Medicago truncatula: M. truncatula jemalong A17 seeds were scarified, sterilized, plated on water agarose medium and transferred to 4°C for 3 days in the dark. Seeds were allowed to germinate in the dark at 23°C for 16 h. Germinated seedlings with similar root emergence were transferred to Broughton & Dilworth (B&D) nutrition medium (with 1% agarose) in “filter paper sandwich” systems and grown under long-day conditions (16-/8-h day/night cycle) with 120 mol m² s^–1 light for 4 days (Breakspear et al., 2014). Four-day old seedlings uniformly germinated with roots 3–4 cm in length were sown in cutouts of Identi-Plug foam (Jaece Industries Inc., NY, United States) and 15 mL falcon tubes with the bottom cone cut away, and placed into aerated hydroponic tanks containing B&D full nutrition medium (Supplementary Table 1 and Figure 1A). Plants were grown under long-day conditions (16 h light, 8 h dark) in Conviron walk-in plant growth rooms at 22°C temperature and 120 mol m² s^–1 light for 11 additional days. Prior to the nutrient uptake experiment, the plants were then transferred to a macronutrient-free nutrient solution (500 μM CaCl₂, 1,000 μM MES) used for pre-treatment with the respective peptide treatment of 1 μM and micronutrients unless otherwise stated, for 48 h before measurement (Figure 1B).

Arabidopsis: A. thaliana Columbia-0 seeds were surface sterilized (Household Bleach diluted to 50% in water followed by 75% ethanol treatment) and plated aseptically on 1/2 Murashige & Skoog [with 0.4% (w/v) Gelzan] and transferred to 4°C. After 2 days the seedlings were transferred to 22°C and grown under long-day conditions (16 h light, 8 h dark) with 120 mol m² s^–1 light for 4 days or till roots were about 1 cm long. These seedlings were then transferred to Falcon 6-well culture plates (Corning, AZ, United States) containing 5 mL of liquid 1/2 MS and grown for 10 days on a shaker (80 rpm, New Brunswick platform shaker). Ten plants were pooled together per well to make a biological replicate. The plants were then transferred to a macronutrient-free solution with 1 μM AtCEP1 peptide treatment (Pepscan, Netherlands), 48 h prior to the nutrient uptake experiment, in order to promote uptake induction without significant developmental change in root length.

For the nutrient uptake experiment, plants were processed following the RhizoFlux ions protocol with modifications (Griffiths et al., 2021). A custom ion uptake analysis assay was used with individual plant hydroponic chamber control of nutrient solutions or treatments. The setup consisted of 24 chambers coupled to two peristaltic pumps for nutrient sampling and aeration (Ismatec ISM944A, Cole-Parmer Instrument Company LLC, IL, United States) (Figure 1C). For the Arabidopsis experiments the plants were grown on a shaker and nutrient sampling was conducted with a pipette. Each chamber was filled with a “procedure” solution containing the respective peptide treatment and macronutrients: (in μM) 100 KNO₃, 100 NH₄Cl, 12.5 Ca(H₂PO₄)₂H₂O, 25 MgSO₄, 487.5 CaCl₂, 1,000 MES (adjusted to pH 6.8 using NaOH). The procedure solution volume used in the Medicago nutrient uptake experiments was 35 mL per plant and for the Arabidopsis experiments 15 mL per pool was used. Two minutes after the macronutrient-starved plants were transferred to the individual chambers the first 650 μL nutrient sample was collected. Nutrient solution samples were taken between 0 and 4 h on a deep-well collection plate and the plate was sealed with a sealing tape plate cover and transferred to 4°C for short term storage or −20°C for long-term storage. After the nutrient uptake experiment, the plants were immediately transferred to 4°C in plastic bags for later root image processing within 2 weeks. Ion concentrations of the collected nutrient solution samples were determined using a Thermo Scientific ICS-5000+ ion chromatographic system (Thermo Fisher Scientific, MA, United States) and the data processed to give nutrient concentrations using the Chromeleon 7.2 SR4 software (Thermo Fisher Scientific, MA, United States) (Figure 1D).

To investigate nutrient responsive effects on the M. truncatula Jemalong A17 transcriptome, plants were first germinated and grown on full nutrient plates with 1% agarose. Three-day old seedlings were then transferred to low nitrate (50 μM NH4NO3), low P (6 μM KH2PO4) and sulfate free B & D liquid media (Supplementary Table 2). After 48 h, root material from 20 seedlings per biological replicate, was harvested and immediately frozen in liquid nitrogen. For RNA sequencing, 3-day old M. truncatula seedlings grown on water agarose (Life Technologies) medium were treated with 1 μM AtCEP1:Hyp4,11 (DFRPTNPGNSPGVGH), MtCEP1D1:Hyp4,11 (AFQPTTPGNSPGVGH), and MtCEP1D2:Hyp11 (EFQKTNPGHSPGVGH) peptide concentrations in water for 3 h (Mohd-Radzman et al., 2015). For all three biological replicates 20–30 seedling roots were used. A RIN number of 9 to 10 (10 being the purest) were used in this study for RT-PCR and RNAseq. Harvested tissue for each biological replicate per treatment per control was collected under liquid nitrogen and immediately frozen. Harvested frozen tissue was ground under liquid nitrogen manually using mortar and pestles to a powder-like texture prior to RNA extraction.

TRIzol reagent was used to extract total RNA (Life Technologies) following the manufacturer’s protocol (Invitrogen GmbH, Karlsruhe, Germany). Total DNA was digested with RNase free DNase1 (Ambion Inc., Houston, TX, United States) and column purified with RNeasy MinElute CleanUp Kit (Qiagen). RNA was quantified using a Nanodrop Spectrophotometer ND-100 (NanoDrop Technologies, Wilmington, DE, United States). RNA integrity was assessed on an Agilent 2100 BioAnalyser and RNA 6000 Nano Chips (Agilent Technologies, Waldbronn, Germany). First-strand complementary DNA was synthesized by priming with oligo-dT₂₀ (Qiagen, Hilden, Germany), using Superscript Reverse Transcriptase III (Invitrogen GmbH, Karlsruhe, Germany) following manufacturer’s protocol. Primers were designed using Primer Express V3.0 software. qPCR reactions were carried out in QuantStudio7 (Thermo Fisher Scientific Inc.). Five microliters reactions were performed in an optical 384-well plate containing 2.5 μL SYBR Green Power Master Mix reagent (Applied Biosystems), 15 ng cDNA and 200 nM of each forward and reverse gene-specific primer. Transcript levels were normalized using the geometric mean of two housekeeping genes, MtUBI (Medtr3g091400) and MtPTB (Medtr3g090960). Three biological replicates were included and displayed as relative expression values. Primer sequences are provided in Supplementary Table 4.

One microgram of total RNA was used to generate RNA-seq libraries using TruSeq Stranded mRNA Library Prep Kit (Illumina Inc.) according to the manufacturer’s protocol. Prior to library construction, RNA integrity and quality were assessed with TapeStation 4200 (Agilent) and only an RNA integrity number (RIN) above nine was used. Size distribution of RNA-seq libraries was analyzed using TapeStation and the libraries were quantified using the Qubit 2.0 Fluorometer (Thermo Fisher Scientific) before being shipped to Novogene Inc., for sequencing at 150 bp paired-end with an Illumina Hiseq4000 (Illumina). The sequencing read depth was estimated to be 30–40×. Data are available on NCBI under SRA number PRJNA764762 (Biosamples SAMN21528496 to SAMN21528507).

Roots were imaged using a flatbed scanner equipped with a transparency unit (Epson Expression 12000XL, Epson America Inc., CA, United States). The roots were cut away from the shoots, spread out on a transparent plexiglass tray (420 mm × 300 mm) with a 5 mm layer of water (400 mL) avoiding root overlap, imaged in grayscale at a resolution of 600 dpi, and the total root length for each image was analyzed using RhizoVision Explorer v2.0.1 (Seethepalli et al., 2021). One root scan was performed per biological replicate for the nutrient uptake experiments, with one scan per Medicago plant, and one scan per pool of Arabidopsis plants. One root scan was performed per biological replicate for the nutrient uptake experiments, with one scan per Medicago plant, and one scan per pool of Arabidopsis plants. For all images using RhizoVision Explorer, image thresholding level was set to 200; filter of particles larger than 1 pixels; edge smoothing set to 2; root pruning threshold set to 1.

For the nutrient uptake rate study, data processing to determine specific nutrient uptake rates was conducted using R version 3.6.0 (Team, 2020)(R Core Team, 2020) with minor modification to the R code available at https://doi.org/10.5281/zenodo.3893945 (Griffiths et al., 2021). The net nutrient depletion and therefore uptake rate by roots was calculated by In = (C_t − C₀) / (t₀ − t) where In is the net influx into the plant; C₀ is the initial concentration of the solution at the start of the experiment t₀; C_t is the concentration at sampling time. The net uptake rate was then divided by the root system length (cm) to calculate the net specific nutrient uptake rate with the units μmol cm^–1 h^–1. Statistical tests were conducted using Graphpad V. 8. The bar in the box plots represents the median values, with each box representing the upper and lower quartiles, and the whiskers representing the minimum and maximum values.

Low quality bases and primer/adapter sequences were removed for quality trimming of each sample using Trimmomatic version 0.36¹. Reads less than 30 bases long after trimming were discarded, along with their mate pair. Using HISAT2 version 2.0.5² default mapping parameters and 24 threads, trimmed reads were mapped to an in-house mapped to an in-house re-annotated version of the M. truncatula genome release 4.0_reanno. Transcripts were assembled and quantified using Stringtie 1.2.4³ with the default assembly parameters. The transcripts identified in control (no peptide) and CEPp treated samples were unified into a single set of transcripts and compared with the reference gene annotation set using Stringtie’s “merge” mode. Differential expression testing was performed using DESeq2⁴ (see quality control plot in Supplementary Figure 4 and reads in Supplementary Table 5). Fold changes were calculated based on average FPKM values and differentially expressed gene’s (DEG’s) selected at a p-value cut-off of 0.05 and below.

The threshold for determining the DEGs was set to a fold change of 1.5 (log2 fold change > | 0.58|) and p-value cut-off < 0.05. For assessing the common DEGs, all the up and down-regulated DEGs along with shared DEGs between AtCEP1, MtCEP1D1, and MtCEP1D2 were analyzed as Venn diagrams using Venny 2.1.0 (Oliveros, 2016). The up and down-regulated DEGs were enriched for Gene Ontology (GO) terms using the online gene discovery platform, Legume IP V3 (Dai et al., 2021). GO term enrichment tool on the platform extracts GO terms from functional descriptions of protein in UniProt and InterproScan annotations. An adjusted p-value of p < 0.05 was used as a cut-off for GO terms to be considered enriched. Unique genes in the top 20 significantly enriched GO terms for all up and down-regulated DEGs were plotted. The up and down-regulated differential expression of known nitrate, phosphate and sulfate transporters were plotted as a heatmap log2FC > | 0.58|), p-value < 0.05. From the DEGs, upregulated kinases and transcription factors were plotted as a heatmap log2FC > 2.0 (corresponding to a fourfold change in expression level), p-value < 0.05. All plots were generated using GraphPad Prism 9.0.0 (GraphPad Software, Inc.) and modified using Adobe Illustrator (Adobe Inc. 2021).

A new platform for evaluating the effect of synthetic peptides on root uptake of multiple nutrient ions was developed (Figure 1). Plants were grown in a hydroponic system (Figure 1A) and peptides of interest were applied to the nutrient solution around the root system 48 h prior to nutrient uptake assays (Figure 1B). Plants were then transferred to small assay tubes containing nutrient solution with defined levels of nutrient ions and the respective peptide, which was sampled over a short-duration into a deep-well collection plate (Figure 1C). The anion and cation concentrations of the collected samples were determined using ion chromatography. A decline in ion concentration in the assay solution over time indicated a linear rate of net uptake of the ions (Figure 1D and Supplementary Figure 1). “Specific” nutrient uptake rates were calculated by dividing ion uptake rate by total root length obtained from image analysis. In a proof-of-concept experiment, exogenous application of Medicago SSP MtCEP1D1 increased the specific rate of nitrate uptake by 70–140% at low external concentrations (100 and 500 μM, p < 0.05 and p < 0.001, respectively) but not higher concentrations (1 and 5 mM) in treated plants compared to non-treated controls (Figure 1E).

To determine if expression of MtCEP1 was regulated by multiple nutrient stresses, we grew M. truncatula seedlings on agarose plates containing nutrients for optimal growth (B & D Full Nutrition medium) for 3 days. Seedlings were deprived of specific macronutrients for 48 h before quantitative RT-PCR estimation of endogenous MtCEP1 transcript abundance. Notably, nitrogen deprivation significantly enhanced MtCEP1 transcript abundance (p < 0.01) but not phosphate and sulfate deprivation (Figure 2A). To elucidate the functions of CEP1 peptides and identify key peptide domains, Arabidopsis CEP1 peptide (AtCEP1) and the M. truncatula CEP1 (MtCEP1) domain 1 and domain 2 peptides (Figure 2) were added to agar media upon which seedlings were grown, after which root system architecture was analyzed. In agarose, exogenous application of either Arabidopsis or Medicago CEP1 domain 1 (AtCEP1 and MtCEP1D1) significantly reduced lateral root number by ∼50% in Medicago as was previously reported (p < 0.01 and p < 0.05, respectively; Figures 2B,C; Imin et al., 2013). The median number of lateral roots in the presence of MtCEP1D2 was lower than control but was not statistically significant. The peptide application used in the nutrient uptake assay had no effect on root architecture as the 48-h exposure was too short for significant developmental change (Supplementary Figure 2).

The ion uptake platform was used to measure root uptake rates of multiple nutrients simultaneously. In addition to enhancing nitrate uptake rates (p < 0.01), application of 1 μM of the Arabidopsis AtCEP1 peptide significantly enhanced phosphate and sulfate uptake in A. thaliana (p < 0.05; Figure 3A). For M. truncatula, both AtCEP1 and Medicago MtCEP1 domain 1 peptide significantly enhanced the nitrate uptake rate (p < 0.05 and p < 0.001, respectively; Figure 3B). Unlike MtCEP1, that enhanced uptake of both phosphate and sulfate in Medicago, AtCEP1 peptide application did not enhance uptake of these two nutrients in M. truncatula, (p < 0.001 and p < 0.1, respectively; Figure 3B). Thus, MtCEP1 had a greater effect than AtCEP1 on uptake rates of nitrate, phosphate, and sulfate in Medicago (Figure 3B). In contrast, the CEP1 peptides had no effect on ammonium or potassium uptake rates in Medicago truncatula (Supplementary Figure 3).

Differential gene expression analysis of RNAseq data from M. truncatula seedling roots treated with the three CEP1 peptide variants revealed that MtCEP1D1 triggered the greatest changes in gene expression, with 2,466 genes affected by MtCEP1D1 application of which 1,349 were induced and 1,117 were repressed (Figure 4B). Application of CEP1D2 resulted in induction of 1,278 genes and repression of 871 genes. Fewer genes in Medicago were affected by treatment of plants with AtCEP1, with only 617 and 482 genes up and down regulated, respectively (Figure 4A and Supplementary Table 1). Interestingly, 322 genes were up-regulated and 116 were down-regulated by all three peptide treatments.

Gene Ontology enrichment analysis, using Legume IPV3, revealed several biological processes that were affected by CEP1 peptide treatments. Both MtCEP1D1 and AtCEP1 induced expression of genes involved in cell division (GO:0008283, GO:0022402, GO:1903047, GO:0000278, GO:0000280) and cell proliferation (GO:0008284). Both MtCEP1 peptides enhanced the expression of genes involved in pollination (GO:0009856), pollen-pistil interaction (GO:0048544, GO:0009875) and cell recognition (GO:0008037). Genes related to photosynthesis (GO:0015979, GO:0019684, GO:0009765, GO:0009768, GO:0009416, GO:0009767, GO:0009773, GO:0010109) and oxidative stress (GO:0000302, GO:0006979, GO:1901700, GO:0045454, GO:0009651, GO:0006970, GO:0042744, GO:0042743) were down-regulated in response to all three peptides. Genes involved in hormone responses were also affected by CEP1 peptides, with auxin response genes being repressed by both MtCEP1 peptide domains. Upregulation of genes responsible for phosphatase activity (GO:001092) and ABA response (GO:0009738, GO:0071215, GO:0009737) following CEP1D2 application were also found (Supplementary Table 2).

Given the observed increase in nitrate, phosphate, and sulfate uptake in response to CEP1 peptides, we looked for changes in the expression of gene families involved in these processes, namely the NITRATE/PEPTIDE TRANSPORTER (NRT/PTR), PHOSPHATE TRANSPORTER (PHT) and SULFATE TRANSPORTER (SULTR) families (Figure 4C and Supplementary Table 3). Of the 117 NRT transporter encoding-genes analyzed, 17 were differentially expressed following application of peptides of which seven were induced by MtCEP1D1. One putative sulfate transporter gene, an ortholog of AtSULTR3;5 (Medtr6g086170), was highly induced by all CEP peptide domains. In our data, we found no PHT phosphate transporter genes significantly induced by the CEP1p application.

Finally, we wanted to identify signaling pathway genes that responded to CEP1 application, which might be interesting targets for breeding crops with enhanced sensitivity to such peptides. We analyzed the top 15% of genes up-regulated by the peptides and focused on those involved in perception and/or relay of signals, especially kinases and transcription factors (Figure 4D and Supplementary Table 4). Four kinases: Cyclin dependent Kinase (Medtr8g461270), Serine/Threonine Kinase (Medtr7g056617) and an LRR receptor like kinase (MT4Noble_051661; Medtr2g019170) were highly induced by all three CEP1 peptides. Four Myb transcription factors were also induced by all three peptides. MADS-box transcription factor gene Medtr6g015975 was highly induced by both MtCEP1 domain 1 and 2 peptides, while Medtr4g109830 was induced only by MtCEP1_D2p.