FANS uncovers WOX2-associated early regulators of auxin-induced somatic embryogenesis in Arabidopsis
Anna M. Wójcik, Weronika M. Buchcik, Viktoriia Verezunska, Ruben Gutzat, Ortrun Mittelsten Scheid, Małgorzata D. Gaj
TL;DR
This study uses FANS to identify genes involved in early stages of plant embryogenesis, revealing a regulatory network centered on WOX2.
Contribution
The study identifies a new transcriptional layer of totipotency control centered on WOX2 in Arabidopsis.
Findings
Genes up-regulated in embryogenic nuclei are enriched in embryo and tissue development processes.
Transcription factors like MYB46 and ZAT14 act downstream of WOX2 to regulate embryogenic competence.
The WOX2-centered network precedes activation of canonical SE regulators like LEC2 and BBM.
Abstract
Somatic embryogenesis (SE), the process by which differentiated somatic cells are reprogrammed to form embryos, represents a unique manifestation of plant cell totipotency. Despite its fundamental and applied importance, the molecular mechanisms that initiate embryogenic reprogramming remain largely unknown, mainly because explant tissues are cell-type heterogeneous and contain only a small fraction of SE-competent cells. Here, we applied fluorescence-activated nuclei sorting (FANS) in Arabidopsis thaliana to isolate cells expressing the WOX2 gene, a marker for early embryogenesis. Comparative transcriptomic analysis of WOX2(+)-positive and WOX2(-)-negative nuclei revealed that genes up-regulated in the embryogenic nuclei were strongly enriched in biological processes related to embryo and tissue development, while down-regulated transcripts were linked to primary metabolism, suggesting…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6
Figure 7
Figure 8
Figure 9- —National Science Centre in Poland
- —The Bekker Programme - NAWA (Polish National Agency for Academic Exchange)
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsPlant Molecular Biology Research · Plant tissue culture and regeneration · Chromosomal and Genetic Variations
Background
Plant somatic embryogenesis (SE) is a hallmark of cellular totipotency: differentiated somatic cells can be reprogrammed to initiate embryo development in response to exogenous stimuli, primarily auxin or stress treatment. SE induced under in vitro culture conditions serves as a powerful model to investigate the mechanisms of cell fate transition and to facilitate clonal propagation and transformation in plant biotechnology [1]. However, the mechanisms that control the reprogramming from differentiated somatic cells into embryos are still poorly understood. Previous experimental approaches towards identifying pluripotency-related genes essential for embryogenic development in vitro have been reported, particularly in the model plant Arabidopsis [2–4]. Multiple layers of regulation contribute to SE, including hormone signaling, stress responses, cell-wall remodeling, transcription factor (TF) networks, and chromatin-based reprogramming. Auxin is central, both as an exogenous inducer and through local biosynthesis/transport, establishing instructive gradients for embryogenic transition and morphogenesis of somatic embryos [5]. Auxin rapidly rewires chromatin accessibility to redeploy a hierarchical TF network for totipotency, directly linking canonical embryogenesis regulators (e.g., LEC1/LEC2/BBM/FUS3) with early embryonic patterning genes such as WOX2/WOX3 (WUSCHEL RELATED HOMEOBOX) [6]. In particular, LEC2 (LEAFY COTYLEDON2), known as one of the SE key TFs [7, 8], has been demonstrated to activate transcription of WOX2/WOX3, mechanistically connecting auxin-responsive TFs with the earliest axes of embryo patterning during SE [6]. Recently, the conserved role of the WOX genes has been indicated within the early embryogenesis program [9]. Within the WOX family, WOX2, which plays a pivotal role in apical domain specification and the initiation of shoot-meristem identity, has been identified as a marker of the early stages of embryogenic transition in somatic cells [6, 9]. A critical and conserved role of WOX2 in the early stages of SE in both mono- and dicotyledonous species - Castanea mollissima [10], Medicago truncatula [11], Picea abies [12, 13], Pinus pinaster [14], and Arabidopsis [14, 15] has been reported. Consistently, WOX2 has been successfully used as a marker of early embryogenic cells in various plant species, including Arabidopsis, and has facilitated the identification of genes that enhance transformation efficiency in maize [6, 16, 17]. These observations place WOX2 at the interface between auxin-driven totipotency modules and early embryo patterning, as a conserved key regulator of SE induction.
During SE induction, explant tissues consist of a mosaic of somatic cell types that differ in developmental identity and embryogenic potential. Only a small subset of cells retains a totipotent-like status and can spontaneously re-enter the embryogenic program; for example, stomatal-lineage cells in the cotyledonary epidermis of Arabidopsis exhibit intrinsic competence for SE, and ectopic expression of LEC2 can induce embryo formation in these cells even in the absence of exogenous auxin [18]. In contrast, most somatic cells within an explant require auxin treatment to acquire embryogenic competence through extensive genetic and epigenetic reprogramming [19–21]. Such auxin-mediated induction involves chromatin remodeling, transcriptional activation of developmental regulators, and metabolic reconfiguration that collectively enable the transition from a differentiated to an embryogenic identity [4, 22]. These early events of cellular reprogramming, especially the transcriptional responses that initiate totipotency, remain poorly understood, largely because explants are cell-type-heterogeneous and contain only a small fraction of SE-competent cells. Separating intact individual plant cells requires protoplasting, at the cost of time-consuming enzymatic treatments that can interfere with gene expression patterns. A more straightforward and fast approach is the INTACT (Isolation of Nuclei TAgged in specific Cell Types) method, which enables affinity-based isolation of nuclei from defined cell types by expressing a biotinylated nuclear envelope protein under a cell-type-specific promoter [23]. This approach provides nuclei of high purity and yield without enzymatic or mechanical disruption, allowing genome-wide analyses of gene expression and chromatin states in distinct plant cell populations [24]. The INTACT method has been used in Arabidopsis to explore the gene networks underlying the indirect SE pathway in callus cells expressing LEC2 [25]. A complementary approach, fluorescence-activated nuclei sorting (FANS), enables rapid isolation of fluorescently labeled nuclei without protoplasting [26]. Using FANS, the few CLV3-marked stem cells in the Arabidopsis shoot apical meristem were isolated across different developmental stages. Sorting purity and transcriptome integrity were validated [26], establishing FANS as a robust approach for cell type-resolved nuclear RNA-seq in Arabidopsis [24].
Here, we leverage FANS in the context of auxin-induced direct SE to overcome explant heterogeneity and interrogate the earliest transcriptional programs in WOX2-positive (WOX2^(+)^) versus WOX2-negative (WOX2^(−)^) nuclei during auxin-induced SE in Arabidopsis. We collected nuclei expressing GFP-labelled histone H2B under the control of the WOX2 promoter and performed RNA-seq to define differentially expressed genes (DEGs). We identified several TF genes expressed and highly up-regulated early in the WOX2-engaged network during the somatic-to-embryogenic transition, including MYB46, ZAT14, MYR2,* TCP19*, MYB98, and GRF7. Their SE-related function was validated by the findings that overexpression of MYB46, ZAT14, MYR2,* MYB98*, and GRF7 might compensate for the auxin treatment required for SE induction and promote the formation of somatic embryos on the explants. Identifying new SE-associated TFs opens further opportunities for better understanding the regulatory network of plant cell reprogramming and for improving SE-based in vitro plant regeneration across different species.
Results
To render the nuclei of cells responsive to SE induction separable from others by the FANS method [24, 26], we used an Arabidopsis line expressing GFP-labeled histone H2B under the control of the SE-specific WOX2 promoter. This line was subsequently used to induce embryogenic development in auxin-treated explants cultured in vitro according to the established SE induction protocol [27]. For transcriptional profiling, we selected the third day (3rd day) after SE induction, based on previous reports indicating early chromatin accessibility and transcriptional competence of WOX2 regulatory regions during this period [6, 15]. While earlier studies detected WOX2 transcripts starting from day 4, our reporter-based analyses revealed WOX2 activation as early as day 3 (Supplementary Fig. S1), suggesting that this stage may represent a critical window for the onset of embryogenic reprogramming. These observations prompted us to focus on day 3 as a potentially informative time point for capturing the earliest transcriptional events associated with the acquisition of totipotency. We isolated nuclei from manually enriched upper parts of explants on the 3rd day after SE induction and applied FANS (Fig. 1) to collect GFP-positive and GFP-negative nuclei (Fig. 2A, B). Microscopy revealed that nuclei sorted into the positive channel appeared intact and displayed green fluorescence, validating the purity of the fraction (Fig. 2C, D, E). The population of embryogenic cells obtained with FANS was used for transcriptome analysis with a protocol suited for limited amounts of RNA (low-input RNA-seq). Expression data were reproducible across replicates, and housekeeping genes [28] were uniformly represented in both GFP-positive and GFP-negative samples. High levels of endogenous WOX2 transcript in GFP-positive (> 170-fold FC) versus control (Suppl. Fig. S2) confirmed enrichment of SE-engaged nuclei.
Fig. 1. The FANS procedure for isolating SE-engaged nuclei. Cotyledons excised from pWOX2::H2B–GFP explants of Arabidopsis cultured for three days on auxin-containing induction medium (E5) were used for nuclei isolation. GFP fluorescence was mainly observed in the distal regions of the cotyledons (A), of the immature zygotic embryos (B). Sorted GFP-positive and GFP-negative nuclei were subsequently used for RNA sequencing (C)
Fig. 2. Gating strategy used for sorting nuclei of SE-engaged cells. Representative FANS plots are shown. Events were gated for GFP(⁺) or GFP(⁻) nuclei in the control (A) and the pWOX2::H2B-GFP line (B). Examples of GFP(⁺) nuclei obtained after sorting are shown in panels (C–E). Scale bar = 5 µm
The data were scored for differentially expressed genes (DEGs) in GFP-positive versus GFP-negative nuclei on day 3 after SE-induction. The results revealed 826 up-regulated and 921 down-regulated genes. These included 11% of highly up-regulated (log2FC > 5; p < 0.05) and 10% highly down-regulated genes (log2FC<−5; p < 0.05) (Fig. 3A). TF genes accounted for 8% of the up-regulated and 10% of the down-regulated DEGs (Fig. 3B).
Fig. 3. Differentially expressed genes (DEGs) on the third day after SE induction. (A) Highly up-regulated (fold change ≥ 2, red) and down-regulated (fold change ≤ 0.5, blue) genes. (B) Fraction of transcription factors (TFs) among up-regulated (fold change ≥ 2) and down-regulated (fold change ≤ 0.5) DEGs (p < 0.05)
To gain functional insight into the molecular programs distinguishing SE-engaged WOX2^(+)^ vs. WOX2^(−)^ nuclei, we performed Gene Ontology (GO) enrichment analysis restricted to the Biological Process (BP) category (Fig. 4). This revealed that up- regulated DEGs are strongly enriched in categories linked to developmental and differentiation processes, including regulation of single organism developmental processes, reproductive processes, tissue/organ development and post-embryonic development, indicating that WOX2^(+)^ nuclei employ transcriptional and regulatory mechanisms essential for embryogenic reprogramming. In contrast, the down-regulated DEGs predominantly cluster into metabolic and structural processes related to carbohydrates, polysaccharides, glucosinolates, and negative regulation of cell growth. The overrepresented terms among the down-regulated DEGs were associated with primary metabolism, including secondary metabolite and carbohydrate metabolic processes, which are linked to cell differentiation. The results suggest that, in contrast to the non-embryogenic state of WOX2^(−)^ nuclei, the embryogenic WOX2^(+)^ nuclei appear less differentiated and more metabolically active (Fig. 4B).
Fig. 4GO enrichment analysis of DEGs. GO enrichment was performed using the PlantRegMap platform for (A) up-regulated (fold change > 2) and (B) down-regulated (fold change < 0.5) genes (p < 0.05) in the “biological process” category. The bubble plots were generated using the SRplot web server [29, 30]. The X-axis indicates the enrichment score; bubble size represents the number of DEGs assigned to each GO term, and bubble color reflects statistical significance (−log₁₀ p-value)
Recognizing the critical importance of TFs in transcriptomic reprogramming associated with somatic-to-embryogenic transition of explant cells, we focused our analysis on those TFs whose transcripts represented a substantial fraction of the DEGs (Fig. 5). We identified 13 TF genes with highly up-regulated expression in WOX2^(+)^ vs. WOX2 ^(−)^ nuclei (log2FC > 5, which equals FC > 32) (Fig. 5A), and six highly down-regulated TF genes (log2FC<−5, equaling FC < 0.031 (around 32x down regulated expression) (Fig. 5B). Among the up-regulated DEGs, MYB46 and ZAT14 stood out due to markedly elevated expression levels log2FC 22.9 and 22.6 respectively (Fig. 5A). The candidate SE-regulator genes belong to different TF families, including the MYB, C2H2, G2-like, TCP, GRF, and C2C2-Dof.
Fig. 5. Differentially expressed TRANSCRIPTION FACTOR genes (TFs) during SE induction. (A) TFs strongly up-regulated (log₂ fold change ≥ 5). (B) TFs strongly down-regulated (log₂ fold change ≤ 0.31) on the third day after SE induction (p < 0.05)
Six TF genes, showing pronounced up-regulation in SE and possessing accessible genetic resources and known/suggested biological functions, were selected for functional characterization during embryogenic transition: MYB46, ZAT14, MYR2, TCP19, MYB98, and GRF7. To this end, the capacity for SE induction, evaluated based on SE efficiency and productivity, was assessed in cultures of transgenic lines with altered expression of the analyzed TFs, including loss-of-function insertional mutants and overexpression lines (Figs. 6 and 7). The results revealed decreased SE efficiency in myb46, zat14, and myb98 mutant cultures (Fig. 6A) and a significant reduction in SE productivity across all analyzed lines (Fig. 6B), suggesting that these genes positively regulate embryogenic potential. Since constitutive TF overexpression may severely impair plant development and fertility, we utilized an inducible system widely applied in TF functional studies, where transgene activation is triggered by β-estradiol treatment [31]. The analysis of embryogenic potential in the overexpression lines provided additional evidence that the candidate TFs may contribute to SE induction. Induced activity of MYB46, ZAT14, TCP19, and MYB98 significantly reduced the embryogenic potential of transgenic explants cultured on the auxin-containing E5 medium (Fig. 7A). In contrast, in the absence of exogenous auxin, explants with enhanced activity of MYB46, ZAT14, MYR2, and MYB98 showed enhanced embryogenic potential and developed embryo-like structures with a frequency of 10–30% (Fig. 7B, Supp. Fig. S3). These embryo-like structures subsequently developed into complete seedlings upon subculture on hormone-free medium, confirming their somatic embryo identity.
Fig. 6. Embryogenic potential of wild type (WT) and myb46, zat14, myr2, tcp19, myb98, and grf7 insertional mutant lines. Explants were cultured under the same conditions as in the FANS experiments. (A) Somatic embryogenesis (SE) efficiency; (B) SE productivity. Statistical significance was determined using one-way ANOVA with the Dunnett multiple-comparison test. Significance levels: p adj < 0.05 (), < 0.01 (), < 0.001 (), < 0.0001 (****)
Fig. 7. Embryogenic potential of β-estradiol-inducible overexpression lines (35S::MYB46-ER, 35S::ZAT14-ER, 35S::MYR2-ER, 35S::TCP19-ER, and 35S::MYB98-ER) and WT (Col-0). Explants were cultured on auxin-containing SE-inducing medium (E5; green bars) or on auxin-free medium (E0; turquoise bars). Transgene overexpression was induced by adding β-estradiol. Solid bars indicate SE efficiency on medium without β-estradiol; striped bars represent SE efficiency after β-estradiol induction. Statistical significance was assessed by two-way ANOVA with Bonferroni’s multiple-comparison test. Significance levels: (p adj < 0.05 (), < 0.01 (), < 0.001 (), < 0.0001 (****))
To identify DEGs common across different SE-transcriptomes, we compared our results with two other transcriptomic studies investigating embryogenic induction in auxin-treated immature zygotic embryos of Arabidopsis thaliana [25, 32]. The compared datasets differ in both the induction systems and the stages of SE that were analyzed (Table 1). The present dataset (FANS, 3 d SE) represents DEGs detected after three days of direct SE induction in WOX2^(+)^ versus WOX2^(−)^ nuclei, while Magnani et al. describe DEGs identified in LEC2^(+)^ versus LEC2^(−)^ callus nuclei after ten days of indirect SE induction (INTACT, 10 d SE) [25]. In contrast, Wickramasuriya and Dunwell examined the SE transcriptome of whole explants induced for five days and compared it with that of leaf tissue (RNAseq, 5 d SE) [32]. Table 1 also includes other TF genes, such as WOX8, WUS, and CLV3, which have been reported to be highly up-regulated during SE in Arabidopsis in other studies [3, 4, 33].
Table 1. Comparison of somatic-embryogenesis-associated DEGs identified in three transcriptomic studies of Arabidopsis thaliana culturesThe datasets represent:– FANS 3 d SE – DEGs identified in the present study after 3 days of direct SE induction (WOX2(⁺) vs. WOX2(⁻) nuclei);– INTACT 10 d SE – DEGs reported by Magnani et al. [25] after 10 days of indirect SE induction in LEC2⁺ vs. LEC2⁻ callus nuclei;– RNAseq 5 d SE – DEGs described by Wickramasuriya & Dunwell [32] after 5 days of direct SE induction in embryogenic explants compared with leaf tissue. Black box – absence of transcript(s) for the corresponding gene in a given dataset. Red box – statistically significant differential expression (fold change ≥ 2; p < 0.05). NSD – no significant difference. # Data without biological replicates; genes considered significantly up-regulated only when transcript levels increased more than fourfold (fold change > 4)
Eight of the SE-related TFs identified in the FANS-based analysis, including ZAT14, JGL, MYR2, TCP19, CDF4, SRS7, GRF7, and GRAS12, were also found to be differentially expressed on the fifth day of SE induction in the dataset reported by Wickramasuriya and Dunwell [32]. This overlap suggests a partial similarity between the direct SE transcriptomes characterized in these two studies, despite differences in the timing of SE induction (3 vs. 5 days) and experimental design (WOX2-expressing nuclei vs. whole explants). In contrast, only one gene, TCP19, showed a convergent expression pattern between the FANS- and INTACT-derived datasets, indicating that the genetic mechanisms controlling direct SE (FANS) and indirect, callus-mediated SE (INTACT) differ substantially. However, we cannot exclude that the observed discrepancies may also result from differences in the analyzed stages of SE (third versus tenth day), differences in culture conditions, or from methodological variations, such as the use of fluorescence-tagged nuclei versus protoplast-derived material (Table 1).
Discussion
Somatic embryogenesis (SE) remains one of the most powerful models for studying plant cell totipotency; yet its molecular basis has long been obscured by cellular heterogeneity within explants. Previous transcriptomic approaches, including the INTACT-based profiling of callus-derived nuclei, have provided essential but limited insights into SE-responsive explant cells due to low labeling efficiency and the confounding effects of protoplasting [25]. The FANS-based approach employed in this study overcomes these limitations by enabling the rapid isolation of intact, WOX2-marked nuclei without the need for enzymatic digestion. Moreover, the use of a labelled histone protein, which is highly expressed and efficiently incorporated into chromatin [34], provided stronger fluorescence signals and greater stability during the sorting process. The reliability of the transcriptomic data was further enhanced by applying a protocol optimized for low-input material, initially developed for studies of zygotic embryogenesis in Arabidopsis, in the downstream RNA-seq analysis of the isolated nuclei [35, 36]. Altogether, these methodological improvements enabled the capture of a transcriptionally homogeneous population of embryogenic nuclei, thereby providing a high-resolution view of the initial molecular events underlying auxin-induced SE.
Bioinformatic analysis of expression profiling of WOX2^(+)-^positive versus WOX2^(-)^-negative nuclear populations revealed gene expression networks specific to embryogenic cells. Gene ontology studies on DEGs showed that the up-regulated genes were strongly enriched in categories related to embryonic and reproductive development, while down-regulated transcripts were associated with metabolic and biosynthetic processes (Fig. 4), similar to what was previously observed during SE induction in Arabidopsis [25, 37]. Notably, the GO terms of genes with inverse regulation patterns support the view that embryogenic competence involves metabolic repression coupled with the activation of developmental and chromatin-regulatory programs, a hallmark of the somatic-to-embryogenic transition [1].
Within SE-specific DEGs, we found around 10% of genes encoding transcription factors (TFs), which play a crucial role in cell reprogramming and initiation of SE by regulating hormonal signaling, chromatin accessibility, and activating embryo-specific genetic pathways [3–5, 38–41]. Guided by this evidence, we focused on TFs to identify early markers of developmental competence and reveal regulatory mechanisms driving pluripotency acquisition during SE induction. We identified MYB46, ZAT14, MYR2, TCP19, GRF7, and MYB98 as new regulators during the early stages of SE. Consistent with this result, the role of the MYB, GRF, and TCP TF family members in SE regulation have been documented [3, 42–44]. Gain-of-function of the MYB factors MYB118/PGA37 and MYB115 triggers a vegetative-to-embryonic switch and production of somatic embryos, and global SE transcriptome profiling shows broad modulation of MYB-family transcripts in Arabidopsis [3, 45]. The miR396-GRF module contributes to the embryogenic response via an auxin-related pathway in Arabidopsis, and elevating GRF activity markedly enhances plant regeneration/SE competence across species, underscoring a conserved GRF-mediated route to embryogenic competence [42, 46]. Consistent with our observations for TCP19, functional studies in dicots identified TCPs as integral nodes, e.g., a LEC1–CKI–TCP15–PIF4 network controlling SE via auxin homeostasis in cotton, and a CsTCP14–IAA4 module promoting SE in citrus [43, 44]. Together with these reports, our identification of TCP19 within the SE-TFs of Arabidopsis points to a conserved TCP role in embryogenic reprogramming of plant somatic cells.
The TF genes of highly up-regulated expression in our analysis, including MYB46, ZAT14, MYR2, TCP19, GRF7, and MYB98, are associated with diverse developmental and stress-related processes [47–52]. MYB46 is a well-established master regulator of secondary cell wall biosynthesis, controlling lignin, cellulose, and hemicellulose deposition in xylem tissues [47, 53]. ZAT14, a C2H2 zinc-finger protein, is a transcriptional repressor in programmed cell death in the root cap [48]. MYB98 is expressed explicitly in synergid cells of the female gametophyte, where it controls the formation of the filiform apparatus and guides pollen tube reception [49]. Members of the TCP family, including TCP19, have been implicated in cell proliferation and differentiation, linking hormonal and developmental signals [50]. GRF TF family members are broadly involved in plant growth regulation and stress response supporting a possible role of GRF7 in SE induction [51]. Although MYR2 function remains less characterized, its negative regulation of flowering time points to the role in developmental phase transition thus supporting the involvement in SE [52].
Insights into the embryogenic responses of mutant and overexpression lines suggest that MYB98, ZAT14, MYB46, and TCP19 act as potent activators of auxin-mediated SE induction. Explants overexpressing these TFs, when cultured on auxin-free medium, promoted somatic embryo formation, whereas their culture on auxin-containing medium resulted in a reduced embryogenic response and excessive callus proliferation (Fig. 7A, B). A similar auxin-dependent modulation of embryogenic capacity has been reported for LEC2 and WOX5, which promote SE by enhancing auxin biosynthesis [54–56]. Supporting the proposed functional association of the candidate TFs with auxin, TCP19, GRF7, and MYB46 are known to regulate lateral root development, vascular cell elongation, and SE, via distinct auxin-related pathways [46, 57, 58].
Altogether, the phenotypes of transgenic lines with altered expression of candidate TFs indicate that excessive TF activation suppresses SE efficiency in the presence of auxin, whereas moderate or inducible expression can trigger SE even in the absence of exogenous auxin. These results reflect the regulatory principles observed for other key SE-related TFs, such as LEC2 and BBM, emphasizing that SE-involved transcriptional reprogramming depends not only on TF identity but also on the timing and intensity of their activation [33].
Moreover, the route toward SE is highly dependent on the cellular context and mode of induction. Our FANS-based transcriptomic analysis captured auxin-responsive WOX2^(+)^ nuclei derived from subepidermal cells located deeper within the explant, which respond to auxin treatment [15, 59]. In contrast, stomatal lineage-related epidermal cells that ectopically express LEC2 can enter embryogenic development without auxin treatment [18]. Comparison of these systems indicates that SE regulatory networks are context-dependent, varying according to cell type (epidermal vs. subepidermal tissues) and inductive cue (LEC2 overexpression vs. auxin treatment). The LEC2-induced epidermal route of SE proceeds through autonomous activation of the auxin pathway, whereas exogenous auxin triggers embryogenic competence in subepidermal cells via WOX2-associated transcriptional cascades. Together, these findings support the view that plant cell totipotency can be achieved through multiple, context-specific transcriptional trajectories that converge on a shared embryogenic state through distinct regulatory entry points.
Comparative analysis with previously published SE transcriptomes [25, 32] revealed limited overlap in differentially expressed genes, underscoring the strong context dependency of embryogenic reprogramming. A partial convergence was observed for eight up-regulated transcription factors detected both at the third day of direct SE induction (FANS) and the fifth day of whole-explant analysis [32], whereas overlap with the callus-derived, indirect SE transcriptome obtained with the INTACT method [25] was minimal. These differences likely reflect distinct developmental stages and analytical resolutions (nucleus-specific versus bulk) as well as different induction routes (direct versus indirect). The lack of significant differential expression of classical SE regulators such as LEC1, LEC2, BBM, and AGL15 among the FANS-identified DEGs, suggests that WOX2-marked nuclei may represent an earlier developmental window preceding activation of the canonical LEC/BBM network. Basal LEC2 activity in immature zygotic embryos may be sufficient to trigger WOX2 expression and initiate the first transcriptional switch toward embryogenic identity [6]. As SE progresses (days 5–10), LEC1/LEC2–BBM and other auxin-related regulators become up-regulated, enhancing endogenous auxin biosynthesis and maintaining embryonic identity [25, 32]. The absence (WUS, CLV3) or non-differential expression of shoot apical meristem marker genes (STM, CLV1,* CLV2*) further supports that the FANS-captured nuclei correspond to a pre-meristematic, early embryogenic state [25, 60]. Collectively, these findings position the WOX2-linked transcriptional network at the onset of direct SE, providing a molecular context for downstream structural reprogramming.
Cell fate transitions, including the onset of somatic embryogenesis, are tightly coupled with remodeling of the cell wall and reorganization of cell–cell communication domains [16]. In this context, two transcription factors identified among early WOX2-associated regulators, MYB46 and ZAT14, may potentially contribute to these structural transitions. MYB46 is a well-characterized activator of secondary cell wall biosynthesis [47, 53, 61, 62], while ZAT14 is involved in programmed cell death and regulation of expansin genes during xylem development [48, 63]. Moreover, ZAT14 has been reported to regulate MYB46 expression [64]. Although their canonical roles are not directly linked to embryogenic induction, their co-activation during SE and reported regulatory interaction (ZAT14→MYB46) raise the possibility that these TFs are repurposed in this context. We hypothesize that they may function in balancing local wall loosening and stabilization processes, which facilitate the spatial isolation of embryogenic cells and the formation of the extracellular matrix surface network [16].
This interpretation is reflected in the proposed WOX2-dependent regulatory model (Fig. 8), which integrates early transcriptional responses with putative cell wall remodeling processes associated with the embryogenic transition, and is intended as a conceptual framework to guide future experimental validation.
Fig. 8. Proposed hypothesized model summarizing the WOX2-mediated regulatory framework underlying the earliest events of somatic embryogenesis (SE). WOX2 acts as an upstream activator initiating the SE program and co-activating two transcription factors, ZAT14 (a C2H2-type transcriptional repressor) and MYB46 (a MYB-type transcriptional activator). ZAT14 represses wall-integrity genes such as PME (PECTIN METHYLESTERASE46), XTH (XYLOGLUCOSYL TRANSFERASE), and EXP (EXPANSIN) [48], promoting local cell-wall loosening, detachment, and programmed cell death (PCD). In parallel, MYB46 activates primary-wall remodeling genes, including CesA (CELLULOSE SYNTHASE), XTH, and PG (POLYGALACTURONASE) [47, 65, 66], contributing to the stabilization of cell-wall plasticity in newly isolated embryogenic cells. Moreover, ZAT14 has been reported to regulate MYB46 expression [64]. Together, the complementary actions of ZAT14 and MYB46 lead to the formation of the extracellular matrix surface network (ECMSN) and the spatial isolation of SE-competent cells, processes associated with the early stages of the embryogenic transition [15, 29]. The proposed module seems to be another way together with LEC/BBM (LEAFY COTYLEDON/BABY BOOM) related auxin biosynthesis pathway to induce somatic embryos development [18, 67]
Conclusions
This study establishes a high-resolution framework for dissecting the earliest transcriptional events of auxin-induced SE in Arabidopsis thaliana. By combining fluorescence-activated nuclei sorting (FANS) with a WOX2-based reporter, we isolated a homogeneous population of embryogenic nuclei and identified a distinct set of early transcription factors initiating cellular reprogramming. Among them, MYB46 and ZAT14 emerged as candidate components of a WOX2-dependent regulatory module. Their early activation suggests a potential involvement in cell‑wall–related processes associated with the formation of the extracellular matrix surface network and the spatial isolation of SE‑competent cells, although their precise roles remain to be experimentally validated. These findings suggest the existence of an early, pre-LEC/BBM transcriptional layer that may contribute to the control of totipotency prior to the activation of canonical embryogenic regulators.
Among them, MYB46 and ZAT14 emerged as candidate components of a WOX2-dependent regulatory module. Their early activation suggests a potential involvement in cell‑wall–related processes associated with the formation of the extracellular matrix surface network and the spatial isolation of SE‑competent cells, although their precise roles remain to be experimentally validated. These findings suggest the existence of an early, pre-LEC/BBM transcriptional layer that may contribute to the control of totipotency prior to the activation of canonical embryogenic regulators. An intriguing aspect for future studies on the roles of MYB46 and ZAT14 in SE is their potential association with auxin, as indicated by the embryogenic effect observed in explants overexpressing these genes, even in the absence of auxin treatment.
The methodological and conceptual framework presented here provides a foundation for future single-nucleus and multi-omic studies of plant cell fate transitions, highlighting potential molecular targets, such as MYB46 and ZAT14, for future exploration aimed at improving regeneration efficiency in recalcitrant species. An intriguing aspect for future studies on the roles of MYB46 and ZAT14 in SE is their potential association with auxin, as indicated by the embryogenic effect observed in explants overexpressing these genes, even in the absence of auxin treatment. Comparative studies of SE across different induction systems, both direct and indirect, and across various plant species will be essential to distinguish conserved regulatory modules from system-specific mechanisms that control embryogenic competence. Furthermore, integrating FANS-based transcriptomic profiling with single-nucleus multiomic approaches offers a promising route to reconstruct the temporal hierarchy of regulatory events, including chromatin accessibility and histone modification dynamics, that underlie the initiation of embryogenic reprogramming at single-cell resolution. Together, these approaches will provide a comprehensive framework for understanding the molecular logic of plant cell totipotency and for translating fundamental insights into improved biotechnological applications.
Materials and methods
Plant material
All experiments were performed with Arabidopsis thaliana (L.) Heynh ecotype Col-0, wild type. For the FANS procedure, transgenic line pWOX2:H2B-GFP was used, kindly provided by Michael Nodine (Wageningen University and Research). For functional analysis, insertional mutants in candidate genes MYB46, ZAT14, MYR2, MYB98, TCP19, and GRF7 (Supplementary Table S2) from NASC (The European Arabidopsis Stock Centre) and β-estradiol-induced overexpression lines 35 S::MYB46-ER, 35 S::ZAT14-ER,* 35 S::MYR2-ER*,* 35 S::MYB98-ER*, and 35 S::TCP19-ER from TRANSPLANTA collection (N2102359, N2102311, N2101927, N2102201, N2102495) were studied [31]. Homozygous mutants were selected following the NASC standard protocol (http://signal.salk.edu/tdnaprimers.2.html).
Plant cultivation and growth conditions
Donor plants for explant collection were cultivated in Jiffy-7 peat pellets (Jiffy, Zwijndrecht, Netherlands) and maintained in a walk-in phytotron with controlled environmental conditions: 22 ± 1 °C with a 16-hour light/8-hour dark photoperiod and an irradiance of approximately 100 µmol m⁻² s⁻¹ provided by white fluorescent lamps. In vitro cultures were incubated in a growth chamber under constant conditions of 22 ± 1 °C, a 16/8 h (light/dark) photoperiod, and a light intensity of 50 µmol m⁻² s⁻¹.
Induction of somatic embryogenesis
Immature zygotic embryos (IZEs) of Arabidopsis, collected at the late cotyledonary stage of development, were utilized as explants for the induction of somatic embryogenesis (SE) in vitro. IZEs were manually isolated from siliques 12–16 days after pollination using a stereomicroscope. Following surface sterilization for 20 min in a 20% (v/v) commercial bleach solution containing sodium hypochlorite, they were rinsed three times in sterile distilled water (10 min per wash).
Explants were placed on a modified B5-based medium solidified with 8 g L⁻¹ agar (Oxoid, UK) and supplemented with 20 g L⁻¹ sucrose. The basal medium (E0) was composed of B5 macro- and micronutrients (3.2 g L⁻¹; Duchefa Biochemie, Netherlands; #G0210) and adjusted to pH 5.8 prior to autoclaving. For SE induction, the E0 medium was supplemented with 5.0 µM 2,4-dichlorophenoxyacetic acid (2,4-D; Duchefa Biochemie, #D0911) to prepare the embryogenic medium (E5). In some experimental variants, auxin-free E0 medium was used as a control.
Ten explants were cultured per 90 mm Petri dish and incubated under controlled environmental conditions (22 ± 1 °C, 16 h light/8 h dark photoperiod, ~ 50 µmol m⁻² s⁻¹). Each treatment was performed in at least three biological replicates, each comprising thirty explants. After 21 days of culture, the embryogenic response was quantified by evaluating (i) SE efficiency, defined as the percentage of explants that had formed somatic embryos, and (ii) SE productivity, calculated as the mean number of somatic embryos produced per responsive explant. Statistical significance was assessed using one-way ANOVA with Dunnett’s multiple comparison test for comparisons of multiple genotypes against a control (Col-0), and two-way ANOVA with Bonferroni’s multiple comparison test for experiments involving two independent variables (e.g., genotype and culture condition), allowing evaluation of both main effects and their interaction. Graphs display the means with standard errors.
To induce MYB46, ZAT14, MYR2, MYB98, and TCP19 overexpression in the 35 S::MYB46-ER, 35 S::ZAT14-ER,* 35 S::MYR2-ER*,* 35 S::MYB98-ER*, and 35 S::TCP19-ER lines, explants were transferred to plates containing E0 or E5 medium supplemented with 10 µM β-estradiol. The same media without β-estradiol were used for controls.
Genotyping of insertional mutants
To determine the genotype in plants of the insertional mutant lines obtained from NASC, genomic DNA of individual T3 seedlings was extracted from leaf tissue using a modified version of the micro-CTAB method [68]. Genotyping was performed using PCR with gene-specific forward and reverse primers, in combination with an insertion-specific primer, designed based on publicly accessible databases (http://signal.salk.edu/tdnaprimers.html). The expected sizes of amplification products allowed discrimination between homozygous, heterozygous, and wild-type alleles for each analyzed line.
Fluorescence-activated nuclei sorting (FANS)
FANS was conducted according to [26] with slight modifications. To isolate nuclei from cells undergoing SE, IZEs of Arabidopsis carrying the pWOX2::H2B-GFP reporter construct were cultured on E5 medium. After 3 days of culture, ~ 150 explants per replicate were collected and immediately transferred onto ice into freshly prepared Galbraith buffer (GB) with RiboLock RNase inhibitor (Thermo Scientific #EO0381; final concentration 1 U/µL) [26]. The hypocotyls were chopped out with sterile blades before the upper parts of the explants were transferred to a new Petri dish on ice with GB buffer. To isolate nuclei, the upper parts of the explants were manually chopped with a double-sided razor blade for 4 min, then for an additional 4 min with a new blade. The resulting suspension was filtered through a 30 μm nylon mesh (Sysmex #04–0042-2316) and centrifuged at 1,500 g for 15 min at 4 °C. The pellet containing the nuclei was resuspended in staining buffer (SB) with DAPI (final concentration: 5 µg/µL) and incubated on ice for 15 min. The stained nuclei were again passed through a fresh 30 μm filter and transferred to cooled FACS tubes on ice (Sarstedt #55.484.001).
FANS was performed using a BD FACS Aria™ III sorter (70 μm nozzle). Gating parameters for DAPI and GFP fluorescence were set using wild-type nuclei as negative controls, ensuring that the GFP-positive gate excluded background fluorescence. Sorting was performed directly into a Bio-Rad HSP-9601 plate with buffer for SMART-Seq v3 library preparation from a small number of cells. GFP-positive and GFP-negative nuclei, isolated under the control of the WOX2 promoter, were then isolated for downstream analysis.
RNA isolation
To assess RNA quality and integrity, three additional samples from each genotype were directly sorted into TRIzol LS (Ambion, #10296028). RNA was prepared according to the manufacturer’s recommendation. Amount and quality of RNA was determined on an Agilent Plant RNA Pico Chip (Bioanalyzer/Agilent Technologies).
Library preparation and sequencing
mRNA-seq libraries were prepared by the VBCF NGS Unit with SMARTer Ultra Low Input RNA Kit for sequencing-v3 (Clontech) according to the manufacturer’s recommendations. The sequencing was performed on an Illumina NovaSeq 2000 SR100. Raw sequencing reads were subjected to adapter trimming using Trim Galore (v0.5.0; Babraham Bioinformatics), which automatically detected and removed adapter sequences. Trimming parameters required a minimum overlap of four nucleotides between read and adapter, and only reads longer than 18 bases were retained for further analysis. Transcript quantification was performed using the pseudoaligner Kallisto (v0.45; [69]). A custom index was built from Arabidopsis thaliana Araport11 annotations, including categories such as protein-coding genes, non-coding genes, pseudogenes, transposable element (TE) genes, and novel transcribed regions. Transcript sequences were extracted from the TAIR10 reference genome using bedtools getfasta (v2.27.1; [70]). Differential expression analysis was performed using DESeq2 (v1.16) [71]. Comparisons between GFP-positive and GFP-negative samples were carried out with the Wald test, and genes with an adjusted false discovery rate (FDR) below 0.05 were considered significantly differentially expressed. Functional enrichment of differentially expressed genes was assessed with the AmiGO2 platform and the PANTHER classification system [72].The bubble plots were generated using the SRplot web server [30].
Supplementary Information
Supplementary Material 1: Figure S1. The expression pattern of WOX2 on the 3rd day of SE in the upper part of the explant (IZE), which was taken for the FANS procedure. Expression of pWOX2::H2B-GFP (green) after 3 days of SE induction in two different representative explants. Scale bars represent 100 µm. c- cotyledon.
Supplementary Material 2: Figure S2. Expression of WOX2, GFP, and housekeeping genes: PP2A (AT1G13320), PPR (AT5G55840), and the PP2AA3 (AT1G13320). Asterisks indicate significantly different expression (Two-way ANOVA with Tukey multiple comparisons test p < 0.05) between positive and negative fraction of nuclei.
Supplementary Material 3: Figure S3. The embryogenic potential of the 35S::MYB46-ER, 35S::ZAT14-ER, 35S::MYR2-ER, 35S::TCP19-ER, 35S::MYB98-ER, and 35S::GRF7-ER lines. The explants were cultured on an auxin-free E0 medium supplemented with β-estradiol to induce the overexpression of MYB46, ZAT14, MYR2, TCP19, and MYB98. The efficiency of SE is represented by % values.
Supplementary Material 4: Figure S4. FANS result separating GFP-labeled nuclei from SE-undergoing cells in the upper part of the IZE explant (left), the corresponding wild type as a negative control (right (A) DAPI profile revealing different ploidy levels; (B) distribution of events; (C) statistical data.
Supplementary Material 5: Table S1. Specification of T-DNA insertion mutant lines for genes potentially involved in somatic embryogenesis induction in Arabidopsis thaliana. All T-DNA lines are in the Col-0 background.
Supplementary Material 6: Table S2. List of differentially expressed genes (DEGs) identified between embryogenic nuclei (pWOX2::H2B-GFP positive) and non-embryogenic nuclei (pWOX2::H2B-GFP negative) isolated by FANS.
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1Krasnoperova EY, Tvorogova VE, Smirnov KV, Efremova EP, Potsenkovskaia EA, Artemiuk AM, et al. Mt WOX 2 and Mt WOX 9-1 effects on the embryogenic callus transcriptome in medicago truncatula. Plants. 2024;13. 10.3390/plants 13010102.10.3390/plants 13010102 PMC 1078091738202410 · doi ↗ · pubmed ↗
- 2Shivani AP, Sharma V, Kaur N, Kaur N, Pandey P, et al. Genome-wide analysis of transcription factors during somatic embryogenesis in banana (Musa spp.) Cv. Grand Naine P Lo S One. 2017;12. 10.1371/journal.pone.0182242.10.1371/journal.pone.0182242 PMC 555228728797040 · doi ↗ · pubmed ↗
- 3Aguilar-Martlnez JA, Sinha N. Analysis of the role of arabidopsis class I TCP genes At TCP 7, At TCP 8, At TCP 22, and At TCP 23 in leaf development. Frontiers in Plant Science. 2013. 10.3389/fpls.2013.0040610.3389/fpls.2013.00406 PMC 379744224137171 · doi ↗ · pubmed ↗