For physiological experiments, we used the A. alpina accession Pajares and pDR5::GUS Pajares lines previously described in Vayssières et al., 2020. Pajares was collected in the Coedillera Cantábrica mountains in Spain at 1,400 m altitude (42°59′32′′ N, 5°45′32′′ W) (Wang et al., 2009). Vernalization experiments were performed as described before (Lazaro et al., 2018; Vayssières et al., 2020; Viñegra de la Torre et al., 2022) except that a constant photoperiod (LD, 16h light: 8h dark) was maintained throughout the experiment to specifically study the effect of long-term cold on adventitious rooting. Hence, the plants were grown in soil in a glasshouse under long-day conditions (LD; 16 h light: 8 h dark) at 20°C/18°C day/night temperatures until they were 8 weeks old, and were then moved to growth chambers maintained at 4°C and LD conditions for 4, 8, 12, 16 and 21 weeks ( Figure 1A ). After cold treatment, plants were moved back to the LD glasshouse conditions ( Figure 1A ). The presence of adventitious roots on internodes of the main stem and on axillary branches of each plant was scored before cold treatment, at the end of each cold treatment, 0 week in LD (+0wLD) and after plants were returned to glasshouse conditions for 1 (+1wLD) and 2 (+2wLD) weeks. The presence of adventitious roots was also monitored on 8 weeks old plants maintained in LD condition for 5 additional weeks (+5wLD) ( Supplementary Figure 1C ). These plants had the same number of internodes as the flowering plants being exposed to 21 weeks in cold and returned to glasshouse conditions for 2 weeks (21w + 2wLD). Only the presence and not the number of adventitious roots on the internodes was scored ( Figure 1E ; picture on the right). The percentage of plant with adventitious root on the internodes or on branches was calculated. Due to the unique spatial presence of adventitious roots on the main stem we decided to characterized it further using the percentage of the presence of adventitious roots per internode position along the main stem. The internodes of the main stem that develop during cold exposure did not elongate and were refer to as “compressed internode zone” compared to the internodes that develop before and after cold exposure that showed elongation and were refer to as “non-compressed internode zone” ( Figure 1E ; picture on the left). We scored the percentage of flowering plants and the total number of leaves at flowering in 8 weeks old plants exposed to different durations (0w, 4w, 8w, 12w, 16w and 21w) at 4°C followed by 4 weeks in long days (+4wLD). At least 9 plants were used for each condition.

For the transcriptome analysis, we harvested internodes 9 – 13, that show specific adventitious root formation in plants after long term cold treatment. Around 2 cm stem section excluding the axillary buds was collected from Pajares plants before cold exposure (0w), after being exposed for 4, 8, 12, 16 and 21 weeks to 4°C (4, 8, 12, 16 and 21w; +0wLD) and after being exposed to 12, 16 and 21 weeks to 4°C and transferred back to a LD glasshouse for 5 days (12, 16 and 21w; +5dLD) ( Figure 1A ). This stem section corresponded to the 5 internodes below the compressed internode zone. Given that plants cultivated in long-day conditions without cold exposure (8wLD + 5LD) did not flower, therefore did not have the same shoot architecture (lacking the compressed zone) and did not share the same age, we didn’t include them in the transcriptomic analysis. Each sample consisted of stems from 10 plants and harvested in three biological replicates. The purified RNA samples were sent to the Max Planck Genome Center, Cologne, Germany, for library preparation and sequencing (https://mpgc.mpipz.mpg.de/home/). 1 µg of total RNA was enriched for poly-A RNA using the NEBNext Poly(A) mRNA Magnetic Isolation Module (New England Biolabs), followed by library preparation using the NEBNext Ultra Directional II RNA Library Prep kit (New England Biolabs). RNA quality and quantity were monitored throughout by capillary electrophoresis (TapeStation, Agilent Technologies) and fluorometry (Qubit, Thermo Fisher Scientific). Sequencing was performed on a HiSeq3000 device (Illumina) generating 150-bp single-end reads. Approximately 12.5 million reads were sequenced from each library. Before further processing, the sequencing data quality was verified using FastQC (https://www.bioinformatics.babraham.ac.uk/projects/fastqc/). Sequences were mapped and aligned to the reference genome using HISAT2 followed by assembly and quantification of expression levels in different samples using STRINGTIE (Pertea et al., 2016). The gene counts of all samples were obtained using a Python script (http://ccb.jhu.edu/software/stringtie/dl/prepDE.py). The quality of the samples was assessed by producing dispersion plots among replicates. The differentially expressed genes with more than 2-fold change and a corrected p-value below 0.05 were obtained using DESeq2 and selected for further analysis. Transcriptome data from this study is deposited in the Gene Expression Omnibus database (https://www.ncbi.nlm.nih.gov/geo/) under the accession number GSE176054. For Gene Ontology (GO) term enrichment, the R-based graphical tool BACA (Bubble chArt to Compare Annotations) generated using DAVID was used to visualize (Dennis et al., 2003; Fortino et al., 2015). A bubble chart was generated summarizing the enriched GO terms (Benjamini corrected p-value < 0.05) among the modulated genes in different samples. The size of each bubble representing a GO term denotes the number of genes in a list of differentially expressed genes associated with the GO term, and the color indicates the direction of modulation (green = downregulation, magenta = upregulation). A minimum of five genes was required for the GO term to be considered for further analysis. The enrichment index of hormone responsive genes was calculated in the following manner: Enrichment = (number of up- or down-regulated genes responding to a hormone/number of all genes responding to that hormone)/(number of all up- or down-regulated genes/number of all genes). We assumed in this case that the A. alpina orthologues respond in an analogous manner. The hormone responsive genes correspond to the respective GO terms collected from AmiGO. The following GO terms were considered: response to abscisic acid (GO:0009737), auxin (GO:0009733), brassinosteroid (GO:0009741), cytokinin (GO:0009735), ethylene (GO:0009723), gibberellin (GO:0009739), jasmonic acid (GO:0009753) and salicylic acid (GO:0009751) ( Supplementary Dataset 1 ). The significance of the enrichment was estimated using Monte Carlo method by testing whether the overlap between the sets of hormone responsive genes and the differentially regulated genes was higher than it would be expected by chance. The test consists of randomly selecting a number of differentially regulated genes from the A. alpina genome 10,000 times. We thus counted the number (n) of times that the intersection size of the random lists was equal to or higher than the intersection observed for two lists of tested genes. A P value of n/10,000 was then calculated. Strigolactone responsive gene enrichment was not studied due to the small number of gene in the GO terms (7 genes). Genes sharing similar expression profiles were identified using Cluster 3.0 (Eisen et al., 1998) using the same parameters as in Vayssières et al., 2020. Genes were selected out of the total of 30510 genes based on the following criteria: log₂(Fold-change) ≥ 2 and the difference between the maximum and the minimum value ≥ 3. Heatmaps were generated for the genes passing this threshold using TreeView based on the values obtained from Cluster 3.0 (de Hoon et al., 2004). A group of genes showing similar distribution in expression pattern were grouped into a cluster. The genes and corresponding GO terms in the clusters representing the distinct expression profiles were further analyzed.

For GUS staining assays, stems from the last extended internodes below the compressed internode zone were harvested from pDR5::GUS A. alpina plants grown for 8 weeks in a LD greenhouse, and after being exposed to 4 °C for 4 weeks and 21 weeks. After harvesting, stem samples were incubated for 1 h in 90% ice-cold acetone and were subsequently washed in 50 mM phosphate buffer (pH 7.0) and submerged in 2.5 mL GUS staining solution under moderate vacuum for 20 min. The GUS staining solution was composed of 1 mM EDTA (pH 8), 5 mM Potassium ferricyanide, 5mM Potassium ferrocyanide, 100 mM Sodium phosphate (pH 7.0), 1% Triton-X-100, 1 mg/mL X-Gluc (Cold Spring Harb Protoc, 2007; https://doi.org/10.1101/pdb.rec10860). Samples were incubated afterwards at 37°C in the dark for a maximum of 16 h and subjected to ethanol washes to remove the chlorophyll. Transverse sections of 70 µm thickness were prepared using a Leica VT1000S vibratome from samples immobilized on 6% (w/v) agarose. Representative photographs from two biological replicates were obtained using the Zeiss Axio Imager microscope with the Zeiss Axio Cam 105 color camera (Zeiss, Jena, Germany) or the Nikon Eclipse Ni-U microscope with the Nikon Digital Sight camera (DS-Fi2) (Nikon, Tokyo, Japan).

The three biological replicates of stem material harvested for the transcriptome analysis were also used for endogenous free IAA quantification. Around 15 mg of plant material (fresh weight) was processed as previously described (Andersen et al., 2008). Free IAA was quantified in the purified samples using combined gas chromatography - tandem mass spectrometry. 500 pg ¹³C₆-IAA internal standard was added to each sample before homogenization and extraction. Values represent the average of three biological replicates and error bars represent the standard deviation of the mean.

For the quantification of GUS expression in pDR5::GUS plants, stem samples were collected following the protocol used for sample collection for the transcriptome analysis. Total RNA was extracted from the samples using the RNeasy Plant Mini kit (QIAGEN) following the manufacturer’s protocol. Residual DNA was removed using the DNA-free DNA removal kit (Invitrogen). The RNA concentration and integrity were determined using a NanoDrop ND-1000 UV-Vis spectrophotometer (Thermo Fisher Scientific). Total RNA (2 µg) was used for the synthesis of cDNA using SuperScript II Reverse Transcriptase (Invitrogen) and oligo dT (18) primers. 3 μL of each cDNA solution (1:5.5 dilution) was used as a template for each quantitative PCR reaction (qPCR) performed in three technical replicates. The CFX Connect Real-Time PCR Detection System (Bio-Rad) and the iQ SYBR Green Supermix (Bio-Rad) were used to perform the real-time quantitative PCR analyses using the primers: GUS-F – CCCTTACGCTGAAGAGATGC, GUS-R – TTCGTTGGCAATACTCCACA, AaPP2A-F – AGTATCGCTTCTCGCTCCAG and AaPP2A-R - AACCGGGTTGGTCGACTATTG. The PCR programme followed was 95°C for 3 min, followed by 40 cycles of 95°C for 15 s and 60°C for 60 s, then a melting curve (55–95°C with a heating rate of 0.5°C/5 s). Values represent the average of three biological replicates and error bars represent the standard deviation of the mean.

To test whether the length of cold exposure influences clonal growth through adventitious roots in A. alpina, we exposed plants of the Pajares accession to different durations of cold treatment ( Figure 1A ) (Wang et al., 2009). We observed that plants that did not experience cold treatment or experienced cold for only 4 and 8 weeks did not show adventitious roots on the main stem or in the branches ( Figures 1B, C ). Plants cold-treated for 21 weeks had a higher propensity to produce adventitious roots on the main stem and axillary branches compared to plants treated with shorter durations of cold (12 and 16 weeks) ( Figures 1B, C ). The length of cold exposure had a quantitative effect on adventitious rooting in the internodes of the main stem and in axillary branches. Also, the longer the plants were exposed to cold, the sooner we observed adventitious roots after plants were returned to greenhouse conditions ( Figures 1B, C ). Overall, these results suggest that long term cold exposure leads to adventitious root formation.

On the main stem, we observed adventitious roots developed preferentially on specific internodes ( Figures 1D, E ). Plants cold treated for 21 weeks produced adventitious roots on the lower internodes of the main stem (33.3% of plants on first internode and 55.6% on the second internode) and on the internodes 9 to 13. Specifically, 11.1% of plants produced adventitious roots on internodes 9, 12 and 13, 33.3% on internode 11 and 22.2% on internode 10, respectively ( Figure 1D ). Adventitious roots on lower internodes appeared also in non-cold treated plants and on plant cold treated for shorter time, whereas the ones at higher internodes were specific to plants after long term cold treatment ( Figure 1D and Supplementary Figure 1D ). We have previously shown that in greenhouse conditions, plants of the Pajares accession occasionally develop adventitious roots in the lower internodes but not in the upper internodes (Mishra et al., 2020). These results suggest that like in other species, adventitious roots in A. alpina at lower internodes develop in response to soil humidity (Stevenson and Laidlaw, 1985). Adventitious roots at higher internodes were mostly observed below a zone of compressed (short) internodes that contain dormant buds that shape the perennial shoot architecture of A. alpina ( Figure 1E ) (Vayssières et al., 2020). Interestingly, internodes 9 to 13 on the main stem were the last to expand before the plants were exposed to cold (purple in Figure 1D ).

As previously noted in A. alpina accession Pajares, plant cold-treated for 4 weeks did not flower and flowered only partially for plant treated for 8 weeks (only 50% of the plant flowered; Supplementary Figure 1A ) (Wang et al., 2009; Lazaro et al., 2018). Plants cold-treated for 12, 16 or 21 weeks flowered and showed the same total leaf number at flowering ( Supplementary Figures 1A, B ). We followed plants that haven’t been treated by cold until they reached the same number of leaves as the plants cold-treated for 12, 16 and 21 weeks (39.7 leaves). In these plants, no adventitious roots were observed on the main stem unlike as observed for the plants cold-treated for 12, 16 and 21 weeks. This result suggests that the total number of internodes does not influence adventitious rooting, but stem aging in cold does ( Supplementary Figures 1C–F , Figure 1D ). This also suggests an age- or position- dependent response to cold treatment for the development of adventitious roots on the main stem.

We performed a transcriptome profiling to understand how prolonged cold triggers the formation of adventitious roots at internodes 9 – 13. We compared stem samples of plants harvested before, during cold exposure for 4, 8, 12, 16 and 21 weeks and after they were transferred back to greenhouse conditions for 5 days (+5dLD) ( Figure 1A ). Of the differentially expressed genes (DEGs) we identified, 73.4% had homologues in A. thaliana ( Supplementary Table 1 ; Supplementary Figures 2A, B ). Comparisons of the transcriptome before (0w) and during cold exposure (4w/8w/12w/16w/21w) indicated that a higher number of DEGs were down- than upregulated during cold exposure ( Supplementary Table 1 ; Supplementary Figure 2A ). There was an overlap of DEGs in stem samples harvested during different durations of cold treatments ( Supplementary Figure 2A ). We then compared the transcriptome of stem samples at the end of cold treatment (+0wLD) with the transcriptome of stem samples after cold treatment (+5dLD) in plants exposed to 12w, 16w and 21w of cold. In these comparisons, we found that more genes were up- than downregulated after cold treatment ( Supplementary Figure 2B ). Expression profiles of samples during cold exposure clustered separately compared to the ones collected in LD greenhouse conditions both before and after cold treatment ( Supplementary Figure 2C ). Among the samples collected during cold, internodes from 21w cold treated plants clustered separately from the rest suggesting that the duration of cold exposure changes the transcriptome of the internodes that have a higher propensity to produce adventitious roots. We observed an enrichment of GO terms associated with lateral and post-embryonic root development in internodes harvested at the end of 8w, 12w, 16w and 21w of cold treatment compared to before treatment (0w) ( Figure 2A ). Accordingly, homologs of transcripts of root development-associated genes were upregulated in the internodes during cold treatment, few of which have been described to be involved in adventitious root development ( Figure 2B ) (Li et al., 2018; Tahir et al., 2021; Zhang et al., 2021). This includes PUCHI, AGAMOUS-LIKE 44 (AGL44), HAESA (HAE), Related to ABI3/VP1 (RAV1), plant U-box 9 (PUB9) and polyol/monosaccharide transporter 5 (PMT5), CUP SHAPED COTYLEDON 1 (CUC1), WRKY DNA-BINDING PROTEIN 75 (WRKY75), EXPANSIN A17 (EXPA17), Translationally Controlled Tumor Protein 1 (TCTP1), NUCLEOSOME ASSEMBLY PROTEIN 2 (NRP2) and NITRATE TRANSPORTER 2.1 (NRT2:1) (Hu et al., 2004; Remans et al., 2006; Zhu et al., 2006, 2019; Devaiah et al., 2007; Hirota et al., 2007; Lee and Kim, 2013; Deb et al., 2014; Li et al., 2018; Alvarez-Buylla et al., 2019; Branco and Masle, 2019). The increase in expression of root-associated genes suggest that adventitious root primordia develop on the internodes during cold exposure.

To verify whether adventitious root primordia formation on the stems initiate during cold treatment, we used the A. alpina lines carrying the synthetic DR5 promoter fused to the reporter gene GUS ( Figure 3 ) (Vayssières et al., 2020). Reporter plants transformed with pDR5::GUS have been used in several species to monitor developing adventitious roots following auxin accumulation (Inukai et al., 2005; Sukumar et al., 2013; Rasmussen et al., 2015). In A. alpina stems, GUS expression under pDR5 during cold exposure was reduced in comparison to before cold exposure suggesting a decreased GUS transcription in cold as described before ( Figure 3 ; Supplementary Figure 3 ) (Lee et al., 2005). Nevertheless, stem sections in internodes 9-13 during different durations of cold treatment demonstrated the presence of adventitious root primordia during but not before cold treatment ( Figure 3 ). In stem sections harvested from plants after 4 weeks of cold treatment, we observed the presence of a meristematic mass, whereas the ones harvested from plants cold treated for 21 weeks had established root primordia ( Figure 3 ). Root primordia in A. alpina were present in the stem vasculature, at the border of xylem and phloem. These results suggest that adventitious roots in the main stem are formed in response to cold exposure and that prolonged duration of cold treatment may be important to ensure the development of adventitious roots. Emergence of adventitious root primordia may occur during or after the cold exposure.

Our experimental design allows us to differentiate gene regulatory networks that participate in cold adaptation, adventitious primordia formation and emergence in cold (in cold-treated plants for 21 weeks) or in adventitious root emergence in response to the transition between cold and warm temperature (in cold-treated plants for 12 and 16 weeks). First, to identify molecular processes affected during cold-induced adventitious rooting in A. alpina, we investigated the transcriptomic changes in stems that take place during cold exposure ( Supplementary Dataset 2 ). We constructed a heat map and looked for co-expression clusters that might explain the different developmental stages of adventitious root formation, including the induction and initiation of adventitious roots ( Figure 4 ; Supplementary Dataset 3 ). To identify genes with strong changes in expression pattern, we selected genes that showed a more than four-fold change between the highest and the lowest expression values. Among all the A. alpina genes, 3540 genes fulfilled this criterion and were included for the co-expression analysis. The heat map was divided into 69 significantly enriched sub-clusters. We focused on 13 clusters that represented 46% of the total clustering ( Figure 4A ; Supplementary Figure 4 ; Supplementary Dataset 4 ). Among these, we identified clusters in which transcript accumulation of genes was downregulated during cold exposure (Group I; highlighted in green; Clusters B and C) and others in which gene expression was upregulated (Group II; highlighted in red; Cluster D). Genes in these groups may be involved in the induction phase of root primordia and play a negative or positive role during subsequent phases of root primordia formation. In Group I, we detected genes associated with auxin homeostasis (GH3.9, CYP83A1 and UGT74B1), response (IAA7, SPL3, SAUR11, SAUR27, SAUR29, SAUR51, SAUR66 and SAUR67) and transport (ABCB19); gibberellic acid degradation (GA2OX1), and signaling (GASA6); brassinosteroid degradation (BR6OX2) and meristem associated genes (LBD33 and WOX1). Group II included genes involved in red/far-red light responses (PIF3, PIL2 and PIL6) or associated with plant morphogenesis (TCP2 and TCP14). While the genes in group I have not been specifically identified previously as transcriptionally regulated by cold, we cannot rule out the possibility that they may be associated with cold response independently of root primordia formation. The other groups exhibited a more specific pattern of expression at various time points during cold exposure. We detected clusters in which gene expression was transiently silenced at early (Group III; highlighted in yellow; Clusters E - L) or late stages of cold treatment (Group IV; highlighted in grey; Clusters M - N). Group III included genes involved in auxin response (DAO2, SAUR1, SAUR6, SAUR16, SAUR50, SAUR52, SAUR53 and IAA19); dividing meristems (ARGOS-LIKE, CYCB2;3, CDKB2;2, MAP65-4 and KINESIN5); cytokinin signaling (ARR15, AHP4 and LOG4); and ethylene signaling (ERF53). Group IV includes genes associated with ethylene signaling (ERF39 and ERF73); suberin biosynthesis (CYP86A1 and CYP86B1); gibberellic acid signaling and homoeostasis (GASA1 and MES9) and stem cell regulator BLUEJAY. Transcript accumulation of genes in Group III might be involved in the induction/initiation phase of root primordium, whereas transcript accumulation of genes in Group IV in the emergence phase of root primordium. These data also suggest a downregulation of auxin response during the early phases of cold exposure as observed for the DR5 auxin response in cold ( Supplementary Figure 5 ; Figure 3 ). Accordingly, free auxin levels on stems were significantly reduced in plants after 4w in cold ( Supplementary Figure 5 ). However, we observed a gradual (but not statistically significant) upregulation of endogenous IAA levels in stems of plants exposed to prolonged cold ( Supplementary Figure 5 ).

To identify genes and processes involved in the emergence phase of adventitious roots, we investigated the transcriptomic changes in stems taking place after cold exposure. Stems of 21w cold-treated plants had fewer genes differentially regulated between samples harvested at the end (+0dLD) and after (+5dLD) cold treatment in comparison to the ones exposed to 12w or 16w of cold treatment ( Supplementary Figure 2B ; Supplementary Table 1 ). This result suggests that after 21w, the transcriptome response of A. alpina to cold treatment was saturated, as was the number of plants producing cold-dependent adventitious roots. We performed a co-expression analysis of the genes detected between stems harvested at the end and after cold treatment ( Figure 5 ). We found that the accumulation of transcripts involved in glucosinolate metabolism (including CYP83A1, CYP79F1, UGT74B1, SOT17, IMD1, IPMI1 and IPMI2) was reduced after 21 weeks of cold treatment, whereas it was comparatively higher after 12 and 16 weeks of cold treatment. Inversely, transcript accumulation of genes participating in cell cycle, cell division and organelle reorganization (among them ATCDC48B, CYCA3;2, EB1C, MCM2, MCM6, MCM9, MKK6, SMC2 and TOPII) were significantly increased in stems of 21w cold treated plants compared to the stems of plants cold-treated for 12 and 16 weeks ( Figure 5 ). Homologs of ethylene biosynthesis genes ACO3 and ACS6, ethylene signaling genes ERF6 and ERF104 and the auxin signaling gene IAA19 were also upregulated after extended cold exposure ( Supplementary Dataset 2 ).

As hormones are instrumental in the regulation of adventitious roots, we examined the enrichment of phytohormone-related genes among the DEGs identified during cold treatment during the formation of adventitious root primordia ( Figure 6 ). We found that abscisic acid, auxin, ethylene, gibberellic acid, jasmonic acid and salicylic acid responsive genes were enriched in the upregulated DEGs detected throughout cold exposure except for the auxin response gene at 4 weeks into cold. Among the downregulated genes, we detected an enrichment of auxin, Cytokinin and gibberellic acid-responsive genes ( Figure 6 ). Notably, brassinosteroid response gene were enriched only in upregulated genes at the beginning of the cold period (4 and 8 weeks) and in down regulated at the second part of the cold exposure (12, 16 and 21 weeks) suggesting an opposite role of this hormone in adventitious root primordia initiation/formation and maturation/emergence in cold ( Figure 6C ). Auxin-responsive genes showed similar enrichment among the upregulated and downregulated genes ( Figure 6B ). However, at the end of 21w of cold exposure the number of auxin responsive genes in the upregulated gene was twice as high compared to the other time points ( Figure 6B ). Concurrently, jasmonic acid responsive genes were overrepresented among the upregulated gene and brassinosteroid responsive genes in the downregulated genes. These results point to changes in auxin, jasmonic acid, brassinostreroid and cytokinin response specifically at 21w at 4°C.

To determine the regulation of hormone responsiveness after cold exposure, we looked for hormone responsive genes in stem samples harvested from plants at the end of cold treatment and after plants had been returned to greenhouse conditions. No gibberellic acid and jasmonic acid responsive genes were identified among the DEGs. Interestingly, genes responsive to abscisic acid, auxin, brassinosteroid, ethylene and salicylic acid were enriched among the down regulated genes in cold-treated plant for 12 and 16 weeks but not 21 weeks ( Figure 7 ). Specifically, in stems of plants cold treated for 21w, we observed an exceptionally low enrichment of DEGs related to hormones compared to the ones in plants cold treated for 12w or 16w. This pattern was also observed for the auxin responsive genes. Nevertheless, endogenous IAA levels in stems increased in plants transferred to greenhouse conditions after 21 weeks of cold treatment ( Figure 7G ). This result suggests that hormones are also important for the later stages of adventitious rooting and explains the lack of hormone responsive DEGs in stems of 21 weeks cold treated plants that develop adventitious roots already during cold treatment. The hormone response for adventitious rooting is already primed in cold by 21 weeks.