In this study, four G. hirsutum varieties, including TM-1, 99668, J14, and ZM24, were chosen, representing null, low, and high conversion rates of callus into EC, respectively. Seeds were surface sterilized with 70% ethanol for 1 min, rinsed at least three times with sterile water, and imbibed in 30% (v/v) H₂O₂ for 24 h. The sterilized seeds were germinated on MS medium (Murashige and Skoog, 1962) in the dark at 28°C for 5 days. Hypocotyls were dissected from the seedlings and cut into 5–7 mm segments and cultivated on callus-induction medium as specified in the “Uniconazole and ABA treatments during callus induction and differentiation stages” section (Zheng et al., 2014).

To determine the effect of uniconazole on callus growth and the conversion rate of callus into EC, the hypocotyl explants were cultured on callus-induction medium (MS salts, B5 vitamins, 0.10 mg/L 3-IAA, 0.10 mg/L kinetin, 0.10 mg/L 2,4-D, 25 g/L glucose, and 2 g/L Gelrite^® at pH 5.8) supplemented with 7 μM uniconazole under standard conditions (28°C and 14 h L/h D) for 2 months. To investigate the interactive effects of exogenous 0.04 μM ABA and 7 μM uniconazole (Sigma. Darmstadt, Germany) on SE, the explants of ZM24 were transferred onto freshly prepared callus-induction medium supplemented with appropriate amounts of ABA and uniconazole.

Synthetic auxin response element (DR5) was cloned into the vector pCambia2300 replacing 35 promoters and fused to a GFP reporter gene, constructing the vector DR5::GFP. The vector was transformed into G. hirsutum ZM24 plants via Agrobacterium tumefaciens (LB4404) as described earlier (Zhan et al., 2021).

Callus proliferation rate (CPR) was calculated as the fold change in weight gained of explants at 15 and 25 days of induction as described by Wang et al. (2017). The embryonic differentiation rate (EDR) was calculated by the percentage of explants with embryonic callus at different time points on the callus-induction medium or those supplemented with uniconazole and ABA.

For scanning electron observation, callus was placed on a fixed plate with a conductive adhesive, prior to being photographed under a HITACHI SU3500 scanning electron microscope (Hitachi, Tokyo, Japan).

Alkaline phosphatase was used as the marker for identifying pluripotent stem cells (Štefková et al., 2015). Alkaline phosphatase staining was performed using the BCIP/NBT Alkaline Phosphatase Staining Kit (Beyotime Biotechnology, Shanghai, China) following the manufacturer’s instruction with modifications. The staining solution was added to a 10-ml centrifuge tube and incubated with NEC and EC for 30 min. Following the removal of the staining solution, the NEC and EC were washed with distilled water once before being transferred into absolute ethanol and photographed.

After being ground to powder in liquid nitrogen, approximately 1 g of the samples were homogenized in 5 ml prechilled 80% methanol solution and vortexed for 90 h at 4°C. The extracts were then centrifuged at 12,000 g for 1 h, and the supernatant was collected and evaporated under the flow of nitrogen gas until being reduced to half of the original volume. The supernatant was purified using the C18 Sep-Pak cartridge that had been activated with 3 ml 100% methanol and preequilibrated with the formic acid solution, and then, the supernatant was dried under the flow of nitrogen gas. The dry residue was redissolved in 200 μl methanol, and the solution was filtrated with 0.22 μm syringe filter and 10 μl filter eluate was taken for the high-performance liquid chromatography (HPLC) analysis. Three biological replicates were performed.

Hypocotyls of ZM24 that were cultured on the callus-induction medium supplemented with or without uniconazole were sampled for RNA-seq analysis. Samples were collected at 7, 15, and 30 days after the initiation of culture, which were used to construct the RNA libraries. RNA-seq, consisting of two biological replicates, was performed using Illumina HiSeq™ 2000 (BGI, Shenzhen, China). The raw reads were subjected to quality control and filtered into clean reads, prior to being mapped to the reference genome of G. hirsutum (Li et al., 2015). Gene expression was calculated by Fragments Per Kilobase of transcript per Million mapped reads (FPKM), and differentially expressed genes (DEGs) was identified by fold change ≥ 2 and false discovery rates (FDR) with adjusted p-value < 0.01 combining three biological replicates. The data are available in the NCBI SRA under the number PRJNA808630.

The total RNA was isolated from cotton callus and leaves using the RNAprep Pure Plant Kit (Tiangen). First-strand cDNA was generated from the total RNA using the PrimeScript RT Reagent Kit (Takara). qRT-PCR was performed using SYBR Premix Ex Taq™ (Takara). GhHISTONE was used as an endogenous reference gene, and the relative expression level was calculated using the 2^–△△CT method as described previously (Xu et al., 2015). The gene-specific primers are listed in Supplementary Table 1.

All statistical analyses were based on at least three biological replicates, and all the values are displayed as the mean ± SD. Statistical analysis was determined using the t-test, and *p < 0.05 and **p < 0.01 were considered statistically significant.

Seven days after spraying with 20 mg/L uniconazole, the growth of cotton seedlings was retarded as shown by significantly reduced plant height relative to the unsprayed control plants (Supplementary Figure 1A). The top and reciprocal second leaves were smaller, wrinkled, and with short petioles (Supplementary Figure 1B). The histological analysis of the longitudinal paraffin sections prepared from the reciprocal fourth petioles illustrated that the cell size, elongation in particular, of the cortex cell layer was conspicuously reduced in the uniconazole-treated seedlings relative to the control (Supplementary Figure 1C), suggesting the inhibition of cell elongation or growth rate on uniconazole treatment.

Given that the uniconazole spraying significantly affected the growth of cotton seedlings via repressing cell elongation, the question arises as to whether uniconazole could affect callus growth and SE in cotton tissue culture. To address such a query, callus derived from four G. hirsutum cultivars with the variable level of SE capability was subjected to uniconazole treatment by supplementing 7 μM uniconazole to callus-induction medium. This concentration of uniconazole was chosen in accordance with a preliminary gradient experiment that suggested 7 μM uniconazole as the optimal level for growth suppression (Supplementary Figure 2). Following 7-day incubation on callus-induction medium supplemented with uniconazole, the hypocotyl explants produced callus mass at their ends, without discernible variation between control and uniconazole treatment (Figure 1A), indicating that uniconazole may not exert a significant inhibition effect on cell growth during the callus initiation process. At 15 days after the initiation of culture, the callus in the uniconazole treatments appeared to be smaller than the control callus and exhibited abnormality in callus appearance showing compactness rather a smooth surface, which is in contrast to the well-developed control callus (Figure 1A). At 30 days in culture, calluses in the control explants were substantially larger in size relative to the uniconazole-treated explants (Figure 1A). The effect of uniconazole on cell growth was further quantified using the CPR in the callus proliferation stage, showing that uniconazole treatment resulted in significant reductions in the CPR at days 15 and 25 independent of the genotype (Figures 1A,B).

Given that uniconazole inhibited callus growth via decreasing the cell size, we then investigated if the regulation function of uniconazole is conserved in seedling and SE. A scanning electron microscope was used to facilitate the measurement of the callus cell size. Strikingly, it was the reduction in the number of callus cells rather than the cell size that could be attributable to relatively smaller callus in the uniconazole treatments relative to the untreated control (Figure 1C). Notably, the large callus cells increased in numbers under uniconazole treatment at a relatively fast pace than control, but the rate of cell division was reduced, resulting in lesser cell proliferation and gross reduction in the callus size as observed in the uniconazole treatments. In essence, the suppressing effect of uniconazole on growth is common between seedling and callus, but through a different route, negatively impacting cell size in the former and cell number in the latter.

Alkaline phosphatase shows lower activity in differentiated cells than pluripotent stem cells; therefore, it could be used to evaluate cell pluripotent properties (Zhao et al., 2009). To investigate whether uniconazole could impact the capacity of callus cells for differentiation into stem cells, the explants of ZM24, J14, 99668, and TM-1 were maintained on the callus-induction medium with or without uniconazole and were sampled for alkaline phosphatase staining at 30 days. In contrast to the rather weak and sporadic bluish staining in a very small fraction of callus on the control medium, distinct blue color was observed in a large fraction of callus derived from ZM24, J14, and 99,668 on the medium supplemented with uniconazole. It was noted that TM-1 callus was only slightly stained under uniconazole treatment (Figure 2A), without conspicuous variation between control and uniconazole treatment.

To detect the relationship of alkaline phosphatase activity and pluripotent stem cells, a part of callus was continuously cultured to induce EC until the 40-day culture. A fraction of EC in ZM24, J14, and 99,668 was obtained under uniconazole treatment, but only an enlarged callus mass was present on the control plates (Figure 2B). In line with the weak activity of alkaline phosphatase, EC was not obtained, except a small fraction of callus displaying EC-like features after uniconazole treatment in TM-1. Therefore, it is evident that uniconazole could promote cell fate specification in the genotypes with EC regeneration capacity, but not in those without such an ability.

For further quantification of the SE-promoting efficiency of uniconazole, the EDR representing the transition rate of callus into EC was then measured in the aforementioned genotypes at 60 days. It is overt that EDR was significantly increased under uniconazole treatment, whereby EDR accounted for 80.12% in ZM24, 19.17% in J14, and 9.19% in 99,668 under uniconazole treatment, which was in sharp contrast to 0.25% in ZM24, 0.17% in J14, and null in 99,668 on the control explants (Figure 2C). As anticipated, no EC was obtained in TM-1 with or without uniconazole (Figures 2B,C). But for the remaining three genotypes, unlike a small proportion of callus that could form EC without the supplementation of uniconazole, most of the calli under uniconazole treatment could differentiate into EC (Supplementary Figure 3). Overall, as alkaline phosphatase activity is positively related to EDR, it could be used as a practical indicator for the potential of pluripotent stem cell acquirement in cotton callus.

It has been reported that uniconazole could change the content of endogenous hormones (Izumi et al., 1988). Considering that auxin and ABA play an important role in SE (Yang et al., 2012; Jin et al., 2014), the concentrations of IAA and ABA in ZM24 callus at 7-, 15-, and 30-day postinduction were measured by HPLC-MS. As described in Figure 3A, the IAA concentration of callus grown on CIM supplemented with uniconazole was decreased significantly compared with the control at all-time points. It is suggested that the reduction of the auxin content in callus was a pivotal factor for inhibiting callus growth on uniconazole treatment. Different from the downward trend of auxin in uniconazole treated callus, the change in ABA concentration is more complex. After 7-day induction in the callus initiation stage, there was a slight decrease, yet not statistically significant, in the ABA content in the uniconazole treatment relative to control. However, at 15 and 30 days at the callus proliferation stage, the ABA contents in uniconazole treated callus rose significantly relative to control (Figure 3B). The reduction in the auxin content coupled with a rise in the ABA level may act in concert in inhibiting callus proliferation on uniconazole treatment.

DR5, a highly active synthetic auxin response element (AuxRE), displays the distribution of auxin in tissues (Ulmasov et al., 1997). Transgenic cotton containing the auxin-responsive reporter DR5::GFP was used for the assessment of the spatial pattern of auxin distribution in cotton callus at both callus initiation and proliferation stages, using the GFP fluorescence signal. As shown in Figure 3, uniconazole treatment reduced the IAA content in the callus and influenced the fate of the callus; therefore, the auxin content and its distribution are both important for cell growth and development. Consistent with the decrease in the endogenous auxin level, the GFP fluorescence intensity of the callus under the uniconazole treatment was overtly weaker than control in both callus initiation and proliferation stages. At 7 days, the initiation of callus formation was observed at one end of a hypocotyl cultivated on the Murashige and Skoog Basal (MSB) medium, without discernible callus formation by naked eyes at the other end on uniconazole treatment (Figure 3C). Clear green fluorescence was observed in the callus formed on hypocotyl in the MSB medium (control), which was in contrast to the uniconazole-treated hypocotyls that did not show visible GFP. Likewise, a stronger and more widely distributed GFP fluorescence signal was exhibited in the control callus relative to uniconazole-treated callus at 15 and 30 days, suggesting the inhibitory role of uniconazole in callus proliferation via reducing the auxin level and its transport (Figure 3C).

The increase in the ABA content in uniconazole-treated callus (Figure 3) was concomitant with the promotion in pluripotent stem cell and EC acquirement, suggesting ABA and uniconazole acting in concert to determine cell fate transition. To facilitate the investigation on such a premise, explants of ZM24 were cultured on callus induction media containing uniconazole with or without (control) exogenous 0.04 μM ABA. The concentration of ABA was chosen according to preliminary gradient experiments (Supplementary Figure 4). Callus presented the same texture and color regardless of the ABA supplement at 15 and 30 days (Figure 4A). Quantitative analysis on callus growth using the CPR did not show significant variation between uniconazole and uniconazole + ABA treatment (Figure 4B). However, the analyses of the EC phenotype and EDR on the cultures between 40 and 60 days appeared to be very different. As shown in Figure 4A, at 40 days, embryonic cell mass grown on the medium supplemented with both ABA and uniconazole was visibly larger than those treated with ABA alone. Such an observation was supported by the quantitative analysis on EC differentiation rate that showed significant increments compared with the control at 40, 50, and 60 days (Figure 4C). By virtue of these data, it is conceivable that ABA and uniconazole displayed additive effects on promoting cell fate transition from callus cell into EC but not on callus proliferation.

To explore the molecular mechanism underpinning the uniconazole regulation on SE in cotton, genome-wide transcript profiling of callus with or without uniconazole treatment was performed. After quality control and filtering, approximately 50 million clean reads were generated in each library. Approximately 93% of reads were mapped to G. hirsutum TM-1 genome sequences (Li et al., 2015), and approximately 90% were mapped to gene sequences (Supplementary Table 2).

Compared with control, 47 genes related to auxin biosynthesis, transport, and signaling pathway were differentially expressed in the callus under uniconazole treatment relative to the untreated control. Six auxin biosynthesis transcripts, encoding cytochrome P450 (CYP71A1 and CYP83B1), L-tryptophan pyruvate aminotransferase 1 (TAA1), and indole-3-pyruvate monooxygenase (YUCCA10), showed a sophisticated expression pattern at different cultural stages. Compared with control, most of the auxin biosynthesis-related transcripts exhibited lower expression levels at 7 and 15 days but upregulated at 30 days on uniconazole treatment (Figure 5A and Supplementary Table 3), corroborating the biochemical analysis of endogenous auxin analysis. Five indole-3-acetic acid-amido synthetase transcripts (GH3.1 and GH3.6) showed a relatively higher expression level than control, except for the GH3.6 transcript (Gh_–A01G072200) that maintained its lower expression level at all three time points. Compared with control, the expressions of GH3 genes were downregulated at 7 and 15 days and before their upregulation at 30 days on uniconazole treatment (Figure 5B). GH3 catalyzed the synthesis of IAA-amino acid conjugates to cope with the presence of excess auxin, and increased auxin level under uniconazole may need high expressed GH3 at 30 days. The auxin transport-associated genes, including three LAX genes and three PIN genes, were differentially expressed. Regardless of uniconazole treatment or not, the expression levels of LAX and PIN genes showed a consistent decrease along with the progress of uniconazole treatment (Figure 5C), indicating the weakening of auxin polar transport during callus growth. Furthermore, compared with control, most of the LAX and PIN genes, with the exception of PIN1A, were downregulated at 7 days, prior to their upregulation by uniconazole treatment at 15 and 30 days (Figure 5C). The differential expression of auxin transport genes may be well attributed to the differential auxin level in both the control and uniconazole treated callus.

A number of auxin-responsive protein genes were differentially expressed. Except for IAA30, two IAA27 members and one IAA26 were downregulated at 7 days before being upregulated at 30 days with uniconazole treatment. Most of the Aux genes, with the exception of two AUX22D genes, were highly expressed at all the three callus stages in the control. However, their gene expression patterns were substantially altered by uniconazole treatment. The response of six SAUR genes to uniconazole treatment was variable. SAUR41 showed a high level of expression at all the time points, which was further upregulated by uniconazole treatment. In contrast, other SAUR members, such as SAUR32 and SAUR50, showed rather a low expression throughout. SAUR32 was induced by uniconazole treatment at 7 days but downregulated at 15 and 30 days, while the opposite was true for SAUR50 (Figure 5D). Taken together, it appears that auxin biosynthesis, transport, and signaling pathway may have participated in the interactive modulation with uniconazole for callus growth and differentiation.

Four transcripts corresponding to the key enzyme ABA catabolism, ABA 8′-hydroxylase CYP707A1, including Gh_–A08G171800, Gh_–D08G170200, Gh_–A12G225800, and Gh_–D12G220400, were all upregulated by uniconazole treatment at 7 and 15 days, before being downregulated at 30 days (Figure 5F). ABA catabolism might need to be appropriately upregulated via negative feedback regulation from 7 to 15 days in order to mitigate the adversity stress stem from excessive endogenous ABA levels in the callus on uniconazole treatment, which are detrimental to cell development. From 15 to 30 days, the concentration of ABA decreased in callus with or without uniconazole treatment. Unlike the control, the high level of endogenous ABA as the result of the downregulation of ABA catabolism following uniconazole treatment is likely favorable for cell fate transition from callus cell to embryogenic cell.

Ten SE-associated transcription factors were differentially expressed at 7, 15, and 30 days. WUSCHEL-related homeobox transcript, WOX11, displayed high expression at 7 days in control, which was downregulated at the same time point in uniconazole treatment (Figure 5E). This is congruent with a previous report that WOX11 positively regulates callus initiation and proliferation (Liu et al., 2014), as the reduced expression of WOX11 on uniconazole may hinder callus growth. Other transcripts, including JACKDAW (JKD), LOW PHOSPHATE ROOT1 (LPR1), ABCB19, and 4-COUMARATE: COA LIGASE1 (4CL1), all exhibited similar expression patterns with WOX11, whereby they were highly expressed at 7 days, but downregulated on the uniconazole treatment (Figure 5E). Three nuclear transcription factor Y subunit transcripts encoding NF-YB9 (Gh_–D13G152700 and Gh_–A13G151400) and NF-YB6 (Gh_–D05G177600) were not expressed at 7 and 15 days but clearly upregulated expression at 30 days, showing significant induction in uniconazole treated callus relative to control. It is particularly intriguing to explore whether the increased expression of NF-YB genes in the callus at 30 days could promote the cell fate transition from callus cell to embryogenic cell, which warrants further study. Furthermore, qRT-PCR was used to validate the RNA-seq results and confirmed that some auxin-, ABA-, and SE-related genes were differentially expressed in the callus under uniconazole treatment (Supplementary Figure 5).

Numerous studies have shown that uniconazole plays an important role in SE in several plant species, accentuating somatic embryo development. For example, uniconazole significantly improved the production of somatic embryos and their further development to plantlets in asparagus (Li and Wolyn, 1995). In geranium and carrots, the number of somatic embryos was increased by uniconazole treatments in the somatic embryo growth and differentiation period (Hutchinson et al., 1997; Tokuji and Kuriyama, 2003). In contrast, less is known about the effect of uniconazole on callus growth and embryonic callus production. It was reported that uniconazole was able to promote callus induction from maize immature embryos and rice mature embryos, improving callus differentiation in rice but not in maize (Fu et al., 2005; Shan and Zou, 2014). It appears that the functional role of uniconazole in promoting callus differentiation and transition to somatic embryos may vary from plant species to plant species and invokes exploration in individual crops before using it as an important research tool in assisting plant regeneration through tissue culture and genetic modification. To our knowledge, no study has hitherto reported the impact of uniconazole on SE in cotton. In this study, we uncovered the role of uniconazole on SE in different cotton varieties (cultivars with low or high regeneration ability and a recalcitrant variety TM-1). Uniconazole inhibited callus proliferation independent of genotypes (Figure 1), whereby decreasing cell numbers resulted in smaller callus in size. The ratio of the callus conversion into embryonic callus was overtly improved along with the increase in the size of embryogenic cell mass in ZM24, J14, and 99,668 lines, but in TM-1, only the texture of callus was changed on uniconazole treatment (Figure 2). It is conceivable that embryogenic cell specification is dependent on multifaceted factors, among which uniconazole plays an assistive role instead of a decisive factor in promoting cotton SE. In essence, uniconazole inhibited callus growth in a genotype-independent manner and in a facilitating role in enhancing EC differentiation rate in the genotypes with regeneration ability.

Callus and embryonic callus induction are essential to cotton SE, which is a prerequisite for producing transgenic cotton. The molecular mechanisms of callus induction and differentiation have been revealed, where auxin, cytokinin, and wound signal played key roles in the reacquisition of embryonic or meristematic fate (Ikeuchi et al., 2013). Roles of uniconazole in regulating plant hormones have been demonstrated in rice, winter rape, soybean, and bioenergy crop duckweed (Landoltia punctata) (Izumi et al., 1988; Zhou and Leul, 1998; Zhang et al., 2007; Liu et al., 2015). The effect of uniconazole on the endogenous IAA level in callus was variable depending on plant species. In winter rape, the endogenous IAA level was decreased following foliar application of uniconazole (Zhou and Leul, 1998, 1999), while in rice seedlings, IAA was not altered on uniconazole treatment (Izumi et al., 1988). In our research, endogenous IAA concentration was reduced significantly from the early stage of callus initiation to the formation of non-EC after uniconazole treatment (Figure 3A). Concomitant inhibition of GFP fluorescence was observed in uniconazole-treated callus (Figure 3C). These results indicated that uniconazole regulates cell fate transition by reducing the content of auxin and disrupting its polar distribution, low auxin level may repress the callus over-proliferated growth and promote EC acquirement, and congruent with the previous report that the endogenous auxin level was significantly lower in EC and somatic embryos than those without embryogenic potential (Yang et al., 2012).

Uniconazole has been shown to regulate the endogenous concentration of ABA by inhibiting the activity of the ABA catabolism enzyme (Todoroki et al., 2008, 2012). Uniconazole treatment gave rise to an increase in the ABA level in plants (Zhou and Leul, 1999; Zhang et al., 2007; Li et al., 2015). In this study, the callus contained significantly higher endogenous ABA levels following uniconazole treatment (Figure 3B), supporting the premise that enhancement in the ABA level could facilitate EC production, which is validated by the cooperative effects of ABA and uniconazole on EC differentiation.

The importance of transcription factors to the process of SE has been demonstrated in an eclectic list of studies. WUSCHEL ELATED HOMEOBOX 11 (WOX11) is involved in the first step of cell fate transition in the de novo root organogenesis process, which shares a similar mechanism with callus formation (Liu et al., 2014). LEAFY COTYLEDON1 (LEC1) can maintain embryonic cell fate at an early development stage in Arabidopsis (Lotan et al., 1998), by regulating the key genes involved in cellular differentiation and embryo development (Kwong et al., 2003; Lee et al., 2003). LEC1 and LEC1-like (L1L1) are partially functionally redundant (Yamamoto et al., 2009), and overexpression of GhL1L1 accelerated embryonic cell formation (Xu et al., 2019). In our study, NF-YB9 (LEC1) was not expressed at 7- and 15-day induction, but became highly expressed at 30 days with uniconazole treatment. Similar expression patterns were observed in NF-YB6 (LEC1-like) (Figure 5E). As previously reported, the middle cell layer of callus is endowed with the quiescent center (QC)-like transcriptional feature, where some marker genes including JKD, LPR1, and ABCB19 were all co-expressed, while 4CL1 was highly expressed in explant vascular and callus founder cells of Arabidopsis thaliana (Lee et al., 1995; Welch et al., 2007; Zhai and Xu, 2021). As shown in Figure 5E, uniconazole affected the expression of these marker genes in the middle cell layer and inhibited the callus proliferation. Altogether, the differential expression of SE-related marker genes induced by uniconazole may contribute to pluripotency acquisition via establishing QC-like stem cell identity in callus (Zhai and Xu, 2021). In summary, we showed that uniconazole promotes the process of SE in cotton by inhibiting callus proliferation and promotes the transformation of callus into EC (Figure 6).