In this study, we utilized the Gabes 2355 variety of Medicago sativa (alfalfa) and the Kl4 line of Hordeum marinum (sea barley). Gabes 2355, a product of a Tunisian breeding program, is recognized for its ability to thrive in saline environments and is widely cultivated in Southern Tunisia (Jabri et al., 2021). Alfalfa is typically planted in April and harvested up to three times annually (Guerchi et al., 2024). The Kl4 line of Hordeum marinum was sourced from the saline region of Sebkhet El Kalbia in Kairouan, central Tunisia. This line, developed over two generations of self-pollination at the Centre of Biotechnology of Borj-Cedria (CBBC), exhibits a high degree of homozygosity, reflecting the self-pollinating nature of H. marinum (Saoudi et al., 2017, 2019). The distinct growth patterns of alfalfa and sea barley allow for efficient resource use, including light, water, and nutrients, due to their temporal and spatial complementarity.

Prior to planting, soil analysis showed a pH of 7.5, available phosphorus of 3.08 ppm, total nitrogen of 0.1%, and exchangeable potassium of 130 ppm. According to standard agronomic thresholds, these values correspond to very low phosphorus (<10 ppm), low nitrogen (≤0.1%), and low potassium (100–150 ppm). Fertilizer treatments were therefore designed to correct these deficiencies and to evaluate plant responses under nutrient limitation. Soil was amended with phosphorus and potassium at three fertilizer levels: nutrient deficiency as 0% fertilization (0 kg P ha⁻¹, 0 kg K ha⁻¹), 50% of fertilization (75 kg P ha⁻¹, 45 kg K ha⁻¹), and optimum condition as 100% fertilization (150 kg P ha⁻¹, 90 kg K ha⁻¹) (Figure 1). Phosphorus was applied as monoammonium phosphate (52% P₂O₅) and potassium as potassium chloride, with rates selected based on local soil test results and regional alfalfa production guidelines.

Seed germination was initiated by placing the seeds on filter paper saturated with distilled water in Petri dishes, which were incubated in the dark at 25°C. After germination, seedlings were transplanted into 50-liter plastic boxes (61.5 cm × 39.5 cm × 33 cm) filled with 46 kg of meadow soil. In both monocropping and intercropping systems, plants were established in pots with a spacing of 5 cm between rows and 7 cm between plants within each row. In the intercropping treatment, rows of M. sativa and H. marinum were alternated within the same pot to ensure uniform distribution and equal access to growth space.

Each 50-liter box contained 15 plants, arranged in four cultivation modes: i) monoculture of M. sativa, ii) monoculture of H. marinum, iii) M. sativa in parallel intercropping, and iv) H. marinum in parallel intercropping. The intercropping setup consisted of one row of H. marinum (5 plants) intercropped with two rows of M. sativa (10 plants), with 5 cm spacing within rows and clear separation between species.

One week after planting, three nitrogen regimes were applied: nutrient deficiency (0 kg N ha⁻¹), 50% of fertilization (60 kg N ha⁻¹), and control (120 kg N ha⁻¹). Nitrogen was applied in the form of urea to minimize NH₃ volatilization. The amount of fertilizer applied per box was calculated from the recommended field dose (kg ha⁻¹). Specifically, the field dose was first converted to g m⁻² and then scaled to the box surface area (0.243 m²). Since plant performance under partial (50%) and control fertilization was comparable in the intercropping system at the first harvest, all subsequent metabolic and nutrient‐uptake analyses focused on the nutrient deficiency and partial fertilization (50%) treatments.

The experiment included three replicates for each cultivation method and treatment, totaling 27 boxes. Plants were irrigated every two days with water equivalent to 40% of the container capacity. The water availability level of 40% container capacity was chosen to simulate semi-arid conditions, which are characterized by limited water availability and frequent drought stress.

The study was conducted at the Centre of Biotechnology Borj-Cedria (CBBC) in Tunisia, located at 36°41′13″ N latitude and 10°22′55″ E longitude, 70 meters above sea level. This coastal region with a semi-arid climate has an average annual rainfall of 450 mm and a mean annual temperature of 18.6°C (Oueslati et al., 2024). The controlled glass-covered environment ensured no interference from rainfall or other external factors, with the experiment running from March to July 2023.

We evaluated three forage harvest timings, conducting each harvest at the onset of the earing stage for H. marinum. Simultaneously, the harvest for M. sativa was performed at the same growth stage, following the guidelines by Guerchi et al. (2024). The first harvest took place three months after sowing, at the end of May, followed by the second and third harvests at monthly intervals, at the end of June and the end of July, respectively.

Biomass yield was measured post-harvest by cutting the plants approximately 5 cm above ground level. After harvesting, plants were separated for morphological trait analysis under different treatments. The plants were uprooted, and the soil around the roots was washed away. They were then divided into root and leaf samples.

The fresh weight of the aerial parts and roots was recorded, and the samples were dried at 60°C for 48 hours in a Memmert UN55 oven (Germany). The dry matter content of the leaves and roots was subsequently determined. At the last harvest, for metabolic analyses, samples from each box were collected, lyophilized, and stored at -20°C until further analysis.

The extraction of soluble sugars was carried out as outlined in Castañeda et al. (2019). Lyophilized root and leaf samples (20 mg and 40 mg, respectively) were subjected to extraction three times with 1 mL of 80% (v/v) boiling ethanol for 30 seconds, followed by a final extraction at room temperature. The combined supernatants were evaporated to dryness in a Turbovap^® LV Evaporator (Zymark, Hopkinton, MA, USA) at 40°C and 1.2 bar pressure. The dried residues were reconstituted in 1 mL of deionized water through a two-step process, with 10 minutes of ultrasonication between steps. Subsequently, the reconstituted solution was centrifuged at 2300 g for 10 minutes at 4°C, and the supernatants were stored at −20°C for later analysis.

The remaining tissue, after extracting ethanol-soluble compounds, was transferred to an oven and dried at 70°C for 24 hours. The dried pellet was then weighed and resuspended in 5 mL of water. The suspension was boiled at 100°C to break the starch grains and then allowed to cool to room temperature. Subsequently, a desalted amyloglucosidase solution (EC 3.2.1.3, 0.5 units/mL acetate buffer, pH 4.5) was added to each tube. The tubes were incubated with shaking at 55°C for 12 hours to digest the starch. After incubation, the samples were centrifuged at 1200 rpm for 5 minutes, and the supernatant was collected for further glucose analysis (Gálvez et al., 2005).

Sucrose, fructose, and glucose levels were quantified using ionic chromatography on a 940 Professional IC Vario Metrohm system, equipped with Metrosep Carb2 guard and Metrosep Carb2 150/4.0 columns, operating at a flow rate of 0.5 mL/min, at 30°C, using 300 mM NaOH and 1 mM sodium acetate as the mobile phase (Calleja-Satrustegui et al., 2025).

Lyophilized samples (approximately 20 mg of roots and 40 mg of leaves) were ground into a fine powder using a mortar and pestle in liquid nitrogen. The resulting powder was homogenized in 1 mL of ultrapure deionized distilled water (DDW) and heated at 90°C for 30 minutes. After heating, the homogenate was filtered using a 0.45 µm filter. Ion levels were measured using ion chromatography on a 940 Professional IC Vario 2, Metrohm. Cations were analyzed through Metrosep C6 150/4.0 Metrohm column (0.9 ml/min; 45°C; HNO₃ 1.76 mM, picolinic acid 1.7 mM), while anions were analyzed using Metrosep A Supp7 150/4.0 Metrohm column (0.7 ml/min; 45°C; Na₂CO₃ 3.6 mM) following the manufacturer’s protocols.

Lyophilized samples (approximately 20 mg of root and 40 mg of leaves) were ground into powder under liquid nitrogen and then homogenized using a mortar and pestle. The extraction of free amino acids will be carried out in methanol/water/chloroform extracts (Hacham et al., 2002). Each ground sample was vortexed five times for 15 seconds in a cold water: chloroform: methanol (3:5:12 v/v) solution and incubated on ice for 10 minutes. The extracts were then centrifuged at 1200 rpm and 4°C for 2 minutes. The supernatant was collected in a 2 ml tube, and the pellet was resuspended in 600 μl of the water: chloroform: methanol solution. This mixture was centrifuged again for 2 minutes at 1200 rpm and 4°C. The supernatant was collected and combined with the previous supernatant. To this combined supernatant, 300 μl of chloroform and 450 μl of water were added. The mixture was centrifuged at maximum speed for 2 minutes. The upper phase (water: methanol) was collected and transferred to a new tube, then evaporated in a vacuum centrifuge.

The dry residue was dissolved in distilled water and stored at –80 °C until analysis. Organic acids were quantified by ion chromatography using a 940 Professional IC Vario system (Metrohm) equipped with a Metrosep A Supp16 150/4.0 column, operated at 1 mL min⁻¹ and 55 °C, and employing a gradient of 20 mM Na₂CO₃, 300 mM NaOH, and water. Detection was performed using a conductivity detector. For amino acid analysis, the samples underwent derivatization using 1 mM FITC dissolved in acetone at room temperature for 15 hours in a 20 mM borate buffer (pH 10). The content of free amino acids was measured using a Sciex MDQ+ (AB Sciex LLC, MA, EE.UU.) equipped with laser-induced fluorescence detection as described by Arlt et al. (2001)Takizawa and Nakamura (1998).

The study’s data were subjected to a two-way analysis of variance (ANOVA) to identify significant interactions between cropping mode and variables related to crop performance or metabolomic analysis. Only those variables showing significant interactions were selected for further detailed statistical examination. To compare the means, the Duncan test was applied at a 5% significance level. All statistical analyses were conducted using SPSS software, version 20.0.

Leaf and root ion concentrations showed clear species- and treatment-specific patterns. In leaves, Hordeum marinum (MHm) accumulated higher Cl⁻ and Na⁺ than Medicago sativa (MMs) under both fertilized (Cl⁻: 17.88 vs 9.89; Na⁺: 19.71 vs 13.25) and unfertilized conditions (Cl⁻: 15.01 vs 2.98; Na⁺: 18.11 vs 17.26), reflecting MHm’s salt-tolerant nature. In contrast, MMs leaves had higher K⁺ and Ca²⁺, particularly under fertilization (K⁺: 13.25; Ca²⁺: 35.11), indicating efficient nutrient assimilation. Intercropping further enhanced nutrient uptake, with IMs and IHm leaves showing higher NO₃⁻ (IMs: 0.; IHm: 0.18) and PO₄³⁻ (IMs: 2.69; IHm: 3.52) compared to monocrops. In roots, MHm maintained higher Cl⁻ and Na⁺ (fertilized IHm: Cl⁻ 12.19, Na⁺ 5.3), while MMs roots contained higher PO₄³⁻ (fertilized MMs: 3.16) and stable K⁺, suggesting complementary nutrient strategies. Fertilization amplified these differences, particularly for NO₃⁻, PO₄³⁻, and Mg²⁺ (IHm roots Mg²⁺ 8.09), highlighting the combined effects of nutrient supply and interspecific interactions (Table 1).

Biomass accumulation reflected these nutrient patterns. Cumulative aerial dry weight (ADW) significantly differed among cropping systems and fertilization levels (Figure 2). Under non-fertilized conditions (0%), M. sativa in intercropping (IMS, 4.0 g plant⁻¹) accumulated more biomass than its monocrop (MMs, 3.0 g plant⁻¹), while H. marinum in intercropping (IMH, 5.2 g plant⁻¹) outperformed its monocrop (MMH, 2.0 g plant⁻¹). At 50% fertilization, ADW markedly increased across treatments, with IMH exhibiting the highest biomass (15.9 g plant⁻¹), followed by MMH (11.3 g plant⁻¹), IMS (8.4 g plant⁻¹), and MMs (7.0 g plant⁻¹). Compared to monocropping, intercropping enhanced biomass by approximately 35% for M. sativa and 40% for H. marinum. Under 100% fertilization, IMH maintained the highest cumulative ADW (15.8 g plant⁻¹), whereas MMH, IMS, and MMs reached 12.0, 8.5, and 8.8 g plant⁻¹, respectively. Despite the increase in biomass with fertilization, the relative gain between 50% and 100% was limited, suggesting that moderate fertilization (50%) was sufficient to sustain optimal growth within the intercropping system.

Overall, ANOVA confirmed that intercropping and fertilization significantly influenced ion homeostasis and biomass production (Supplementary Tables S1, S2). These findings demonstrate that intercropping improved resource use efficiency and nutrient acquisition, particularly under moderate fertilization (50%), enhancing growth and stress resilience in the M. sativa–H. marinum association.

Carbohydrate analyses revealed distinct accumulation patterns in Medicago sativa and Hordeum marinum under varying nutritional regimes and cropping systems.

In M. sativa leaves, intercropping under nutrient deficiency significantly enhanced the accumulation of key osmoprotectants and soluble sugars. Myo-inositol increased to 12.38 µmol/g DW under stressed intercropping, compared to 8.09 µmol/g in fertilized monocropping. Sorbitol rose by 76.5%, while glucose and fructose increased by 6.2% and 99.9%, respectively, compared to intercropping in control. Notably, sucrose remained high across all intercropping conditions. In the roots of M. sativa, intercropping buffered the negative effects of low nutrition. Although myo-inositol and pinitol declined under stress, their levels were better preserved in intercropping than in monocropping. Mannitol showed a slight increase, and sucrose peaked at 200.37 µmol/g DW under intercropping with fertilization significantly higher than in all monocropping conditions. Interestingly, starch accumulation reached 463.66 µmol/g DW under low-intercropping, suggesting improved energy storage capacity under stress (Figure 3).

In H. marinum leaves, intercropping under stress induced strong accumulation of stress-related sugars. Trehalose increased 9-fold (from 0.17 to 1.65 µmol/g), and raffinose was newly detected at 21.74 µmol/g. Fructose and mannitol increased by 81.1% and 35.1%, respectively, compared to stressed monocropping. In roots, H. marinum under intercropping in stressed condition exhibited the highest levels of glucose (148.0%), fructose (102.7%), and sucrose (42.0%) compared to fertilized condition. Raffinose, absent in control conditions, accumulated under low-intercropping, supporting its role in stress tolerance (Figure 3).

According to the two-way ANOVA (Supplementary Table S3), these trends were statistically supported, with significant interaction effects among species, organ, and treatment. This confirms that the observed changes reflect species-specific metabolic adjustments to intercropping under varying nutritional conditions.

Organic acid profiles revealed marked modulation of key metabolic intermediates in response to intercropping and nutrient availabilityIn M. sativa leaves, intercropping with no fertilization significantly boosted citric acid (64.5% vs. fertilized-monocropping; 64.5% vs. unfertilized-intercropping) and succinic acid (66.7% vs. low-monocropping). Malic acid remained stable under intercropping but declined under monocropping. These patterns suggest that intercropping supports energy metabolism and osmotic regulation during stress. In roots, citric acid levels were 40.1% higher under unfertilized-intercropping than in unfertilized-monocropping and only 13.9% lower than in fertilized-intercropping. Succinic acid rose by 87.7%, and oxalic acid by 88.2%, suggesting enhanced ion balance and TCA cycle activity under nutrient deficiency (Figure 4).

In H. marinum leaves, stressed intercropping induced increases in citric acid (55.4%), succinic acid (121.7%), and malic acid (60.7%) compared to stressed monocropping. These compounds are central to energy generation and stress resilience. In roots, intercropping led to higher levels of citric acid (36.4%), succinic acid (77.6%), and oxalic acid (63.5%) when no fertilizer was added compared to stressed monocropping. Compared to intercropping with fertilization, levels declined modestly (14% to 18%), indicating metabolic stability under stress. Overall, intercropping maintained organic acid profiles that support both energy metabolism and detoxification (Figure 4).

According to the two-way ANOVA (Supplementary Table S4), these metabolic shifts were statistically significant, with strong interaction effects among species, organ, and treatment, confirming that intercropping and nutrient availability jointly shape organic acid metabolism in both species.

Amino acid analysis showed that intercropping modulates nitrogen and stress-related metabolism depending on species, organ, and nutritional status.

In M. sativa leaves, intercropping with no fertilization increased proline by 59.4% compared to fertilized intercropping, underlining its osmoprotective function. GABA rose by 33.3% and glutamine by 28.6% compared to stressed monocropping, whereas glutamic acid declined by 17.2%, indicating nitrogen flux toward protective metabolism. In roots, intercropping sustained GABA (44.9%) and glutamine (35.2%) under stress. Aspartic acid decreased under stressed monocropping (18.5%) but remained stable with intercropping, demonstrating a stabilizing effect of intercropping on amino acid balance under stress (Figure 5).

In H. marinum leaves, proline increased by 77.8% (vs. fertilized-monocropping), GABA by 51.4%, glutamine by 60.0%, and aspartic acid by 42.9% under unfertilized-intercropping. These responses suggest enhanced nitrogen redistribution and osmoprotection. In roots, arginine increased by 64.3%, GABA by 55.6%, and glutamine by 46.2% under unfertilized-intercropping vs. unfertilized-monocropping (Figure 5).

According to the two-way ANOVA (Supplementary Table S5), these patterns were statistically significant, with strong interaction effects among species, organ, and treatment, confirming that amino acid metabolism was tightly regulated by both intercropping and nutrient regime.

Our result showed that the intercropping system consistently produced greater biomass yields, regardless of the treatment applied, highlighting a synergistic interaction between alfalfa and sea barley. The increase in cumulative dry biomass, under optimal intercropping, underscores the complementary resource use between M. sativa and H. marinum (Guerchi et al., 2024). Legume–grass mixtures often enhance overall productivity through resource partitioning and stress buffering (Tang et al., 2021; Salinas-Roco et al., 2024). Legumes such as alfalfa fix atmospheric N₂ and exude organic acids and protons that mobilize soil phosphorus, thereby improving soil fertility for companion cereals (Tang et al., 2021). For example, wheat–legume intercrops achieve higher nitrogen accumulation and land‐equivalent ratios under low-N than high-N regimes (Salinas-Roco et al., 2024), and wheat–chickpea mixtures in Mediterranean climates boost wheat yield over monocultures (Kherif et al., 2021). Meta analyses further confirm that cereal–legume systems generally outperform monocultures, especially when nutrients are scarce (Rodriguez et al., 2020). This beneficial effect of intercropping was also reflected in enhanced nutrient uptake in both species. Under mixed cropping, leaves and roots accumulated higher concentrations of NO₃⁻, SO₄²⁻, and Cl⁻, indicating improved nutrient acquisition process and root-soil interactions. For example, H. marinum roots under low-nutrient intercropping showed notably greater SO₄²⁻ and Cl⁻ levels, suggesting that its ion absorption mechanisms are stimulated by proximity to M. sativa (Ma et al., 2017; Zhao et al., 2022). Its extensive root system and prolonged lifecycle likely amplify these effects by increasing root contact with nutrient zones (Wang et al., 2020). Similarly, M. sativa in intercrops took up more NO₃⁻, NH₄⁺, and Mg²⁺ which are critical elements for symbiotic nitrogen fixation and photosynthesis (Zhu et al., 2020; Ahmed et al., 2023). Intercropping also alters the rhizosphere: root-released enzymes (e.g., acid phosphatase, phytase) and organic acids mobilize phosphorus and other minerals, boosting soil and plant nutrient levels (Wang et al., 2014; Tang et al., 2022). Under fertilization, sea barley in the intercrop accumulated more NH₄⁺ and K⁺, reflecting effective nitrogen transfer from alfalfa and enhanced uptake by the cereal partner (Latati et al., 2014). In line with these results, wheat–soybean and alfalfa–maize intercropping studies report higher N content and uptake in both crops when adequately fertilized (Betencourt et al., 2012; Nasar et al., 2020).

These nutritional improvements were accompanied by significant shifts in carbohydrate metabolism, highlighting enhanced osmoprotection and stress buffering (Abid et al., 2016). Under low-N monoculture, M. sativa leaves accumulated large starch reserves (up to ~235 mg g⁻¹ DW), indicating a carbon sink when growth is limited. In contrast, intercropped alfalfa stored less starch and maintained higher levels of soluble sugars (glucose, fructose, sucrose) and sugar alcohols (myo-inositol, pinitol). This shift suggests that intercropped alfalfa sustained active metabolism using sugars to fuel growth and N assimilation instead of sequestering carbon as starch (Sun et al., 2002). H. marinum showed the opposite pattern: under low N, intercropping increased leaf starch (from ~22 to 41 mg g⁻¹ DW) and hexose pools, reflecting enhanced carbon fixation when supplied with legume-derived N (Chapagain and Riseman, 2014). Carbohydrate profiling further underscored this metabolic synergy. Under nutrient limitation, alfalfa intercrops exhibited 6–100% increases in soluble sugars and sugar alcohols, enhancing osmoprotection and redox balance (Afzal et al., 2021). In contrast, sea barley accumulated additional starch and hexoses, an energy reserve enabled by the extra N provided by the legume (Amoah et al., 2012). Comparable responses have been documented in other intercropping systems: peanut roots exude more soluble sugars under sugarcane/peanut intercropping (Tang et al., 2024), and same for Bletilla striata when plant alongside companion plants, it build up extra sugars and amino acids that fuel growth and support vital metabolic processes, which not only boost their growth but also make them tougher and more productive (Deng et al., 2023). Same for intercropping Urochloa spp. and Megathyrsus maximus with soybean it positively influences the nutritional profile by adjusting the fractions of proteins and carbohydrates, thereby potentially enhancing their digestibility and value as forage for animals (Dias et al., 2025).

The shifts in carbohydrate metabolism were matched by parallel changes in organic acid accumulation. Under intercropping, both species accumulated higher levels of citrate and isocitrate in leaves compared to monocultures, based on a study by Finkemeier et al. (2013) when a plant is under stress, citrate, an intermediate of the tricarboxylic acid (TCA) cycle, starts to build up because parts of this metabolic pathway are slowing down. That excess of citrate then acts as a specific signal that triggers changes in gene expression related to photosynthesis, defense, and nutrient management, helping the plant adapt to stress conditions. Moreover, root profiles showed that intercropped plants, particularly H. marinum, built up malate, succinate, malonate, and citrate under N limitation. For example, the rise in succinic acid levels in Hordeum marinum leaves and roots suggests that the plant may be converting GABA into succinic acid to fuel its metabolism via TCA cycle. This process likely helps the plant stay active and resilient when facing stresses like cold, heat, drought, or salinity (Hijaz and Killiny, 2019; Wu et al., 2025).

Free amino acid profiling completed the picture of stress adaptation by revealing distinct strategies. Under nutrient limitation, H. marinum leaves accumulated very high levels of γ-aminobutyric acid (GABA) and asparagine (Asn), despite low proline, indicating that the halophyte channels nitrogen into these pools rather than into proline under N stress. Barakat (2010) demonstrated that under stress conditions, plants quickly accumulate GABA, which plays a protective role by limiting oxidative damage, modulating hormonal signaling, and maintaining energy production via the GABA shunt, particularly when TCA cycle activity is compromised. Stress-induced accumulation of asparagine may result from underlying mineral deficiencies (Lea et al., 2007), Asparagine accumulation under stress seems to be an adaptive strategy for plants, especially during times of disease or nutrient deficiencies. It likely helps store and transport nitrogen and also aids in detoxifying excess ammonia. When stress causes an increase in proteolysis and ammonia levels, plants convert this surplus ammonia into asparagine using the GS/GOGAT pathway and asparagine synthetase (Oddy et al., 2020).

In contrast, M. sativa leaves under low nutrients showed elevated proline. Proline plays a vital role in helping plants cope with stress by acting as an osmoprotectant. It helps stabilize cell structures and enzymes, neutralize harmful reactive oxygen species (ROS), and maintain the redox balance. Its accumulation is often associated with improved levels of certain mineral nutrients (Cacefo et al., 2021). The decline in glutamate may reflect its use as a precursor for other protective compounds, highlighting a reorientation of nitrogen metabolism (Forde and Lea, 2007). Overall, intercrops displayed improved N status: the H. marinum partner’s surge in GABA and Asn indicates enhanced N assimilation, while alfalfa’s maintenance of glutamate and other amino acids supports sustained symbiotic N fixation. These complementary metabolic shifts mirror observations in other legume–cereal intercrops, where legumes upregulate amino-acid metabolism when grown in mixtures. Studies by Gu et al. (2025) and Liu et al. (2024) show that intercropping, whether Camellia oleifera with peanuts or tobacco with maize or soybean, stimulates amino acid metabolism and transport in plant roots. This includes an increase in key compounds like L-tryptophan and proline, which play vital roles in nutrient assimilation, stress tolerance, and plant growth. Additionally, the modulation of root exudates helps attract beneficial soil microbes, further supporting plant health and resulting in improved photosynthesis and higher yields.

Figure 6 presents the distribution of carbon metabolites. Metabolite profiling revealed distinct species- and tissue-dependent responses to intercropping, as indicated by the log₂(intercrop/monocrop) ratios. In Medicago sativa, root concentrations of proline, threonine, and aspartate markedly increased under the 0% nutrient level, whereas in Hordeum marinum, leaf aspartate significantly declined under the 50% treatment. Organic acids, particularly succinate, consistently decreased in both organs of M. sativa, with a similar but less pronounced trend in H. marinum. Conversely, glucose and fructose accumulated in the roots of H. marinum under control conditions, reflecting an adjustment in carbon metabolism.

Taken together, these results illustrate how intercropping enhances metabolic flexibility and nutrient economy, enabling both species to better withstand nutrient limitations. The complementary resource use, combined with improved biochemical resilience, underlines the ecological advantage of intercropping systems in low-input conditions. Such findings support the strategic use of intercropping for sustainable forage production in semi-arid environments, offering a pathway to mitigate the effects of nutrient stress while maintaining crop productivity and quality.