Uncovering the endophytic bacterial diversity and crop improvement potential under salinity stress in the halophyte Halocnemum strobilaceum Authors: Elena Romano-Rodríguez, Noris J. Flores-Duarte, Eloísa Pajuelo, Ignacio D. Rodríguez-Llorente, Enrique Mateos-Naranjo, Susana Redondo-Gómez, Salvadora Navarro-Torre Published: 2026 ABSTRACT Highlights•76 endophytic bacteria were isolated from Halocnemum strobilaceum from Murcia (Spain).•Most isolates (96.05%) exhibited plant growth-promoting traits in vitro.•Selected strains in a consortium enhanced Medicago sativa growth under salinity stress.•Four novel species for agricultural biotechnological applications were described. Introduction Salinity is one of the main problems in modern agriculture worldwide because most crops are glycophytes and are therefore salt-sensitive, with limited tolerance to saline conditions (Navarro-Torre et al., 2023a; Zörb et al., 2019). This abiotic stress affects >10% of the world’s land area, resulting in a reduction in yield and significant economic losses for farmers (Mishra et al., 2023). Different sustainable strategies are used to solve this problem, and one of them is inoculation with plant growth-promoting bacteria (PGPB) (Navarro-Torre et al., 2023a). Halophilic or halotolerant PGPB are commonly applied to enhance crop development under saline conditions, resulting in significant improvements in crop performance (Jiang et al., 2025; Navarro-Torre et al., 2023b; Pérez-Inocencio et al., 2023; Teo et al., 2022). These bacteria not only exhibit plant growth-promoting (PGP) traits that enhance nutrient and water uptake and confer protection against salt stress but also grow under saline conditions while maintaining these beneficial properties. Halophytes are an important source of halophilic and halotolerant PGPB (Jiang et al., 2025; Teo et al., 2022) as they grow in saline environments. One of the mechanisms that allows them to grow under high salt concentrations is salt accumulation in their plant tissues (Meng et al., 2018), which selects for salt-tolerant associated microbiota. Halocnemum strobilaceum is a halophilic plant belonging to the family Amaranthaceae that grows in saline soils in the Mediterranean, Central and West Asia, Arabia, Iran, Pakistan, and Mongolia (Nasernakhaei and Zahraei, 2021). The bacterial population of the rhizosphere of H. strobilaceum has been described in several studies (Behairi et al., 2022; Bekkaye et al., 2023; Li et al., 2018; Marasco et al., 2016), but the endospheric bacteria are poorly described (Begali et al., 2021). In both cases, the isolated strains showed PGP traits and promoted the growth of H. strobilaceum and cotton (Begali et al., 2021; Marasco et al., 2016) and durum wheat under saline conditions (Bekkaye et al., 2023). These studies demonstrated that H. strobilaceum is a good PGPB source. In addition, strains isolated by Bekkaye et al. (2023) belonged to the genera Halomonas and Kushneria (family Halomonadaceae), which are the principal genera of halotolerant PGPB that are associated with halophytic plants (Meinzer et al., 2023). Previous studies in other halophytes, such as Arthrocnemum macrostachyum, demonstrated that novel species of both genera can be isolated from them (Navarro-Torre et al., 2020, 2018), increasing the knowledge of the huge diversity in the bacterial microbiota of halophytic plants. This study aimed to explore the endophytic microbiota of H. strobilaceum to identify halotolerant PGPB with the potential to enhance plant development under saline conditions. The following objectives were pursued to achieve this aim: (i) to characterize the diversity of cultivable halophilic and halotolerant endobacteria from different H. strobilaceum populations, (ii) to determine the taxonomic identity and genomic features related to plant growth-promoting (PGP) traits and salinity tolerance of selected isolates, (iii) to evaluate the in vitro PGP traits and enzymatic activities of the isolates and identify compatible strains for consortium formation within the family Halomonadaceae, (iv) to assess whether selected Halomonadaceae isolates represent novel species based on phylogenomic, phenotypic, and biochemical analyses, and (v) to evaluate the effects of selected strains, applied individually and as a consortium, on growth performance and salinity tolerance of Medicago sativa under greenhouse conditions. Materials and methods Study areas and collection of plants Aerial parts of H. strobilaceum were collected during the autumn of 2021 from three different populations in the Mediterranean salt steppes of Murcia, southern Spain (Supplementary Figure 1). The populations were in Calarreona (TE) (37°23′06.5″N, 1°37′23.4″W), which is characterized by low coverage of small individuals; Guadalentín salt marshes (GE) (37°48′47.8″N, 1°25′09.8″W); and Alhama de Murcia (EE) (37°49′35.78″N, 1°23′56.71"W), which are more similar in appearance, with greater coverage and larger individuals. Fresh plant samples were transported in plastic bags under cold conditions and stored at −80 °C until use. In addition, the electrical conductivity (EC), redox potential, and pH of different locations were determined using a Crison-522 conductivity meter and a Crison pH/mVp-506 portable meter (Crison Instrument, S.A., Spain) (Mesa et al., 2015), and the Bouyoucos hydrometer method was used for soil texture (Bouyoucos, 1936). Supplementary Table 1 summarizes the physicochemical properties of the soils of each population. Isolation of culturable halophilic and halotolerant endophytic bacteria Endophytic bacteria were isolated from surface-disinfected aerial parts of H. strobilaceum using 70% ethanol and 3% (v/v) sodium hypochlorite as follows in Navarro-Torre et al. (2016). For endobacteria extraction, six plant fragments (2–3 cm length) obtained from a pool of five different individuals of each population (TE, GE, and EE) were homogenized separately with a sterile mortar and pestle with 2 mL of sterile 0.9% (w/v) NaCl solution, to avoid osmotic stress and guarantee representativity. Then, aliquots (100 μL) of serial dilutions of the extracts (10–1 to 10–4) were plated onto tryptic soy agar (TSA; Intron Biotechnology, Korea) medium plates supplemented with 0.5 M NaCl from SW30 solution stock (Mateos-Naranjo et al., 2020), which was selected as the optimal concentration to ensure the isolation of halophilic and halotolerant bacteria (Ventosa et al., 2021). The plates were incubated at 30 °C for 72 h. Negative controls of air in the sterile chamber and plating of disinfected tissues were also performed to confirm the effective disinfection process. For isolation of bacteria as pure cultures, single colonies were selected according to the different morphological characteristics such as shape, color, and size, followed by Gram staining. A total of 76 pure bacterial cultures were isolated and stored at −80 °C in a sterile 30% (v/v) glycerol solution for long-term preservation. DNA extraction and 16S rRNA gene sequencing for isolates identification The genomic DNAs of the 76 isolated strains were extracted using the G-spin™ Genomic DNA Extraction Kit for Bacteria (Intron Biotechnology, Korea) according to the manufacturer’s protocol. PCR amplifications of the 16S rRNA genes of the 76 strains were performed using the universal primers 16F27 (5′-AGAGTTTGATCMTGGCTCAG-3′) and 16R1488 (5′-CGGTTACCTTGTTAGGACTTCACC-3′) as described in Lane (1991). PCR products were purified with the enzyme ExoSAP-IT (Affymetrix, USA) following the manufacturer’s instructions, then sequenced by the company StabVida (Portugal), and subsequently manually edited and assembled using the Blastn tool from NCBI and BioEdit free software (Hall, 1999). The resulting 16S rRNA gene sequences were pairwise aligned with closely related phylogenetic neighbors using EzBioCloud’s 16S database (Chalita et al., 2024). All 16S rRNA gene sequences obtained in this work were deposited in the National Center for Biotechnology Information (NCBI) GenBank database under the accession numbers listed in Supplementary Table 2. Since isolated bacteria were identified, the proportions of isolates per genus were calculated to characterize the composition of the cultured community, allowing for qualitative comparisons between study areas. Description of the novel bacterial species Whole genome sequencing and analysis of novel strains Several isolates exhibited low 16S rRNA gene sequence identity (<98.7%) compared with the type strains of the most closely related species, suggesting that they may represent potential novel species (Chun et al., 2018). Whole-genome sequencing was performed by MicrobesNG (Birmingham, UK) using a hybrid approach combining short-read Illumina (NovaSeq 6000) and long-read Oxford Nanopore (GridION) technologies to further investigate these strains. Raw reads were adapter-trimmed using Trimmomatic version 0.30 (Bolger, Lohse and Usadel 2014), applying a sliding window with a Q15 quality cutoff. The genome was de novo assembled using SPAdes v3.7 (Bankevich et al., 2012), and hybrid assemblies were generated using Unicycler v0.4.0 (Wick et al., 2017). The whole draft genomes of the seven possible novel species discovered in this study were deposited in GenBank/EMBL/DDBJ under the accession numbers: GCA_965650455 (TE3T), GCA_968048975 (EE22T), GCA_968193635 (EE7T), GCA_968199205 (EE24), GCA_968206495 (EE27), GCA_968212445 (GE22), GCA_970647075 (TE19T), and GCA_977968315 (EE4). Genome annotation was performed using Prokka version 1.11 (Seemann, 2014), and basic statistics about the genomes were extracted using the RAST server (Aziz et al., 2008), QUAST version 5.3.0 (Gurevich et al., 2013), and CheckM v1.2.3 (Parks et al., 2015). Pairwise genome comparisons and digital DNA-DNA hybridization (dDDH) were performed using the Type Strain Genome Server (TYGS) database (Meier-Kolthoff and Göker, 2019), and the ANI was calculated using the JSpeciesWS web server (Richter et al., 2016) for the identification and classification of truly novel strains. Finally, the phylogenetic tree of the potential novel species was inferred using FastME 2.1.6.1 (Lefort et al., 2015) from the GBDP intergenomic distances calculated from the genome sequences. The branch lengths were scaled according to the GBDP distance formula d5 (Meier-Kolthoff et al., 2013) and inferred based on 100 pseudo-bootstrap replicates. The phylogenomic tree was also inferred using FastME 2.1.6.1 from whole-proteome-based GBDP distances. The branch lengths are scaled using the GBDP distance formula d5. Branch values are GBDP pseudo-bootstrap support values > 60% from 100 replications, with an average branch support of 75.6%. The tree was midpoint-rooted (Farris, 1972). Both trees were constructed using the TYGS database. Phenotypic and biochemical characterization We followed the recommended minimal standards for describing new taxa of the family Halomonadaceae to characterize the possible novel strains in this study (Arahal et al., 2007). Salt tolerance of the novel strains was determined in tryptic soy broth medium (TSB; Intron Biotechnology, Korea) with NaCl supplementation ranging from 0 to 4.5 M after 24–48 h of incubation at 30 °C. The following phenotypic and biochemical characterization tests were performed under the established optimal salinity growth condition for the strains (0.5 M NaCl). To study the optimal temperature range, we evaluated bacterial growth in TSA plates at different temperatures (5, 10, 20, 25, 30, 37, 40, 45 °C) after 24–96 h of incubation. Regarding the pH range and the optimum, 100 µL of bacterial culture was suspended in 1 mL TSB with the pH adjusted from 2.0 to 12.0. After 24–96 h of incubation at 30 °C, absorbance at 600 nm was recorded to determine bacterial growth, with three replicates for each condition. The morphological characteristics of the colonies, including size, color, and shape, were evaluated after 48 h of incubation at 30 °C under 0.5 M NaCl supplementation (optimal growth conditions) using a stereoscopy microscope (Olympus SZ61). Cell morphology was analyzed using an optical microscope (Olympus CX41) using a 100 × objective. To study the motility of cells, a drop of liquid culture was observed under optical microscopy using a 40 × objective. Biochemical characterization was performed using the API 20NE gallery (bioMérieux, Italy), following the manufacturer’s instructions. Since the strains did not grow in the standard API 20NE medium after 48 h of incubation, a sterile 0.5 M NaCl solution was added. Bacterial cells were resuspended in this adjusted medium and used for analysis. Oxidase activity was tested by applying a drop of 1% N, N,N′,N′-tetramethyl-p-phenylenediamine solution (Becton, Dickinson and Company) onto bacterial biomass placed on absorbent paper. The development of blue coloration indicated a positive result. Similarly, catalase activity was assessed by adding a few drops of 3% (v/v) H₂O₂ directly onto isolated bacterial colonies, with bubble formation indicating a positive reaction. Plant growth-promoting traits and enzymatic activities In vitro PGP traits and enzymatic activities related to plant development and beneficial effects were evaluated by adjusting the different media to the optimal salt concentration for bacterial growth (0.5 M NaCl). Indole-3-acetic acid (IAA) production was evaluated via a colorimetric method in 96-well microplates (Sarwar and Kremer, 1995). Strains were cultured in TSB supplemented with 0.5 M NaCl for 3 days at 30 °C under continuous shaking. IAA present in 50 μL bacterial culture supernatant reacted with 200 μL of Salkowski reagent in the dark for 20 min. The color intensity was measured at 535 nm on a plate reader (UVM340; Asys Hitech GmbH, Austria). The produced IAA concentration was calculated using a pattern curve with known IAA concentrations. Siderophores production was detected by orange halos formation around the bacterial growth in CAS agar plates (Schwyn and Neilands, 1987). Phosphate and K solubilization were evaluated in NBRIP (Nautiyal, 1999) and Aleksandrov (Setiawati and Mutmainnah, 2016) media. A clear halo formation around the bacterial biomass on agar plates indicates positive results. In these three tests, the halo size was measured after 5 days of incubation at 30 °C. Nitrogen fixation capability was presumptively examined by bacterial growth in nitrogen-free bromothymol blue (NFb) medium (Baldani et al., 2014). 1-aminocyclopropane-1-carboxylate (ACC) deaminase activity was evaluated according to the method of Penrose and Glick (2003) by quantifying the formation of α-ketobutyrate as a product of the enzymatic hydrolysis of ACC, the immediate ethylene precursor in plants. Briefly, α-ketobutyrate was detected after bacterial protein extraction using 2 M NaOH and the absorbance was measured at 540 nm using a spectrophotometer (Lambda25; PerkinElmer, USA). ACC deaminase activity was expressed as μmol of α-ketobutyrate per hour per milligram of bacterial protein. The Bradford protocol was used for protein quantification (Bradford, 1976). Biofilm formation was assessed by incubating bacteria in TSB medium supplemented with 0.5 M NaCl at 30 °C for 4 days. To observe biofilm formation, 0.01% (w/v) crystal violet staining was performed (Mesa-Marín et al., 2019). The hydrolytic activities were evaluated by measuring the diameter of clear halos formed around bacterial colonies on agar plates after 5 days of incubation, as reported by Navarro-Torre et al. (2016). Lipase and protease activities were determined in Tween80 and casein agar media, respectively (Harley and Prescott, 2002). For chitinase activity, we used an M9 minimal medium supplemented with 1.5% (w/v) colloidal chitin (Mesa et al., 2015). DNase activity was recorded after adding 1 N HCl solution to DNA agar (Scharlab, Spain) plates. For amylase, 0.4% (w/v) Lugol solution was added to starch agar medium (Scharlab, Spain) after incubation to observe the positive strains. Finally, the activities of cellulase and pectinase were evaluated in carboxymethyl cellulose (CMC) and ammonium mineral agar (AMA) medium, respectively (Elbeltagy et al., 2000). Effect of selected bacterial strain inoculation on plant performance under salinity stress conditions In a greenhouse experiment, the potential of selected individual bacterial PGP strains and their consortium was evaluated on Medicago sativa plant growth and physiological performance under salinity stress conditions. Selection of bacterial strains and inoculum preparation After identifying the isolates, we focused on the strains belonging to the Halomonadaceae family, a taxonomic group known for harboring numerous halotolerant bacteria with PGP capabilities (Meinzer et al., 2023). We prioritized strains representing potentially PGP new species from recently described and less-explored genera such as Kushneria and Vreelandella (de la Haba et al., 2023) within this family. The strain selection criteria were based on a combination of functional, taxonomic, and compatibility considerations. Strains were selected according to their in vitro PGP traits performance, considering both the number and intensity of positive traits, as well as their 16S rRNA gene sequence similarity to described species. The compatibility among strains was initially assessed by co-culturing them on TSA plates supplemented with 0.5 M NaCl for 24 h at 30 °C. In addition, each strain was first grown individually in TSB supplemented with 0.5 M NaCl at 30 °C for 24 h to further evaluate the compatibility and stability of the consortium. Subsequently, equal volumes (50 µL) of each culture were added to fresh TSB medium supplemented with 0.5 M NaCl and incubated under the same conditions. After 24 h, serial dilutions were performed and plated to determine the colony-forming units (CFU), allowing the assessment of the relative abundance of each strain in the mixed culture. Furthermore, the PGP traits were re-evaluated using the combined inoculum to confirm their maintenance in the consortium. Following these criteria, strains TE3T, EE22T, EE4, and EE7T were selected to evaluate their PGP effects on M. sativa tolerance to salinity stress, both individually and in combination (consortium treatment). Bacterial inoculum for plant treatments was prepared by cultivating each strain individually in TSB supplemented with 0.5 M NaCl for 24 h. Bacterial cells were harvested by centrifugation, washed with 0.9% (w/v) NaCl solution to remove residual medium, and individually adjusted to an optical density of 1.0 at 600 nm (OD₆₀₀ = 1; ∼107 CFU·mL-1). Comparable CFU values among strains at this OD₆₀₀ were confirmed by serial dilution plating under identical growth conditions. Equal proportional volumes of the individual bacterial suspensions were mixed after centrifugation to obtain a final consortium inoculum adjusted to OD₆₀₀ = 1. Plant and bacterial preparation M. sativa seeds were first surface-disinfected. Briefly, the seeds were soaked in 70% (v/v) ethanol for 10 min, rinsed three times with sterile distilled water, and then immersed in a 3% (w/v) sodium hypochlorite solution with gentle agitation for 30 min, followed by six thorough rinses with sterile distilled water. After disinfection, batches of 100 seeds were bacterized by immersion in 1 mL of the respective bacterial culture suspension for 1 h under agitation according to their assigned inoculum treatment (TE3T, EE22T, EE4, EE7T, and consortium). In contrast, non-inoculated control seeds (Control treatment) were immersed in a sterile 0.9% (v/v) NaCl solution and agitated for 1 h. Finally, disinfected seeds were sown on sterile 0.9% (w/v) agar plates and pre-incubated at 4 °C in darkness for 24 h to synchronize their growth. Subsequently, they were incubated at 30 °C for 24 h in the dark to induce germination. Greenhouse experimental conditions In this experimental phase, we evaluated the effect of previously isolated and characterized bacterial strains on the tolerance of M. sativa to high salinity. Thus, 180 M. sativa seeds were surface-disinfected and bacterized following the procedure described previously. Pre-germinated alfalfa sprouts were transplanted into disinfected plastic pots (11 cm², 12 cm height) containing perlite as a substrate, which was previously sterilized by three successive autoclaving cycles at 121 °C for 20 min. Pots were placed in a greenhouse under a day/night regime of 14 h of light with a maximum photosynthetic photon flux density (PPFD) incident on leaves of 1000 μmol·m-2·s-1 at 25 °C and 10 h of darkness at 18 °C and 40%–50% relative humidity (RH). Plants were divided into six inoculation treatments: non-inoculated control (Control), inoculation with individual strains (TE3T, EE22T, EE4, and EE7T), and the joint of all strains in the consortium (treatment Consortium). Each pot was sown with three seeds, and 10 pots were used per condition (n = 30, N = 180). The pots were placed in disinfected plastic trays, separated by condition, and watered with sterile Hoagland’s solution (Hoagland and Arnon, 1950). A constant irrigation level of 1 cm was maintained by regular replenishment to ensure stable saline concentration. The plants were initially subjected to saline irrigation with 85 mM NaCl, which was gradually increased to a final concentration of 171.1 mM until the end of the experiment. For bacterial inoculation, each seedling was inoculated weekly with 5 mL of ∼107 CFU·mL-1 bacterial suspension, whereas plants of the Control treatment were irrigated with 0.9% (w/v) NaCl solution. Just before plant harvesting, plant growth and physiological performance parameters were also determined. Thus, fully developed leaves (n = 7) from different plants grown in different pots were randomly selected for measurements using an infrared gas analyzer (LI-6400, LI-COR Inc., USA) equipped with a fluorescence leaf chamber (LI-6400–02B, Li-Cor Inc.). Net photosynthetic rate (AN), stomatal conductance (gs), intercellular CO2 concentration (Ci), intrinsic water use efficiency (iWUE), photochemical quantum efficiency of PSII (ΦPSII) and electron transport rate (ETR) were recorded under the following leaf chamber conditions: a photosynthetic photon flux density (PPFD) of 1000 μmol photons·m-2·s-1, a vapor pressure deficit of 2.0–3.0 kPa, 50 ± 5% relative humidity (RH), CO2 concentration surrounding the branches (Ca) of 400 and air temperature of 25 °C. All measurements were corrected by photosynthetic area, using Easy Leaf Area software (Easlon and Bloom, 2014). Additionally, plant height was measured. On day 40 of the experiment, all plants from each treatment were harvested and divided into biomass fractions, which were then dried at 60 °C for 48 h and weighed to obtain estimates of the dry mass content of the leaf (LDMC) and root (RDMC). Statistical analysis Statistical tests were performed using the Statistica v. 10.0 software package and RStudio v. 2024.12.1. The model assumptions were evaluated by testing the normality of residuals (Shapiro–Wilk test) and homogeneity of variances (Levene’s test). A one-way analysis of variance (ANOVA) followed by Fisher’s least significant difference (LSD) post hoc test was used to analyze the effect of bacterial inoculation on plant growth and physiological parameters under greenhouse salinity conditions. The total number of PGP traits and enzymatic activities related to plant health detected for each bacterial isolate was compared among the three populations using the Kruskal–Wallis test, as the data were discrete and not normally distributed. Pairwise Wilcoxon rank-sum tests with Benjamini–Hochberg correction were used for multiple comparisons. Results Isolation and identification of cultivable halophilic and halotolerant bacteria A total of 76 endophytic bacterial strains were isolated by culture methods from H. strobilaceum aerial parts of the three differently located populations (GE, TE, and EE; Supplementary Figure 1). The isolates were subjected to 16S rRNA gene sequence analysis for taxonomic identification (Supplementary Table 2). Eight bacterial genera were identified, with notable differences in their relative taxonomic profile across populations with respect to the cultivable fraction of endophytic bacteria (Fig. 1). The most abundant taxonomic family found in the study was the family Halomonadaceae (67.1%), represented by the genera Vreelandella and Kushneria, both found in the same proportions (31.6%), and Halomonas in less quantity (3.9%). Moreover, the genus Marinilactibacillus was found in 22.4% of the total isolates, followed by the Bacillaceae family (6.6%) and the genera Bacillus (5.3%) and Priestia (1.3%). The lowest represented genera were Salinicoccus (2.6%) and Agrococcus (1.3%), exclusively from EE and TE populations, respectively. In contrast, Marinilactibacillus, Kushneria, and Vreelandella were common in all three populations, although they showed different proportional distributions across the study sites. The Alhama de Murcia (EE) population was dominated by Marinilactibacillus (44.4%) and Vreelandella (29.6%), whereas the Calarreona (TE) population was dominated by the genus Kushneria (57.7%). Population Guadalentín (GE), conversely, presented a more equitable distribution of taxa, as shown in Fig. 1.Fig. 1Genera distribution of cultivable endophytic bacterial isolates from the aerial tissues of H. strobilaceum, shown for (a) all samples and for each population: (b) Guadalentín salt marshes (GE), (c) Calarreona (TE), and (d) Alhama de Murcia (EE). Percentages (%) of the relative taxonomic profile are displayed.Fig 1 dummy alt text Genera distribution of cultivable endophytic bacterial isolates from the aerial tissues of H. strobilaceum, shown for (a) all samples and for each population: (b) Guadalentín salt marshes (GE), (c) Calarreona (TE), and (d) Alhama de Murcia (EE). Percentages (%) of the relative taxonomic profile are displayed. Detection of novel bacterial species According to 16S rRNA gene sequence analysis, some of the isolates showed low percentage identity with the type strains present in the database (Supplementary Table 2), indicating that they could be non-described species. Based on these results, we selected strains TE3T, TE19T, EE4, EE7T, EE22T, EE24, EE27, and GE22 (belonging to the family Halomonadaceae) for taxonomical characterization. A phylogenetic tree based on the 16S rRNA gene sequence (Supplementary Figure 2) showed that the closest species for strains TE3T and EE4 was Kushneria indalinina CG2.1T With 98.01% and 99.25% identity, respectively. In the case of TE19T, GE22, EE7T, EE24, EE22T, and EE27, Vreelandella azerica TBZ9T was the most closely related species, with percentages of identity of 97.18% for strains TE19T, GE22, EE7T, and EE24, 97.76% for strain EE22T, and 97.25% for EE27. Phylogenomic analyses Genome analyses were performed and basic statistics of the genomes were determined to be more accurate in the taxonomic affiliation of these strains (Supplementary Table 3). The phylogenomic tree based on whole proteomes (Fig. 2) showed that TE3T and EE4 branched in the same cluster as K. indalinina DSM 14324T and Kushneria pakistanensis KCTC 42082T, confirming that both strains belong to the genus Kushneria. Strains TE19T, EE7T, EE22T, EE24, EE27, and GE22 form a well-supported clade with other Vreelandella species, being the most closely related to Vreelandella malpeensis YU-PRIM-29TFig. 2Phylogenomic tree inferred from GBDP distances calculated from whole proteomes showing the phylogenetic position of strains TE3T, TE19T, EE4, EE7T, EE22T, EE24, EE27, and GE22 relative to the closest species. The branch lengths are scaled according to the GBDP distance formula d5. The numbers above the branches are GBDP pseudo-bootstrap support values of >60% from 100 replications, with an average branch support of 75.6%. The tree was rooted in the midpoint (Farris, 1972).Fig 2 dummy alt text Phylogenomic tree inferred from GBDP distances calculated from whole proteomes showing the phylogenetic position of strains TE3T, TE19T, EE4, EE7T, EE22T, EE24, EE27, and GE22 relative to the closest species. The branch lengths are scaled according to the GBDP distance formula d5. The numbers above the branches are GBDP pseudo-bootstrap support values of >60% from 100 replications, with an average branch support of 75.6%. The tree was rooted in the midpoint (Farris, 1972). Pairwise genome comparisons, dDDH analysis, and ANI values were obtained to determine the novelty of these strains (Table 1). All the studied strains showed dDDH and ANI values lower than the established threshold to describe new bacterial species (<95–96% ANI and <70% dDDH; Riesco and Trujillo, 2024). These results suggested that these eight isolates belong to a new bacterial species of the genera Kushneria and Vreelandella. Additional comparisons were performed among the eight strains to determine whether they are new species (Table 1). Based on these comparisons, strains TE3T and EE4 were the same new species, and the proposed name was Kushneria halocnemicola sp. nov., with TE3T as the type strain. Of the strains belonging to the genus Vreelandella, TE19T was the same new species as GE22, and EE7T was the same as EE24 and EE27. The strains were named Vreelandella mediterranea sp. nov. and Vreelandella halophytica sp. nov., where strains TE19T and EE7T were the type strains, respectively. Finally, strain EE22T represented the fourth new species, Vreelandella xerophytica sp. nov.Table 1Digital DNA-DNA hybridization (dDDH) and average nucleotide identity (ANI) based on MUMmer (ANIm) and BLAST+ (ANIb) values from more similar pairwise comparisons of overall genomes with their most closely related species. Similar species detection is highlighted in bold.Table 1 dummy alt textStrainsMost closely related speciesdDDH (%)ANIm (%)ANIb (%)TE19TV. malpeensis YU-PRIM-29T22.8084.5579.76GE2279.2097.8797.45EE7T66.0096.1395.57EE2466.0096.1395.57EE2765.7096.0895.48EE22T25.4086.0482.10GE22V. malpeensis YU-PRIM-29T22.8084.5579.92EE7T66.2096.2095.65EE2466.2096.2095.65EE2766.7096.1895.61EE22T25.5086.2082.23EE7TV. malpeensis YU-PRIM-29T22.9084.5779.75EE24100.0099.99100.00EE2781.3098.0797.33EE22T25.6086.0882.10EE24V. malpeensis YU-PRIM-29T22.9084.6079.78EE2781.3098.0697.37EE22T25.6086.0882.04EE27V. malpeensis YU-PRIM-29T23.0084.6780.07EE22T25.7086.2182.07EE22TV. malpeensis YU-PRIM-29T22.6084.5679.74TE3TK. indalinina DSM 14324T35.2089.3188.14EE484.0898.4198.13EE4K. indalinina DSM 14324T35.5089.3988.26 Table 1Digital DNA-DNA hybridization (dDDH) and average nucleotide identity (ANI) based on MUMmer (ANIm) and BLAST+ (ANIb) values from more similar pairwise comparisons of overall genomes with their most closely related species. Similar species detection is highlighted in bold.Table 1 dummy alt textStrainsMost closely related speciesdDDH (%)ANIm (%)ANIb (%)TE19TV. malpeensis YU-PRIM-29T22.8084.5579.76GE2279.2097.8797.45EE7T66.0096.1395.57EE2466.0096.1395.57EE2765.7096.0895.48EE22T25.4086.0482.10GE22V. malpeensis YU-PRIM-29T22.8084.5579.92EE7T66.2096.2095.65EE2466.2096.2095.65EE2766.7096.1895.61EE22T25.5086.2082.23EE7TV. malpeensis YU-PRIM-29T22.9084.5779.75EE24100.0099.99100.00EE2781.3098.0797.33EE22T25.6086.0882.10EE24V. malpeensis YU-PRIM-29T22.9084.6079.78EE2781.3098.0697.37EE22T25.6086.0882.04EE27V. malpeensis YU-PRIM-29T23.0084.6780.07EE22T25.7086.2182.07EE22TV. malpeensis YU-PRIM-29T22.6084.5679.74TE3TK. indalinina DSM 14324T35.2089.3188.14EE484.0898.4198.13EE4K. indalinina DSM 14324T35.5089.3988.26 Digital DNA-DNA hybridization (dDDH) and average nucleotide identity (ANI) based on MUMmer (ANIm) and BLAST+ (ANIb) values from more similar pairwise comparisons of overall genomes with their most closely related species. Similar species detection is highlighted in bold. Phenotypic and biochemical characterization The physiological, phenotypic, and biochemical characteristics of the eight strains were evaluated under the optimal growth conditions determined in this study, using TSA medium supplemented with 0.5 mM NaCl and incubated for 48 h at 30 °C. All strains were able to grow at temperatures ranging from 5 °C to 37 °C after 2–5 days with an optimum at 30 °C, except for strain TE3T, whose maximum growth temperature was 30 °C. None of the isolates grew at 40 °C. Regarding NaCl tolerance, strains EE7T, EE24, TE19T, TE3T, GE22, EE27, and EE4 grew at NaCl concentrations of up to 4 M, whereas strain EE22T exhibited a smaller tolerance range, with a maximum at 2.5 M. Strains EE22T, GE22, and EE4 were unable to grow in TSB medium without NaCl supplementation, confirming their halophilic nature. The remaining strains were able to grow in the absence of NaCl supplementation and were therefore classified as halotolerant bacteria. All strains grew at a pH value ranging from 6 to 10, but strains EE22T and EE4 presented weak and good growth at pH 12, respectively. Strains belonging to the genus Vreelandella (TE19T, GE22, EE22T, EE7T, EE24, and EE27) generally showed phenotypic characteristics similar to those of their closest relative species in terms of NaCl tolerance, temperature, and pH requirements, although some of the strains described in this study required higher minimum NaCl concentrations for growth (Supplementary Table 4). In contrast, the Kushneria strains TE3ᵀ and EE4 differed from their closest relatives by tolerating slightly lower temperatures and NaCl concentrations, whereas only strain EE4 tolerated higher pH values (Supplementary Table 5). The colony morphology and cell characteristics were also examined under optimal growth conditions. All strains were Gram-negative, non-spore-forming, motile rods with slightly shorter cell sizes but generally comparable to those of their closest related species: 1 × 1.30 µm (TE19T), 1 × 1.33 µm (GE22), 1 × 1.5 µm (EE22T), 1 × 1.13 µm (EE7T), 1 × 1.67 µm (EE24), 1 × 1.23 µm (EE27), 1 × 2.0 µm (TE3T) and 1 × 1.5 µm (EE4) (Supplementary Tables 4 and 5). All colonies exhibited a butyrous texture and a smooth, shiny appearance. Colonies of strains TE19ᵀ and GE22 were yellow-pigmented, slightly raised, smooth, and bright, with undulate and slightly translucent margins; TE19ᵀ colonies were irregular, whereas those of GE22 were circular and lighter in color. Strains EE7ᵀ and EE27 formed bright yellow, circular colonies with slightly raised elevation and undulate translucent margins, although EE27 colonies were smaller. Strains EE24 and EE22ᵀ also produced circular, yellow colonies with entire margins, with EE22ᵀ colonies being smaller and brighter than EE24 colonies. Strains TE3ᵀ and EE4 formed coral- to salmon-colored colonies that became pinker after prolonged incubation. TE3ᵀ colonies were circular, slightly raised, opaque, and had entire margins, whereas the EE4 colonies were more intensely pigmented and exhibited translucent margins. All strains were strongly positive for catalase activity and negative for oxidase activity. API 20NE biochemical tests showed positive nitrate reductase activity and negative d-glucose fermentation activity for all strains. Arginine dihydrolase activity was only detected in EE27, distinguishing it from the remaining strains. β-Galactosidase activity and aesculin hydrolysis were observed exclusively in TE3ᵀ and EE4, which differed in d-maltose and d-mannitol utilization; these were positive for TE3ᵀ but negative in EE4. Gelatin hydrolysis was observed only in EE24 and EE27 strains. None of the strains used l-tryptophan, capric acid, phenylacetic acid, or N-acetylglucosamine as the sole carbon source, whereas all strains assimilated trisodium citrate and potassium gluconate. Regarding Vreelandella strains, none produced indole from l-tryptophan or utilized d-maltose as the sole carbon source, in agreement with V. malpeensis YU-PRIM-29ᵀ and V. populi MCCC 1K03942T (Kämpfer et al., 2018; Xu et al., 2022). All strains shared positive assimilation of trisodium citrate with V. malpeensis YU-PRIM-29ᵀ and were negative for aesculin hydrolysis, consistent with V. populi MCCC 1K03942ᵀ. Differences between Vreelandella strains and their closest relative species are shown in Supplementary Table 4. Regarding Kushneria strains TE3ᵀ and EE4 shared these biochemical traits with their closest related species, including negative oxidase activity, both positive catalase and β-galactosidase activities, and the utilization of d-mannose, d-glucose, l-arabinose, malic acid, and trisodium citrate, as reported for K. indalinina DSM 14,324ᵀ, K. pakistanensis KCTC 42,082ᵀ, and K. avicenniae DSM 23,439ᵀ (Bangash et al., 2015; Cabrera et al., 2007; Soto-Ramírez et al., 2007). These strains or their closest relatives did not use capric acid, adipic acid, and phenylacetic acid. The differences between Kushneria strains and related species are shown in Supplementary Table 5. Genome-based identification of PGP traits and salinity tolerance genes The whole genomes of strains TE3T, TE19T, EE4, EE7T, EE22T, EE24, EE27, and GE22T were sequenced to perform phylogenomic analyses, and the genomes were analyzed to find the genes involved in PGP traits (Supplementary Table 6). Genes involved in phosphate, sulfate, and iron solubilization and metabolism were annotated in all strains in terms of nutrient acquisition facilitation. Phosphate solubilization mechanisms commonly found in rhizospheric bacteria were identified, such as gluconic acid (GA) synthesis by glucose-1-dehydrogenase (gcd) and oxidoreductase (gdhB), with pyrroloquinoline quinone (PQQ) acting as a cofactor for both enzymes, which confirms our in vitro results. Both strain TE3T and EE4, along with strains EE7T and EE24, presented both gcd/gdhB genes together with pqq operon genes required for GA synthesis. All strain genomes presented complete metabolic pathways for inorganic phosphate uptake, including high-affinity ABC-type transporters (pstABC, pstS) and low-affinity transporters (pit), phosphate-sensing systems (phoR/phoB), phosphate homeostasis and intracellular mobilization (ppx, ppk), organic phosphate mineralization (phoA, ppa, nudJ), and utilization of alternative phosphorus sources such as phosphonates (phnA, phnCDE). In terms of sulfur metabolism, genes for sulfate/thiosulfate intake were encoded in all strains (cysA, cyst/W, cysZ, cysP) together with genes for assimilatory reduction of sulfite, including both the classic NADPH-dependent pathway (cysJI) and the alternative ferredoxin-dependent pathway (sir), as well as sulfotransferases involved in intracellular redistribution (glpE). Alkanesulfonates and aromatic sulfonates represented the main sources of sulfur in agricultural lands, for which strains presented genes associated with their transport and, in some cases, genes for their partial utilization and aromatic sulfonate catabolism (tsaM1/tsaC1). Regarding iron acquisition, all strains presented metabolic pathways associated with iron scavenging, such as ABC siderophore transporters, permeases, TonB-dependent receptors, processing enzymes, and catechol-type exporters (enterobactin/bacillibactin), which confirm our in vitro test results. Strains TE3T and EE4 also encoded well-known genes associated with catechol-type siderophore biosynthesis (ent/dhb cluster genes). No nitrogenase complex genes or ACC deaminase gene were detected. Other found genes related to PGP traits included motility-associated genes, such as structural and motor components of the flagellar apparatus, regulatory elements, and flagellar biosynthesis genes, along with genes associated with chemotactic responses. Genes encoding poli-N-acetylglucosamine (PGA) biofilm formation were identified in strains EE7T and EE22T, while all strains encoded genes associated with alginate synthesis, biofilm dispersion, and regulation. All strains also encoded genes for tryptophan biosynthesis, the metabolic precursor of IAA (trpA-E), the indole-3-acetamide (IAM) pathway for auxin biosynthesis (amiE, iaaH), and the last conversion of indole-3-acetaldehyde into auxin (IAA) (aldhT/puuC). Multiple genetic traits commonly found in halophilic and halotolerant bacteria, such as genes related to salinity stress tolerance and osmotic stress mitigation, were detected in all of our novel species strains (Supplementary Table 7). These included heat-shock proteins and chaperones (dnaK, dnaJ, grpE, groES/groEL, and clpB), proteases (clpP, clpX, clpA, htpX, and htrA), and genes involved in DNA/RNA protection and translation stability (dps, hfq, lepA, hslR, and mazG). In addition, genes related to ion homeostasis and intracellular pH conservation were widely distributed across the eight strains, such as K+ uptake high-affinity systems (kdp genes) and low-affinity ones (kup/trkA, trkH), K+ efflux and channel proteins, together with Na+/H+ antiporters. Genes involved in osmolyte synthesis and accumulation were also broadly conserved, indicating a strong capacity for osmotic stress tolerance. These included the biosynthesis of proline (proA, proB, and proC), glycine betaine (betA and betB), and ectoine (ectA, ectB, and ectC), as well as different trehalose biosynthetic pathways. All strains encoded the treY/treZ and treS pathways, whereas the otsA/otsB pathway was specific for strains TE3T and EE4. In addition, the polyamine synthesis pathways were identified. All strains encoded genes for putrescine biosynthesis (spe genes), whereas genes involved in spermidine biosynthesis were only present in Vreelandella strains, and genes for nospermidine biosynthesis (nspC) were encoded in Kushneria strains. Finally, a wide distribution of genes encoding superoxide dismutases involved in ROS detoxification (sodA, sodB, sodC), catalases (katE, katA/katB, katG), peroxidases, peroxiredoxin (osmC), and glutaredoxin enzymes involved in redox homeostasis (grxA-D) was observed. Screening of the in vitro PGP traits and enzymatic activities of the isolates All isolates from the three different populations were tested for plant growth-promoting traits and plant growth-related enzymatic activities. Most of the isolates (96.05%) (n = 73) presented at least one of the PGP traits tested: nitrogen fixation, K and P solubilization, IAA production, siderophores, biofilm formation, or ACC deaminase (Supplementary Table 8). Conversely, nine isolates presented six of the seven PGP traits tested, making them good candidates for further selection for biofertilization strategies. Possible nitrogen fixation was detected in 60.5% of the isolates. Almost all isolates from Calarreona (TE) exhibited this capability (96.2%), whereas only 21.7% of isolates from the Guadalentín salt marshes (GE) and 59.3% from the Alhama de Murcia (EE) populations showed this activity (Fig. 3).Fig. 3PGP traits and enzymatic activities in the endophytic bacterial isolates from the aerial tissues of Halocnemum strobilaceum: N fixation (N fix), P solubilization (P solub), K solubilization (K solub), auxin production (IAA prod), siderophores production (siderophores), biofilm formation (biofilm), protease, lipase, amylase, DNase, and cellulase activities. Different populations are indicated by different colors (Alhama de Murcia (EE), Guadalentín salt marshes (GE), and Calarreona (TE)). Activities with no positive results are not shown. For IAA production, only strains able to produce ≥ 0.9 mg·L-1 have been considered.Fig 3 dummy alt text PGP traits and enzymatic activities in the endophytic bacterial isolates from the aerial tissues of Halocnemum strobilaceum: N fixation (N fix), P solubilization (P solub), K solubilization (K solub), auxin production (IAA prod), siderophores production (siderophores), biofilm formation (biofilm), protease, lipase, amylase, DNase, and cellulase activities. Different populations are indicated by different colors (Alhama de Murcia (EE), Guadalentín salt marshes (GE), and Calarreona (TE)). Activities with no positive results are not shown. For IAA production, only strains able to produce ≥ 0.9 mg·L-1 have been considered. K and P solubilization was less frequent, occurring in 34.2% and 36.8% of isolates, respectively, across all populations. Nevertheless, both activities were considerably more prevalent in the TE population, with 61.5% and 65.4% of isolates exhibiting K and P solubilization, respectively (Fig. 3). IAA production was particularly notable among the isolates from the Guadalentín (GE) population, with 82.6% of the isolates exhibiting this activity. This proportion was considerably lower in the other populations, with 42.3% in TE and only 14.8% in EE (Fig. 3). However, in terms of IAA concentration, the highest production values (mg·L⁻¹) were recorded in the EE population, ranging from 3.513 mg·L⁻¹ in strain EE22 to 11.713 mg·L⁻¹ in strain EE24, with the latter being the highest IAA production observed (Supplementary Table 8). The GE population showed values ranging from 0.923 mg·L⁻¹ in strain GE21 (the lowest recorded value) to 5.051 mg·L⁻¹ in strain GE16. Finally, in the TE population, IAA production values ranged from 1.069 mg·L⁻¹ to 6.792 mg·L⁻¹ in strain TE11 (Supplementary Table 8). Siderophores production and biofilm formation were generally widespread among all isolates, being detected in 57.9% and 75.0% of them, respectively. The highest prevalence of both traits was observed in the TE population, with 76.9% and 84.6% of isolates exhibiting siderophore production and biofilm formation, respectively (Fig. 3). The observed differences in PGP traits among populations were confirmed by the Kruskal–Wallis test (χ² = 16.65, df = 2, p < 0.001) and pairwise Wilcoxon rank-sum tests, which revealed that the Calarreona (TE) population displayed significantly higher numbers of PGP traits compared with both Alhama de Murcia (EE) and Guadalentín (GE) populations (p < 0.001 for both comparisons), whereas no significant differences were detected between EE and GE populations (Supplementary Figure 3). ACC deaminase activity was also assessed, but none of the endophytic isolates exhibited this activity. The isolates were further evaluated for enzymatic activities, including protease, lipase, amylase, DNase, cellulase, chitinase, and pectinase, which are beneficial to plant nutrition and health. DNase activity appeared to play a particularly relevant role in the EE population, which not only presented the highest proportion of isolates with this activity (81.5%) (Fig. 3) but also produced the largest halo diameters during the in vitro assays (Supplementary Table 8). Protease and amylase activities were detected in all three populations, whereas lipase activity was detected in only one isolate from each population (Fig. 3). No positive results were obtained for chitinase or pectinase. Additionally, all isolates showed at least one positive functional attribute when considering both PGP traits and enzymatic activities related to plant health. Effect of bacterial inoculation on M. sativa growth and physiological performance under greenhouse conditions Four isolates were selected to assess their impact on the growth and physiological performance of M. sativa under salt stress. The selected bacteria exhibited all the PGP traits observed in this study (Table 2). After a compatibility test, the results showed that the four strains could grow together stably in the same medium and to maintain the PGP traits.Table 2Plant growth promoting (PGP) traits and enzymatic activities of the selected strains (TE3T, EE22T, EE7T, and EE4) and the consortium.Table 2 dummy alt textTreatmentTE3TEE22TEE7TEE4ConsortiumPGP traitsN fixation++++-P solubilization+--++K solubilization+-+++++IAA production(mg l-1)1.4623.513--15.95Siderophores production(cm)0.80.20.60.80.2Biofilm formation++++++++Enzymatic activitiesProtease+-++-n. dChitinase----n. dLipase----n.dPectinase----n. dAmylase++++n. dDNase-++++-n. dCellulase+-++++n. d+, positive activity; ++, moderate activity; +++, strong activity; n.d, no data.IAA values < 0.9 mg l-1 are considered non-active.The values of siderophores produced express the diameter of the halo in cm. Table 2Plant growth promoting (PGP) traits and enzymatic activities of the selected strains (TE3T, EE22T, EE7T, and EE4) and the consortium.Table 2 dummy alt textTreatmentTE3TEE22TEE7TEE4ConsortiumPGP traitsN fixation++++-P solubilization+--++K solubilization+-+++++IAA production(mg l-1)1.4623.513--15.95Siderophores production(cm)0.80.20.60.80.2Biofilm formation++++++++Enzymatic activitiesProtease+-++-n. dChitinase----n. dLipase----n.dPectinase----n. dAmylase++++n. dDNase-++++-n. dCellulase+-++++n. d+, positive activity; ++, moderate activity; +++, strong activity; n.d, no data.IAA values < 0.9 mg l-1 are considered non-active.The values of siderophores produced express the diameter of the halo in cm. Plant growth promoting (PGP) traits and enzymatic activities of the selected strains (TE3T, EE22T, EE7T, and EE4) and the consortium. +, positive activity; ++, moderate activity; +++, strong activity; n.d, no data. IAA values < 0.9 mg l-1 are considered non-active. The values of siderophores produced express the diameter of the halo in cm. Bacterial inoculation had a significant overall positive effect on M. sativa plant growth under excess salinity at the end of the experiment (40 days), with the magnitude of this effect varying depending on the specific inoculation treatment applied (ANOVA, p < 0.05). Thus, compared with the non-inoculated control, plants inoculated with the bacterial consortium showed increases of 60%, 84%, and 105% in plant height, LDMC, and RDMC, respectively, whereas those inoculated with strain EE22T exhibited increases of 21%, 44%, and 96%, respectively (Fig. 4). In contrast, individual inoculation with TE3T and EE7T only increased RDMC by approximately 50% for both treatments, whereas inoculation with EE4 had no significant effect on M. sativa growth compared with their non-inoculated counterparts (ANOVA, p > 0.05; Fig. 4).Fig. 4Plant height (A), leaf dry matter content (LDMC; B), and root dry matter content (RDMC; C) of Medicago sativa plants subjected to different bacterial inoculation treatments after 40 days. Treatments: Control (non-inoculated), Consortium (combined inoculation with strains TE3T, EE22T, EE4, and EE7T), and individual inoculations of each strain. Values represent means ± SE (n ≥ 7). Different letters indicate statistically significant differences among means (LSD test, p < 0.05).Fig 4 dummy alt text Plant height (A), leaf dry matter content (LDMC; B), and root dry matter content (RDMC; C) of Medicago sativa plants subjected to different bacterial inoculation treatments after 40 days. Treatments: Control (non-inoculated), Consortium (combined inoculation with strains TE3T, EE22T, EE4, and EE7T), and individual inoculations of each strain. Values represent means ± SE (n ≥ 7). Different letters indicate statistically significant differences among means (LSD test, p < 0.05). Bacterial inoculation had an overall positive impact on M. sativa gas exchange and chlorophyll fluorescence traits under salinity excess (ANOVA, p < 0.05; Fig. 5), regardless of the type of inoculation treatment used. However, the highest values of AN, gs, ΦPSII and ETR were recorded for the consortium treatment compared with the inoculation treatments with the individual strains (Fig. 5A, B, E, and F).Fig. 5Photosynthetic parameters of Medicago sativa: net photosynthetic rate (AN; A), stomatal conductance (gₛ; B), intercellular CO₂ concentration (Cᵢ; C), intrinsic water-use efficiency (iWUE; D), PSII quantum efficiency (ΦPSII; E), and electron transport rate (ETR; F). Measurements were taken from fully developed leaves of plants from different pots (n = 7) after 40 days. Treatments: Control (non-inoculated), Consortium (combined inoculation with strains TE3T, EE22T, EE4, and EE7T), and individual inoculations of each strain. Data are means ± SE (n = 5). Different letters denote significant differences (LSD test, p < 0.05).Fig 5 dummy alt text Photosynthetic parameters of Medicago sativa: net photosynthetic rate (AN; A), stomatal conductance (gₛ; B), intercellular CO₂ concentration (Cᵢ; C), intrinsic water-use efficiency (iWUE; D), PSII quantum efficiency (ΦPSII; E), and electron transport rate (ETR; F). Measurements were taken from fully developed leaves of plants from different pots (n = 7) after 40 days. Treatments: Control (non-inoculated), Consortium (combined inoculation with strains TE3T, EE22T, EE4, and EE7T), and individual inoculations of each strain. Data are means ± SE (n = 5). Different letters denote significant differences (LSD test, p < 0.05). Discussion The salinization of agricultural lands is a global problem that causes yield loss because most crops cannot grow in saline environments (Mishra et al., 2023). Biofertilization using halophilic or halotolerant PGPB is one of the most sustainable solutions (Navarro-Torre et al., 2023b). The first objective in this work was to search for this kind of bacteria in the microbiota associated with H. strobilaceum. H. strobilaceum was selected because it is a halophyte and is a good source of halophilic and halotolerant bacteria (Jiang et al., 2025; Navarro-Torre et al., 2023b; Teo et al., 2022). This plant species has evolved pronounced shoot succulence as a salt-tolerance mechanism, sequestering high concentrations of salts in the vacuoles of aerial tissues while maintaining cytoplasmic homeostasis (Mann et al., 2023; Nasernakhaei and Zahraei, 2021). Therefore, the search for halophilic and halotolerant PGPB was performed by isolating bacteria from aerial tissues. The microbiota of the shoot endosphere of H. strobilaceum included the most abundant genera, such as Kushneria, Vreelandella, and Marinilactibacillus, with some variation among growing areas. These results are expected because the plant microbiota can be modulated based on environmental conditions, including abiotic stresses and soil properties (Furtado et al., 2019; Mora-Ruiz et al., 2016; Navarro-Torre et al., 2017, 2016; Zhang et al., 2024). The fact that the most abundant bacteria belonged to the Kushneria and Vreelandella genera is also expected as they have been described by several authors as the principal genera associated with halophytes (Meinzer et al., 2023; Navarro-Torre et al., 2016) due to their high salt tolerance. It is important to highlight that many isolates from both of these genera exhibited a low percentage of identities with the 16S rRNA gene of the species deposited in public databases, indicating the possibility of being new species. The gold standard for species delineation has shifted from dDDH to genome-to-genome distance and average nucleotide identity: strains sharing ≥95% ANI and ≥70% dDDH belong to the same species (Riesco and Trujillo, 2024). Accordingly, four new species were identified (strains TE3T, TE19T, EE7T, and EE22T) and named Kushneria halocnemicola sp. nov., Vreelandella mediterranea sp. nov., Vreelandella halophytica sp. nov., and Vreelandella xerophytica sp. nov., respectively. The isolation of new species from halophytes has been reported previously (Navarro-Torre et al., 2023b), and now H. strobilaceum is included as an important specialized ecological niche of novel bacterial species. Regarding characterization, most isolates showed PGP traits, such as nitrogen fixation, K and P solubilization, IAA production, siderophores, and biofilm formation, in addition to lytic enzymatic activities, such as protease, lipase, amylase, DNase, and cellulase. These properties are involved in the increment of nutrient acquisition by plants because limiting factors such as K, P, Fe, and N are solubilized or chelated from inorganic compounds or fixed from the atmosphere, making them available for the plant (Maciel-Rodríguez et al., 2025). Although some strains were able to grow in a nitrogen-free minimal medium, we cannot confirm their capacity for nitrogen fixation because this medium could present trace amounts of nitrogen. Therefore, growth under these conditions should be considered as a preliminary screening for putative nitrogen-fixing bacteria and requires additional studies, such as the detection of nitrogenase genes in their genomes. Based on genomic analyses, we can conclude that strains TE3T, TE19T, EE4, EE7T, EE22T, EE24, EE27, and GE22 are not nitrogen fixers, as the nifH gene and other alternative genes were not found. According to the beneficial effects provided by these properties, all the isolated bacteria in this study have the capacity to promote plant growth, and they could be used as biofertilizers. Several authors have demonstrated that bacterial strains isolated from the microbiota of H. strobilaceum promote the growth of this plant species and cotton (Begali et al., 2021; Marasco et al., 2016). Moreover, Bekkaye et al. (2023) tested strains BSSM328 and BSSM27 as biofertilizants to enhance the growth of durum wheat under saline conditions (100 and 200 mM NaCl), and the results showed that both strains improved plant growth, decreasing the salt inhibitory effect. This protective effect against excess salt could be because these two strains were halotolerant PGPB of the genera Halomonas and Kushneria (Halomonadaceae family) (Göker et al., 2025). In our work, the most promising bacteria of the Halomonadaceae family were selected for the in planta assay according to the PGP properties and compatibility results. In addition to the PGP traits, the strains’ novelty was considered for selection to extend the known species applied to sustainable agriculture. Alfalfa was selected as a model plant in this experiment because it is a legume usually used to recover the fertility and quality of agricultural soils between crops (intercropping) (Dai et al., 2019; Raza et al., 2020). In addition, M. sativa is a high‑productivity forage legume closely related to the model species Medicago truncatula, which is widely used to extrapolate findings to other agriculturally important legumes (Bell et al., 2001; Zhang and Wang, 2025). Plants inoculated with the consortium showed the highest values in all the measured parameters under saline conditions. Bacterial consortium increased the size and weight of plants, improving leafiness under a final concentration of 171.1 mM NaCl. Single inoculation with strain EE22T also showed higher statistically significant values in plant height and weight, suggesting that the plant growth promotion under saline conditions by the consortium could be due to the role of EE22T principally, and the other three strains support this effect. Although consortia of bacteria are generally reported to provide significantly greater benefits to plants than single-strain inoculations (Timofeeva et al., 2023), both the single inoculation with strain EE22T and the consortium inoculation significantly enhanced alfalfa growth under saline conditions. Plant promotion could be due to improved nutrient acquisition mediated by the IAA production because this phytohormone plays a key role in root development and acts as a signalling molecule that modulates gene expression in the response of plants against stress (Maciel-Rodríguez et al., 2025). In addition, biofilm formation could be implicated with a dual effect: improvement of nutrient acquisition by the accumulation of nutrients in the exopolysaccharides and protection of plants against abiotic stress, avoiding the entry of toxic compounds such as metals or salt, among others (Li et al., 2024). In addition to the PGP traits demonstrated in vitro and confirmed in silico for the bacterial inoculants used in the experiments, the enhanced plant growth observed under saline stress conditions may also be related to the different salt-tolerance mechanisms identified in the selected strains through in silico analyses. These mechanisms include high- and low-affinity K+ uptake systems, K+ efflux and channel proteins, Na+/H+ antiporters, osmolyte synthesis and accumulation (proline, glycine betaine, ectoine, and trehalose), and an antioxidant defense systems involving superoxide dismutases, catalases, peroxidases, peroxiredoxins, and glutaredoxins. Together, these adaptations likely enable the bacteria to survive and remain metabolically active under high salinity conditions where many other bacterial species cannot persist (Gunde-Cimerman et al., 2018; Wang et al., 2023; Zamanzadeh-Nasrabadi et al., 2023). However, the exact mechanisms underlying our results require further investigation through transcriptomic analyses and mutant assays. Additionally, field experiments should be conducted to determine whether the observed beneficial effects persist under real environmental conditions and to evaluate the behavior of the selected bacteria in natural soils across different seasonal periods. Conclusion In this study, H. strobilaceum was shown to be a valuable source of halotolerant and halophilic PGPB, as the bacterial isolated obtained displayed PGP traits in vitro. These traits were further confirmed in silico and in vivo in the strains evaluated in the plant experiments. Both individual and consortium inoculation experiments with a selection of the strains showing the most promising PGP traits promote the growth of M. sativa under a stressful salt concentration. This plant promotion under saline conditions could be due to the capacity of Vreelandella and Kushneria strains to protect the plant against salt through different mechanisms found in their genomes. However, the exact implicated mechanism should be thoroughly studied in further experiments. The positive effect of the bacterial inoculation in alfalfa suggests that these strains could be used as biofertilizers to assist the development of different legumes with agricultural interest in agricultural soils affected by salinity, an important problem around the world. Despite these results, field experiments should be conducted to confirm them. Finally, some of the isolated strains (including the selected ones) were described as new species of bacteria of the genera Vreelandella and Kushneria, indicating that H. strobilaceum is also a source of new species with biofertilization potential. Description of Kushneria halocnemicola sp. nov. Kushneria halocnemicola sp. nov. (ha.lo.cne.mi’co.la. N.L. neut. n. Halocnemum, the scientific name of a botanical genus; L. masc./fem. n. suff. -cola, inhabitant; N.L. fem. n. halocnemicola, inhabitant of Halocnemum, isolated from stems of Halocnemum strobilaceum). Cells are Gram-negative, non-spore forming rod or short oval-shaped (1.0 × 2.0 μm) and motile. Colonies grown on optimal conditions (TSA supplemented with 2.92% NaCl (w/v) at pH 7.0 at 30 °C for 48 h) are light salmon, smooth and shiny in appearance, opaque, circular, and slightly convex with entire margins and butyrous texture. Moderately halophilic, growing at NaCl concentrations in the range 0–23.38% (w/v) with optimum growth at 2.92% (w/v). Growth occurs at 5–30 °C (optimum, 30 °C) and the pH range for growth is 6.0–10.0 (optimum, pH 7.0). The test for catalase activity is positive, and the oxidase is negative. Nitrate reduction is positive, while d-glucose fermentation and l-arginine dehydrolase and indole production from l-tryptophan are negative. Hydrolysis of gelatin is negative, and hydrolysis of aesculin is positive. According to API 20NE, several compounds are utilized as a sole carbon source by strain TE3T, among them are d-glucose, l-arabinose, d-mannose, d-mannitol, d-maltose, potassium gluconate, malic acid, and trisodium citrate. N-acetyl-glucosamine, capric acid, adipic acid, and phenylacetic acid are not utilized as a sole carbon source. The genome of the type strain TE3T has 3567539 bp assembled in 1 contig with a mean coverage of 113.7 × and an N50 value of 3567539 bp. The G + C content is 60.92%. The type strain is TE3T (= CECT 31276T = LMG 34081T) and was isolated from the internal tissues of the halophyte Halocnemum strobilaceum, which grows in the south of Spain. The GenBank accession number for the 16S rRNA gene sequence is PV916401. The GenBank/EMBL/DDBJ accession number for the draft genome is GCA_965650455. Description of vreelandella halophytica sp. nov. Vreelandella halophytica sp. nov. (ha.lo.phy'ti.ca. Gr. masc. n. hals, halos, salt; Gr. neut. n. phyton, plant; N.L. fem. adj. halophytica, pertaining to salt plants, i.e. halophytes). Cells are Gram-negative, non-spore forming, short rod-shaped (1.0 × 1.13 μm) and motile. Colonies grown on optimal conditions (TSA supplemented with 2.92% NaCl (w/v) at pH 7.0 at 30 °C for 48 h) are yellow-pigmented, smooth and bright in appearance, opaque, circular and raised with entire/slightly undulate margins and butyrous texture. Growth occurs between NaCl concentrations in the range 0–23.38% (w/v) with optimum NaCl concentration at 2.92% (w/v). The temperature range for growth is 5–37 °C (optimum, 30 °C) and the pH range for growth is 6.0–10.0 (optimum, pH 6.5–7.0). The test for catalase activity is positive, and the oxidase is negative. Nitrate reduction is positive, while d-glucose fermentation, l-arginine dehydrolase, β-galactosidase (PNPG), and indole production from l-tryptophan are negative. Hydrolysis of gelatin is positive, and hydrolysis of aesculin is negative. Positive results for utilization as a sole carbon source in API 20NE tests are d-glucose, l-arabinose, potassium gluconate, malic acid, and trisodium citrate. d-mannose, d-mannitol, N-Acetyl-glucosamine, d-maltose, capric acid, adipic acid, and phenylacetic acid are not used as a sole carbon source. The genome of the type strain EE7T has 3860807 bp assembled in 1 contig with a mean coverage of 130.5 × and an N50 value of 3860807 bp. The G + C content is 60.46%. The type strain is EE7T (= CECT 31281T = LMG 34082T) and was isolated from the internal tissues of the halophyte Halocnemum strobilaceum, which grows in the south of Spain. The GenBank accession number for the 16S rRNA gene sequence is PV916403. The GenBank/EMBL/DDBJ accession number for the draft genome is GCA_968193635. Description of vreelandella mediterranea sp. nov. Vreelandella mediterranea sp. nov. (me.di.ter.ra'ne.a. L. fem. adj. mediterranea, inland, referring to the Mediterranean). Cells are Gram-negative, non-spore-forming, short rod-shaped (1.0 × 1.30 μm) and motile. Colonies grown on optimal conditions (TSA supplemented with 2.92% NaCl (w/v) at pH 7.0 at 30 °C for 48 h) are pale yellow-pigmented, smooth, shiny, butyrous, opaque, a bit raised, circular with irregular and translucent margins. Requires NaCl for growth, from 1.0–23.38% (w/v) with optimum NaCl concentration at 1.17–2.92% (w/v). The temperature range for growth is 5–37 °C (optimum, 30 °C) and the pH range for growth is 6.0–10.0 (optimum, pH 6.0–7.0). The test for catalase activity is positive. Oxidase test is negative. Nitrate reduction is positive, while d-glucose fermentation, l-arginine dehydrolase, β-galactosidase (PNPG), and indole production from l-tryptophan are negative. Hydrolysis of gelatin and aesculin is negative. Positive results for utilization as a sole carbon source in API 20NE tests are d-glucose, l-arabinose, potassium gluconate, malic acid, and trisodium citrate. d-mannose, d-mannitol, N-Acetyl-glucosamine, d-maltose, capric acid, adipic acid, and phenylacetic acid are not used as a sole carbon source. The genome of the type strain TE19T has 3723011 bp assembled in 2 contigs with a mean coverage of 105.5 × and an N50 value of 3721309 bp. The G + C content is 60.26%. The type strain is TE19T (= CECT 31275T = LMG 34086T) and was isolated from the internal tissues of the halophyte Halocnemum strobilaceum, which grows in the south of Spain. The GenBank accession number for the 16S rRNA gene sequence is PV916402. The GenBank/EMBL/DDBJ accession number for the draft genome is GCA_970647075. Description of vreelandella xerophytica sp. nov. Vreelandella xerophytica sp. nov. (xe.ro.phy'ti.ca. Gr. masc. adj. xeros, dry; Gr. neut. n. phyton, plant; N.L. fem. adj. xerophytica pertaining to dry plants, i.e. xerophytes). Cells are Gram-negative, non-spore-forming short rods (1.0 × 1.50 μm) and motile. Colonies grown on optimal conditions (TSA supplemented with 2.92% NaCl (w/v) at pH 7.0 at 30 °C for 48 h) are light yellow-pigmented, smooth, bright, butyrous, opaque, a bit raised, circular with entire margins. Requires NaCl for growth, from 1.0–14.61% (w/v) with optimum NaCl concentration at 2.92% (w/v). The temperature range for growth is 5–37 °C (optimum, 30 °C), and the pH range for growth is 6.0–12.0 (optimum, pH 6.0–7.0). The test for catalase activity is positive. Oxidase test is negative. Nitrate reduction is positive. d-glucose fermentation, l-arginine dehydrolase, β-galactosidase (PNPG), and indole production from l-tryptophan are negative. Gelatin and aesculin can’t be hydrolyzed. According to API 20NE tests, assimilation of adipic acid, malic acid, and trisodium citrate is positive. Negative results are obtained in assimilation of d-glucose, l-arabinose, d-mannose, d-mannitol, N-acetyl-glucosamine, d-maltose, potassium gluconate, capric acid, and phenylacetic acid. The genome of the type strain EE22T has 3785665 bp assembled in 1 contig with a mean coverage of 204.5 × and an N50 value of 3785665 bp. The G + C content is 59.83%. The type strain is EE22T (= CECT 31280T = LMG 34087T) and was isolated from the internal tissues of the halophyte Halocnemum strobilaceum, which grows in the south of Spain. The GenBank accession number for the 16S rRNA gene sequence is PV916404. The GenBank/EMBL/DDBJ accession number for the draft genome is GCA_968048975. STATEMENTS & declarations Data availability 16S rRNA gene sequences of the strains were deposited in DDBJ/EMBL/GenBank databases under the following accession numbers: PV891966, PV892438–PV892451, PV892989–PV892999, PV893000–PV893021, PV916240–PV916256, PV916335–PV916336, PV916401, PV916402, PV916403, PV916404, PV916405, PV916406, PV929847–PV929848, and PV932119. The whole genomes were deposited in the DDBJ/EMBL/GenBank databases with the following accession numbers: GCA_965650455 (TE3T), GCA_968048975 (EE22T), GCA_968193635 (EE7T), GCA_968199205 (EE24), GCA_968206495 (EE27), GCA_968212445 (GE22), GCA_970647075 (TE19T), and GCA_977968315 (EE4). Raw data of plant physiological and growth parameters are available at https://hdl.handle.net/11441/183281 and https://doi.org/10.12795/11441/183281. Funding This work is part of project PID2024–160601NB-I00, funded by MICIU/AEI/10.13039/501100011033 and the European Research and Development Fund/EU. Additional funding was provided by project PID2021–124750NB-I00, funded by MICIU/AEI/10.13039/501100011033 and the European Research and Development Fund (ERDF/EU). The University of Sevilla contributed to funding this work through the Plan Propio de Investigación y Transferencia grant PPIT-EPD-683. Author contributions Conceptualization: I.D.R.-L., S.R.-G., E.M.-N., and S.N.-T.; methodology: E.R-R. and N.J.F.-D.; software: E.R-R.; validation: I.D.R.-L., E.M.-N., S.R.-G., and S.N.-T.; investigation: E.R-R., N.J.F.-D., S.R.-G., and E.M.-N.; writing—original draft preparation: E.R-R. and S.N.-T.; writing—review and editing: S.N.-T.; supervision: S.R.-G., E.M-N., and S.N.-T.; funding acquisition: S.R.-G., E.M.-N, and E.P.; resources: I.D.R.-L. and E.P. All authors have read and approved the published version of the manuscript. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.