Effect of the Inoculation Method on the Potential Plant Growth-Promoting Activity of a Microbial Synthetic Consortium

Abstract

Plant microbiomes represent a reservoir of adaptive traits that can enhance plant resilience and productivity. In this study, we investigated the potential of a synthetic microbial consortium (SynCom), composed of five nitrogen-fixing Bacillaceae derived from the phyllosphere of Pistacia lentiscus L., to promote the growth of Lactuca sativa L. under nutrient-limiting availability. The short-term effects of the SynCom were evaluated in a hydroponic system, and four inoculation treatments were compared. The root and leaf inoculation (RL) significantly increased total plant biomass, driven by higher above-ground (+33%) and below-ground (+31%) biomass. The number of leaves per plant and leaf mass per area were also significantly enhanced in RL (+12% and +34%, respectively). While root-only inoculation (R) did not alter plant biomass, structure, or leaf traits, foliar spraying (L) significantly decreased the total leaf area and increased root biomass and the root-to-shoot ratio compared to non-inoculated plants, suggesting a direct influence of microbial metabolites on root growth and nutrient uptake. Compared to the individual R and L treatments, the synergistic effect observed in RL highlights the complex interplay between plant–microbe interactions in the rhizosphere and phyllosphere and the importance of ensuring adequate nutrient availability to nitrogen-fixing bacteria to achieve their growth-promoting potential fully. These findings suggest the potential for utilizing SynComs as bioinoculants to promote plant growth, emphasizing the need to optimize application strategies, considering both the SynCom composition and the host plant’s nutritional status to ensure efficacy.

1. Introduction

Agricultural productivity enhancement through high chemical inputs (particularly nitrogen [N] and phosphorous [P]) represents an economically and environmentally unsustainable approach to reaching future food demand goals [1]. Implementing advanced technological solutions can mitigate fertilization-induced environmental impacts and economic burdens in agricultural production. These solutions focus on optimizing nutrient delivery, irrigation regimes, and soil cultivation techniques and utilizing novel crop varieties with enhanced nutrient use efficiency [2]. Furthermore, integrating microbial biotechnology, specifically harnessing beneficial plant–microbe and microbiome interactions, can significantly improve the efficacy of these on-farm mitigation strategies [3]. Thus, the development of effective microbial biofertilizers, i.e., products containing beneficial microorganisms (mainly fungi and bacteria) that can be applied to the plant [4], represents a promising and sustainable approach for enhancing soil fertility and promoting plant growth, contributing to minimize excessive synthetic fertilizer application, thus mitigating greenhouse gas emissions, water pollution, and soil degradation [5].

Microbial biofertilizers are primarily derived from the plant microbiome, a dynamic ecological community whose composition and function are modulated by a complex interplay of factors, including host plant genotype, edaphic properties, plant developmental stage, inter-microbial resource competition, and responses to abiotic stress [4,5,6,7,8]. Within the diverse plant microbiome, a select subset of microorganisms, known as plant growth-promoting rhizobacteria (PGPR), exhibit beneficial interactions with their host. PGPR has been extensively documented to improve plant nutritional status through various mechanisms, including direct nutrient provision, for example, through N fixation [9,10,11], the transformation of nutrients into more readily assimilable forms, like solubilizing inorganic soil phosphates [12] or iron (Fe) through siderophores [13,14], and modulation of root architecture to enhance nutrient uptake efficiency [15,16,17].

Synthetic Microbial Communities (SynComs) are rationally designed microbial consortia engineered to replicate key interactions and functionalities observed in natural microbiomes, offering a simplified model system with reduced complexity [18]. Low-complexity synthetic microbial consortia, composed of a limited number (1–10) of microbial strains with documented plant growth-promoting activities, represent a promising approach for enhancing plant growth [19,20,21]. Compared to a single strain’s application, the employment of SynCom aims at (1) increasing the abundance of functionally similar microbial populations: by introducing multiple strains with similar metabolic capabilities (e.g., N fixation, phosphate solubilization), the overall efficiency of these PGP traits can be amplified, potentially leading to increased nutrient availability for the plant; (2) providing the plant with a diverse range of PGP functionalities: combining strains with different PGP mechanisms (e.g., one strain for auxin production, another for siderophore production, a third for biocontrol, etc.) offers a more holistic approach to plant growth promotion, addressing multiple needs simultaneously [5,22]. Harnessing this potential, SynCom, composed of strains selected for specific plant growth-promoting traits and stress tolerance, can be developed as an effective inoculant to facilitate plant growth even in adverse environments.

Isolation of new strains to be harnessed as PGPR is essential to enhance the diversity of available strains for various environmental conditions and plant species and increase the understanding of plant–microbe interactions. Plants grown in marginal areas, where climate change increasingly challenges plant survival, may harbor unique microbial communities that can serve as reservoirs of microorganisms adapted to drought, high salinity, and nutrient deficiency [23]. These members of the plant microbiome might strengthen the host’s ability to withstand extreme conditions, such as those found in the Mediterranean Basin [24]. The different plant organs offer many habitats varying in nutrient availability and surface area to adhere and grow, which host microbial communities of variable composition and density [5]. The phyllosphere, comprising the aerial surfaces of plants, constitutes Earth’s most extensive biological interface [25]. This ecological niche is a less nutrient and water-rich plant habitat for microorganisms than the rhizosphere [26]. Phyllospheric microorganisms must face ultraviolet radiation and climatic fluctuations [27]. Despite the challenging conditions for epiphytic bacterial survival, this habitat supports a remarkable diversity of microorganisms [28]. Nevertheless, fewer studies have analyzed plant phyllosphere as a source of novel microorganisms than the rhizosphere [29]. Among PGPR, evidence suggests that the leaf surface provides a crucial habitat for establishing free-living diazotrophic bacteria capable of N fixation [30]. Plant species-specific variations in free-living N fixation rates are often correlated to variations in P concentration, indicating that this element and Fe are important in diazotrophic activity [27,30,31].

This study examined the influence of inoculation methods (leaf, root, or combined inoculation) on the PGP potential of a SynCom. The SynCom was constructed using bacterial isolates obtained from the phyllosphere of Pistacia lentiscus L. growing in marginal areas of Sardinia. P. lentiscus, a widespread evergreen sclerophyllous shrub characteristic of the Mediterranean maquis, exhibits a high degree of drought tolerance and is predicted to increase in abundance under future aridity scenarios [32]. This species demonstrates a rich phyllosphere microbiome, supporting a greater abundance of bacterial cells than co-occurring species [33,34]. Furthermore, P. lentiscus harbors a diverse and abundant community of free-living N-fixing bacteria on its leaf surface [35], representing a promising reservoir of microorganisms with potential plant growth-promoting (PGP) traits. Leaf samples for this study were collected from diverse vegetation types in Northern Sardinia during late summer following a prolonged period of drought. The collection of samples within the Mediterranean macchia during the peak of summer likely enriched the phyllosphere with microbial communities possessing adaptive traits to high temperatures, water stress, and elevated UV radiation. Such traits may confer enhanced survival for PGPR following foliar or root application under challenging environmental conditions. Bacterial isolates were screened for key PGP traits, including N fixation, phosphate solubilization, and salt tolerance. Recognizing the critical role of nitrogen in plant growth and development and the environmental effects associated with intensive use of synthetic N-fertilizers [36], the SynCom was constructed using free-living diazotrophs with varying salt tolerance and phosphate solubilizing capacities, which can represent a key natural source of N in natural and agricultural ecosystems [37]. The SynCom was applied to Lactuca sativa L. (lettuce) grown in a hydroponic system under nutrient-limiting conditions. This experimental setup allowed for the precise evaluation of SynCom’s efficacy in promoting plant growth under controlled conditions. The impact of SynCom on lettuce growth was assessed by analyzing plant biomass, structure, and leaf traits.

2. Materials and Methods

2.1. Isolation of Culturable Aerobic Bacteria from the Phyllosphere of Mediterranean Plants

Pistacia lentiscus branchlets were collected in August 2023 from plants grown in four marginal areas of Sardinia (Italy): (1) a Mediterranean garrigue (G), characterized by a multi-stratified vegetation composed of Pinus pinea L., Juniperus phoenicia L., Rosmarinus officinalis L., P. lentiscus L. and Cistus monspeliensis L.; (2) an experimental pure P. lentiscus plantation (PP); (3) a mixed plantation (MP) of J. phoenicia, R. officinalis, and P. lentiscus; and (4) a heap of slag and discarded working material remaining from a former mining activity (M) characterized by a low maquis principally composed of P. lentiscus and C. monspeliensis. Three branchlets per plant and three plants per site were considered. The branchlets were inserted in a tube with water (1:10 w/v) under sterile conditions and shaken at room temperature for two hours at 150 rpm. The extracted microbiomes were then serially diluted in peptone water (1 g L⁻¹ peptone, 0.5 g L⁻¹ NaCl). A total of 100 µL of the dilution 10⁻³ was spread plated on Luria Bertani Agar (LBA: tryptone 10 g L⁻¹, yeast extract 5 g L⁻¹, NaCl 5 g L⁻¹, agar 15 g L⁻¹) Petri dishes and incubated at 30 °C for 48 h. Spatially separated and morphologically different colonies of predominant culturable aerobic bacteria were selected and streaked on fresh LBA plates. The plates were incubated at 30 °C for 24 h.

2.2. Phosphate Solubilization, N Fixation Capacity, and Salt Tolerance Determination

Twenty-eight morphotypes were obtained and tested for in vitro PGP (phosphate-solubilizing and N-fixing) activities and salt tolerance. Each colony was sampled with a sterile pin and respectively deposited on Pikovskaya (yeast extract 0.5 g L⁻¹, glucose 10 g L⁻¹, (NH₄)₂SO₄ 0.5 g L⁻¹, MgSO₄·7H₂O 0.1 g L⁻¹, Ca₃(PO₄)2 5 g L⁻¹, KCl 0.2 g L⁻¹, MnSO₄·2H₂O 0.002 g L⁻¹, FeS-O₄·7H₂O 0.02 g L⁻¹, agar 15 g L⁻¹), Ashby (mannitol 20 g L⁻¹, K₂HPO₄ 0.2 g L⁻¹, NaCl 0.2 g L⁻¹, MgSO₄·7H₂O 0.2 g L⁻¹, K₂SO₄ 0.1 g L⁻¹, CaCO₃ 5 g L⁻¹, bromophenol blue 0.4% (w/v) stock solution in ethanol 10 mL L⁻¹, agar 15 g L⁻¹), LBA, and LBA plates amended with 5%, 7.5%, and 10% NaCl (Figure 1). Pikovskaya’s medium is a selective culture medium used to isolate and enumerate soil microorganisms capable of solubilizing inorganic phosphate [38]. The medium contains tricalcium phosphate (Ca₃(PO₄)₂) as the insoluble phosphate source. Microbial phosphate solubilization is indicated by forming a clear halo zone around the colonies, signifying the breakdown of tricalcium phosphate. Ashby’s Agar is a specialized growth medium devoid of nitrogen, facilitating the isolation and cultivation of nitrogen-fixing bacteria capable of utilizing atmospheric nitrogen as a source of this macro-element [39]. Luria Bertani (LB) agar is a nutrient-rich growth medium for cultivating bacteria [40]. When supplemented with NaCl, it creates an environment that favors the growth of salt-tolerant bacteria while inhibiting the growth of bacteria sensitive to higher salt concentrations.

The plates were incubated at 30 °C for 48 h. The colony diameter of the morphotypes was measured on each growth medium. The reported colony diameter represents the average of three independent replicates. N-fixing capacity was evaluated by detecting the formation of colonies on Ashby plates, while colony and halo formation were considered distinct traits of phosphate-solubilizing bacteria. The solubilization index (SI) was calculated as SI = (colony diameter + halo zone diameter)/colony diameter [41].

Five N-fixing bacteria were selected, genotypically characterized, and inoculated in lettuce plants.

2.3. Molecular Identification of Culturable Bacteria

The five N-fixing isolates were inoculated in LB and grown overnight at 30 °C for the molecular analysis. The cells were pelleted by centrifugation, and DNA was extracted using the DNeasy PowerSoil Pro kit (Quiagen, Hilden, Germany), based on manufacturer protocol. DNA purity and quantity were examined using Qubit fluorometers (Thermo Fischer, Waltham, MA, USA). The DNA was used as a template for amplification of the 16 S rRNA gene. Universal primer 8F (5′-AGAGTTTGATCCTGGCTCAG-3′) and 1389R (5′-ACGGGCGGTGTGTACAAG-3′) were used to generate amplicons (about 1400 bp) that were cloned into the pGEM-Teasy vector (Promega, Madison, WI, USA) and sequenced using BioFab commercial service (Rome, Italy). All 16S rRNA sequences from isolates with the same morphotype were identical, and only sequences that could be shown to be derived from independent templates were analyzed. The 16S rRNA sequences were compared with those of all known bacterial species available in the EzBioCloud database [42] to identify potential phylogenetic relationships. All sequences were aligned using Clustal Omega [43], and the phylogenetic tree was constructed using Geneious Prime software (ver. 2025.0.1; Biomatters Ltd., Auckland, New Zealand). The confidence values of the branches were determined by performing a bootstrap analysis based on 100 replicates.

2.4. Nucleotide Sequence Accession Numbers

Sequences from independent templates/clones and amplicon libraries were deposited in GenBank under accession numbers PQ473532, PQ475936, PQ475940, PQ476008, and PQ483140.

2.5. Experimental Set-Up

Lettuce (Lactuca sativa L.) plantlets grown under nonsterile conditions for one week were divided into four blocks (n = 5). The initial plant fresh weight (TFW_i) was determined to achieve an equal plant size distribution among the treatments. The plants were then transplanted in four hydroponic systems filled with two liters of sterilized and deionized water, added with 0.1% (v/v) 0.2 µm filtered nutrient solution Hydro A (B’cuzz, Atami B.V., Rosmalen, The Netherlands) containing K₂O 4.7%, CaO 3.8%, MgO 1.3%, SO₃ 0.11%, Fe 0.04%, and N 4.9% (calcium and ammonium nitrate salts); 0.1% (v/v) filtered nutrient solution Hydro B (B’cuzz, Atami B.V., Rosmalen, The Netherlands) was also added, containing P₂O₅ 4.1%, K₂O 5.71%, B 0.01%, Mn 0.03%, Mo 0.001%, Zn 0.039%. Only one nutrient application was performed during the whole experiment. Electrical conductivity (EC), a proxy for nutrient concentration, was determined using an EC probe (EC400, Apera Instruments, Columbus, OH, USA). The EC was 1.2 ± 0.2 dS m⁻¹ during the growth period, which is, on average, 20% less than the minimal optimum value (1.5–2.5 dS m⁻¹ [44]) and about 25–30% less compared to the recommended nutrient supply (1.6–1.7 dS m⁻¹). The pH of the growth solution was measured daily with a portable pH meter (PH400, Apera Instruments, OH, USA). A buffering agent (pH- Terra Aquatica, Fleurance, France) was applied to adjust the pH value and keep it in the range of 5–6.5, optimal for plant and microbial growth.

The hydroponics were placed in a grow tent (Mars Hydro, Europe, Ginsheim-Gustavsburg, Germany) equipped with the lamp Mars Hydro Smart FC 3000 Samsung LED Grow Light powered by a Samsung LM301B LED. The PPFD was set at 400 µmol photon m⁻² s⁻¹ (14/10 h, light/dark). The temperature was set to 26 °C, and a ventilation system (DF150A, Inline Duct Fan, Mars Hydro, Europe) guaranteed an air exchange in the tent.

2.6. PGPR Treatment

The five selected N-fixing strains were grown overnight in LB medium at 30 °C, under agitation (150 rpm). To prepare the high-density SynCom stock culture, appropriate aliquots of each culture were mixed to achieve a final density of 4 × 10⁸ cells mL⁻¹. Plants were divided into four treatment groups: Control (C), not inoculated plants; Treatment R, root-inoculated plants; Treatment L, leaf-inoculated plants; Treatment RL, root- and leaf-inoculated plants (Figure 1). In both R and RL, 10 mL of the high-density SynCom stock culture (4 × 10⁸ cells mL⁻¹) were added to the hydroponic tank; thus, the final SynCom concentration in the hydroponic tanks was equal to 2 × 10⁶ cells ml⁻¹. This SynCom concentration was 10-fold higher than the total aerobic bacterial population in the C tank (2 × 10⁵ cells mL⁻¹). For foliar application in L and RL, the SynCom high-density stock culture was diluted to achieve a final concentration of 2 × 10⁶ cells mL⁻¹, and each plant was sprayed with 1 mL of this suspension.

2.7. Plant Biomass Production, Allocation, and Plant Structure

On day 9 after inoculation (DAT9), all plants were removed from the hydroponic systems. Subsequently, the plants were separated into leaves and roots. The fresh weight of each portion was recorded to obtain root fresh weight (RFW_f), leaf fresh weight (LFW_f), and total final fresh weight (TFW_f) for every plant.

Leaves from all plants were carefully arranged on a flatbed scanner and digitized. The resulting images were analyzed using ImageJ software version 1.53t, Wayne Rasband and contributors, National Institutes of Health, Bethesda, MD, USA) to determine the total leaf area (TLA) and the number of leaves (NL) per plant. The average leaf area (ALA = TLA/NL) was calculated (Figure 1).

Root and leaf subsamples were oven-dried, using a Sartorius MA 100 moisture analyzer (Göttingen, Germany) to determine root dry weight (RDW) and leaf dry weight (LDW). Significant correlations were observed between RFW and RDW in C (RDW_C = 0.023 × RFW_C + 0.0117; R² = 0.4017), L (RDW_L = 0.053 × RFW_L + 0.0023; R² = 0.7662), RL (RDW_RL = 0.0354 × RFW_RL − 0.0037; R² = 0.9595), and between LFW and LDW in C (LDW_C = 0.0632 × LFW_C + 0.0141; R² = 0.565), L (LDW_L = 0.0783 × LFW_L + 0.0289; R² = 0.8232), RL (LDW_RL = 0.0878 × LFW_RL + 0.0343; R² = 0.7335). No significant correlations were observed between RFW and RDW, and between LFW and LDW in R. Consequently, RDW and LDW in this treatment were estimated from the general root and leaf regressions, considering all the dataset (RDW = 0.0438 × RFW-0.0025; R² = 0.4705; LDW = 0.0697 × LFW + 0.0178; R² = 0.6768). The obtained equations were subsequently employed to estimate RDW_f, LDW_f, and total final dry weight (TDW_f = RDW_f + LDW_f) for plants subjected to each treatment. Each plant’s root-to-shoot ratio (R/S) was calculated to assess biomass allocation as the estimated RDW_f to LDW_f ratio. The average leaf biomass (ALW = LDW/NL) and leaf mass per area (LMA; the ratio of LDW to TLA) were then determined (Figure 1).

2.8. Statistical Analysis

To test the effect of the PGPR inoculation method on structural and biomass parameters, the Kruskal–Wallis test was performed on the website www.socscistatistics.com. Pairwise comparisons between groups were conducted using Dunn’s test, with statistical significance determined at p ≤ 0.05.

3. Results

3.1. Phenotypic Characterization of Culturable Bacteria Isolates

According to the colony morphology, twenty-eight morphotypes were identified in the epiphytic population of P. lentiscus plants collected in the four different sites: eight isolates were collected in the mixed plantation (MP), five in the pure plantation (PP), eight in the garrigue (G) and seven in the old mine (M) (

Table 1. Phenotypic characteristics of the strains isolated from the phyllosphere of P. lentiscus plants grown in a pure (PP) and a mixed plantation (MP), in garrigue (G) and an old mine (M), located in North Sardinia. The salt tolerance was tested on LBA amended with 0, 5, 7.5, and 10% NaCl, measuring the colony diameter. The phosphate-solubilizing activity (PS) was determined by calculating the phosphate solubilizing index (PI). The N-fixing (NF) ability was marked with a + symbol. The minus symbol (−) refers to the absence of growth (no colony formation).

Site	Strain	Colony Diameter (cm)				PS *	NF
		NaCl (%)				PI	+/−
		0	5	7.5	10
MP
	1	0.85	0.96	−	−	1.87	−
	2	0.91	0.87	0.74	0.74	1.50	−
	3	1.61	0.94	−	−	−	+
	4	1.11	0.92	−	−	1.60	+
	21	0.78	0.64	0.58	0.55	1.40	−
	22	0.77	0.77	0.63	0.51	1.88	−
	23	1.23	0.64	−	−	−	−
	24	0.74	0.57	0.66	0.51	1.48	+
PP
	5	0.85	1.58	0.93	0.73	1.88	−
	6	1.18	1.24	1.01	0.68	−	−
	7	1.18	0.92	0.86	0.72	−	−
	8	1.11	1.07	0.90	0.64	−	−
	16	1.03	3.21	0.69	0.58	−	−
G
	17	0.95	0.67	0.58	0.47	1.54	−
	18	0.60	0.65	0.55	0.58	−	−
	19	0.73	0.74	0.74	0.52	2.00	−
	20	1.09	3.22	0.71	0.54	−	−
	29	1.14	0.75	0.60	−	−	−
	30	1.27	2.12	1.88	−	−	+
	31	0.60	0.81	0.71	0.62	−	−
	32	0.94	0.69	0.70	0.54	1.37	−
M
	33	0.76	0.65	0.62	0.68	−	−
	34	0.99	0.85	0.65	0.57	1.21	+
	35	0.76	0.75	0.60	0.59	−	−
	57	0.62	0.67	0.56	0.49	−	−
	58	0.71	0.92	0.61	0.61	−	−
	59	1.08	0.96	0.84	0.65	1.17	−
	60	0.86	0.70	0.64	0.61	−	−

Strains MP3, MP4, MP24, G30, and M34 (in bold) have been selected for lettuce inoculation and identified. * Phosphate solubilization index.

).

Phenotypic evaluation of the 28 epiphytic isolates revealed a high salt tolerance among the strains. All bacteria produced visible colonies on LBA plates amended with 5% NaCl (Table 1). Among these strains, 11 produced colonies with a larger mean colony diameter than LBA (Table 1). Approximately 85% of the isolates (24 out of 28) exhibited growth on LBA plates amended with 7.5% NaCl, three of which (PP5, G30, and G31) formed colonies with larger mean diameters compared to LBA (Table 1). The salt-tolerant strains were 22, with strain G31 forming larger colonies on LBA plates amended with 10% NaCl than on LBA (Table 1). All isolates from pure P. lentiscus plantation and from the heavy-metal contaminated site formed colonies on the most saline plates, while only 4 out of 8 mixed plantation strains were tolerant to such a high salt concentration (Table 1).

Phosphate solubilization activity was observed in 12 out of 28 strains. Solubilization indices (SI) ranged from 1.17 to 2.0 (Table 1). Most phosphate-solubilizing bacteria were found in the mixed plantation site (6 out of 8), while only 1 out of 5, 3 out of 8, and 2 out of 7 were respectively found in the pure plantation, in the garrigue and the mine sites (Table 1).

Only 5 strains out of 28 could fix N and thus formed colonies on Ashby plates (Table 1). Most N-fixing bacteria (3 out of 5) were isolated from the mixed plantation. None of the diazotrophs were found in the pure plantation, while one strain per site was found in the garrigue and the mine (Table 1).

All N-fixing strains (MP3, MP4, MP24, G30, and M34) were used to inoculate lettuce plants; among these, strains MP3 and MP4 only formed colonies on LBA amended with 5% NaCl (Table 1). Strain G30 formed colonies of higher dimension on LBA with 5 and 7.5% NaCl than on LBA plates (diameter ratios equal to 1.67 and 1.48, respectively) but did not grow on LBA amended with 10% NaCl, while strains MP24 and M34 formed colonies in all salt-amended plates (Table 1). The highest phosphate-solubilizing activity was observed in strain MP4, followed by strain MP24 and strain M34 (SI indexes equal to 1.60, 1.48, and 1.21, respectively), while strains MP3 and G30 did not show phosphate solubilizing activity (Table 1).

3.2. Molecular Characterization of Culturable Bacteria

All five isolates were selected for their N-fixing capacity to create the SynCom and test its efficacy in promoting plant growth under nutrient limitation. For each morphotype, at least two independent colonies were characterized to the genus/species level, using the 16S rRNA gene as a DNA barcode (see Section 2 “Materials and Methods”). All strains exhibiting the same morphotype had identical 16S rRNA gene sequences. Sequence data of 16S fragments were used to generate a phylogenetic tree to evaluate the genetic relatedness among these bacteria and known species. All five morphotypes belonged to Bacillaceae and were affiliated with taxa belonging to three representative clades of this family: Subtilis Clade (strains G30 and MP24), Cereus Clade (strain MP3), Megaterium Clade (strains MP4 and M34). The phylogenetic relationship between the isolates and the closely related taxa is shown in Figure 2.

3.3. Plant Biomass

The SynCom significantly increased lettuce biomass when inoculated in roots and leaves (p < 0.05; Figure 3). The total plant biomass in RL was 31.1% (0.62 ± 0.03 g) higher (p < 0.05) compared to non-inoculated plants (0.47 ± 0.04 g; Figure 3a). On the other hand, the application of the PGPR synthetic consortium to roots or leaves did not affect biomass production. The total dry weight in R (0.47 ± 0.03 g) and L (0.45 ± 0.05 g) was similar to non-inoculated plants and significantly lower (p < 0.05 in both treatments) compared to RL (Figure 3a). Although inoculation with the PGPR synthetic consortium did not affect total dry weight in R and L, significant changes in biomass partitioning between leaves and roots occurred. While leaf biomass in R (0.37 ± 0.02 g), L (0.32 ± 0.04 g), and RL (0.51 ± 0.03 g) was similar to non-inoculated plants (0.39 ± 0.03 g), inoculation methods differently affected leaf biomass (Figure 3b): in R and L, leaf dry weight was significantly lower (p < 0.05 and p < 0.01 respectively) compared to RL. Root biomass significantly increased (p < 0.001) in L (0.127 ± 0.01 g) compared to non-inoculated plants (0.078 ± 0.01 g; Figure 3c). The root biomass observed in R and RL inoculated plants was respectively equal to 0.102 ± 0.01 g and 0.097 ± 0.01 g, not dissimilar from C and L (Figure 3c). Inoculation with the SynCom significantly affected biomass allocation (p < 0.01; Figure 3d). The root-to-shoot ratio was significantly (p < 0.01) higher in L (0.40 ± 0.02 g g⁻¹) compared to non-inoculated plants (0.20 ± 0.01 g g⁻¹). Moreover, the ratio was significantly lower (and similar to C) in RL (0.19 ± 0.01 g g⁻¹) compared to L (p < 0.001) and R (0.28 ± 0.02 g g⁻¹; p < 0.01) (Figure 3d).

3.4. Plant Above-Ground Structure and Leaf Traits

The number of leaves was significantly higher in RL (10.8 ± 0.37 leaves per plant) compared to non-inoculated lettuces (9.6 ± 0.24 leaves per plant; p < 0.05) and L (9.6 ± 0.24 leaves per plant; p < 0.001), while the number of leaves in R (10 ± 0.32 leaves per plant) was similar to C, L, and RL (

Table 2. The number of leaves per plant (NL), average leaf area (ALA), total leaf area (TLA), and leaf mass per area (LMA) were measured or calculated in non-inoculated lettuces (C) and in lettuces where the SynCom was applied to the roots (R), to the leaves (L), or both roots and leaves (RL). Mean values and standard errors (s.e.) are reported. Distinct uppercase letters denote statistically significant variations (p < 0.05) among treatments for individual parameters, assessed independently for each parameter (n = 5).

	NL		ALA		TLA		LMA
	No. Leaves Per Plant		cm2		cm2		g m−2
C	9.60	B	18.43	A	177.38	A	21.98	B
R	10.00	AB	15.79	AB	157.36	AB	23.08	B
L	9.60	B	13.38	B	128.79	B	24.95	B
RL	10.80	A	16.36	AB	176.20	A	29.46	A
s.e.
C	0.24		0.84		10.93		0.54
R	0.32		0.74		5.95		0.82
L	0.24		1.10		12.28		1.72
RL	0.37		0.72		7.02		0.90

). The average leaf area was affected by the inoculation procedure only in L-inoculated plants (13.4 ± 1.1 cm²), being on average significantly lower (p < 0.01) compared to non-inoculated plants (18.4 ± 0.8 cm²) (Table 2). In contrast, in R (15.8 ± 0.7 cm²) and RL (16.4 ± 0.7 cm²), intermediate values were observed (Table 2). The total leaf area was significantly lower in L (128.8 ± 12.3 cm²) compared to non-inoculated plants (177.4 ± 11.0 cm²; p < 0.05) and RL (176.2 ± 12.3 cm²; p < 0.05) but was not affected by inoculation in R (157.36 ± 5.95 cm²) (Table 2). The leaf mass per area was significantly higher in RL (29.5 ± 0.9 g m⁻²) compared to non-inoculated plants (22.0 ± 0.5 g m⁻²; p < 0.001), to R (23.1 ± 0.8 g m⁻²; p < 0.05) and L (25.0 ± 1.7 g m⁻²; p < 0.05) (Table 2).

4. Discussion

The implementation of PGPR in plant production systems can be achieved through various methods, including seed coating, root dipping, leaf spraying, and soil drenching; the choice of method depends on the crop species, the specific SynCom composition, and the environmental conditions [45]. The main limiting factors affecting the effectiveness of PGPR inoculants in enhancing plant nutrition or acting as biocontrol agents are related to (1) the competition or predation of the PGPR strains with the already established microbiome [46,47,48], (2) the survival of the PGPR strains used to inoculate plants [49,50] and/or (3) the loss of the efficacy of the promoting traits under stress conditions [51,52]. In this study, we isolated five free-living N-fixing (diazotrophic) Bacillaceae and evaluated their ability to promote the growth of a model plant (L. sativa L.) as SynCom. Although the high selectivity of host plants limits successful colonization of the N-fixing bacteria [30,53], biofertilization through leaf spraying of diazotrophs has been shown to increase crop yield [9,53,54,55,56] and plant resistance to pathogens [57]. Evaluation of the efficacy and cost-effectiveness of diazotrophic biofertilizers requires a comparative analysis of soil and foliar application methods. This approach mirrors the evaluation strategies used for traditional fertilization methods, where optimization of nutrient delivery and economic considerations are paramount. For example, Ferrari et al. [58], who studied the effects of N fertilization methods on wheat, showed that the foliar application of organic (urea) and inorganic (ammonium nitrate) N allowed an increase in grain protein content comparable to that obtained with soil fertilization but with a lower dose of fertilizer. This finding highlighted the potential for foliar application of N fertilizers to maximize nutrient use efficiency and minimize environmental impact. This concept can be applied to the inoculation of PGPR diazotrophs, considering that, among their traits, they can enhance the plant’s access to N.

In the present study, we used a soilless culture system to reduce the potential for competition or predation by indigenous rhizosphere microbial communities and three different inoculation methods to evaluate better the efficacy of the PGPR: root, foliar, or root and foliar application. Microorganisms applied to the leaf surface, similarly to the phyllosphere microbiome, encounter and must survive fluctuating environmental conditions, including variations in temperature, humidity, wind speed, and radiation, which cause significant physiological stress [27,28]. The leaf surface is an oligotrophic environment with limited nutrient availability to support the growth of the epiphytic microbial community [25,26]. The spatial distribution of carbon (C), N, and micronutrients follows the distribution of water availability, as water droplets are effective sinks for the outward diffusion of photosynthates through the cuticle [59]. The leaf cuticle, a hydrophobic layer composed of cutin and waxes overlying the epidermal cells, contributes to the nutrient-limited conditions characteristic of the phyllosphere by restricting the transport of water and polar molecules, thereby limiting nutrient availability for microbial colonization [60]. These issues are particularly pronounced in Mediterranean ecosystems, where factors like drought, high salinity, and nutrient scarcity select for stress-tolerant microbial communities [23].

Moreover, sclerophyllous Mediterranean species are typically characterized by a thick cuticle, which has been proven to be negatively correlated with a bacterial population size of the phyllosphere [60]. We focused on the leaf microbiome of P. lentiscus, a representative Mediterranean evergreen sclerophyll species with an abundant culturable epiphytic bacterial population [34], to evaluate its potential to serve as a reservoir for PGPR strains suitable for foliar application. Leaf samples were collected in different vegetation types of Northern Sardinia in the late summer after a prolonged dry period, which could have potentially enriched the leaf in strains tolerant to high temperature, water stress, and UV radiation. Independently from the vegetation type, all strains isolated from the phyllosphere of P. lentiscus were able to grow in a medium amended with 5% (w/v) NaCl, and most of them tolerated a NaCl concentration of up to 10% (w/v) (Table 1). Two strains from the garrigue and one from the pure plantation developed larger colonies when grown in the 7.5% NaCl (w/v) amended medium, exhibiting a halophilic behavior. The highest salt tolerance was observed among isolates (100%) from the pure plantation (PP) and the heavy metal-contaminated site (M) (Table 1). This could be related to the lower canopy stratification of these sites compared to the garrigue (G) and mixed plantation (MP), where P. pinea L. (G), J. phoenicea L. (G and MP), and R. officinalis (G and MP) were present in the dominant layer. This vertical stratification decreased light penetration and attenuated maritime winds on P. lentiscus in the understory, creating a microclimate with reduced environmental stressors (UV radiation, thermal stress, and salt deposition) compared to the PP and M sites. These conditions may have affected the differential distribution of halotolerant strains in the P. lentiscus L. phyllosphere, as shown in Table 1.

Analysis of the phosphate-solubilizing and N-fixing activity of the different isolates also revealed an effect of the vegetation type on the enrichment of PGPR (Table 1). Interestingly, the highest percentage of strains with PGPR traits was observed in the mixed plantation, a site in which it has been shown that there is a heterogeneous distribution of the soil N contents, with lower values under P. lentiscus canopies compared to the other co-occurrent species (J. phoenicea L. and R. officinalis) [61]. This result agrees with Idbella et al. [35], who studied the effect of encroachment by six Mediterranean plant species, namely P. lentiscus L., J. phoenicea L., Myrtus communis L., R. officinalis, Olea europaea L., and Euphorbia dendroides L., on soil microbial communities. Soils associated with P. lentiscus L. had the lowest N content and highest abundance of free-living N-fixing bacteria.

We isolated N-fixing diazotrophic bacteria from leaves sampled in three collection sites: MP, G, and M. The abundance of N-fixing bacteria varied among sites, with higher MP frequency than G and M. Over half of them (3 out of 5) can solubilize phosphate. At the same time, only 25% of the phosphate solubilizing isolates were N-fixers (Table 1). Molecular analysis indicated that the N-fixing isolates belonged to spore-forming species affiliated with Bacillaceae (Figure 2). Maurice et al. [62] reported that spore-forming Actinobacteria and Firmicutes are common in dryland ecosystems. Similar results were reported by Postiglione et al. [63], who analyzed the phyllosphere microbiome of Quercus ilex L. from Southern Italy.

Several authors reported that the composition of a plant microbiome is profoundly influenced by the host plant [64] and, within the same species, by the environment [63] and the plant–plant interaction. As shown by Wang et al. [65], on peanut and maize, and by Newberger et al. [66], on alfalfa, brassica, and fescue, intercropping of different species affects the plant microbiome composition determining an increase in specific taxa, which are not present in monocultures. In agreement, our results on the phenotypic characterization of the isolates from the different sampling sites indicate that plant–plant and plant–environmental interactions contribute to shaping the P. lentiscus-associated microbiome.

A SynCom comprising the five N-fixing bacteria was used to evaluate their ability to promote plant growth when applied at root (R), leaf (L), or root and leaf (RL) levels. Optimizing inoculation methods is crucial for maximizing the PGPR’s beneficial effects on crops. A comparison of PGPR inoculation methods has already been analyzed in cereals and vegetable crops. In maize, although higher compared to non-inoculated plants, foliar application of Azotobacter chroococcum B002, Bacillus subtilis B004, and Bacillus megaterium, individually or as SynComs reduced leaf biomass and grain yield compared to soil application [67]. A similar study on kale foliar and soil applications of A. chroococcum, B. subtilis, Priestia megaterium (formerly B megaterium), and a mixture of A. chroococcum and B. subtilis demonstrated that the application method did not significantly affect above-ground biomass production, but the specific PGPR strain did [68]. In contrast to Kordatzaki et al. [68], our results highlight the importance of the inoculation method on leaf vegetables. Inoculation with the SynCom differently affected lettuce biomass production depending on the method of application: an increase in total plant biomass compared to non-inoculated plants was only observed when the SynCom was inoculated in both roots and leaves (RL) compared to non-inoculated and to R- and L-inoculated plants (Figure 3a). The higher total plant biomass resulted from a 33% increase in leaf biomass compared to non-inoculated lettuce (Figure 3c). This can be due to the higher leaf mass per area and number of leaves per plant in RL-inoculated lettuce plants compared to non-inoculated controls (Table 2). The heightened LMA indicates increased leaf density, likely caused by the production of cells characterized by thicker cell walls and a higher proportion of structural tissues, conferring enhanced resistance to physical and chemical stressors [69] or a palisade/spongy ratio modification. The leaf investment in structural components is usually associated with reduced nutrient concentrations and lower rates of both photosynthesis and respiration [70]. While the higher leaf dry weight in RL-inoculated plants did not result in a statistically significant increase in above-ground biomass compared to the non-inoculated group (Figure 3b), the combined effect of increased leaf and root biomass (+31%) accounts for the observed statistically significant difference in total plant biomass. Exclusive foliar SynCom application decreased total leaf area and stimulated root biomass accumulation without affecting leaf biomass or leaf mass per area (Figure 3b; Table 2). This plant response increased the root-to-shoot ratio compared to the non-inoculated and RL-inoculated plants (Figure 3d), suggesting that foliar-applied PGPR triggers a signaling cascade that promotes root development. The differential stimulation of leaf and root biomass production was also observed by Fukomi et al. [71] in wheat, where foliar spray inoculation of the diazotroph Azospirillum brasilense resulted in increased root systems and N accumulation in the shoots. The authors attributed these effects to microbially derived hormonal signaling, primarily involving indole acetic acid, jasmonic acid, and salicylic acid. Notably, jasmonic and salicylic acids induced an up-regulation of oxidative stress and plant defense genes in leaves while concurrently down-regulating these genes in roots [53,71]. We can hypothesize that similar mechanisms took place in our experiment, which can be coupled to the low nutrient concentration provided to plants. Together, these results highlight insufficient nutrient availability to support above-ground growth, considering the costs of plant responses to the oxidative stress induced by inoculation. Thus, no increase in leaf biomass was observed despite the increased root biomass conferred by leaf PGPR inoculation. Interestingly, root inoculation significantly alters above- and below-ground biomass only if combined with applying the SynCom at the leaf level, indicating a complex interplay between PGPR inoculation, root development, and leaf growth (Figure 3).

In conclusion, to optimize the benefits of PGPR in sustainable agriculture, application strategies should be tailored to the composition of the microbial consortium and the host plant’s nutritional status. When nutrient availability is high, foliar application of N-fixing diazotrophs that stimulate root growth may be sufficient to determine an increase in aboveground biomass. However, under nutrient-limited conditions, a combined root and leaf application of N-fixing PGPR may be more beneficial. Root application of diazotrophs can also be advantageous when utilizing their additional traits, such as biocontrol activity, independent of their N-fixing capabilities. The choice of an appropriate approach ensures that PGPR applications are targeted and efficient, maximizing their potential to promote plant growth and health while minimizing environmental impact. Further research into the molecular mechanisms underlying these interactions, particularly the role of microbial signaling molecules and their impact on plant hormone regulation and nutrient assimilation, is essential to refine PGPR inoculation practices and realize their full potential in enhancing crop productivity. While this study focused on a hydroponic system to evaluate SynCom’s efficacy under controlled conditions, it is important to investigate its performance in soil-based systems, considering its application in different plant systems. Further research is therefore needed to fully understand the long-term effects of the SynCom on plant health and productivity in soil-based cultivations and soil biota and the implications of inoculation on the overall sustainability of agricultural practices.