A G. hirsutum cultivar, Xuzhou 142, was planted in the greenhouse with a 16 h light, 30°C/8 h dark, 30°C cycle, as previously reported (He et al., 2017). For phytohormone treatment, 0 DPA fresh ovules were collected from cotton bolls, sterilized, and cultured in previously reported liquid culture medium (Shi et al., 2006), which added with 5 μM 1-Naphthylacetic acid (NAA, Sigma) and 1 μM gibberellin acid (GA3, Sigma) for the indicated time (He et al., 2019), respectively. After treatment, the ovules were collected for quantitative real-time (qRT-PCR) experiments. For RNA extraction, fresh cotton seed fibers were harvested from 0, 5, 10, 15, 20, and 25 DPA, and then immediately frozen in liquid nitrogen.

The genome sequences of G. hirsutum L. genome (NDM8), G. raimondii (JGI_v2.1), G. arboreum (CRI_v3.0), G. anomalum (NSF_v1), G. stocksii (NSF_v1), G. longicalyx (NSF_v1), and G. rotundifolium (HAU v1) were downloaded from CottonGen¹ (Ma et al., 2021). The genome sequence of Arabidopsis thaliana was downloaded from The Arabidopsis Information Resource (TAIR²) database (Lamesch et al., 2012). The HXK sequences from O. sativa (Cho et al., 2006), Phyllostachys edulis (Moso Bamboo) (Zheng et al., 2020), and Manihot esculenta (Cassava) (Geng et al., 2017) were downloaded from the Nucleotide database.³ The genome sequence (Chalhoub et al., 2014) of Brassica napus was downloaded from the Brassicaceae Database (BRAD⁴). The genome size, sequences and taxonomy ID of Ostreococcus lucimarinus, Chlamydomonas reinhardtii, Volvox carteri, Coccomyxa subellipsoidea, Chlorella variabilis, and Selaginella moellendorffii were downloaded from the Genome database of NCBI.⁵

Two HXK Pfam domains (PF03727 and PF00349) were used to search against the G. hirsutum L., G. raimondii, G. arboreum, G. anomalum, G. stocksii, G. longicalyx, and G. rotundifolium genomes using the hidden Markov model (HMM) with HMMER 3.0 (Prakash et al., 2017). The candidate GhHXKs, GaHXKs, GrHXKs, GanHXKs, GstHXKs, GloHXKs, and GroHXKs were submitted to the SMART software (Letunic et al., 2021)⁶ and the Conserved Domain Database (Lu et al., 2020) (CDD⁷) to confirm that all candidate HXK proteins contained the Hexokinase domain.

We used the general feature format (GFF) file of the genomes to determine the relative position of HXKs on chromosomes, and visualized the locations with the online software MG2C (Jiangtao et al., 2015). Furthermore, the gene structures of HXKs were also analyzed according to the GFF files, and the “exon-intron” structure was shown by the Gene Structure Display Server (Hu et al., 2015) (GSDS 2.0⁸).

Protein motif analysis was performed using MEME⁹ with a maximum of eight motifs and using other default parameters.

The physicochemical properties, including molecular weight (MW), isoelectric point (pI), instability index, and grand average of hydropathicity (GRAVY), were analyzed using the online software ExPASy ProtParam tool (Artimo et al., 2012)¹⁰ in GhHXKs, GaHXKs, GrHXKs, GanHXKs, GstHXKs, GloHXKs, and GroHXKs, respectively.

The subcellular localization of the candidate HXKs were predicted by the online software, WoLF PSORT (Horton et al., 2007).¹¹

The HXK protein sequences of G. hirsutum L., G. raimondii, G. arboreum, A. thaliana, O. sativa, P. edulis, M. esculenta, and B. napus were aligned using ClustalW, and the evolutionary tree was constructed using the neighbor-joining method with MEAG 7.0 (Kumar et al., 2016). To evaluate the reliability of the phylogenetic tree, the bootstrap value was set as 1,000.

The duplication types of GhHXKs, GaHXKs, and GrHXKs were analyzed using the Multiple Collinearity Scan (MCScanX) toolkit under the Linux system (Yupeng et al., 2012). The orthologous- and homologous-gene pairs were visualized by the CIRCOS software (Krzywinski et al., 2009). The synonymous substitution rate (Ks), non-synonymous substitution rate (Ka), and Ka/Ks ratios were calculated using the KaKs_Calculator software (Wang et al., 2010). The divergence time between the homologous- and orthologous–gene pairs was calculated according to previously used methods (Yang et al., 2006).

The sequence 2,000 bp upstream of the initiation codon was extracted as the candidate promoters with the “fastacmd –d database –s chromosome –L start location, end location –o result” using the local BLAST software (Camacho et al., 2009). The Cis-elements in the candidate promoter sequence were analyzed by Plant Cis-acting Regulatory Element (Plant CARE¹²) (Lescot et al., 2002).

The allotetraploid cotton cultivar, Xuzhou 142, was grown in Shaanxi Normal University under controlled conditions (He et al., 2017). A total of 30 ovules were used for each phytohormone and were performed in triplicate. Cotton ovules were collected at one DPA, sterilized with sodium hypochlorite (NaClO, 10%), and cultivated as previously reported (Shi et al., 2006). Five μM 1-Naphthylacetic acid (NAA, Sigma, Germany) and 1 μM GA₃ (Sigma, Germany) were added to the culture medium. The ovules treated with phytohormones were used to perform RNA-seq, while the data was conserved in our lab (He et al., 2019).

To illustrate the spatial and temporal expression patterns of GhHXKs, the transcriptomes of various tissues (stamen, anther, seed, fiber, ovule, petal, calycle, torus, leaf, stem, root, cotyledon, stigma, and pistil) and a successive fiber developmental stages (0, 5, 10, 15, 20, 25, 30, and 35 DPA) were downloaded from NCBI (accession NO. PRJNA680449) (Ma et al., 2021). The expression data were normalized and visualized using Omicshare tools.¹³

The total RNA extraction was performed according the instructions for the RNAprep Pure Plant Plus Kit (Code No. DP441, TIANGEN, China), and the cDNA was reverse-transcribed from 2 μg total RNA (Xiao et al., 2016). The qRT-PCR was conducted with three biological and three technical replicates as the following reaction parameters: 95°C for 30 s, followed by 40 cycles of 95°C for 5 s, 60°C for 15s, and 72°C for 20 s. A melting curve was generated from 65 to 95°C. The ubiquitin gene GhUBQ7 (GenBank accession no. AY189972) was used as the internal control for each qPCR experiment. Primers for qRT-PCR experiments were listed in Supplementary Table 1.

To identify HXK genes in cotton species, two hexokinase domains (PF03727 and PF00349) were used as the query domains with the HMMER 3.0 software (on a Windows system) to search against the genomes of G. hirsutum L. (NDM8), G. arboreum (CRI_v3.0), G. raimondii (JGI_v2.1), G. anomalum (NSF_v1), G. stocksii (NSF_v1), G. longicalyx (NSF_v1), and G. rotundifolium (HAU_v1). There were 17 GhHXKs, 9 GaHXKs, 8 GrHXKs, 8 GanHXKs, 8 GstHXKs, 7 GloHXKs, and 8 GroHXKs retrieved from G. hirsutum L., G. arboreum, G. raimondii, G. anomalum, G. stocksii, G. longicalyx, and G. rotundifolium, respectively (Table 1 and Supplementary Table 2).

The amino acids ranged from 137 (GhHXK14) to 504 (GhHXK8 and GhHXK17). The molecular weight of GhHXKs ranged from 15.07 kDa (GhHXK14) to 54.94 kDa (GhHXK17). According to isoelectric point (pI) analysis, 14 GhHXKs had pI less than 7.0 (with an average of 6.15) and were acidic proteins. In contrast, three GhHXKs were predicted to encode proteins more than 7.0 (average of 7.15) and were basic. Grand average of hydropathicity (GRAVY) analysis found that 12 GhHXKs with GRAVY scores less than zero were hydrophilic proteins; and that five GhHXKs with GRAVY scores more than zero were hydrophobic proteins. Based on the instability index analysis, 14 GhHXK proteins have instability index values less than 40.0 and three GhHXK proteins have instability index values greater than 40.0 (GhHXK2, GhHXK8, and GhHXK17). The detailed physicochemical properties of GaHXKs, GrHXKs, GanHXKs, GstHXKs, GloHXKs, and GroHXKs are listed in Supplementary Table 2.

According to the GFF files of G. hirsutum L. (NDM8), the 17 GhHXKs are distributed on 12 G. hirsutum L. chromosomes. The 17 GhHXKs genes were named GhHXK1 to GhHXK17 from chromosomes At01 to Dt13 based on their relative chromosomal locations from the chromosome top to bottom (Figures 1, 2). There are eight GhHXKs distributed on six At_subgenomes (At05, At06, At09, At10, At11, and At13) and nine GhHXKs distributed on six Dt_subgenomes (Dt05, Dt06, Dt09, Dt10, Dt11, and Dt13). The GhHXK genes are evenly distributed on At_ and Dt_subgenomes, except for GhHXK14. Nine GaHXKs are distributed on seven G. arboreum genomes. The distribution of GaHXK genes across chromosomes was similar to that of GhHXKs on At_subgenome in G. hirsutum L., while there was an extra on Ga_Chr02 of G. arboreum. Eight GrHXKs were distributed on six chromosomes, which was similar to the distribution of GhHXKs on the Dt_subgenome in G. hirsutum L. At the same time, there is one more gene distributed on the chromosome Gr_Chr07 and one lost gene on the chromosome Gr_Chr05 in G. raimondii (Supplementary Figure 1). This indicated that their gene loss or duplicated evince existed in the G. hirsutum L. genome.

Furthermore, combined with GFF annotation files for other cotton species, the eight GanHXKs were distributed on six G. anomalum (2n = 2X = 26, BB) chromosomes, including B05, B06, B09, B10, B11, and B13, while there were eight GstHXKs distributed on six G. stocksii (2n = 2X = 26, EE) chromosomes, including E05, E06, E09, E10, E11, and E13. The seven GloHXKs were distributed on the identical chromosomes of G. longicalyx (2n = 2X = 26, FF) and were also distributed across the identical chromosomes of G. rotundifolium (2n = 2X = 26, KK), except for K05 (Figure 2). The distribution analysis demonstrated that the HXKs were conservatively distributed on the 5th, 6th, 9th, 10th, 11th, and 13th chromosomes among cotton species. According to the detailed distribution and transcription direction of HXKs between cotton species (Figure 2 and Supplementary Figures 1, 2), inversion and segmental duplication existed in the chromosomes of these cotton species’.

In general, nucleic acid sequences are more variable than protein sequences. To well illustrate the evolutionary relationships among GhHXKs, a CDS phylogenetic tree was constructed by MEGA 7.0 (Figure 3A). According to the phylogenetic tree in Figure 3A, the GhHXK genes could be clustered into three groups.

The gene structures of GhHXKs were determined by assessing the annotation information of the GFF files in the G. hirsutum L. genome (NDM8), which were visualized using the GSDS 2.0 online software. The results demonstrated that most GhHXKs contained nine exons and eight introns, and four GhHXKs contained eight exons and seven introns, including GhHXK1, GhHXK9, GhHXK11, and GhHXK13. GhHXK14 contained two exons and one intron. The 13 GhHXKs genes include both 5′- and 3′-UTRs, two GhHXKs contain 3′-UTR, while the remaining two genes (GhHXK13 and GhHXK14) have no UTR region (Figure 3B). Studies assessing the gene structure of other cotton species (Supplementary Figure 3B) demonstrated that most HXKs in cotton species are conservative and have “intron-exon” structures.

To illustrate the phylogenetic relationships between the HXKs proteins in G. hirsutum L. and those of other species, including G. arboreum, G. raimondii, O. sativa, A. thaliana, P. edulis, M. esculenta, and B. napus, an unrooted neighbor-joining tree was created using the MEGA 7.0 software based on their entire length of the amino acid sequences.

According to the phylogenetic tree of HXKs from multiple species (Figure 4 and Supplementary Figure 4), the HXKs protein sequences were divided into four groups (Clade I, II, III, and IV). Two GhHXKs were classed into Clade I, four GhHXKs was classed into Clade II, 11 GhHXKs were grouped in Clade IV. However, no GhHXK were grouped in Clade III; only HXKs from monocotyledons O. Sativa and P. edulis were grouped into Clade III.

The protein sequence of GhHXKs was aligned using ClustalW software to characterize the protein structures. The amino acid sequence alignment showed 39–99% identity between GhHXKs members (Supplementary Figure 4), based on previous work analyzing HXK proteins in A. thaliana, O. sativa, P. edulis, M. esculenta, and B. napus (Cho et al., 2006; Geng et al., 2017; Zheng et al., 2020). The adenosine phosphate binding domain (Supplementary Table 3) and glucose-binding domain were found in most GhHXKs (Supplementary Figure 4 and Supplementary Table 4). The core glucose-binding domain of GhHXKs was conservative as “I/L-GFT-F/V-S-F/S-P/G-V/D” (Figure 5A). There is no glucose-binding domain in GhHXK14 (Supplementary Figure 4), while an intact adenosine phosphate binding domain exists in GhHXK14 (Figure 5B). The adenosine phosphate binding domain has a conserved motif of “RX₂R-V/L-X₃GX₃-I/L/V” in GhHXKs, except for GhHXK9, GhHXK11, and GhHXK13 (Figure 5B). Sequences alignment showed that most of the GhHXKs are conservative with adenosine phosphate and glucose-binding domain.

Furthermore, the GhHXK protein motif characteristics were analyzed using the MEME online software, and ten conservative motifs were identified in the GhHXK gene family (Figure 5C). The majority of GhHXK proteins contain at least eight motifs, except for GhHXK11 and GhHXK14, which have seven and one motifs, respectively.

By searching the HXK domain against the genomes from chlorophyta to lycophytes plant species (Supplementary Figure 5), we found that HXK gene family members increased from low to high plant species. The chlorophyta species have less than two HXKs; however, lycophytes plants have more HXKs numbers than five.

Therefore, to illustrate the duplication events of the HXKs gene on chromosome segments, the evolution of GhHXK genes was analyzed in G. hirsutum L., G. arboreum and G. raimondii, respectively, using MCScanX software. The results demonstrated that 16 GhHXK genes were derived from segmental duplication (accounting for 94.12%) of GhHXK gene family members, while only GhHXK14 was derived from dispersed distribution on the chromosomes. Five GaHXKs were derived from segmental duplication, accounting for 55.56% of the total gene family members. Three GaHXKs were derived from dispersed distribution (accounting for 33.33%), and one GaHXK was a singleton gene. There are four GrHXKs derived from segmental duplication (accounting for 50%), three GrHXKs derived from tandem duplication events, and only one derived from dispersed distribution (Figure 6 and Supplementary Table 5). Duplication analysis demonstrated that segmental duplication is the leading cause of HXK genes duplication in cotton species.

Selection pressure refers to the evolutionary force of natural selection, which dictates the survival and reproduction of adaptive organisms. We further analyzed the Ka, Ks, and Ka/Ks ratios of the orthologous gene pairs in G. hirsutum L., paralogous gene pairs between G. arboreum and G. hirsutum L., and G. raimondii and G. hirsutum L. (Supplementary Table 6). The Ka/Ks ratios for the GhHXKs versus GaHXKs orthologous pairs ranged from 0.0784 to 0.819 and the Ks ranged from 0.00571 to 0.0510, suggesting that the orthologous pairs diverged 1.10 million years ago (MYA). The Ka/Ks ratios for the GhHXKs versus GrHXKs orthologous pairs ranged from zero to 1.917 and the Ks ranged from 0.0055 to 0.0317, suggesting that the orthologous pairs diverged from 1.06 MYA.

We further analyzed the cis-regulatory elements in the promoter regions of GhHXKs. The cis-acting elements that we identified in GhHXKs promoters were classified into three categories, including light-, hormone- and abiotic stress-responsive promoters (Figure 7). Light-responsive elements were identified in all GhHXKs’ promoters. Of them, G-box was the most abundant (54) and was found in the promoters of 17 GhHXKs. Analysis of hormone-related elements demonstrated that the number of abscisic acids (ABA)-responsive elements were highest (43), followed by methyl jasmonate (MeJA)-responsive elements (32). Except for GhHXK9, all GhHXK promoters contain ABA-responsive elements (ABRE). Cis-acting elements involved in MeJA (TGACG-motif and CGTCA motif) were found in the promoters of 12 GhHXKs. The promoters of 11, 6, and 3 GhHXKs contain SA-, GA-, and Auxin-responsive elements, respectively. Additionally, all GhHXK promoters had at least two hormone-responsive elements, and GhHXK14 contained all five hormone-responsive elements.

To illustrate the spatial expression patterns of GhHXKs genes, we analyzed the transcriptomes of various tissues (stamen, anther, seed, fiber, ovule, petal, calycle, torus, leaf, stem, root, cotyledon, stigma, and pistil) in G. hirsutum L. Transcripts of GhHXKs were detected in all tissues (Supplementary Figure 6), while their expressions exhibit a tissue-specific expression pattern in G. hirsutum L.

Gossypium hirsutum L. is one of the most important textile crops in the world. Considering its importance, we investigated the expression profiles of GhHXK genes during the fiber developmental stages at 0, 5, 10, 15, 20, 25, 30, and 35 DPA (Figure 8A). According to the expression patterns of GhHXKs during the fiber development process, GhHXKs expression patterns were classified into three groups: (i) secondary cell wall synthesis, where GhHXK6, GhHXK7, GhHXK11, GhHXK15, and GhHXK16 were highly expressed 20–45 DPA, the; (ii) the elongation process during fiber development, where GhHXK4, GhHXK7, GhHXK10, GhHXK12, and GhHXK16 had higher expression levels in fibers from 10 DPA to 20 DPA; and (iii) the fiber initiation and elongation process, where GhHXK1, GhHXK2, GhHXK3, GhHXK5, GhHXK8, GhHXK9, GhHXK14, and GhHXK17 were highly expressed at 0 and 5 DPA. These transcriptome data were also verified by qRT-PCR experiments in Figure 9. GhHXKs have similar expression pattern both in qRT-PCR experiments and transcriptome data during fiber developmental stages.

The promoters of GhHXK1, GhHXK4, GhHXK5, GhHXK9, GhHXK10, and GhHXK14 contain GA-responsive cis-elements, while the promoters of GhHXK2, GhHXK14, and GhHXK17 have auxin-responsive cis-elements that are highly expressed from 5 to 20 DPA (Figure 8B). These genes can also be induced by GA and auxin treatment (Figure 10). Our results indicated that GhHXKs are involved in regulating the fiber development process and that the promoters of GhHXKs (i and ii) contain auxin- and GA-responsive elements.

Gossypium hirsutum L. (2n = 4X = 52) is an allotetraploid cotton species. It originated approximately 1–2 MYA after the hybridization of two diploid cotton species, G. arboreum (2n = 2X = 26) and G. raimondii (2n = 2X = 26) (Galau and Wilkins, 1989). In this work, we identified 17, 9, and 8 HXKs from G. hirsutum L., G. arboreum, and G. raimondii, respectively. The total numbers of GaHXK and GrHXK equal that in G. hirsutum L.

Duplication events are the primary reason for the expansion of each gene family member. Analysis of the synteny and phylogeny of HXKs in the G. hirsutum L. genome demonstrated that GhHXKs, GaHXKs, and GrHXKs duplicated due to segmental duplication. The Ks and Ka were more significant in paralogous gene pairs (GhHXKs) than in orthologous gene pairs (GhHXKs vs. GaHXKs and GhHXKs vs. GrHXKs), and in the divergence time in paralogous gene pairs (GhHXKs) than between orthologous gene pairs. This indicates that duplication events in GaHXKs, GrHXKs, and GhHXKs occurred before the divergence of G. raimondii and G. arboreum. Additionally, the HXK sequences were conservative among cotton species.

Hexokinases in higher plants typically contain nine exons, such as PeHXKs (Zheng et al., 2020), OsHXKs (Cho et al., 2006), and MeHXKs (Geng et al., 2017). These nine exons were also found in most GhHXKs. Phylogenetic analysis of HXK proteins found more clade numbers among monocotyledons and fewer clade numbers among dicotyledons. The GhHXKs, GaHXKs, GrHXKs, AtHXKs, MeHXKs, and BnHXKs were clustered into three clades (I, II, and IV), while OsHXKs, PeHXKs were clustered into four clades (I, II, III, and IV). This indicates that the HXKs of monocotyledonous plants had a higher mutation level than that of dicotyledons.

Sucrose is the primary photosynthesis produce and is transported to growing cells, such as fiber cells. Sucrose is a disaccharide made up of glucose and fructose, and functions as an osmotic substance and raw material for the synthesis of cell wall cellulose. Sucrose synthase (SuSy) and invertase are involved in the first step of sucrose degradation by cleaving the glycosidic bond between glucose and fructose (Cabello et al., 2014). SuSy helps break down sucrose into fructose and UDP-glucose for cellulose biosynthesis. A SuSy protein, SusC, is highly expressed during the synthesis of the secondary cell wall in fibers and the cell wall fraction. The subcellular location of the protein demonstrated that SusC is localized on the cell wall, which could indicate the presence of UDP-glucose function in cellulose and callose synthesis (Brill et al., 2011). Jiang et al. (2012) demonstrated that over-expressing GhsusA1, a cotton SuSy gene, increased the thickness of the secondary cell wall and overall fiber strength, which indicates that a sucrose signal is involved in controlling cellulose biosynthesis in the development of cotton fiber (Jiang et al., 2012). When this synthetic SuSy gene is overexpressed in cotton, the transgenic cotton plants showed longer fiber length, enhanced fiber strength, and increased cellulose contents (Ahmed et al., 2020). The cell wall invertase (CWIN) is responsible for sucrose cleaving into fructose and glucose, while the expression levels of GhCWIN are significantly more highly expressed at 5 and 10 DPA than 15 and 20 DPA (Wang and Ruan, 2012). During the fiber development process, the sucrose content decreased from fiber initiation (0 DPA) to fiber elongation (12 DPA), and was accompanied by increasing in glucose and fructose in fiber content (Sun et al., 2019). Both SuSy and CWIN can catalyze sucrose into monosaccharides and contribute to cotton fiber development by providing component hexoses for cellulose synthesis. Additionally, the production of hexoses can increase the content of osmotic substances content, which contributes to turgor pressure for fiber elongation (Weschke et al., 2003).

The synthesis and elongation of cotton fiber cell could be required to obtain higher energy. Studies have found significantly higher ATP synthase activity in 10 DPA wild-type fiber cells than in ovule samples and leaf samples. Additionally, exogenously applying the inhibitors of ATP synthase, piceatannol (PA), and oligomycin (OM) decreased fiber length and lowered the ATP/ADP ratio (Pang et al., 2010). Other studies demonstrated that phosphorylated glucose participated in the pentose phosphate pathway, which provides NADPH for cellular respiration (Lu et al., 2016), and that HXKs catalyzes the irreversible step of glycolysis, which provides energy for cell growth (Aguilera-Alvarado and Sanchez-Nieto, 2017).

Transcriptome analysis demonstrated that the hexokinase inhibitor NAG, which repressed cotton fiber elongation, depends on the glucose signal transduced by HXKs (Li et al., 2021). Glucose phosphorylation is also involved in the synthesis of inositol, a signal molecule (Loewus et al., 1982), which positively regulates cotton fiber length. The inositol synthase enzyme, myo-inositol-1-phosphate synthase, positively regulates fiber elongation. The GhMIPS1D gene was ectopically expressed in the Arabidopsis mips1 mutant showed longer root cells and a higher plant height (Ma et al., 2019). RNAi MdMIPS1/2 in apple promoted programmed cell death and necrosis, while apple necrosis was directly caused by the excessive accumulation of reactive oxygen species. Therefore, apple necrosis could be associated with salicylic acid, which increased the polysaccharide-mediated cell wall (Hu et al., 2020).

Gibberellin and auxin are two plant hormones that promote fiber elongation. Analysis of promoter cis-elements and expression data of expression profile and qRT-PCR experiments demonstrated that some GhHXKs are regulated by GA and IAA.

Gibberellin (GA) plays two roles when regulating the content of intracellular glucose. GA₃ treatment can promote the accumulation of sugar in potato tubers under low-temperature conditions by inducing changes in the expression of genes involved in sugar accumulation, ADP-glucose pyrophosphorylase (AGPase) (Xie et al., 2018). In the presence of glucose, the GA synthesis enzyme, GA20ox1, can significantly up-regulated by KNO₃ (Ikeya et al., 2020).

In the daytime, cytochrome C (Cyt C)-deficient Arabidopsis accumulates glucose with lower levels of GA, while GA treatment complements this reduction of glucose accumulation in Cyt C-deficient plants (Racca et al., 2018). GA synthesis was suppressed by glucose, and the application of mevalonic acid could break down this suppression. Therefore, the key enzyme of the isoprenoid pathway was the target of C-catabolite suppression (Bruckner, 1992). GA₃ repressed the transcriptional levels of HXK1 and HXK2, which negatively interfered with the transduction of glucose signals, depending on hexokinase phosphorylation in grape berries (Zhang et al., 2014).

Abscisic acids negatively regulates cotton fiber development, and other studies demonstrated that ABA crosstalked with glucose signal transduction. During the germination process of rice seeds, high glucose concentrations delayed seed germination by repressing ABA catabolism (Zhu et al., 2009). The enzyme UGT73C14 utilized UDP-glucose as sugar donors for ABA glycosylation in G. hirsutum L., and the UDP-glucose can be synthesized by UTP and phosphorylated glucose (Glc-Pi) (Gilbert et al., 2013).

The glucose sensor HXK mutant gin 2 is also resistant to exogenous auxin (Moore et al., 2003). Glucose affects most of the genes regulated by auxin metabolism (Mishra et al., 2009). High glucose concentrations reduced the root meristem zone by repressing the auxin transporters, PIN1 accumulation, and reducing auxin levels in Arabidopsis roots (Yuan et al., 2014). Most IAA-regulated genes were transcriptionally regulated by glucose alone; however, glucose antagonistically functions on IAA-regulated genes (Gupta et al., 2009).

Above all, glucose functions as a molecular signal that crosstalks with IAA, GA, and ABA, while HXK-catalyzed glucose-phosphate is the core of glucose signal transduction. Analysis of the promoter cis-elements analysis and RNA-seq data demonstrated that GhHXKs contain GA- and IAA-related cis-elements can also be regulated by these phytohormones. This indicates that various hormones can crosstalk HXKs with sugar signals when regulating the development of cotton fiber (Figure 11).