Parasponia PanNSP2 was studied previously, revealing it represents a single-copy gene essential for rhizobium-induced nodulation (van Zeijl et al., 2018). To enable ectopic expression of a stable PanNSP2 transcript, we first identified the putative miR171h target site (Lauressergues et al., 2012). Such a putative site was identified in the coding region and subsequently removed by introducing synonymous DNA mutations ( Supplementary Figure S1A ). An additional alteration was included in the PanNSP2 coding region, which may enhance expression stability; the insertion of an intron ( Supplementary Figure S1B ) (Feike et al., 2019). Ultimately, two miR171h resistant PanNSP2 versions (mNSP2) were used for Parasponia transformation, a construct with and one without an engineered intron, both driven by the constitutive L. japonicus UBIQUITIN1 promoter (pLjUBQ1) ( Supplementary Figure S1B ) (Maekawa et al., 2008).

Six transgenic Parasponia lines were selected, displaying a gradient level of PanNSP2 ectopic expression ranging from a 6 to 95-fold increase in root tissue ( Figure 1A ). In the shoots of these transgenic lines, mNSP2 was also expressed; in tissues where PanNSP2 transcripts are normally not detected ( Supplementary Figure S2A ). Noteworthy, lines possessing the mNSP2intron construct generally have a lower transgene expression when compared to constructs possessing mNSP2 with only an adapted miR171h target site.

To analyze whether morphogenic phenotypes are associated with mNSP2 ectopic expression in Parasponia, we selected lines 1, 3, and 6 that possess a 6-, 51-, and 95-fold enhanced NSP2 expression in root tissue compared to empty vector control plants. First, the root systems of 34-day-old plantlets grown in vitro on EKM medium containing low levels of ammonium nitrate (0.375 mM) and high levels of phosphate (3 mM) were analyzed. Control plants transformed with an empty vector showed a highly branched root system with up to 80 secondary lateral roots. This contrasts with what was observed in m NSP2ox lines 3 and 6, in which the root system was less branched, with only ~30 secondary lateral roots, whereas the primary root length is unaffected ( Figures 1B, C ). Such root phenotype was not observed on m NSP2ox line 1. As this line has only moderately levels of transgene expression compared to lines 3 and 6, it suggests that repression of lateral root growth has a threshold of PanNSP2 expression. Next, we analyzed the shoot architecture of 60-day-old plantlets grown on potting soil. In the case of shoot branching, control plants transformed with an empty vector exhibited an average of 11 branches. In contrast, mNSP2ox lines showed increased shoot branching, with an average of 14 to 16 branches, respectively ( Figure 1D ). This increased number of lateral branches is associated with an increased number of internodes ( Figure 1E ). This phenotype is also observed in mNSP2ox line 1, indicating that NSP2-controlled developmental effects on plant architecture in root and shoot might have a different threshold. Quantification of internode diameter suggests that ectopic PanNSP2 expression may also increase stem thickness, even though the trees were all similar in size ( Supplementary Figure S3 ). Taken together, the data show that ectopic mNSP2 expression affects shoot and root development in Parasponia.

Over-expression of an NSP2 miR171h resistant allele in M. truncatula roots enhances arbuscular mycorrhizal colonization under phosphate-limited conditions (Lauressergues et al., 2012). Furthermore, a more recent study showed that in M. truncatula and barley miR171h resistant NSP2 over-expression promotes arbuscular mycorrhizal root colonization even under exogenous phosphate conditions that are less amenable for the mycorrhizal symbiosis. This response is associated with the up-regulation of the strigolactone biosynthesis pathway (Li et al., 2022). We questioned whether Parasponia m NSP2ox lines also possess enhanced mycorrhization colonization levels. We tested two phosphate conditions; 20 μM and 3 mM exogenous Pi, and inoculated all six m NSP2ox lines with Rhizophagus irregularis DOAM197198 spores. Mycorrhization levels were scored 6 weeks post inoculation. Empty vector control plants can be colonized under both conditions, though the high exogenous phosphate concentration affects arbuscular mycorrhizal infection levels, most notably the number of vesicles formed. In the m NSP2ox lines, the number of hyphae, arbuscules, and vesicles was significantly increased under both phosphate conditions when compared to the empty vector control line ( Figure 2 ). At high exogenous phosphate conditions, the level of mycorrhizal colonization is associated with the level of PanNSP2 expression. Thus, we conclude that ectopic expression of PanNSP2 in Parasponia promotes mycorrhization, which is consistent with reported data in M. truncatula and barley.

To identify genes that are potentially directly or indirectly regulated by NSP2 in Parasponia, we investigated the transcriptional effect of increased mNSP2 expression under two distinct nutrient conditions. First plants were grown for 6 weeks on nutrient-rich medium containing 24.72 mM NO₃⁻, 2.6 mM NH₄⁺, and 2.6 mM PO₄ ^3- and subsequently transferred for 3 weeks to nutrient starved medium (no nitrogen (N) or phosphate (P) source). This treatment will enhance the mycorrhizal responsiveness of the plants. In the second treatment, the plants were transferred from the nutrient-rich medium to nodulation medium, which contains high phosphate levels (0.375 mM NH₄NO₃ and 3 mM PO₄ ^3-) (see Materials & Methods). The Pearson correlation test revealed that under nodulation permissive conditions, 2.73% (873 genes) exhibited a positive correlation with NSP2 expression, while only 0.95% (305 genes) showed a negative correlation (p-value, BH < 0.05). Conversely, under nutrient starved conditions, 1.70% (525 genes) were positively correlated with NSP2, and 1.74% (536 genes) were negatively correlated ( Supplementary Table S1 ). Most genes displayed a non-significant correlation with NSP2 under either condition ( Figure 3A ).

We performed a hypergeometric test on the overlap between the positive and negative correlated gene sets under both nutrient conditions. The results indicated a p-value of 0 for the overlap in positively correlated genes and a p-value of 1.250571e^-96 for the negatively correlated gene set. These findings led us to conclude that NSP2 consistently regulates a core set of genes, regardless of nutrient conditions, with a subset of genes showing differential regulation attributable to the nutrient environment ( Figure 3B ).

Studies in M. truncatula, rice, and barley suggest that ectopic expression of NSP2 enhances the colonization of plant roots by arbuscular mycorrhizal fungi, potentially through the modulation of the strigolactone biosynthetic pathway (Liu et al., 2011; Li et al., 2022). We investigated the activation of the methylerythritol phosphate (MEP), carotenoid, strigolactone, zaxinone, and abscisic acid (ABA) biosynthetic pathways in roots of Parasponia m NSP2ox lines ( Figure 4A ; Supplementary Table S1 ). This revealed a positive correlation between the expression of NSP2 and all genes encoding enzymes in the carotenoid, strigolactone, and zaxinone pathways ( Figure 4B ; Supplementary Table S1 ). The encoded enzymes of these genes convert geranylgeranyl pyrophosphate at the start of the carotenoid pathway into strigolactones (Okada et al., 2000; Ablazov et al., 2023). In addition, the expression levels of genes required for the biosynthesis of the apocarotenoid zaxinone correlate with the NSP2 expression. Zaxinone may act in a positive feedforward loop towards strigolactone biosynthesis, as was shown in rice (Votta et al., 2022). Taken together it is most probable that NSP2 ectopic expression leads to an increased biosynthesis of strigolactones in Parasponia root tissue.

Next, we questioned whether the transcriptional activation of the strigolactone biosynthetic pathway also occurs in the shoot of mNSP2ox lines. To test this, we used real-time polymerase chain reaction (qRT-PCR) to quantify the expression of three Parasponia genes, DWARF27 (PanD27), CAROTENOID CLEAVAGE DIOXYGENASE7 (PanCCD7), and PanCCD8, in shoots of lines 1, 3, and 6 under both nutrient conditions ( Supplementary Figure S2 , Supplementary Table S4 ). This revealed that NSP2 can also activate the expression of these genes in shoot tissue.

Together, these observations lead us to conclude that enhanced NSP2 expression in Parasponia markedly enhances the expression of genes essential for strigolactone biosynthesis and the apocarotenoid zaxinone across varying nutrient conditions.

We also analyzed the expression dataset for symbiosis genes of which the expression correlates with mNSP2 expression. This revealed that the expression of only a limited subset of the symbiosis-related genes positively correlated with mNSP2 expression ( Figure 5 ; Supplementary Table S1 ). Among the positively correlated genes, the transcription factor PanCYCLOPS, the two transporters STUNTED ARBUSCULE2 (PanSTR2) and YELLOW STRIPE-LIKE 1 (PanYSL1), and ANNEXIN 1 (PanANN1) were shared between both growth conditions. Intriguingly, the genes encoding a nuclear envelope localized cation channel CASTOR and the VAMP-associated protein VAPYRIN (PanVPY) show only positive correlation with NSP2 expression under nodulation permissive conditions, whereas studies in legumes indicated they are also essential for mycorrhizal symbiosis (Imaizumi-Anraku et al., 2005; Pumplin et al., 2010; Murray et al., 2011). Conversely, under nutrient starved conditions, a negative correlation with NSP2 expression was observed with several transcription factor genes, including NIN-LIKE PROTEIN1 (PanNLP1) and several NUCLEAR FACTOR Y genes: PanNFYA5, PanNFYA6, PanNFYA7, PanNFYB6, PanNFYB8, and PanNFYC4 ( Supplementary Table S3 ). This indicates a complex regulatory network where enhanced NSP2 expression differentially influences the expression of a few symbiosis-related genes depending on the nutrient status of the plant. This observation suggests a nuanced role of NSP2 in the regulatory network governing symbiotic permissiveness.

To investigate the gene expression patterns during mycorrhization and nodulation in Parasponia and their interplay with NSP2 expression, we plotted the Log₂fold changes of genes in mycorrhizal roots and young developing nodules ( Supplementary Table S3 ) against their Pearson correlation coefficients with PanNSP2 expression under both conditions ( Figure 6 ). This revealed that PanSTR2 not only exhibits upregulation in mycorrhizal roots, but also is transcriptionally induced young developing Parasponia nodules. When focusing on the carotenoid and strigolactone biosynthetic pathways, only a single gene stood out; namely PHYTOENE SYNTHASE1 (PanPSY1). Parasponia possesses three gene copies encoding phytoene synthases ( Supplementary Figure S4 ), of which only one correlates with NSP2 expression. PanPSY1 is transcriptionally enhanced in mycorrhizal roots (0.58 Log₂fold (1.5 fold) upregulated) and young nodules (6.47 Log₂fold (89.7 fold) upregulated) ( Supplementary Table S3 ). Phytoene synthase catalyzed the first step in carotenoid biosynthesis and is a major rate-limiting enzyme (Zhou et al., 2022). This suggests that NSP2-controlled of PanPSY1 expression represent a critical step in the biosynthesis of downstream products, including strigolactones in Parasponia roots that associate with arbuscular mycorrhizal fungi or nitrogen-fixing rhizobia.

Given that NSP2 is critical for both mycorrhization and nodulation, we investigated whether overexpression of mNSP2 could enhance symbiotic interactions with rhizobia. To test that, we grew all six mNSP2ox lines on nodulation medium (0.375 mM NH₄NO₃, 3 mM PO₄ ^3-) and scored the number of nodules at 8 weeks post inoculation with Mesorhizobium plurifarium BOR2. We observed fewer nodules on lines with the highest mNSP2 expression ( Figure 7A ). We examined the cytoarchitecture of these nodules, revealing a wild-type phenotype where infection threads and fixation threads were formed ( Figures 7B–D ). A surprising observation was the occurrence of abnormal cell divisions on roots of plants highly expressing mNSP2. To determine if this cell proliferation was spontaneous due to the level of PanNSP2 expression or a response to rhizobium application, we grew plantlets of three PanNSP2 over-expressor lines with fold change in gene expression of 6, 51, and 95 on nodulation medium either non-inoculated or inoculated with M. plurifarium BOR2. Under the non-inoculated condition, all m NSP2ox lines showed no such epidermal cell divisions, whereas under inoculated conditions, m NSP2ox lines with 51 and 95-fold increase in mNSP2 expression, but not 6-fold, exhibited foci of massive cell divisions ( Figures 7E–G ). This shows that high levels of NSP2 promote cell proliferation in response to rhizobium inoculation. Sections of these aberrant structures revealed mitotic activity in epidermal, cortical and endodermal cell layers, ( Figures 7H, I ). We concluded that increased NSP2 expression in Parasponia negatively regulates nodule formation and causes abnormal proliferation of root tissue in response to rhizobium inoculation.

The tissue culture propagation and maintenance of Parasponia plants were done on agar plates supplemented with Parasponia propagation medium. This medium contains 3.2 g/L SH-basal salt, 1 g/L SH-vitamin mixture, 10 g/L sucrose, 1 mL/L BAP (1 mg/mL), 100 μL/L IBA (1 mg/mL), 3 mL/L MES (1 M, pH 5.8), and 8 g/L Daishin agar (van Zeijl et al., 2018; Wardhani et al., 2019). Performance of nodulation and mycorrhization assays was carried out on platelets vegetatively grown and rooted following previously published procedures (van Zeijl et al., 2018; Wardhani et al., 2019).

Rooted plantlets were grown in 1 liter crystal-clear polypropylene pots (OS14BOX, Duchefa Biochemie, Netherlands) half-full, with a mixture (1:2 weight ratio) of river sand and agri perlite (Massmond-Westland, Netherlands). The mixture was watered with modified EKM medium [15 μM Fe-Citrate, 3 mM MES (pH 6.6), 6.6 μM MnSO₄, 4.1 μM Na₂MoO₄, 2.08 mM MgSO₄, 0.70 mM Na₂SO₄, 1.4 mM CaCl₂, 0.375 mM NH₄NO₃, 1.5 μM ZnSO₄, 0.88 mM KH₂PO₄, 1.6 μM CuSO₄, 2.07 mM K₂HPO₄, and 4 μM H₃BO₃] (Becking, 1983) and inoculated with Mesorhizobium plurifarium BOR2 (OD600 = 0.025). Inoculated plants were then incubated in a controlled climate room at 85% humidity, under a 16/8 h day/night growth condition for eight weeks. Plants were removed from pots and washed under running tap water followed by quantification of total nodule number and harvest of some nodules for sectioning. Subsequently, harvested nodules were fixed in 0.1 M phosphate buffer (pH 7.2) containing 0.5% glutaraldehyde followed by vacuum application for 60 minutes. Next, nodules were embedded in infiltration plastic, Technovit 7100 (Heraeus-Kulzer, Germany), following the protocol provided by the manufacturer. A Leica RJ2035 microtome was used to make 4.5 μm plastic sections which were stained with 0.5% toluidine blue. A Leica DFC425c camera was used to make high resolution images of the cytoarchitecture of the fixed nodules.

Rooted plantlets were placed in pots half-filled with a mixture (5:1 weight ration) of river sand and commercial potting soil and watered with half-strength Hoagland’s medium containing 20 μM phosphate (low P condition) or 3 mM (high P condition). Plants then were incubated in a controlled climate room at 85% humidity, under a 16/8 h day/night regime for two weeks followed by inoculation with 300 spores of Rhizophagus irregularis fungi (Agronutrion-DAOM197198, Carbonne, France) and subsequent incubation for 6 weeks. Afterwards, plants’ roots were harvested and treated with 10% (w/v) KOH and boiled for 30 minutes at 90°C and then stained in a staining solution contains 5% acetic acid and 3% ink (Waterman Vulpen, Zwolle, the Netherland) for 15 minutes at 90 °C. Next, fungal colonization and formation of arbuscules and vesicles were quantified, using the gridline intersect method (Giovannetti and Mosse, 1980), and high-resolution images were taken, using a Leica CTR66000 microscope.

All vectors used in this study were propagated in Escherichia coli strain DH5α and assembled, using Golden Gate Cloning (GGC) as was previously described (Engler et al., 2009). Level one and two acceptors and binary vectors used for the assembly were obtained from the GGC toolbox (Engler et al., 2014). Parasponia clone of NSP2 coding sequence (CDS) with and without insertion of intron 10 of Arabidopsis thaliana UBQ10 gene (intron) in the first putative splicing site in NSP2 CDS, including silent mutation in microR171h putative site and golden gate BsaI or BpiI restriction sites was synthesized as Level zero vectors. Subsequently, level zero vectors of NSP2, Lotus japonicus UBQ1 (LjUBQ1) promoter, C-terminal Myc tag, and Agrobacterium tumefaciens nopaline synthase terminator (T-NOS) were recombined into level one acceptor. Finally, level one vectors of NSP2 with or without (intron), including Kanamycin resistant Level one vectors were recombined into Level 2 binary vectors. All binary GGC vectors were validated by restriction enzymes and sequencing before transformation.

Level 2 binary vectors were transformed into Agrobacterium tumefaciens AGL1 strain, using Eppendorf Eporator (Eppendorf SE, Hamburg Germany). Transformed AGL1 were then grown on lysogeny broth (LB) agar plate supplemented with appropriate antibiotics and incubated at 28°C for 2 days. Parasponia plant transformation was carried out as previously described (van Zeijl et al., 2018; Wardhani et al., 2019), so young stems and petioles were collected from Parasponia trees grown in controlled greenhouse conditions at 85% humidity and under a 16/8 h day/night regime. The surface of the stems and petioles was sterilized with 2.5% bleach with a few drops of Tween 20, and subsequently stems and petioles washed six times with sterile water followed by a co-cultivation with AGL1 strains carrying constructs of interest for 2 days. Afterwards, transformed stems and petioles were transferred to callus induction medium supplemented with cefotaxime (300μg/ml) and kanamycin (50μg/ml) antibiotics for one week followed by transferring to plant propagation medium supplemented with cefotaxime (300μg/ml) and kanamycin (50μg/ml) antibiotics, refreshed every 1-2 weeks until transgenic shoots were formed. Genotyping was performed using a pair of primers designed to anneal to the L. japonicus UBIQUITIN 1 promoter (pLjUBQ1) and the C-terminal Myc tag, with amplification carried out using the Phire Plant Direct PCR kit (Thermo Fisher, F130WH), following the manufacturer’s protocol.

Parasponia plant lines, consisting of three NSP2 overexpression mutants (mNSP2ox1, mNSP2ox3, and mNSP2ox6), a nsp2 knockout (nsp2-9), and control line 46 with an empty vector were selected. The NSP2 overexpression mutants displayed NSP2 expression levels of 6, 51, and 95-fold higher than empty vector control. Selected plants were cultivated for 6 weeks on nutrient rich medium containing 24.72 mM NO₃⁻, 2.6 mM NH₄⁺ and 2.6 mM PO₄ ^3-. Next, cultivated plants were transferred for 3 weeks to either nutrient starved medium (no N and P resource), or the nodulation permissive medium (0.375 mM NH4NO3 and 3 mM PO4^3-).

Subsequently, healthy plantlets were transferred to pots containing a 2:1 weight ratio of perlite to sand. Each pot was supplemented with 150 mL of EKM medium, with or without nitrogen and phosphorus, referred to as nodulation permissive (low nitrate, high phosphate) and nutrient starved (low nitrate and low phosphate) treatments, respectively. Pots were populated with three plantlets of a single genotype and grown under a 16-hour light/8-hour dark photoperiod for 21 days. Following this period, roots from each pot were pooled for RNA extraction.

For each treatment, five root samples weighing 50 mg each were collected. Tissue lysis was performed using a QIAGEN TissueLyser LT, followed by sequential phenol:chloroform and chloroform extractions. Nucleic acid precipitation was performed using sodium acetate, followed by sequential washing steps with isopropanol and ethanol to purify the samples. Subsequently, the samples were treated with DNase to eliminate genomic DNA contamination. A second round of phenol:chloroform and chloroform extractions was conducted. The RNA was then re-precipitated in ethanol and re-suspended in nuclease-free water. The concentration of the isolated RNA was quantified using both a Biochrom SimpliNano™ Spectrophotometer and a Qubit 2.0 Fluorometer, utilizing Broad Range Assay Kits. RNA integrity was confirmed through gel electrophoresis. To validate NSP2 expression levels, quantitative polymerase chain reaction (qPCR) was employed. ACTIN and ELONGATION FACTOR1 ALPHA (EF1α) served as the housekeeping genes for normalization.

For each harvested sample, mRNA was isolated from total RNA utilizing poly-T oligo-attached magnetic beads. The resulting mRNA served as the template for first-strand cDNA synthesis, which was initiated with random hexamer primers, and subsequently followed by second-strand cDNA synthesis. Sample quality and nucleic acid concentration were quantitatively assessed using Qubit and qPCR assays, using EF1α and ACTIN for normalization genes. Size distribution of the samples was further confirmed via bioanalyzer analysis. Three biological replicates for each sample exhibiting the best characteristics (quantity and quality) for sequencing were selected for library preparation and sequencing, conducted at Novogene UK (Cambridge, United Kingdom). The resulting libraries were pooled and subjected to paired-end sequencing on an Illumina NovaSeq 6000 platform. Each biological replicate generated approximately 25 million reads, with each read consisting of 150 base pairs.

The quality of the sequenced reads was assessed using FastQC and subsequently visually inspected through MultiQC. Remaining adapters and low-quality bases were trimmed using Trimmomatic with settings TRAILING:3, SLIDINGWINDOW:4:15, MINLEN:50, HEADCROP:12. Transcript abundances were quantified with Kallisto, and the resulting files were loaded into R Studio version 2022.07.2 using R version 4.1.1. Differential gene expression analysis was conducted using DESeq2 for each condition (Nodulation permissive and nutrient starved) separately ( Supplementary Table S2 ).

Pearson correlation analysis was performed for each nutrient condition (nodulation permissive and nutrient starved) to assess the expression correlation with NSP2. The analysis was conducted in R, calculating the correlation, p-value, and False Discovery Rate (FDR, or adjusted p-value). An FDR cutoff of 0.05 was applied and used to categorize the correlation as positive, non-significant, or negative. The Pearson correlation graphs were plotted using the ggplot2 package. Overlapping genes between nutrient conditions were visualized using the VennDiagram package in R.

Prominent pathways featuring upregulated genes under NSP2 overexpression conditions, as well as genes of interest related to nodulation and mycorrhization in Parasponia, were selected for detailed investigation. The first set, named the “Symbiosis Network,” consists of a gene set in Parasponia previously annotated ( Supplementary Table S1 ). The second set includes genes involved in the carotenoid biosynthesis pathway in Parasponia. Orthologous genes in Parasponia involved in carotenoid biosynthesis were identified using gene IDs from Arabidopsis thaliana, as reported in the study by (Li et al., 2022) ( Supplementary Table S1 ). Corresponding gene IDs for Parasponia were identified through a reciprocal best BLAST hit approach, using the Arabidopsis gene IDs as references. BLASTP was performed with the following parameters: E-value cutoff of 1e-5, BLOSUM62 scoring matrix, gap penalties of 11 for existence and 1 for extension, and a word size of 3. The top BLAST hits in Parasponia were then used for a reverse search against the TAIR Araport 11 protein set to verify orthologous relationships.

Parasponia wild type young nodule data were previously published under PRJNA272473. Additionally, an experiment was conducted on Parasponia plants inoculated with Rhizophagus irregularis (MYC+) or left uninoculated (MYC-) [(van Velzen et al., 2018; van Zeijl et al., 2018; Bu et al., 2020; Rutten et al., 2020; Alhusayni et al., 2023)]. Parasponia plants were grown under controlled greenhouse conditions with a 16-hour light/8-hour dark photoperiod at 28°C and 85% humidity. The growth medium consisted of a mixture (5:1 ratio) of river sand and commercial potting soil and watered with half-strength Hoagland solution containing 20 μM PO₄^3- every other week. After placing the plants in pots, they were allowed to recover for 2 weeks before being inoculated with 300 R. irregularis spores. Root samples were harvested 42 days after inoculation.

Total RNA was extracted from collected samples using the RNeasy Plant Mini Kit (Qiagen). mRNA was isolated from total RNA using poly-T oligo-attached magnetic beads. The isolated mRNA served as the template for first-strand cDNA synthesis, initiated with random hexamer primers, followed by second-strand cDNA synthesis. Sample quality and nucleic acid concentration were assessed using Qubit and qPCR assays. Size distribution of the samples was confirmed via bioanalyzer analysis. Three biological replicates per sample exhibiting the best characteristics (quantity and quality) were selected for library preparation and sequencing.

Sequencing was conducted at Novogene UK (Cambridge, United Kingdom) using an Illumina NovaSeq 6000 platform. Each biological replicate generated approximately 25 million paired end reads of 150 base pairs. The quality of the sequenced reads was assessed using FastQC and subsequently visually inspected through MultiQC.

To analyze the RNA-seq dataset from nodule stages 1, 2, and 3 versus roots, as well as mycorrhization data of Parasponia wild type MYC- and MYC+ data, the following procedures were performed using the R programming language. Transcript abundances were quantified using Kallisto. The resulting files were imported into R Studio version 2022.07.2 with R version 4.1.1 for subsequent analysis using DESeq2 as previously described ( Supplementary Table S2 ). The differential expression results were integrated with existing annotated datasets for nodulation permissive and nutrient starved conditions. Ggplot2 was used to plot Log₂fold change versus the NSP2 correlation ( Supplementary Table S3 ).

Box plot graphs were generated using BoxPlotR, an online tool developed by (Spitzer et al., 2014). Bar charts were created in Microsoft Excel for Mac (Version 16.89). The Kruskal–Wallis test, followed by Tukey’s post-hoc test, was employed for statistical analysis. Statistical significance was established at a threshold of p < 0.05. Additionally, the quantification of internode diameter was assessed using a Student’s t-test, with statistical significance also defined as p < 0.05.