Figure 1 shows the life cycle of the annual Ae. arabicum, which is characterized by the production and dispersal of dimorphic diaspores. The M⁺ seed diaspore is dispersed by dehiscence (pod shatter), and upon imbibition, it produces a mucilaginous layer during germination ( Figure 1B ). The IND fruit diaspore is dispersed by abscission and constitutes an indehiscent fruit in which the single M⁻ seed is covered by the pericarp (fruit coat), which confers coat dormancy and prevents or delays germination ( Figure 1C ). Earlier work with Ae. arabicum TUR (Chandler et al., 2024) revealed the molecular and morphological mechanisms underpinning the distinct dormancy and germination responses of the dimorphic diaspores to different imbibition temperatures. A comparison of the maximal germination percentages (G_max) and the germination rates (speed) of germination (GR₅₀; i.e., the inverse of the time of the diaspore population to reach 50% radicle protrusion) identified 14°C as the optimal temperature for the highest germination speed of the M⁺ seed morphs, the bare M⁻ seeds (manually extracted from IND fruits by pericarp removal), and the IND fruit morphs ( Supplementary Figure 1 ). The IND fruit morph, however, exhibited a degree of pericarp-imposed dormancy across the entire temperature window. Thus, our initial physiological experiments investigating the effects of abiotic stress factors on seedling growth focused on comparing M⁺ and M⁻ seedlings derived from germinated M⁺ and bare M⁻ seeds, respectively. These experiments, involving the removal of the pericarp to obtain exposed M⁻ seeds, align with dimorphic systems (see the Introduction) characterized by true seeds of distinct colors as diaspores. In such systems, the resulting seedlings exhibit differential responses to salinity, with the unique seedling reactions having already been established during dimorphic seed development (Xu et al., 2011; Zhang et al., 2021; Cao et al., 2022; Song et al., 2023). In our subsequent comparative biomechanical and transcriptome analyses with Ae. arabicum TUR, we compared the seed–seedling transition phase and early growth of seedlings derived from germinated M⁺ seeds, bare M⁻ seeds, and IND fruits. Adult plants grown from M⁺ and M⁻ seedlings are indistinguishable from each other and produce dimorphic diaspores ( Figure 1 ).

To study the effects of thermal stress on M⁺ and M⁻ seedling growth independently from the temperature effects on dimorphic diaspore germination ( Supplementary Figure 1 ), M⁺ and M⁻ seeds were imbibed at an optimal germination temperature, and seeds that had just completed germination (1-mm radicle protrusion visible) were transferred to agar plates for conducting the seedling growth assay ( Figure 2 ). M⁺ and M⁻ seedlings grown on vertical agar plates at 14°C, 20°C, 24°C, 30°C, and 35°C were compared for their root and shoot lengths at the times indicated [in “hat” (hours after transfer)], and seedling growth rates were calculated. Though the total seedling length differed significantly in several cases between the morphs, the difference was very small, and there was no overall morph-specific physiological response. Similarly, seedling growth rates did not differ between the M⁺ and M⁻ morphs ( Figure 2A ). The fastest seedling growth rate was observed at 72 hat, and 30°C was identified as the temperature for fastest growth leading to the longest seedlings at 240 hat ( Figure 2B ; Supplementary Figure 2 ). Fresh and dry weights of separated root and shoot tissue at the end of the experiment (10 days) revealed a strong temperature response (p < 0.001), with the highest masses observed at 30°C ( Supplementary Figure 2 ). Together with seedling lengths, growth rates, and fresh weights, the most optimal seedling growth condition considered was thus 30°C, likely corresponding with the maximum climatic temperatures of sites in Turkey from which Ae. arabicum TUR was collected (Arshad et al., 2019). At 35°C, there were slight indications that M⁺ seedlings had greater vigor, with a higher mean shoot fresh and dry weight than M⁻ seedlings ( Supplementary Figure 2 ). Temperature, therefore, had a profound effect on M⁺ and M⁻ seedling growth, in particular on post-germination pre-emergence growth until 72 hat ( Figure 2 ).

To study the effect of osmotic stress on the growth of Ae. arabicum M⁺ and M⁻ seedlings, their growth was analyzed at lowered water potentials using high-molecular-weight polyethylene glycol (PEG). After more than 3 weeks of vertical growth, seedling morphs did not differ in their total length and growth rates under three different concentrations of PEG ( Supplementary Figure 3 ). Taken together, no differences in the physiological responses of M⁺ and M⁻ seedlings derived from M⁺ seeds and bare M⁻ seeds to temperature and reduced water potential were identified. This suggests that during the seed–seedling transition, the observed dimorphic diaspore trait differences observed in abiotic stress responses (Lenser et al., 2016, 2018; Arshad et al., 2019; Wilhelmsson et al., 2019; Bhattacharya et al., 2019a, b; Arshad et al., 2020, 2021; Fernandez-Pozo et al., 2021; Chandler et al., 2024) may be reset to a large extent. To test this hypothesis, we conducted transcriptome analysis.

Figure 3A depicts the experimental design for the RNA-seq analysis in which we compared the transcriptomes during the late germination phase with those during early seedling growth. Samples were collected from imbibed Ae. arabicum M⁺ seeds, M⁻ seeds, and IND fruits at T_1%, i.e., the time for the onset of the completion of germination of a seed population, and at T_100%, i.e., the time when the entire population had completed germination by visible radicle protrusion ( Figure 3A ). It is known from previous work (Wilhelmsson et al., 2019; Arshad et al., 2021; Chandler et al., 2024) that the pericarp of imbibed IND fruits is dead tissue that does not contain any RNA and that dry seed and germination transcriptomes until T_1% differ considerably between the dimorphic diaspores. The physiological sampling times for T_1% and T_100% of M⁺ seeds, bare M⁻ seeds, and IND fruits are indicated in Figure 3A . Diaspores that had just completed germination (1-mm radicle protrusion) were transferred to agar plates at T_100% (0 hat) for conducting the seedling growth assay. Root and shoot tissues were harvested from seedling samples at key physiological stages (72 hat and 240 hat) during early seedling growth ( Figure 3 ).

The 72-hat pre-emergence growth samples correspond to the maximal growth rate, and 72 hat was associated with the onset of cotyledon greening by chlorophyll accumulation ( Figure 3B ). Between 6 and 10 days, the growth rates of M⁺ and M⁻ seedlings derived from germinated M⁺ seeds and bare M⁻ seeds, respectively, remained roughly equal. M⁺ and M⁻ seedlings derived from germinated M⁺ seeds and bare M⁻ seeds had straight hypocotyls. In contrast to this, seedlings derived from IND fruits often exhibited a bent lower hypocotyl connected with an overall slower shoot and root growth ( Figure 4A ). To investigate if morph caused a long-lasting effect on the biomechanical properties of the hypocotyls, we conducted comparative hypocotyl tensile tests of 15-day-old seedlings ( Figure 4B ). No significant differences in the hypocotyl breaking force were evident between the seedlings derived from germinated M⁺ seeds, bare M⁻ seeds, and IND fruits ( Figure 4C ). The observed difference in hypocotyl shape (bent versus straight) and the slower shoot and root growth of IND fruit-derived seedlings, therefore, did not affect the hypocotyl biomechanical properties at this stage of seedling growth.

To provide insights into the association between RNA-seq samples, datasets were visualized using principal component analysis (PCA), based on the 500 genes with the highest variance. Replicate RNA-seq samples clustered tightly by diaspore and by organ (root and shoot) type ( Figure 5A ). As expected, three distinct clusters were separated primarily by the derivation of the samples from seed, root, or shoot tissue based on the first two components, explaining 48% and 44% of the variability. In this combined analysis ( Figure 5A ), IND samples at T_100% (germinated diaspores) remain distinct from all other tissues and time points. Separation of samples involved in key processes of germination (T_1% and T_100%), root growth (72 and 240 hat), and shoot growth (72 and 240 hat) revealed comparative transcriptome profiles between the morphs in greater detail ( Figure 5 ). Clear differences were observed with IND samples, such that both T_1% and T_100% samples clustered separately from M⁺ and M⁻ seed samples. M⁺ and M⁻ seed samples, however, showed tight correlations throughout the course of germination ( Figure 5B ). As seedlings, differences between morphs appear to be smaller. M⁺ and M⁻ samples clustered together during the two time points during root ( Figure 5C ) and shoot ( Figure 5D ) growth. IND root samples remain distinct from M⁺ and M⁻ samples. However, transcriptional profiles of IND shoot tissue suggest that while samples at 72 hat remain as a separate cluster, by 240 hat, there is a tendency toward greater similarity of M⁺ and M⁻ samples.

Cleaned RNA-seq reads mapped to the Ae. arabicum genome were obtained for 23,594 genes ( Supplementary Dataset 1 ). To make the transcript abundance data easily and publicly accessible, a gene expression atlas was generated, which was implemented in the Ae. arabicum genome database (DB) (Fernandez-Pozo et al., 2021) at https://plantcode.cup.uni-freiburg.de/aetar_db/index.php. The Ae. arabicum gene expression atlas includes the transcriptome results of this work and work published earlier (Merai et al., 2019; Wilhelmsson et al., 2019; Arshad et al., 2021; Chandler et al., 2024) and allows adding future transcriptome datasets. The transcript abundance data for the 23,594 Ae. arabicum genes were further investigated, and differentially expressed genes (DEGs; Supplementary Dataset 1 ) were detected in a strict consensus (overlap) approach using an adjusted p-value cutoff set to 0.001 (Wilhelmsson et al., 2019). Pairwise comparisons of M⁺ seeds vs. bare M⁻ seeds (seeds only), M⁺ seeds vs. IND fruits (natural dispersal units), and M⁻ vs. IND (pericarp effect) allowed transcriptome exploration of the dimorphic syndrome during seed germination and seedling growth. Comparisons showed that M⁺ and M⁻ seed transcriptomes became remarkably similar during the completion of germination ( Figure 6A ; Supplementary Table 1 ). A total of 180 and 55 DEGs were detected at T_1% and T_100%, respectively, while after 72 hat, the seedling transcriptomes were almost identical in the root (three DEGs) and shoot (two DEGs). By contrast, comparisons between the natural diaspores, M⁺ seed vs. IND fruit, showed a much higher number of DEGs during germination. A total of 2,041 DEGs during T_1% increased to 2,682 by T_100%, thereafter reducing 10-fold (277 root DEGs) and 16-fold (164 shoot DEGs) by 72 hat. By 240 hat, the time of true leaf emergence for seedlings, differences between M⁺ seed vs. IND fruit-derived seedlings were only evident from 60 DEGs in root samples and 10 DEGs in shoot samples ( Figure 6B ; Supplementary Table 1 ).

Interestingly, a similar but more pronounced pattern of DEGs was observed in the comparison of the pericarp effect (bare M⁻ seed vs. IND fruit) on M⁻ seed germination and seedling growth ( Figure 6C ; Supplementary Table 1 ). The number of DEGs at T_1% was similar to that detected in the M⁺ vs. IND comparison (2,041 vs. 1,938). The greatest differences were observed at T_100% (3,228 DEGs) at the onset of pre-emergence growth. Here, the number of upregulated DEGs in M⁻ seeds increased 1.4-fold (to 1,220 DEGs), while the number of downregulated DEGs increased 1.9-fold (to 2,008 DEGs). As seedlings progressed through root and shoot growth, the pericarp imposed a total of 347 root-specific DEGs at 72 hat, while shoot-specific DEGs were lower (64). By 240 hat, differences in shoot samples comprised four up- and four downregulated DEGs, while 62 up- and 69 downregulated DEGs were detected in the root tissues ( Figure 6C ; Supplementary Table 1 ). Taken together, these results suggest the tendency toward transcriptional “resetting” of seedling morphs mainly during the post-germination pre-emergence growth phase.

In the ecologically relevant comparison (M⁺ seed vs. IND fruit), shoot resetting occurred earlier (by 72 hat), and root resetting occurred later (differences still evident at 240 hat). Comparison of M⁺/IND DEG lists ( Supplementary Dataset 1 ) between the T_100% (germinated diaspores) and the T_1% (ungerminated diaspores) time points revealed that about one-third of the DEGs are overlapping and two-thirds are unique to either T_100% or T_1% ( Figure 7A ). Comparison of the M⁺/IND and M⁻/IND DEG lists at T_100% revealed that while the majority of the DEGs are overlapping, there is also a considerable number of DEGs unique to either M⁺/IND or M⁻/IND ( Figure 7A ). For the ecologically relevant comparison (M⁺/IND), the comparison between the T_100% seed and the 72-hat seedling state delivered root and shoot DEGs common and unique for pre-emergence seedling growth ( Figure 7B ). Comparison of M⁺/IND DEG lists during pre-emergence seedling growth revealed that most of the DEGs at 72 hat are unique for root and shoot, a finding that strongly suggests that the two compartments are distinct ( Figure 7C ). To investigate the effect of the pericarp on the pre-emergence seedling growth at M⁺/IND and M⁻/IND, DEG lists for root and shoot at 72 hat were compared and delivered DEG lists representing pericarp-dependent and pericarp-independent mechanisms ( Figure 7B ; Supplementary Dataset 1 ). These comparative DEG lists (M⁺/IND vs. M⁻/IND) contained for roots a total of 190 up- and 226 downregulated genes and for shoots a total of 43 up- and 138 downregulated genes.

To gain insight into the processes differing during the transition from germinated diaspores to early seedlings derived from M⁺, M⁻, and IND, Gene Ontology (GO) term enrichment analysis of DEG lists was performed comparing M⁺ seeds versus IND fruits and M⁻ seeds versus IND for the T_100% (germinated diaspores) and the 72 hat (root and shoot separately) stages ( Figure 8 ; Supplementary Figure 4 , Supplementary Dataset 2 ). Broadly, selected identified GO terms were categorized by key processes identified as differing between seedlings originating from the different diaspores ( Figure 8 ; Supplementary Figure 4 ). For example, genes up in roots of seedlings 72 hat derived from M⁺ or M⁻ seeds compared to IND diaspores were significantly enriched in GO terms related to nitrates (e.g., nitrate assimilation), cell wall (e.g., cellulose catabolic process), and transport (e.g., regulation of ion transport) (clusters 1 and 5). Interestingly, some cell wall and nitrate-related terms (e.g., cell wall pectin metabolic process and nitrate transport) were enriched in genes more highly expressed in 72-hat seedlings derived from IND compared to M⁺ and M⁻ (clusters 8 and 5). A pronounced effect of pericarp at the T_100% stage was evident in the enrichment of abiotic stress-related GO terms “response to oxidative stress”, “response to salt stress”, and “anaerobic respiration”. Terms identifying hormone signaling (e.g., ethylene, auxin, gibberellin, and ABA) were also identified ( Figure 8 ; Supplementary Figure 4 ; Supplementary Dataset 2 ). Overall, it was evident that pericarp presence at the diaspore stage was a driver of differences in gene expression, with similar differences evident when comparing M⁺ or M⁻ seeds against IND diaspores. However, some contrasts were more evident when comparing 72-hat seedlings derived from M⁺ to IND (rather than M⁻ to IND), suggesting seed morph-specific responses that were not dependent on pericarp presence at imbibition, particularly evident in the enrichment of GO terms “flavonol biosynthetic process” and “anthocyanin-containing compound biosynthetic process”. The pre-emergence seedling growth DEGs from these lists of enriched GO categories ( Figure 8 ; Supplementary Figure 4 ) and a comparison to early seed germination and dormancy (Chandler et al., 2024) are the focus of the following analysis into the resetting of dimorphic expression patterns.

In earlier work, we identified morph-specific expression patterns of hormone-related genes during seed and fruit development (Arshad et al., 2021) and dormancy and germination of the dimorphic diaspores (Chandler et al., 2024). This revealed the importance of ABA metabolism and signaling in dimorphic diaspore development and germination, and key roles for ABA and hypoxia in imposing pericarp-imposed dormancy in IND fruits. Figure 9A shows that when IND fruits were compared to M⁺ and M⁻ seeds, the distinct expression patterns of hypoxia-responsive genes in seeds were reset during the transition from the germination (T_100%) to the early seedling (72 hat) phase. This was evident for the hypoxia-regulated transcription factor (TF) genes, such as AearNAC102 and AearERF71/73, and downstream genes, such as the ethanolic fermentation enzyme gene AearADH1a (Chandler et al., 2024). The Ae. arabicum genes presented in Figure 9A were identified as part of the core hypoxia-responsive gene list derived from hypoxia-treated Arabidopsis thaliana seedlings (Christianson et al., 2009; Gasch et al., 2016; Lee and Bailey-Serres, 2019), and the resetting of their expression patterns, therefore, also indicates the absence of hypoxia once the M⁻ seedlings have emerged from the IND pericarp. Most genes involved in ethylene biosynthesis, 1-aminocyclopropane-1-carboxylic acid (ACC) oxidase (ACO) and ACC synthase (ACS), and many genes encoding ethylene response factor (ERF) TFs also exhibited resetting during pre-emergence seedling growth ( Figure 9A ; Supplementary Figure 5A ). However, for genes encoding the ACC oxidases AearACO2 and AearACO4, the ethylene receptor AearETR2, and several ERF TFs (AearERF2, AearERF11, AearERF113/RAP2.6, and AearRAP2.11), distinct transcript abundances were retained in 72-hat shoots or roots ( Supplementary Figure 5A ). The A. thaliana homologs of these ERF TFs are known to be involved in the control of seedling growth by modulating ABA, ethylene, gibberellin, and auxin signaling (Kim et al., 2012; Zhou et al., 2016; Lorrai et al., 2018; Zhu et al., 2020; Templalexis et al., 2022). AtRAP2.11 is, in addition, known as a major regulator of potassium and nitrate transporters in responses to low-nutrient conditions (Kim et al., 2012; Meng et al., 2016; Templalexis et al., 2022).

In contrast to ethylene-related genes, for all ABA-related metabolism and signaling genes, as well as for the PYR/PYL/RCAR-type ABA receptor genes, resetting of their expression was complete in 72-hat seedlings ( Figure 9B ; Supplementary Figure 5B ). An exception was the ABA-responsive element (ABRE)-binding protein/factor (ABF) AearAREB3b for which higher transcript abundances prevailed in 72-hat M⁻ seedling shoots, but this difference disappeared in 240-hat seedlings ( Supplementary Figure 5B ). Morph-specific expression patterns for ABA biosynthesis [e.g., 9-cis-epoxycarotenoid dioxygenase (NCED) genes] and ABF-type TFs, and ABA contents were a hallmark during the dimorphic seed/fruit development (Lenser et al., 2018; Arshad et al., 2021) and in imbibed dimorphic diaspores (Chandler et al., 2024). Differences in ABA relations were, therefore, most important between dimorphic diaspores but became less important between M⁺ and M⁻ seedlings since they were reset during pre-emergence seedling growth. In contrast to this, distinct expression patterns for gibberellin (GA)-related genes prevailed in 72-hat seedlings ( Figure 9B ; Supplementary Figure 6A ). The transcript abundances for the GA 3-oxidase (biosynthesis of bioactive GA) gene AearGA3ox1 and the AearGID1b gene encoding a GA receptor were higher in 72-hat M⁻ seedling shoots derived from IND fruits compared to M⁺ seedling shoots, suggesting that GA biosynthesis and sensitivity differ between the morphs during early seedling growth. Similarly, the transcript abundances of GASA (“GA-stimulated Arabidopsis”) genes known to be stimulated by GA and regulated by DELLA repressor proteins (Zhang and Wang, 2017) were higher in 72-hat M⁻ seedling roots derived from IND fruits compared to M⁺ seedling roots, while genes for DELLA repressor proteins did not differ between the morphs ( Figure 9B ; Supplementary Figure 6A ). In general, ABA has an inhibitory and GA has a promoting role in the complex hormonal control of seedling shoot and root growth (Ahammed et al., 2020). We conclude from the Ae. arabicum results that for most genes, resetting occurs during the post-germination pre-emergence growth phase (T_100% seeds to 72-hat seedlings) but also that several ethylene and GA-related genes involved in hormonal interactions are among the DEGs in 72-hat seedlings for which resetting occurs later during seedling growth ( Figure 7 ).

That Ae. arabicum shoot and root development changed during pre-emergence seedling, which was also evident from the chlorophyll accumulation ( Figure 3 ), the DEG list comparison ( Figure 7D ), and the shoot-specific induction of chlorophyll-related genes ( Supplementary Figure 7A ). Interestingly, these chlorophyll-related genes were DEGs in 72-hat shoots with lower expression in M⁻ seedlings derived from IND fruits. Ethylene, GA, and ABA interact with auxin to regulate seedling growth differently in shoots and roots (Belin et al., 2009; Hu et al., 2017; Ahammed et al., 2020). Auxin/indole-3-acetic acid (Aux/IAA) proteins (IAAs) repress auxin-inducible genes by inhibiting auxin response TFs (ARFs). In Ae. arabicum, several ARFs and Aux/IAA proteins were DEGs during late germination with lower (ARFs) and higher (IAAs) transcript abundances in germinated IND fruits (T_100%) when compared with germinated M⁺ and M⁻ seeds ( Figure 10A and Supplementary Figure 6B ). With the exception of AearARF6, which was more highly expressed in IND fruit-derived 72-hat and 240-hat seedling shoots, these expression differences had disappeared in 72-hat seedlings. The A. thaliana dormancy/auxin-associated protein is known to be involved in the auxin sensitivity of seedlings (Johnson et al., 2015); its Ae. arabicum homolog was more highly expressed in germinated IND fruits ( Figure 10B ). Small auxin-upregulated RNA (SAUR) genes are auxin-responsive genes involved in cell elongation growth and other processes (Stortenbeker and Bemer, 2019). An example in A. thaliana seeds is AtSAUR11, which accumulates eightfold in the endosperm as compared to the embryo (Dekkers et al., 2013). In general, the differential expression of Ae. arabicum SAUR genes during late germination (T_100%) was reset in 72-hat seedlings ( Figure 10B ; Supplementary Figure 6B ). A notable exception was AearSAUR11 for which the transcript abundance was high in germinated IND fruits (T_100%) and increased further >10-fold in 72-hat M⁻ seedling shoots derived from IND fruit as compared to M⁺ seedling shoots ( Figure 10B ).

Polar auxin transport and auxin homeostasis are key to seedling shoot and root growth and are facilitated by tightly regulated efflux [PIN (PIN-FORMED proteins) and PILS (PIN-LIKES)] and influx (AUX1) carriers, which coordinate cell type-specific asymmetric subcellular auxin localization and local auxin gradients across tissues (Belin et al., 2009; Hu et al., 2017; Yi et al., 2021; Bogaert et al., 2022; Feraru et al., 2022; Nakabayashi et al., 2022; Waidmann et al., 2023). Several auxin carriers including AearAUX1, AearABCG36 ( Figure 10B ), AearPILS5, and AearPILS6 ( Figure 11A ) were identified as DEGs during late germination (T_100%), but their expression difference was reset during seedling pre-emergence growth. In contrast to these auxin transporters, AearPILS3, which was also a DEG in germinated diaspores (T_100%), remained as a DEG in shoots at 72 hat ( Figure 11A ). AearPILS7 was not expressed during diaspore germination, but it was induced afterward and identified as a DEG in roots during pre-emergence seedling growth. The transcript abundances for AearPILS7 were lower in 72-hat and 240-hat seedling roots derived from germinated IND fruits as compared to seedling roots derived from germinated M⁺ and M⁻ seeds ( Figure 11A ). In A. thaliana seedlings, AtPILS7 is involved in fine-tuning stress-responsive root auxin signaling in response to phosphate availability and regulation of phosphate transporter gene expression (Yi et al., 2021). In phylogenies across the plant kingdom, the PILS protein sequences of embryophytes plus green algae form a clade distinct from the PIN proteins, and the seven A. thaliana and other embryophyte PILS protein sequences are distributed over two PILS subclades 2 and 3 (Feraru et al., 2012; Bogaert et al., 2022). Figure 11B shows a family-wide phylogenetic analysis of Brassicaceae PILS protein sequences with the five identified Ae. arabicum protein sequences distributed across the two subclades (one in subclade 2 and four in subclade 3). In addition to PIN and PILS efflux carriers, which are specific for indole-3-acetic acid (IAA) transport, ABCG36 is known to act as a plasma membrane located exporter for the IAA precursor indole-3-butyric acid (Geisler et al., 2017; Aryal et al., 2019; Nakabayashi et al., 2022). Based on the AearABCG36 expression patterns, it is a germination DEG that is reset during pre-emergence growth ( Figure 10B ).

Auxin and other hormones regulate the transition from germination to seedling growth by affecting the expression patterns of water and ion transporters (Yi et al., 2021; Tan et al., 2022; Templalexis et al., 2022). This includes aquaporins, e.g., plasma membrane intrinsic proteins (PIPs) and tonoplast intrinsic proteins (TIPs), transporting water, ammonia (NH₃), and other solutes (Loqué et al., 2005; Footitt et al., 2019; Hoai et al., 2020). Figure 12A shows that the transcript abundances of AearPIP3A/2;7 and AearPIP1E/1;4 were more highly expressed in IND fruits during late germination (T_100%) and remained higher in 72-hat seedling roots derived from IND fruits as compared to seedling roots derived from M⁺ and M⁻ seeds. AearTIP1,4, for which the A. thaliana homolog is regulated by ABA in seeds (Footitt et al., 2019), and AearTIP1/1,1 were reset during pre-emergence growth ( Figure 12A ). In contrast to this, expression of the NH₃ transporter genes AearTIP2;3a and AearTIP2;3b was higher in IND fruits during late germination (T_100%) and remained higher in 72-hat seedling roots derived from IND fruits as compared to seedling roots derived from M⁺ and M⁻ seeds ( Figure 12A ). In A. thaliana seedlings, root high-affinity nitrate (NO₃ ⁻) transporters such as NRT2 interact with polar auxin transport (Wang et al., 2023a). AearNTR2 and AearNTR3.1 were not expressed during Ae. arabicum germination, and their root-specific induction in 72-hat seedlings was lower in roots derived from IND fruits ( Figure 12B ).

Nitrate and potassium transporter gene expression in A. thaliana seedlings is mediated by auxin and the ERF TF RAP2.11 in responses to low-nutrient conditions (Kim et al., 2012; Meng et al., 2016; Templalexis et al., 2022). The expression patterns of AearRAP2.11 ( Supplementary Figure 5A ) and the transporters for nitrate AearNTR2 ( Figure 12B ) and potassium AearKUP3 ( Supplementary Figure 7A ) showed reduced transcript abundances in 72-hat roots derived from IND fruits, which suggest a role of RAP2.11 in the regulation of potassium and nitrate relations during pre-emergence seedling growth. Key enzymes for nitrate assimilation include nitrate reductase (NR) and nitrite reductase (NRI) for which the expression is regulated by hormones and abiotic stresses (Tang et al., 2022). As for the nitrate transporters, also the genes for nitrate assimilation enzymes AearNR1, AearNR2, and AearNIR1 were not expressed during germination, and their induction in 72-hat seedling roots was lower in roots derived from IND fruits ( Figure 12B ). This suggests that nitrate transport and assimilation differ in 72-hat roots of seedlings derived from IND fruits and M⁺ seeds and that instead of resetting, a distinct reprogramming occurred.

The dimorphic expression patterns of genes encoding enzymes of the flavonoid biosynthesis pathway and proanthocyanidin (PA) accumulation are examples of root-specific DEGs during Ae. arabicum pre-emergence growth at 72 hat for which the transcript abundances were higher in seedlings derived from M⁺ seeds as compared to seedlings derived from M⁻ seeds or IND fruits ( Figure 13 ). In A. thaliana, mutations in many of the flavonoid biosynthetic genes as well as in genes of the MYB-bHLH-WDR (MBW) protein complex regulating flavonoid biosynthesis lead to “transparent testa (tt)” mutant phenotypes with reduced seed dormancy (Lepiniec et al., 2006; Macgregor et al., 2015; Xu et al., 2015). A. thaliana TT19 encodes a glutathione-S-transferase (GST)-like protein that functions as a carrier to transport anthocyanin and PA precursors and is involved in the accumulation of PAs in the seed coat (Kitamura et al., 2010; Sun et al., 2012). The AearGSTF12/TT19 transcript abundances during seed germination were high in IND fruits, while there was no expression during M⁺ and M⁻ seed germination ( Figure 13A ). This expression pattern changed during seedling pre-emergence growth associated with the induction of the flavonoid biosynthetic pathway enzymes, which led to higher expression in 72-hat roots of seedlings derived from M⁺ seeds.

The induction of the flavonoid biosynthesis pathway ( Figure 13B ) genes by the MBW complex was associated with enhanced expression of the WDR protein AearTTG1 ( Figure 13C ). Its expression pattern during germination supports a role in the enhanced expression of AearGSTF12/TT19 in IND fruits, but it does not explain the higher expression of AearGSTF12/TT19 and the flavonoid biosynthesis pathway enzymes in M⁺ seed-derived seedling roots at 72 hat ( Figure 13A ). No bHLH and no MYB TF with enhanced expression in M⁺ seed-derived seedling roots at 72 hat were identified. Several Ae. arabicum MYB TFs were identified as DEGs during late seed germination (T_100%) with higher expression in IND fruits compared to imbibed M⁺ and M⁻ seeds, but for all of them, resetting occurred during pre-emergence growth ( Supplementary Figures 8A, B ). Among these DEGs is MYB30, which is a key TF in A. thaliana seeds and seedlings integrating ABA, ethylene, and reactive oxygen species (ROS) signaling (Mabuchi et al., 2018; Khedia et al., 2019; Maki et al., 2019; Nie et al., 2022; Zhang et al., 2023). Other DEGs are presented and discussed in Supplementary Figure 7B .

Dormancy and maturation gene expression usually decline during the germination of non-dormant seeds. In A. thaliana and other species, the transcript abundances for the Delay of Germination 1 (DOG1) and Seed Dormancy 4-Like (SDR4L) decline in imbibed non-dormant seeds (Graeber et al., 2014; Wu et al., 2022), while Dormancy-associated protein Like 1 (DLY1) and Non-Yellowing 1/Stay-Green 1 (NYE1/SGR1) exhibit more complex expression patterns during seed imbibition (Rae et al., 2014; Wilhelmsson et al., 2019). The transcript abundances of AearDOG1, AearSDR4L, Aear DLY1, and Aear NYE1/SGR1 were higher in imbibed IND fruits as compared to imbibed M⁺ and M⁻ seeds and declined for AearDOG1 and AearSDR4L ( Figure 14A ). The observed expression difference during germination for all four genes disappeared during pre-emergence growth, and this resetting led to roughly equal transcript abundances in 72-hat seedlings.

Germin (GER) and germin-like proteins are expressed in seeds, but their functions are largely unknown (Membre et al., 2000). In Ae. arabicum, AearGER3 is a DEG during late germination, but not in seedlings, where it is expressed in a shoot-specific manner ( Supplementary Figure 9A ). Late embryogenesis-abundant (LEA) proteins accumulate late in seed development and play major roles in desiccation tolerance (Hundertmark and Hincha, 2008; Zinsmeister et al., 2020; Smolikova et al., 2022). Most genes encoding LEA proteins are known to be ABA-induced, and their transcript abundances decline during seed germination. Consistent with the ABA inducibility and higher ABA contents in IND fruits (Chandler et al., 2024), the transcript abundances of Ae. arabicum LEA genes were higher in imbibed IND fruits as compared to M⁺ and M⁻ seeds ( Supplementary Figure 9 ). Expression patterns of specific LEA and Heat Shock Protein (HSP) genes are presented in detail in Supplementary Figure 9 .

For most Ae. arabicum TFs, which were identified as DEGs between IND fruit and M⁺ seed diaspores during the germination phase, resetting of the expression pattern was observed during the post-germination pre-emergence phase ( Supplementary Figure 8 ). Examples where the expression differences during late germination (T_100%) persisted or were even intensified into the seedling phase include the auxin and ethylene-related TF genes AearARF6 and AearERF113 presented earlier ( Figure 10A ; Supplementary Figure 5A ). Resetting during the post-germination pre-emergence phase was also observed for NAC (NAM, ATAF, and CUC) TFs involved in hypoxia responses (AearNAC102, Figure 9A ) and for AearNAC40 and AearNAC60 ( Figure 14A ), which, in A. thaliana, are functionally redundant in the inhibition of dormancy (Song et al., 2022). Homologs of AearNAC3/59 ( Figure 14B ) and AearNAC87 ( Figure 14B ) are known for being involved in the primary root growth of A. thaliana seedlings, for being expressed in the endosperm during germination (Dekkers et al., 2013), and for controlling programmed cell death during root growth (Huysmans et al., 2018). In contrast to these NAC TFs, higher transcript abundances in IND fruit diaspores during germination were maintained for AearNAC2 and AearNAC32, and this difference was further intensified at the 72-hat seedling state ( Figure 14B ). In A. thaliana, NAC2 has a role in integrating environmental and hormone (auxin and ethylene) stimuli during seedling root growth (He et al., 2005) and integration of auxin signaling (Park et al., 2011), and NAC32 has a role in upstream TF in the control of seedling root elongation and ROS signaling (Maki et al., 2019).

The WRKY9 and WRKY40 TFs are known as central repressors of ABA signaling during A. thaliana seed germination and seedling growth (Wang et al., 2021) and of ROS and stress signaling (Shin and Schachtman, 2004; Van Aken et al., 2013; Arjmand et al., 2023). In Ae. arabicum, AearWRKY9 and AearWRKY40 were identified as DEGs more highly expressed in germinating IND fruits as compared to M⁺ seeds, and this expression pattern was retained in 72-hat seedlings ( Figure 14B , Supplementary Figure 8C ). WRKY51 and WRKY56 belong to a subgroup of WRKY TFs involved in controlling auxin transport during A. thaliana root development (Templalexis et al., 2022). AearWRKY51 and AearWRKY56 are DEGs with higher expression in IND fruits during the germination of the Ae. arabicum dimorphic diaspores ( Figure 14A , Supplementary Figure 8C ). While resetting of this expression difference during pre-emergence growth was observed for AearWRKY51, it was retained and further intensified for AearWRKY56 in 72-hat seedling roots ( Figure 14B ). Expression patterns of other TFs including homeobox TFs controlling seed-to-seedling phase transition node regulators and the plant-specific AT-rich sequence zinc-binding protein (PLATZ) TFs are presented in detail in Supplementary Figure 8 . Taken together, both resetting and intensification of differential expression patterns were observed during the seed-to-seedling transition for TFs from several distinct gene families. For the TFs that were identified as DEGs at the seedling stage expression, the transcript abundances were always higher in seedlings derived from IND fruits and lower in seedlings derived from M⁺ seeds.

Cell expansion growth is driven by water uptake, which is restricted unless cell wall loosening is achieved by the action of cell wall remodeling proteins (CWRPs) or apoplastic ROS (Finch-Savage and Leubner-Metzger, 2006; Steinbrecher and Leubner-Metzger, 2017; 2018). Expansins are CWRPs that disrupt non-covalent bonds that tether cell wall matrix polysaccharides to the surface of cellulose microfibrils or each other. Expansins are required for endosperm weakening and embryo elongation during germination and seedling growth (Voegele et al., 2011; Graeber et al., 2014; Boron et al., 2015; Cosgrove, 2016; Ilias et al., 2019). In agreement with the slower germination of Ae. arabicum IND fruits as compared to M⁺ seeds, the expression of AearEXPA2 and most other expansins was lower in imbibed IND fruits ( Figure 15A ; Supplementary Figure 10A ). Resetting during post-germination pre-emergence growth was observed for most expansins, which were DEGs during germination, but other expansins also exhibited differential expression in seedling roots. Xyloglucan remodeling enzymes involved in seed germination include α-xylosidase (αXYL) (Shigeyama et al., 2016), β-xylosidase (βXYL) (Arsovski et al., 2009), and xyloglucan endotransglucosylase/hydrolase (XTH) (Voegele et al., 2011; Endo et al., 2012; Graeber et al., 2014; Steinbrecher and Leubner-Metzger, 2017; Holloway et al., 2021). In agreement with roles in imbibed Ae. arabicum dimorphic diaspores, AearαXYL1, AearβXYL1, and AearβXYL2 are DEGs during germination with lower expression in IND fruits ( Figure 15A ). Resetting during post-germination pre-emergence growth leads to roughly similar expression in seedlings derived from germinated IND fruits and M⁺ seeds.

Earlier work by Chandler et al. (2024) demonstrated that during early germination (until T_1%), the cumulative expression of all XTHs in imbibed Ae. arabicum dimorphic diaspores is lower in IND fruits compared to M⁺ and M⁻ seeds. Interestingly, the transcript abundances of many AearXTH genes increased in IND fruits during the late germination (until T_100%), and these were, therefore, identified as germination DEGs ( Figure 15B ; Supplementary Figure 10A ). In general, resetting during post-germination pre-emergence growth was observed for XTH genes, but AearXTH12, AearXTH23, AearXTH31, and AearXTH26 were DEGs during seedling growth, and their transcript abundances were lower in 72-hat seedling roots derived from germinated IND fruits compared to seedling roots derived from M⁺ or M⁻ seeds. In A. thaliana seedlings, AtXTH23 is known to be involved in root development and adaptation to salt stress (Xu et al., 2020), gene induction is induced by touch and darkness (Lee et al., 2005), and AtXTH12 and AtXTH26 exhibit only endotransglucosylase activity (Maris et al., 2010). In contrast to AearXTH2, AearXTH23, AearXTH26, and AearXTH31, the transcript abundances of AearXTH24 were higher in 72-hat seedling roots derived from germinated IND fruits ( Figure 15B ; Supplementary Figure 10A ). In A. thaliana seedlings, overexpression of AtXTH24 is known for its role in promoting hypocotyl growth of etiolated seedlings (Dhar et al., 2022). In A. thaliana seedlings, AtXTH6 is known to be regulated by ABA and auxins (Overvoorde et al., 2005; Huang et al., 2007). In Ae. arabicum seedlings, AearXTH6 is specifically expressed in seedling shoots and lower in 72-hat and 240-hat seedling shoots derived from germinated IND fruits ( Figure 15B ).

β-Galactosidases have β-1,4-galactose in xyloglucan side chains as targets (Steinbrecher and Leubner-Metzger, 2017; Moneo-Sanchez et al., 2019; Steinbrecher and Leubner-Metzger, 2022). In Ae. arabicum, resetting during post-germination pre-emergence growth was not observed for AearβGAL4, and the expression remained higher in 72-hat seedlings derived from IND fruits ( Figure 15B ). In A. thaliana seedlings, βGAL4’s involvement in cell wall changes is associated with the cessation of cell elongation and increased rigidity (Moneo-Sánchez et al., 2016). β-Galactosidases also have β-1,4-galactan in cell wall pectin (rhamnogalacturonan) as a target. Pectins are complex cell wall matrix polysaccharides characterized by α-1,4-linked galacturonic acid and a dynamic degree of methylesterification, and β-galactosidase and various other cell wall remodeling enzymes target pectin (Cao, 2012; Daher and Braybrook, 2015; Scheler et al., 2015). Their expression patterns in Ae. arabicum are presented in detail in Supplementary Figure 10 . Taken together, CWRP gene expression in the dimorphic diaspore model Ae. arabicum revealed either resetting or distinct expression patterns (as DEG in seedlings) during the seed–seedling transition.

The hormonal interactions during seed germination/dormancy (GA-ABA antagonism dominant) and early seedling growth (auxin-ethylene interactions dominant) differ fundamentally (Finch-Savage and Leubner-Metzger, 2006; Linkies and Leubner-Metzger, 2012; Hu et al., 2017; Ahammed et al., 2020; Smolikova et al., 2022; Wang et al., 2023b). Seedlings during pre-emergence growth may encounter increasing post-germination stress in the soil environment, and consequently, seeds/seedlings often fail to establish during this stage (Moles and Westoby, 2006; Finch-Savage and Bassel, 2016; Gardarin et al., 2016). It is clear from many dimorphic diaspore systems that the germination/dormancy traits of the two distinct morphs differ in their responses to environmental cues including abiotic stresses and that the underpinning molecular (hormonal, transcriptomic, and epigenetic) mechanisms differ (Xu et al., 2011; Zhou et al., 2015; Lu et al., 2015a, b; Lenser et al., 2016; Wilhelmsson et al., 2019; Arshad et al., 2021; Gianella et al., 2021; Zhang et al., 2021; Cao et al., 2022; Loades et al., 2023; Song et al., 2023; Chandler et al., 2024). Whether or not seedlings derived from dimorphic diaspores differ in their molecular responses has been far less investigated. Very little is known about if and when these differences disappear by resetting during seedling growth and whether or not these seedling differences were already induced during the dimorphic seed development on the mother plant and then retained during imbibition.

Comparative analysis of seedlings derived from Atriplex and Suaeda dimorphic seeds that were distinct in seed color revealed that seedlings were also distinct in responses to salinity and certain molecular features (Xu et al., 2011; Zhang et al., 2021; Cao et al., 2022; Song et al., 2023). Among the enriched functional gene categories that differed in the seedling transcriptomes were genes encoding inorganic ion transport, hormone metabolism, transport and signaling, TFs and signal transduction components, CWRPs, lipid metabolism, redox, and osmotic homeostasis. In these systems, dimorphic seeds differing in color, size, and dormancy were dispersed by pod shattering from dehiscent fruits. A conclusion from these publications is that the observed differences in salinity tolerance of the dimorphic seedlings were already initiated during seed development on the mother plant. Compared to this, the situation in the Ae. arabicum dimorphic diaspore system is different in that seedlings derived from either M⁺ or bare M⁻ seeds did not show any obvious differences in a range of constant temperature or osmotic stress ( Figure 2 ; Supplementary Figures 3 , 4 ). What is different in the Ae. arabicum M⁺ seed and the IND fruit morph system is the presence of the pericarp, which is known to impose coat dormancy and delay the germination of imbibed IND fruits (Lenser et al., 2016; Chandler et al., 2024).

Pericarp removal experiments with monocarpic species demonstrated that beyond seed traits, the pericarp can also affect seedling establishment and performance (Hu et al., 2009; Mamut et al., 2014; Zhou et al., 2015; Lu et al., 2015b; Lu et al., 2017b; a; Ignatz et al., 2019). Interestingly, in Ae. arabicum, the pericarp restraint led to different growth of hypocotyls derived from imbibed IND fruits as compared to M⁺ and M⁻ seedlings derived from M⁺ and bare M⁻ seeds ( Figure 4 ). This pericarp effect on pre-emergence seedling growth delivered a 5–20-fold higher number of DEGs from the M⁺/IND and M⁻/IND comparisons ( Figures 7 , 8 ) as compared to the 23 DEGs obtained in the M⁺/M⁻ seedling comparison. The pericarp, therefore, plays an important role in the phenotypic plasticity of the Ae. arabicum dimorphic diaspores and has a significant downstream effect on the root and shoot transcriptomes of M⁻ seedlings derived from IND fruits. Details about the identified Ae. arabicum DEGs were already described in the Results section. The following discussion, therefore, focuses on general aspects and selected major functional DEG categories of the transcriptome resetting process.

For the majority of the 1,900–2,000 M⁺/IND and M⁻/IND DEGs at the completion of germination (T_100%), resetting occurred during pre-emergence seedling growth and is completed in 72-hat seedlings ( Figures 7 , 8 ). Examples of this include all hypoxia-responsive and many hormone-related genes (see Section 2.3), as well as dormancy, maturation, LEA, and HSP genes (see Section 2.5). However, a considerable number of ethylene, GA, and auxin-related genes either persisted as a DEG in 72-hat seedling or developed into a DEG during pre-emergence seedling growth (T_100% to 72-hat seedling roots or shoots). This includes enhanced transcript expression for the ethylene-forming enzyme gene AearACO2 in 72-hat seedlings derived from IND fruits ( Supplementary Figure 5 ), GA-related genes including encoding the bioactive GA₄-forming enzyme AearGA3ox1, the GA receptor AearGID1b ( Figure 9B ), and several auxin-related genes including AearSAUR11 ( Figure 10B ). Ethylene, GA, and ABA interact with auxin to regulate seedling growth differently in shoots and roots (Belin et al., 2009; Hu et al., 2017; Ahammed et al., 2020; Wang et al., 2023b). ACO2 is known for its role in counteracting ABA effects in seeds (Linkies et al., 2009; Linkies and Leubner-Metzger, 2012) and in promoting apical hook formation in seedling shoots (Wang et al., 2023b), and ethylene mediates the ABA inhibition on seedling root growth (Ahammed et al., 2020). Flavonoid biosynthesis pathway genes ( Figure 13 ) exhibited a distinct expression pattern from most other DEGs with a root-specific upregulation in 72-hat M⁺ seedlings derived from germinated M⁺ seeds. The role of flavonoids in M⁺ seedlings as compared to M⁻ seedlings and the distinct regulation by MBW protein complexes remain to be elucidated by future research.

ERF and ARF TFs are involved in the control of seedling growth by modulating ABA, ethylene, gibberellin, and auxin signaling (Kim et al., 2012; Zhou et al., 2016; Lorrai et al., 2018; Zhu et al., 2020; Templalexis et al., 2022). In agreement with the role of distinct hormonal signaling in M⁺ and M⁻ Ae. arabicum seedlings, transcripts for AearARF6 and several ERF TFs are DEGs in 72-hat seedling roots or shoots ( Figure 9 ; Supplementary Figure 5 ). ARF and ERF TFs were also identified as DEGs in seedlings derived from the dimorphic (black versus brown) seeds of Suaeda aralocaspica (Cao et al., 2022). Resetting for the majority of the Ae. arabicum TF DEGs during pre-emergence seedling growth was completed at 72 hat ( Figure 14 ; Supplementary Figure 8 ). Notable exceptions were two NAC (AearNAC2 and AearNAC32) and two WRKY (AearWRKY9 and AearWRKY40) TF genes for which higher transcript abundances in IND fruit diaspores during germination persisted at the 72-hat seedling state ( Figure 14B ). In A. thaliana seedling growth, these TFs are involved in integrating environmental and hormonal stimuli (see Section 2.5). Taken together, transcriptomes of M⁺ and M⁻ Ae. arabicum seedlings derived from germinated M⁺ seeds and IND fruits differ in the expression of genes involved in ethylene and GA metabolism, hormone signaling, and for various TFs with roles in integrating environmental and hormonal stimuli.

Precise auxin distribution patterns and polar transport are required for the control of root and hypocotyl growth and development (Teale et al., 2006; Hu et al., 2017; Wang et al., 2023b). The PILS auxin carriers are known to be involved in intracellular auxin homeostasis (Feraru et al., 2012, 2019; Bogaert et al., 2022; Feraru et al., 2022). In Ae. arabicum, several auxin carriers were DEGs during late germination (T_100%), but their expression difference was reset during seedling pre-emergence growth ( Figures 9 , 10 ). AearPILS3, in addition, remained a DEG at 72 hat where it was more highly expressed only in seedling shoots derived from IND fruits. AearPILS7 became a root-specific DEG at 72 and 240 hat, where it was more highly expressed only in seedling roots derived from IND fruits. Seedlings derived from IND fruits differed from M⁺ and M⁻ seedlings derived from germinated M⁺ seeds and bare M⁻ seeds, respectively, in that their lower hypocotyls were often bent and not straight and that their hypocotyl and root growth was slower ( Figure 4 ). We speculate that differences in auxin biosynthesis, signaling, and transport play a major role in this altered seedling growth phenotype. This phenotype was, however, not connected with a difference in hypocotyl mechanics (fracture force) of 15-day-old Ae. arabicum seedlings ( Figure 4 ). Microtensile measurements of A. thaliana seedlings (Saxe et al., 2016) demonstrated that hypocotyls differed mechanically between early (day 4) and later (day 5–7) seedlings, but there were no differences among the later seedlings (Saxe et al., 2016). It seems, therefore, that despite the slower growth and the bent hypocotyls of IND fruit-derived Ae. arabicum seedlings, possible mechanical differences between pre-emergence (72 hat) have already disappeared in later (>240 hat) seedlings.

Downstream of the hormonal regulation are distinct expression patterns in seedlings derived from IND fruits for water and ion transporters ( Figure 12 ) potentially affecting the turgor pressure and for CWRP genes ( Figure 15 ) potentially affecting the cell wall extensibility. The rapid and uniform seedling growth depends on cell expansion, which requires cell wall loosening by the action of CWRPs or apoplastic ROS (Finch-Savage and Leubner-Metzger, 2006; Steinbrecher and Leubner-Metzger, 2017; 2018). Among the CWRP DEGs in Ae. arabicum seedlings were genes encoding enzymes that target xyloglucan (βGAL4 and XTHs) and pectin ( Figure 15 , Supplementary Figure 10 ). AearXTH12, AearXTH23, and AearXTH6 were less expressed in 72-hat and 240-hat seedlings derived from IND fruits as compared to M⁺ and M⁻ seedlings derived from germinated M⁺ and M⁻ seeds. XTHs are known to affect seedling vigor (Ducatti et al., 2022) and hypocotyl growth (Miedes et al., 2013; Dhar et al., 2022), and AtXTH6 is known to be regulated by ABA and auxins in seedlings (Overvoorde et al., 2005; Huang et al., 2007). Morph-specific differential expression of other xyloglucan remodeling enzymes was observed during Ae. arabicum fruit and seed development (Steinbrecher and Leubner-Metzger, 2022), dimorphic diaspore germination (Chandler et al., 2024), and dimorphic seedling growth (this work). CWRP gene expression in the dimorphic diaspore model Ae. arabicum, therefore, revealed either resetting or distinct expression patterns (as DEG in seedlings) during seed–seedling transition. Taken together, we conclude that the transcriptomes of seedlings derived from the dimorphic diaspores, M⁺ seeds and IND fruits, undergo transcriptional resetting during the post-germination pre-emergence growth transition phase from germinated diaspores to growing seedlings.

Plants of Ae. arabicum (L.) Andrz. ex DC. were grown from accessions TUR ES1020 (from Turkey) as described by Chandler et al. (2024). Mature M⁺ seeds and IND fruits were harvested, further dried over silica gel for 1 week, and stored for a few months at −20°C in air-tight containers. For germination assays, dry mature seeds (M⁺ or M⁻) or IND fruits were placed in 3-cm Petri dishes containing two layers of filter paper, 3 mL deionized water (dH₂O), and 0.1% Plant Preservative Mixture (Plant Cell Technology, Washington, DC, USA). Temperature response profiles ( Supplementary Figure 1 ) were obtained by incubating plates on a GRD1-LH temperature gradient plate device (Grant Instruments Ltd., Cambridge, UK). Subsequent germination assays were conducted by incubating plates in MLR-350 Versatile Environmental Test Chambers (Sanyo-Panasonic, Bracknell, UK) at the indicated imbibition temperature as described by Chandler et al. (2024). Seed germination, scored as radicle emergence, of three biological replicates of 30 seeds or fruits was analyzed.

Seedling growth assays were conducted using just-germinated seeds (1-mm radicle protrusion visible; obtained from surface-sterilized seeds germinated in darkness at 9°C), which were selected for transfer to 12 cm × 12 cm plates containing media based on 1-mm protrusion of the radicle. As medium autoclaved 1% (w/v) agar in 1/10 Murashige and Skoog (MS) basal medium (M5519, Sigma, Darmstadt, Germany) was used. For the seedling growth assays, plates were incubated vertically in constant white light (170 µmol·m⁻²·s⁻¹) in MLR-350 Versatile Environmental Test Chambers (Sanyo-Panasonic, Bracknell, UK) at a constant temperature, as indicated ( Figure 2 ; Supplementary Figure 2 ). For seedling growth assays during osmotic stress ( Supplementary Figure 3 ), water potentials were lowered using high-molecular-weight polyethylene glycol (PEG6000; 26603.293, VWR, Radnor, PA, USA) using an overlay method (Van Der Weele et al., 2000; Verslues and Bray, 2004). Seedling growth assay constant temperatures were 14°C for osmotic stress, and temperatures for the thermal stress experiments were between 14°C and 35°C.

The chlorophyll content of seedlings grown horizontally at 30°C was determined after extracting pigments from leaf tissues homogenized in methanol at room temperature for 15 min while shaking at 1,000 rpm on a thermomixer (S8012-0000, Starlab, Milton Keynes, UK). The obtained extracts were centrifuged for 5 min at 14,000 g. The absorbance of the supernatants was determined at 750 nm, 665 nm, and 652 nm using a microplate reader (Spark^® 10M, Tecan, Zürich, Switzerland) and subsequently used to calculate chlorophyll contents (Porra et al., 1989). For the biomechanical analyses (Steinbrecher and Leubner-Metzger, 2017), just-germinated Ae. arabicum seeds (M⁺ and M⁻) and IND fruits were transferred to agar plates as described above and grown for 15 days under constant white light at 30°C. IND fruits were manually split open just after the completion of germination (radicle protruding the pericarp) at 0 hat. To conduct biomechanical analyses of hypocotyls, seedlings were clamped, leaving a 7-mm gap between the jaws of a single-column tensile testing machine (Zwick Roell ZwickiLine Z0.5, Ulm, Germany). A constant speed for separation was set at 5 mm/min. Force-displacement data were obtained, and hypocotyl breaking forces were calculated ( Figure 4 ).

A sampling of imbibed M⁺ seeds, M⁻ seeds, and IND fruits for molecular analyses was described by Chandler et al. (2024). Three biological replicates of samples each corresponding to 20-mg dry weight of seed material were pulverized in liquid N₂ using mortar and pestle. Extraction of total RNA was performed as described by Graeber et al. (2011). Sampling and RNA extraction were performed of root and shoot tissue from seedlings grown at 30°C (grown from pre-germinated seeds at 9°C in darkness). Tissue was homogenized at 6,500 rpm using a Precellys 24 (Bertin Instruments, CNIM Group, Paris, France). Seed and shoot total RNA was isolated using a protocol modified by Chang et al. (1993). After the addition of RNA extraction buffer [2% (w/v) hexadecyltrimethylammonium bromide (CTAB), 2% (w/v) polyvinylpolypyrrolidone (PVP), 100 mM Tris-HCl pH 8.0, 25 mM ethylenediaminetetraacetic acid (EDTA), 2 M NaCl, and 2% (v/v) β-mercaptoethanol], samples were incubated at 65°C for 10 min with intermittent vortexing. Chloroform:isoamylalcohol (24:1) extractions were repeated three times. After the addition of 10 M LiCl to a final concentration of 2 M, RNA was precipitated overnight at 4°C and then dissolved in NaCl-Tris-EDTA (STE) buffer [1 M NaCl, 0.5% (w/v) sodium dodecyl sulfate (SDS), 10 mM Tris-HCl pH 8.0, and 1 mM EDTA]. Three further chloroform:isoamylalcohol (24:1) extractions were then performed before precipitation in 100% (v/v) ethanol overnight at −80°C. Samples were then centrifuged for 20 min at 4°C. After the removal of the aqueous phase, the RNA pellets were washed with 70% (v/v) ethanol. Samples were centrifuged for 20 min, the ethanol was carefully removed, and the RNA was subsequently dissolved in RNase-free water. Genomic DNA was removed by DNase-I (QIAGEN, Valencia, CA, USA) digestion in solution, followed by additional purification using columns (QIAGEN RNeasy Kit). Shoot tissue RNA was isolated using the RNeasy Plant Mini Kit (QIAGEN) and manufacturer’s instructions. RNA quantity and purity were determined using a NanoDrop™ spectrophotometer (ND-1000, ThermoScientific™, Wilmington, DE, USA) and an Agilent 2100 Bioanalyzer with the RNA 6000 Nano Kit (Agilent Technologies, Santa Clara, CA, USA) using the 2100 Expert Software to calculate RNA Integrity Number (RIN) values.

Transcriptome assembly, data trimming, filtering, read mapping, feature counting, and DEG detection were performed using the pipeline previously described by Wilhelmsson et al. (2019). PCA was performed using R (R Core Team, 2021) and the Bioconductor package DESeq2 (Love et al., 2014) and plotPCA on log(x + 1)-transformed RPKM values with non-zero values in at least one sample. GO term enrichment in DEG lists was calculated with R package topGO using the elim method with Fisher’s exact test (Alexa and Rahnenfuhrer, 2021). Gene identifiers and symbols are according to earlier publications of the Ae. arabicum genome (version 2.5) and transcriptome (Wilhelmsson et al., 2019; Arshad et al., 2021; Chandler et al., 2024), and the Ae. arabicum web portal (https://plantcode.cup.uni-freiburg.de/aetar_db/index.php) links this to the current (Fernandez-Pozo et al., 2021) and future genome DB and gene expression atlas.

To identify Brassicaceae PILS genes ( Figure 11 ), the A. thaliana sequences (Yi et al., 2021; Bogaert et al., 2022; Waidmann et al., 2023) were used, and BLAST analyses were conducted via Phytozome (Goodstein et al., 2012). The combined information of the BLAST analyses was used to conduct the phylogenetic analysis with known and putative PILS amino acid sequences aligned using ClustalW (BLOSUM cost matrix, Gap open cost 10, Gap extend cost 0.1), and Neighbor-Joining trees (Saitou and Nei, 1987) were built using Geneious 8.1.9 Tree Builder (Geneious, San Diego, CA, USA) using Jukes-Cantor distance. Consensus support (minimum 20%) was determined using bootstrap (1,000).

Data are expressed as mean ± 1 SEM. Statistical analysis of experiments was performed using the GraphPad Prism software (v.7.0a; San Diego, CA, USA) for the analysis of variance (ANOVA) and unpaired t-test procedures. For studies examining abiotic stress effects, data were analyzed by two-way ANOVA, with seedling morph and seedling age (time) as between-group factors. Multiple comparisons were performed using Sidak’s post-hoc correction in GraphPad Prism. Results were considered statistically significant if the p-value was less than 0.05.