Introduction

Plants are exposed to constantly changing environments that often exert stressful or adverse effects on their growth, development, and reproduction (Zhang et al. 2018). As they lack roots in hydroponic systems, they must contend with a variety of abiotic stresses. Among these, temperature extremes, salinity, and water availability are principal environmental factors limiting plant productivity (Ren et al. 2014; Zhu 2016). Low temperature is a critical environmental stressor that severely impacts plant growth and reproduction, reduces crop yield and quality, and has become a significant factor limiting the geographic distribution of hydroponic plants (Wang et al. 2018; Liu et al. 2023). Currently, two types of low-temperature stress are recognized: chilling stress (0–15 °C) and freezing injury (below 0 °C) (Han et al. 2016; Hwarari et al. 2022). Under low-temperature conditions, plants undergo a series of physiological and biochemical changes as part of a defense system. Chilling stress adversely affects plants by compromising the structural stability and normal function of cell membranes, diminishing the osmotic regulation capacity of the cytoplasm, and inhibiting photosynthesis and the synthesis of various substances (Zhu et al. 2007; Sadura and Janeczko 2021). However, the threats posed by freezing stress are often fatal to numerous plant species, as freezing can cause rapid crystallization within the intercellular spaces, resulting in mechanical injuries that disrupt normal physiological processes and ultimately lead to cell death (Ritonga and Chen 2020). To overcome these challenges or adverse environments, plants have evolved a range of adaptive mechanisms to enhance their cold tolerance (Zhang et al. 2018).

In recent years, numerous studies have demonstrated that cold resistance is a genetically complex quantitative trait. This resistance is closely associated with antioxidant metabolites, osmotic regulatory substances, alterations in gene expression, and variations in metabolite concentrations at different levels (Theocharis et al. 2012). Plants can respond to oxidative stress through endogenous antioxidant enzymes, such as superoxide dismutase (SOD), peroxidase (POD), and catalase (CAT). These enzymes act synergistically to detoxify reactive oxygen species (ROS) produced internally, maintaining cellular oxygen radicals at lower levels and thus mitigating cold stress (Radyuk et al. 2009; Zhang et al. 2011). An increasing body of research has provided evidence for changes in osmotic regulatory substances—such as melatonin, proline, soluble sugars, and ascorbic acid—during plant adaptation to cold stress. These studies highlight the critical role of osmotic regulatory substances in enabling plants to withstand cold conditions (Li et al. 2010; Moustakas et al. 2011). Furthermore, gene expression changes induced by cold stress are essential components of the molecular mechanisms by which plants adapt to environmental challenges. Many biological processes are involved in transcriptional regulation, including photosynthesis, carbohydrate metabolism, MAPK signaling pathways in plants, starch and sucrose metabolism, and plant hormone signaling (Rolland et al. 2006; Han et al. 2016; Hwarari et al. 2022; Liu et al. 2023). For instance, numerous studies have confirmed that genes involved in carbohydrate metabolism are differentially expressed under cold stress (Wang et al. 2013; Calzadilla et al. 2016). Additionally, genes related to sugar metabolism, such as the β-amylase gene (BAM), which is involved in starch breakdown, are upregulated, facilitating sugar accumulation in tea plants (Yue et al. 2015).

The sea purslane (Sesuvium portulacastrum L.) is a perennial facultative halophyte that thrives in sandy environments and is widely distributed across tropical and subtropical coastal areas around the world (Lokhande et al. 2009). Numerous studies have demonstrated sea purslane possesses strong tolerance to salt, drought, and heavy metal stresses, alongside significant environmental adaptability (Ghnaya et al. 2005; Slama et al. 2008; Liu et al. 2023). Due to its salt tolerance and environmental resilience, Sesuvium presents unique advantages for marine environmental remediation. Recent research has revealed that floating beds of this plant can effectively accumulate inorganic salts, heavy metals, and organic pollutants from seawater, offering a green and eco-friendly solution for purifying aquaculture wastewater (Mariem et al. 2014; Senff et al. 2020). The plant primarily propagates through cuttings, yet its root development is sensitive to low temperatures, which significantly hampers the exploitation of its advantages due to its intolerance to cold. Our current research indicates that plants mitigate cold stress by enhancing the activity of antioxidative enzymes and increasing osmoregulatory substances; however, cold stress leads to the rupture and deformation of stem node cells, inhibiting root development (publishing). To date, transcriptomic technologies have been applied exclusively to explore the effects of salt stress on Sesuvium physiology and metabolism (Wang et al. 2022; Liu et al. 2023). The molecular mechanisms underlying root formation under cold stress, however, have yet to be reported. Previous studies utilizing RNA-seq techniques have investigated gene expression profiles in various plant species subjected to cold stress (Moustakas et al. 2011; Wang et al. 2018; Zhang et al. 2018), thereby demonstrating the effectiveness of this technology in analyzing plant gene regulation and differential gene expression networks at the transcriptional level. For instance, in tobacco cultivars, this approach revealed 6,892 differentially transcribed genes (DTGs) induced by low temperature acclimation (Luo et al. 2022). Similarly, a transcriptomic analysis of highbush blueberry (Vaccinium corymbosum L.) identified more than 14,970 genes that were up- or down-regulated during the formation of adventitious roots in response to low temperature stress (An et al. 2020). Thus, in this study, we employed the RNA-Seq platform based on Illumina technology to characterize the transcriptome response to low temperature by comparing the transcriptomes of low-temperature-acclimated S. portulacastrum L. plants, aiming to gain deeper insights into the molecular basis of this physiological adaptation. Furthermore, the identified candidate genes will contribute to elucidating the plant’s capacity to adapt to low temperatures and the root formation mechanism, potentially enhancing productivity and expanding its geographical distribution.

Materials and methods

Plant material

In this experiment, sea purslane strains that had been selectively bred for natural cold tolerance over three consecutive years at our research group’s production base were utilized. The trial was conducted at the pilot base of Ningbo University (29°46'38"N, 121°57'14"E), where sea purslane plants with consistent growth conditions and physiological ages were selected. Stem segments, measuring 12.15 ± 0.87 cm in length and including four nodes, were excised. The leaves from the two nodes furthest from the tip were removed, and the segments were then washed three times with seawater. Subsequently, they were disinfected in a saturated Ca(ClO)2 solution for three minutes, followed by an additional three rinses with seawater.

Experimental design

The stem segments were cultured in square plastic pots supported by a carrier made from polyethylene foam board. Each pot contained 10 L of a modified Hoagland nutrient solution. The constituents of 1 L of the modified Hoagland’s nutrient solution were as follows: 101.11 mg KNO3, 53.49 mg NH4Cl, 98.59 mg MgSO4·7H2O, 27.22 mg KH2PO4, 0.56 mg H3BO3, 0.4 mg MnCl2·4H2O, 0.05 mg ZnSO4·7H2O, 0.014 mg CuSO4·5H2O, 0.02 mg Na2MoO4·2H2O, 1.11 mg FeSO4·7H2O, and 1.5 mg EDTA-Na2. To minimize the impact of microalgae, the water used for cultivation was disinfected under an ultraviolet lamp for four hours. It was ensured that during the experiment, the two nodes from which leaves had been removed remained submerged in the nutrient solution. Preliminary experiments determined that the critical temperature for stem rooting was 9 °C; below this temperature, the plants desiccated, failed to take root, and perished. Consequently, the present experiment was conducted with three temperature groups: 9 °C (low temperature, LT), 17 °C (medium temperature, MT), and 25 °C (high temperature, HT). The pots were placed in a GXZ-260 C type light incubator (Ningbo, China) for cultivation. Each treatment group consisted of three replicates, with each replicate containing 15 sea purslane stem segments. Other cultivation conditions were maintained as follows: a light intensity of 90.76 ± 2.44 µmol· m− 2 ·s− 1, a photoperiod of 12 h light and 12 h dark, and a relative air humidity of 73.35 ± 3.76%. Continuous aeration was provided, and the nutrient solution was renewed weekly. The duration of the experiment was 30 days.

Sample collection and determination of hormones indicators

At the end of the experiment, stem node specimens were immediately submerged in an RNA preservation buffer (#R0118, Beyotime, China), rapidly frozen in liquid nitrogen, and subsequently stored at -80 °C for transcriptome sequencing and hormone concentration analysis.

For the assays of hormone indicators, frozen stem node samples weighing 0.5 g were homogenized in 5 mL of PBS buffer (0.01 M, pH 7.4) containing a protease inhibitor. The homogenate was then centrifuged at 12,000 × g for 20 min. We collected the supernatant from this process for experimental analyses, ensuring that the extraction procedures were conducted at temperatures between 0 and 4 °C. We determined the concentrations of five hormones—3-indole butyric acid (IBA), abscisic acid (ABA), auxin (IAA), and 1-naphthaleneacetic acid (NAA)—using a double antibody sandwich enzyme-linked immunosorbent assay (ELISA), in accordance with the manufacturer’s instructions. The operational steps were as follows: (1) Addition of the standards: Introduce standards into the designated wells, varying the concentrations as required; (2) Sample application: Allocate blank and test sample wells, adding the various samples accordingly; (3) Enzyme introduction: Administer enzyme-labeled reagent to each well, excluding the blank wells; (4) Incubation: Seal the plate with sealing film and incubate at 37 °C for 30–60 min; (5) Preparation of the wash solution: Dilute concentrated wash buffer 20–30 times with distilled water and reserve; (6) Washing: Carefully remove the sealing film, discard the liquid, and shake off excess; fill each well with wash buffer, leave for 30 s, then discard; repeat this step 5 times, and tap the plate to dry; (7) Color development: Add the prescribed amounts of color development agents A and B to each well, mix gently, and allow the color to develop at 37 °C for 10–15 min; (8) Reaction termination: Introduce 50 µL of stop solution to each well to halt the reaction (this action will cause a visible color change); (9) Measurement: Using the blank wells to zero the instrument, measure the absorbance (OD value) at a wavelength of 450 nm. Absorbance values were measured for all hormones at a wavelength of 450 nm and expressed in µmol/L.

RNA extraction, library construction and illumina sequencing

Total RNA was extracted from the stem node using TRIzol® Reagent (Plant RNA Purification Reagent for plant tissue) according the manufacturer’s instructions (Invitrogen) and genomic DNA was removed using DNase I (TaKara). The quality of the RNA was then assessed using a 2100 Bioanalyzer (Agilent), and quantification was performed with an ND-2000 (NanoDrop Technologies). Only high-quality RNA samples—characterized by an OD260/280 ratio of 1.8 to 2.2, an OD260/230 ratio of 2.0 or greater, a RIN (RNA Integrity Number) of at least 6.5, a 28 S:18 S ratio of 1.0 or greater, and a total amount exceeding 1 µg—were used for library construction. The RNA-seq transcriptome library was prepared with 1 µg of total RNA using the TruSeq™ RNA Sample Preparation Kit from Illumina (San Diego, CA). Messenger RNA was initially isolated using the polyA selection method with oligo(dT) beads and then fragmented in a fragmentation buffer. Double-stranded cDNA synthesis was carried out using the SuperScript Double-Stranded cDNA Synthesis Kit (Invitrogen, CA) with random hexamer primers (Illumina). The synthesized cDNA underwent end-repair, phosphorylation, and “A” base addition according to Illumina’s library construction protocol. The libraries were size-selected for cDNA target fragments of 300 bp on 2% Low Range Ultra Agarose and amplified by 15 PCR cycles using Phusion DNA polymerase (NEB). The libraries were quantified with a TBS380, and the paired-end RNA-seq sequencing library was sequenced on an Illumina HiSeq X Ten sequencer (2 × 150 bp read length).

Differential expression analysis and functional enrichment

Differential expression analysis was performed to identify differentially expressed genes (DEGs) between samples. Expression levels for each transcript were calculated using the Transcripts Per Million (TPM) method. Gene abundances were quantified using RSEM (http://deweylab.biostat.wisc.edu/rsem/). DEGs with an absolute log2 fold change (log2FC) greater than 1 and a Q value of 0.05 or less (DESeq2 or EdgeR) / Q value of 0.001 or less (DEGseq) were considered significantly differentially expressed (Robinson et al. 2010; Love et al. 2014). Furthermore, functional enrichment analyses, including Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG), were conducted to identify significant enrichment among the DEGs in GO terms and metabolic pathways using a Bonferroni-corrected P-value threshold of 0.05 or less. GO functional enrichment and KEGG pathway analyses were executed using Goatools (https://github.com/tanghaibao/Goatools) and KOBAS (http://kobas.cbi.pku.edu.cn/home.do) (Xie et al. 2011). In the GO analysis, genes and gene products were categorized into three domains: molecular function, cellular component, and biological process. The KEGG database was used to identify significant pathways for predicted target genes with an adjusted P-value threshold of 0.05 or less. This study revealed the expression patterns of genes under three temperature treatments, determining the up-regulated and down-regulated genes by comparing the groups.

RT‑qPCR verification of expression patterns

Table S1 lists nineteen differentially expressed genes (DEGs) that were selected from four signaling pathways involved in the response to low temperature. These genes were chosen to verify the reliability of RNA-seq data using RT-qPCR. Primer sequences for these genes, detailed in Table S2, were designed using Primer Premier software (version 5.0). The PrimeScript™ RT reagent Kit with gDNA Eraser (TaKaRa) facilitated cDNA synthesis, following the manufacturer’s instructions. RT-qPCR analyses were conducted using a Step One real-time PCR System (Roche, USA) with iTaq Universal SYBR Green Supermix (BIO-RAD). The glyceraldehyde-3-phosphate dehydrogenase (GAPDH) gene functioned as an endogenous control. Each reaction was replicated three times. We calculated the expression levels of the target genes using the 2–ΔΔCT method, as described by Livak and Schmittgen (2001). For normalization, the expression level of the GAPDH gene in the control group was set to 1.

Statistical analysis

RT-qPCR experiments were performed in triplicate, and results were expressed as the mean ± standard deviation (S.D.). We used one-way ANOVA for statistical analysis. Expression histograms were generated using GraphPad Prism version 8.0.2 (https://www.graphpad.com/scientific-software/prism/). To elucidate the overlapping results among comparison groups, we created multi-dimensional Venn diagrams with UpSetR plots.

Results

Overview of the transcript sequencing, functional annotation and classification of unigenes

We performed triplicate analyses for each temperature treatment. The total raw read counts for all nine samples ranged from 20,688,136 (MT1) to 46,825,334 (LT2), as shown in Table 1. After quality control measures, which included the removal of adapters and low-quality reads, the cleaned dataset ranged from 20,219,748 (MT1) to 45,840,590 (LT2) reads (Table 1). The GC content of the raw data varied between 43.22% and 45.45%, with an average of 43.78%.

Table 1 Overview of the sequencing and assembly
Full size table

The unigene sequences of S. portulacastrum were compared with a general functional database. Annotated unigenes were obtained as follows: 76,049 (44.42%) from the NR database, 34,177 (19.96%) from NT, 31,014 (18.11%) from KO, 51,370 (30%) from Swiss-Prot, 59,486 (34.74%) from PFAM, 58,625 (34.24%) from GO, and 22,320 (13.03%) from KOG (Supplementary Fig. 1A). KOG is a database of orthologous gene families. Among the S. portulacastrum unigenes, a total of 22,320 were annotated to 25 KOG pathways. The most dominant cluster for the stem segment of S. portulacastrum was that of translation, ribosomal structure, and biogenesis (4,065 genes, 18.21%), followed by posttranslational modification, protein turnover, chaperones (3,459 genes, 15.49%), and signal transduction mechanisms (1,678 genes, 7.51%) (Supplementary Fig. 1B). Additionally, all GO-annotated genes were classified into 43 GO terms across the three main categories: molecular function, biological process, and cellular components (Supplementary Fig. 1C). The biological process category was the most frequently annotated, with cellular process (34,558 genes, 58.94%) and metabolic process (29,283 genes, 49.94%) as the most common subcategories. The molecular function category ranked second in annotation frequency, led by binding (29,353 genes, 50.07%) and catalytic activity (24,210 genes, 41.29%). Within the cellular component category, the subcategories of cellular anatomical entity (25,896 genes, 44.17%) and intracellular (15,538 genes, 26.5%) were most common. Furthermore, the biological pathways of genes annotated in the KEGG database were analyzed, assigning 39,325 genes to 388 pathways (Figure S1). In this KEGG pathway analysis, categories such as cellular processes, environmental information processing, genetic information processing, metabolism, and organismal systems were identified, with 5, 2, 5, 10, and 9 subgroups each, respectively. The prominent categories included genetic information processing, signal transduction, and metabolism (Supplementary Fig. 1D).

Identification and functional annotation of differential expression gene (DEGs)

We identified 25,748 DEGs from the nine stem segment samples of S. portulacastrum (Figure S2). These results clearly differentiated the expression patterns associated with each temperature treatment, demonstrating consistent replication within each treatment (Fig. 1A). Additionally, principal component analysis (PCA) plots of the nine stem segment samples from S. portulacastrum, represented by circles for each temperature treatment, underscored the distinctions among treatments. Sample profiles were grouped into three clusters corresponding to the temperature treatments (Fig. 1B).

Fig. 1
figure 1

The gene expression pattern of DEGs of nine stem segment samples of S. portulacastrum. A: Hierarchical clustering for the DEGs among the nine stem segment samples. The blue color in the matrix represents downregulation of gene, the red color in the matrix represents upregulation of gene. B: PCA of nine stem segment samples in different temperature treatments

Full size image

The DEGs volcano analysis for the following comparison groups: LT_vs_MT, MT_vs_HT, and LT_vs_HT, is shown in Fig. 2 and Figure S3. These analyses indicated that the expression patterns of DEG genes in all three comparison groups were significantly different. There were 2,126, 10,695, and 8,708 significantly up-regulated genes, respectively, and 876, 5,694, and 1,795 significantly down-regulated genes, respectively (Fig. 2D). Additionally, a Venn diagram analysis revealed unique and overlapping sets of DEGs among the comparison groups. Figure 2E and F, along with Figure S4, indicated that there were 93 and 109 genes differentially expressed in all comparisons, suggesting that these genes are involved in the S. portulacastrum response to both chilling and freezing stresses. These genes are of great value for studies on low-temperature tolerance.

Fig. 2
figure 2

The analysis of expression differences in different comparison groups. A, B, C: the three comparisons cluster volcano of DEGs. One dot in the volcano represents one gene, red dots represent upregulated genes and green dots represent down-regulated genes, black dots indicate the undifferentiated genes. The smaller the qValue, the larger the -log10 (padj), the more significant the difference of the DEGs. (D) The bar charts of the DEGs numbers in three comparisons. (E) and (F) are Venn diagrams of up-regulated and down-regulated genes of three comparisons (LT_VS_MT, MT_VS_HT, LT_VS_HT), respectively

Full size image

Gene ontology and pathway enrichment analysis of DEGs

DEGs were categorized and enriched in GO and KEGG functional classes to identify sets of genes with differential expression. The GO classification of DEGs revealed their distribution across three primary categories. Notably, the number of DEGs in the LT vs. MT (5073 DEGs) and MT vs. HT (7879 DEGs) comparison groups were substantially greater than in the LT vs. HT (2188 DEGs) group (Figure S3). The up-regulated DEGs in the biological process category were primarily associated with cellular modified amino acid metabolic processes and cilium organization. In the cellular component category, up-regulated DEGs were predominantly related to the cytosol, extracellular region, and extracellular matrix. With respect to molecular function, the up-regulated DEGs were mainly involved in DNA binding and receptor ligand activity. Furthermore, most GO terms in the MT vs. HT and LT vs. HT comparisons were dominated by up-regulation, whereas the DEGs in the LT vs. MT comparisons were primarily down-regulated, indicating an increase in certain GO terms from MT to HT. Notably, in the biological processes category of MT vs. HT, DEGs related to structural molecule activity and ribosome biogenesis constituted 8.56% and 7.96%, respectively, and increased to 28.86% and 24.79% in LT vs. HT, respectively (Fig. 3B, C). In contrast, the down-regulated DEGs in the biological process category, cellular component category, and molecular function terms were ribosome biogenesis, ribosome, and structural molecule activity, respectively, for the LT vs. MT comparisons (Fig. 3A and Figure S3).

Fig. 3
figure 3

GO enrichment analysis of DEG comparison groups in the stem segment of S. portulacastrum. A, B and C are GO enrichment of the LT_ VS_ MT, MT_ VS_ HT, and LT_ VS_ HT, respectively. The red part of the plate is presented as GO enrichment of up-regulated genes. The bule part of the plate is presented as GO enrichment of down-regulated genes

Full size image

KEGG pathway enrichment analysis was also performed for each comparison group. A total of 3365 DEGs were enriched into 123 KEGG pathways in LT vs. MT. In MT vs. HT, 3589 DEGs were grouped into 124 KEGG pathways, whereas in LT vs. HT, there were 3868 DEGs across 128 KEGG pathways. The top 12 significantly enriched KEGG pathways in each comparison group are shown in Fig. 4. The top 12 most significantly enriched KEGG pathways (q-value < 0.05) in each comparison group are summarized in Supplementary Figure S5. Integrating these three comparisons, the major pathways involved in the cold response mechanism were identified as “Carbon fixation in photosynthetic organisms”, “Starch and sucrose metabolism”, “Plant hormone signal transduction”, and “Glycolysis/Gluconeogenesis”.

Fig. 4
figure 4

The analysis of KEGG pathway functional enrichment. The ordinate represents the top twelve enriched KEGG pathways, the rich factor is the proportion of the DEGs annotated in the KEGG pathway to the total annotated genes in the KEGG pathway. The redder color of the Q value, the more significant the enrichment of the KEGG pathway. The significant number of DEGs in the KEGG pathway is represented by the size of the circle. A, B and C are the KEGG enrichment up-regulated DEGs of the LT_ VS_ MT, MT_ VS_ HT, and LT_ VS_ HT, respectively

Full size image

K-means clustering analysis of DEGs

DEGs from three comparison groups were clustered using K-means analysis to elucidate the molecular mechanisms underlying the different temperature stresses in S. portulacastrum. Out of the eight subclusters, six (1, 2, 4, 5, 6, and 7) represented up-regulated genes and one (3) represented down-regulated genes across the different temperature stresses (Fig. 5A and Figure S6). The expression trends of subclusters 1 (one gene), 2 (four genes), 4 (nine genes), 5 (nineteen genes), and 6 (ninety-nine genes) continuously increased from low temperature (LT) to high temperature (HT), whereas subcluster 3 (five genes) showed a sharp decreasing trend. In contrast, subclusters 8 (1,584 genes) and 9 (9,904 genes) exhibited gene expression patterns from LT to HT that were relatively constant, as depicted by a horizontal line. GO enrichment analysis revealed that “membrane organization,” “GTPase activity,” “lipid metabolic process”, “antioxidant activity”, and “carbohydrate metabolic process” were annotated in all clusters and were distinctly altered in clusters 2, 3, 4, 5, and 6 (Fig. 5B). Other GO biological processes, such as “GTPase activity” in cluster 6 and “extracellular matrix” in cluster 3, were also identified based on the K-means clustering analysis.

Fig. 5
figure 5

K-means clustering analysis of DEGs obtained among the LT, MT, and HT. A: A total of nine clusters were identified using the K-means clustering analysis. B: Functional annotation of the DEGs in the nine clusters

Full size image

Gene expression analysis of biological processes involved in the response to low temperature

In the present study, we discovered that DEGs in the treatments comparing three different conditions were primarily involved in four signaling pathways: “carbon fixation in photosynthetic organisms,” “starch and sucrose metabolism,” “plant hormone signal transduction-cell division,” and “light harvesting in the chlorophyll protein complex” (Fig. 6). Specifically, the majority of genes encoding enzymes responsible for carbon fixation in photosynthetic organisms—including phosphoenolpyruvate carboxylase (ppc), pyruvate, phosphate dikinase (ppdK), ribulose 1,5-bisphosphate carboxylase (rbcL), and transketolase (tktA)—were downregulated in the stem segment of S. portulacastrum under low-temperature (LT) treatment compared to the other two treatments. Moreover, five DEGs associated with the light-harvesting proteins of photosystem I (LHCA1), light-harvesting complex I chlorophyll a/b-binding proteins 2 and 4 (LHCA2 and LHCA4), the light-harvesting chlorophyll-a/b protein of photosystem 3 (LHCA3), and the light-harvesting complex I chlorophyll a/b-binding protein 5 (LHCA5) were altered in the stem segment of S. portulacastrum under LT treatment, thereby affecting the light-harvesting chlorophyll protein complex. Additionally, the study found that six DEGs involved with enzymes such as UTP-glucose-1-phosphate uridylyltransferase (UPG2), glucose-6-phosphate isomerase (GPI), fructokinase (fruK), and isoamylase (ISA1/ISA2) were upregulated under high-temperature (HT) treatment, while sucrose-phosphate synthase (SPS) and trehalose 6-phosphate synthase (TPS) were downregulated under LT treatment. This influenced starch and sucrose metabolism along with other indirect processes, including amino sugar and nucleotide sugar metabolism, and glycolysis. Furthermore, in plant hormone signal transduction related to cell division, the SAUR family protein (AHK2_3_4) and the two-component response regulator ARR-A family (ARR-A) were upregulated, while the expression of histidine-containing phosphotransfer protein (AHP) and two-component response regulator ARR-B family (ARR-B) genes were downregulated in the stem segment of S. portulacastrum under LT treatment compared to the other treatments.

Fig. 6
figure 6

The differential gene expression of genes involved in relate signaling pathway in each comparison groups under cold stress. The four signal pathways are shown in the figure: carbon fixation in photosynthetic organisms, starch and sucrose metabolism, plant hormone signal transduction-cell division, and light harvesting protein complex. In the carbon fixation in photosynthetic organisms signaling pathway, the left side of legend A is the read count heatmap of each sample gene in the signaling pathway, and the right side of legend A is the change of Log2FC value of genes in a comparison group, changes in the expression level of the heatmap are indicated by changes in color, with blue indicating a lower expression level, whereas red indicating a higher expression level; B is the schematic diagram of signal path; C is the relative expression levels of ppc, ppdK, rbcL and tkti genes in stem segments treated at different temperatures stress. The data are presented as the means ± S.D relative expression level. One-way ANOVA followed by Tukey’s multiple comparison test; different lowercase letters indicated P < 0.05 among different temperatures stress. Other signaling pathways are described in the same way

Full size image

DEGs involved in root primordium formation

In this study, 73 DEGs associated with plant hormone signal transduction were identified across the three comparison groups. Among these DEGs, 40, 22, 4, and 7 transcripts were involved in the auxin signaling pathway—AUX/IAA domains, adventitious root formation (ARF), lateral root formation (LRF), and shoot system development (new genes), respectively (Fig. 7 and Figure S7). Within the auxin signaling pathway, AUX1 and TIR1 factors were regulated by only one and two genes, respectively, and both were up-regulated (with fold changes ranging from 1.17 to 2.27) in the LT_ VS_ MT treatment. For IAA and ARF factors, gene expression was predominantly up-regulated (with fold changes ranging from 0.67 to 2.32) in the LT_ VS_ MT and LT_ VS_ HT treatments. Moreover, multiple early auxin signal transduction-related genes, such as SAURs and GH3, were involved, with most genes being up-regulated (with fold changes ranging from 0.78 to 2.71) in the LT_ VS_ MT treatment. Additionally, a significant number of transcripts annotated as ARF were detected and showed up-regulated expression in the LT_ VS_ MT and LT_ VS_ HT treatments. Notably, DEGs Cluster-11434.89373 and Cluster-11434.6564 exhibited upregulation in the LT_ VS_ MT and LT_ VS_ HT treatment, respectively. Conversely, DEGs Cluster-11434.36225 and Cluster-11434.93279 showed significant downregulation in the LT_ VS_ MT treatment and MT _ VS_ HT comparisons. Furthermore, seven new genes related to shoot system development were detected, including five genes (Cluster-11434.131698, Cluster-11434.125754, Cluster-11434.9818, Cluster-11434.32270, Cluster-11434.127065) with up-regulated expression in the LT _ VS_ MT treatment. Two genes (Cluster-11434.104361, Cluster-11434.89186) demonstrated down-regulated expression, while other comparison treatments did not show obvious expression patterns.

Fig. 7
figure 7

The gene expression related to root formation signaling pathways or associated transcription factors in stem segment of S. portulacastrum. A: Auxin signaling pathway-Aux/IAA domains in plants; B: Dynamic circle heatmap of Adventitious root formation (ARF); C: Dynamic circle heatmap of Lateral root formation (LRF); D: Dynamic circle heatmap of Shoot system development (new genes). Changes in expression level are indicated by a change in color, blue indicates a lower expression level, whereas red indicates a higher expression level

Full size image

Correlation analysis of root formation and hormones, and root development under different temperature stress

In the present study, we observed root development at the stem nodes of plants subjected to different temperature stresses. Roots in the low temperature treatment group were stunted (less than 2 cm in length) and sparse (no more than five in number). Conversely, in the medium temperature treatment group, roots were well-developed, numerous, and the longest root measured 15 cm. As temperature increased, the high temperature treatment group not only exhibited a substantial increase in root number, with the longest root reaching 27 cm, but also a proliferation of lateral roots, as shown in Fig. 8A. Furthermore, we quantified hormone levels in stem segments under various temperature stresses, revealing significant differences in hormone concentrations among the treatment groups. Hormone levels escalated with rising temperatures, with cytokinin (CTK) and indole-3-acetic acid (IAA) concentrations surpassing 100 µmol/L, the peak concentration being 282.04 µmol/L in the high temperature treatment group, as depicted in Fig. 8B. To elucidate potential correlations between root formation gene expression and hormone levels in plants treated at low temperatures, we conducted a correlation analysis. This analysis encompassed three hormonal signals pertinent to adventitious root formation (ARF), lateral root formation (LRF), and shoot system development, alongside five hormonal indices, as illustrated in Fig. 8C and Figure S8. Notably, of the 22 examined genes, only one (Cluster-11434.71074) exhibited a positive correlation with ARF, while the remainder showed negative correlations. Among these, four genes (Cluster-11434.20509, Cluster-11434.113634, Cluster-11434.98700, and Cluster-11434.107967) displayed a strong negative correlation, with seven genes (second circle) having Log2 Fold Change (Log2FC) values exceeding 2.5. Regarding transcription factors associated with LRF, there was an even split: two genes showed positive correlations and two negatives. Only the Cluster-11434.932790 gene had a Log2FC value greater than 2.5. Additionally, of the seven newly identified genes, four exhibited positive correlations with hormone indices, while the other three showed both positive and negative correlations with different hormones. Notably, only Cluster-11434.131698 and Cluster-127,065 genes presented Log2FC values greater than 2.5.

Fig. 8
figure 8

The network analysis of stem segment hormone concentration and root formation genes. A: The root morphology of plants under different temperature stress, the lower right is an enlarged view of the root system; B: Hormone concentration at stem segment under different temperature stress; C: The network analysis of stem segment hormone concentration and root formation genes based on the PCC values. The red and bright blue dots indicate hormone concentration and root formation genes, respectively. The red and blue lines indicate positive and negative correlation, respectively. The DEGs border width is directly related to Log2FC value. The size of the hormone dots is directly related to hormone concentration level

Full size image

Discussion

Plants must adjust to multiple levels of change, including physiological, metabolic, and molecular alterations, to adapt to low-temperature stress. For many tropical and subtropical plant species, exposure to low temperatures over a certain period increases their tolerance to freezing conditions, a phenomenon known as “cold acclimation” (Zhu et al. 2007). In our study, S. portulacastrum L., originating from Hainan, China (19°29'4"N, 110°6'13"E), exhibited tolerance to low-temperature stress with a stem segment rooting threshold of 13.5 °C prior to cold acclimation. After three consecutive years of natural overwintering and cold acclimation, this rooting tolerance threshold decreased to 9 °C. For hydroponically grown plants, the initiation of root formation in stem cuttings is an indication of environmental adaptation, facilitating the plants’ ability to perform photosynthesis and to acquire nutrients from their surroundings, which are essential for growth and metabolic needs. We further compared the transcriptome profiles of stem nodes subjected to various temperature stresses—extreme stress at low temperature (LT, 9 °C), the upper limit of cold stress (MT, 17 °C), and high temperature (HT, 25 °C)—to elucidate plant adaptability to low temperatures and root formation mechanisms. There was considerable overlap in the differentially expressed genes (DEGs) among the treatments, but also significant divergence (Fig. 2). The diversity of GO and KEGG pathway annotations, which were enriched in the LT_VS_MT and LT_VS_HT comparisons, highlighted the complexity and divergence of plant responses to temperature stress.

Quantitative analysis of the DEGs in each comparison revealed numerous genes induced by low temperatures, with the number of upregulated or downregulated genes involved in regulation varying inconsistently between the LT_VS_MT and LT_VS_HT comparisons (Fig. 2D). In LT_VS_MT, there were more downregulated genes than upregulated; conversely, more genes were upregulated in LT_VS_HT than in LT_VS_MT. This suggests that the cold acclimation process may contribute to increased plant tolerance to cold conditions (Chen et al. 2015). A similar trend was observed in rice, where a greater number of upregulated DEGs were detected at low temperatures compared to medium temperatures (Zhang et al. 2018). GO analysis revealed that terms such as “cytosol”, “antioxidant activity”, “thylakoid”, and “extracellular matrix” were commonly enriched among DEGs (Fig. 3), indicating that these processes are part of the general response to low-temperature stress. However, the enrichment of pathways in LT_VS_MT and LT_VS_HT was more specific. For instance, terms like “ribosome”, “ribosome biogenesis”, and “structural molecule activity” were exclusively upregulated and enriched in LT_VS_HT, with Log2FC values exceeding 17. The enrichment of “structural molecule activity,” involving 504 unigenes, signifies that acclimated plants may activate specific defense responses to cold stress (Figure S3). In Pinus koraiensis, genes associated with ‘stress response,’ ‘ADP binding,’ and ‘phosphotransferase activity’ were specifically upregulated in the low-temperature treatment group (5.3 °C) relative to the group exposed to moderate temperature stress (13.4 °C) (Wang et al. 2020). To survive under cold stress, plants have developed a series of complex signaling pathways that regulate metabolism and various biological processes. As Fig. 4 and S5 illustrate, Sesuvium differentially expressed genes (DEGs) related to biosynthesis and metabolic pathways were more enriched under low temperature (LT) conditions than under moderate temperature (MT) conditions. This suggests that low-temperature stress may induce a greater number of biosynthesis and metabolic genes than moderate temperature. Furthermore, the most significantly enriched KEGG pathway in the LT vs. MT comparison was “ribosome”, whereas “plant hormone signal transduction” was most enriched in the LT vs. high temperature (HT) comparison, indicating a pivotal role for plant hormone signal transduction in the abiotic stress response of Sesuvium. This finding is consistent with observations in salt-treated S. portulacastrum (Wang et al. 2022). Additionally, the KEGG pathways significantly enriched in the LT vs. MT comparison included “photosynthesis”, “MAPK signaling pathway-plant”, “amino acid metabolism”, and “starch and sucrose metabolism”. In contrast, the pathways enriched in the LT vs. HT comparison were “nitrogen metabolism”, “carbon fixation in photosynthetic organisms”, and “amino acid metabolism”. This indicates that a larger number of genes involved in signal sensing and transduction pathways are induced by LT treatment than by MT treatment, thereby enhancing the response to cold stress damage.

Enrichment analysis revealed that the DEGs common to all three comparisons were associated with“ribosome”, “photosynthesis”, “starch and sucrose metabolism”, and “brassinosteroid biosynthesis”, signifying distinctive changes in protein and carbohydrate metabolism. Notably, K-means analysis identified “GTPase activity”, “lipid metabolic process”, “antioxidant activity”, and “carbohydrate metabolic process” as highly enriched across all clusters (see Fig. 5). Generally, plants require more energy to mount stress response processes; hence, variations in carbohydrate content and energy consumption under cold stress are crucial for metabolic readjustment. Supporting this notion, similar studies have demonstrated that fluctuating temperatures can lead to unique transcriptional profiles implicated in cell membrane function, antioxidant activity, and various carbohydrate and lipid metabolic processes in tobacco cultivars (Luo et al. 2022). Moreover, chrysanthemum (Dendranthema grandiflorum) may undergo a multitude of metabolic alterations under low-temperature stress, including elevated levels of reactive oxygen species, increased energy demand, and upregulation of antioxidative enzymes and soluble sugar synthetases. Therefore, energy metabolism is essential for Sesuvium to withstand various stresses, and under conditions of low temperature and drought, this metabolism is enhanced. (Li et al. 2023).

When the temperature drops, the membrane lipids of the bio-membrane first undergo a phase transition, which leads to increased membrane permeability, enhanced electrolyte leakage, and inhibition of carbon fixation as well as photosynthesis (Wang et al. 2018). Carbon fixation is a critical primary metabolic process in plants that encompasses pathways such as the Calvin cycle, pyruvate metabolism, glycolysis, the TCA cycle, and fatty acid synthesis, and it is intricately linked with essential metabolites including lipids and carbohydrates (Raines 2006; Valenzuela et al. 2012). Variations in gene expression related to the signal cascade mechanism can influence the expression of genes involved in carbon fixation during photosynthesis, such as ppc, ppdK, rbcL, and tkti. In this study, ppc, ppdK, and rbcL are identified as key factors responsive to carbon fixation signal transduction in Sesuvium, with numerous differentially expressed genes (DEGs) participating in these pathways. Moreover, ppc serves as a critical response element in the C4-dicarboxylic acid cycle signal transduction, contributing to the tricarboxylic acid cycle, gluconeogenesis, and phosphorylation processes (Hügler et al. 2007). Furthermore, when compared to the moderate (MT) and high-temperature (HT) treatments, the expression levels of the ppc, ppdK, and rbcL genes in S. portulacastrum stem segments were significantly down-regulated under low-temperature (LT) treatment. This suggests that low-temperature stress disrupts the balance of the C4 carbon fixation pathway, potentially by restricting carbon fixation in the Calvin cycle, thereby impeding photosynthesis (Koh et al. 2023). Intriguingly, in the light-harvesting chlorophyll protein complex signaling pathway, the transcript levels and gene expression of LHCA1, LHCA2, LHCA3, and LHCA4 were significantly higher under LT treatment compared to MT and HT treatments, whereas LHCA5 showed a contrary trend. The majority of photons that are converted into biochemical energy and biomass through photosynthesis are collected by the major light-harvesting chlorophyll a/b-binding antenna complex, light-harvesting complex II (LHC II) (Vetoshkina et al. 2022), while LHC I show differential expression in this study. However, under various stress conditions, phycobilins carry out state transitions, an adaptive mechanism that involves rapid physiological changes, where light-harvesting antennas move between the two photosystems to balance excitation pressure between Photosystem I (PSI) and Photosystem II (PSII) (Fristedt et al. 2009; Simkin et al. 2022). Therefore, it is speculated that these two photosystems undergo transitions to adapt to low-temperature stress in this study, and further investigations are necessary for a comprehensive understanding of the molecular basis of cold response in S. portulacastrum. Starch serves as the primary storage metabolite in plants, and it is well-documented that cold stress can trigger starch degradation, typically associated with enhanced tolerance (Thalmann and Santelia 2017). The relationship between freezing injury and sugar content has been confirmed in several plants (Krasavina et al. 2014; Zhao et al. 2021, 2022). In the present study, we identified several key genes involved in starch metabolism and investigated their expression patterns under cold stress. Notably, UGP2, which is a pivotal regulator of starch hydrolysis into maltose (Zeeman et al. 2010), plays an essential role in the response to cold stress (Kaplan et al. 2006). Our findings reveal that transcription and gene expression patterns of UGP2 were consistent across three treatments. However, the expression levels under the low-temperature (LT) treatment were significantly lower than those observed in the other two treatments. Conversely, transcription levels of TPS were significantly higher in the LT treatment compared to the other groups, although gene expression levels did not differ significantly. Cold stress results in decreased total carbon assimilation, while concurrently increasing the content of water-soluble carbohydrates, thereby enhancing the cells’ resistance to adverse environments (Li et al. 2023). This suggests that low-temperature conditions exert a significant regulatory effect on starch-sugar metabolism in Sesuvium. Additionally, plant hormones play a pivotal role in regulating stress responses and activating downstream cold pathways. Differential expression of genes involved in the signaling cascade can influence the expression of genes participating in the formation of plant hormone signal transduction and cell division, such as AHK_2_3_4, AHP, ARR-B, and ARR-A. Plant growth and developmental processes have evolved complex adaptive mechanisms to various environmental changes. In our study, genes encoding auxin-induced proteins (AHP, ARR-B) exhibited higher expression levels in moderate temperature (MT) and high temperature (HT) treatments compared to LT, suggesting that gene expression promotes plant development. A similar response to low temperature stress has been reported in Saccharum spontanum (Dharshini et al. 2020). Given the complexity of hormone biosynthesis, the important genes identified in this study will be pivotal for uncovering metabolic pathways related to cold stress. While this study is limited to transcriptome data, comparing these data with other datasets (such as proteins and physiology) will reveal potential genes associated with cold stress tolerance.

Low temperatures can impact photosynthesis, inhibiting CO2 assimilation, which may lead to cellular structural damage and metabolic disruption (Foyer et al. 2002; Savitch et al. 2011). Similarly, root development is affected by low temperatures, inhibiting their ability to absorb nutrients and water (Hund et al. 2007). Plant root development is a multifaceted process with two developmental pathways: the formation of the primary root during embryogenesis and the production of new root primordia following germination. This latter category includes lateral root formation (LRF) and the generation of adventitious roots (ARs) (Laskowski 2013; Bellini et al. 2014; Sheng et al. 2017). The specific genes involved in these pathways, particularly those initiating root primordia in Sesuvium, remain to be fully elucidated. In our study, we have identified several auxin stress response signals that play a role in the root development of Sesuvium. These include auxin response genes (Aux/IAA and SAUR proteins) and root formation response factors, which regulate root formation by binding to auxin-responsive elements and modulating auxin gene expression (see Fig. 7). The plant hormone auxin is recognized to stimulate lateral root formation and is essential for initiating lateral root primordia to form autonomous LRs, as reviewed by Casimiro et al. (2003) and Sheng et al. (2017). In Arabidopsis, auxin is the principal hormone that promotes the growth of lateral and adventitious roots, inducing the dedifferentiation of cortical and parenchyma cells into initial cells and initiating cell division (Liu et al. 2005). Furthermore, low-temperature stress has been found to affect auxin biosynthesis in the meristematic zones of primary and lateral roots, an essential aspect of plant developmental activities (Ljung et al. 2005). Our study also implicates ARF (adventitious root formation), LRF (lateral root formation), and shoot system development genes in root formation (refer to Fig. 7). Notwithstanding, research has indicated variability in the auxin response of Arabidopsis primary roots when grown in different culture media, suggesting that diverse environmental cues may lead to distinct auxin responses in roots (Sheng et al. 2017). Our results further demonstrated that the pattern of root formation in stem segments of Sesuvium typically begins with the emergence of adventitious roots, subsequently followed by the development of lateral roots (shown in Fig. 8A). This finding aligns with other reports asserting that root systems form through a repetitive process where lateral roots are produced along the length of other roots (Deak and Malamy 2005). Considering the intricacies of hormone biosynthesis, our study measured relevant hormone concentrations (depicted in Fig. 8B) and constructed a network diagram to delineate the relationship between hormones and key genes, thus allowing further exploration of the interplay between hormones and gene regulation related to root formation under low-temperature stress (illustrated in Fig. 8C). This was done to elucidate the regulatory pathways of root formation related to low-temperature stress. In this study, hormone levels of CTK and IAA were maintained at a relatively high level, showing significant correlations—either positive or negative—with key genes. Previous research has demonstrated that IAA and IBA affect adventitious root (AR) formation by regulating the balance between auxin and other hormones (An et al. 2020). However, AR formation is a complex process that is regulated not only by differentially expressed genes (DEGs) in the auxin signaling pathway but also by numerous genes and transcription factors associated with various biological pathways; thus, genes from other pathways may also play a role in regulating AR formation. In future studies, we aim to perform a comprehensive analysis of transcriptome data to enhance our understanding of the mechanisms underlying low-temperature tolerance in Sesuvium and the regulation of AR formation. This will encompass physiological metabolism-related factors, including antioxidants and reactive oxygen species, transcription factors, and DEGs involved in the biosynthesis, distribution, transport, and degradation of auxin.

Conclusion

Overall, through comprehensive transcriptomic analysis and network assessment under cold stress, we have identified signaling pathways and key genes that contribute to differential cold tolerance and root formation in S. portulacastrum L. Our study thus offers a valuable transcriptomic resource for investigating cold adaptation mechanisms in this species and will assist in the identification of pivotal genes involved in the cold response, paving the way for genetic modifications to enhance Sesuvium’s yield and extend its geographical distribution to higher latitudes.