a low-light cooled charge-coupled device imaging apparatus (Tanon 5200, Shanghai, China). The LUC/REN ratio was then determined using a dual-luciferase reporter gene analysis system (E2490, Promega, Madison, WI, USA). For each combination, three biological replicates were evaluated. Primers used for the vector construction are listed in Supplemental Table S4. Means and standard deviations were calculated from three biological replicates.