Favorable habitats of endangered halophytes harbor unique soil properties and bacterial communities, with a dominant root-associated genus promoting host plant growth

Understanding how edaphic factors and root-associated bacterial communities interact to shape plant growth is essential for effective conservation of halophytic species in coastal ecosystems. We investigated relationships among bulk soil, rhizosphere, and root bacterial microbiota; soil physicochemical properties; and Salicornia europaea growth spanning three distinct areas (Lakeside Park, Ubaranai and Misaki) within a brackish lake in Japan. Among the three study areas, S. europaea individuals from Lakeside Park exhibited the highest mean values for shoot length, fresh weight, dry weight, and number of primary branches, which were 1.9-, 11.2-, 12.1-, and 7.2-fold higher than those at Misaki in 2021, and 1.8-, 4.7-, 6.7-, and 4.7-fold higher than those at Misaki in 2022. Bacterial diversity and community composition associated with S. europaea varied in response to differences in plant growth and edaphic conditions. Soil texture, electrical conductivity, pH, total organic carbon, total nitrogen, C/N ratio, and water-soluble Ca 2+ were identified as key factors underlying microbial community structures. Based on heatmap and LEfse analyses, six bacterial genera, Bacteroidetes BD2-2, Candidatus Thiobios, SEEP-SRB1, Calditrichaceae, Bacteroidetes VC2.1 Bac22, and Luteibacter, were significantly enriched in areas where S. europaea exhibited vigorous growth. Among these, Luteibacter was consistently identified as the unique and dominant genus in root samples from the most favorable habitats. Inoculation experiments further demonstrated that L. aegosomatis and L. jiangsuensis significantly enhanced S. europaea growth, with shoot length, fresh weight, and dry weight being 1.4- and 1.2-, 2.1- and 1.6-, and 1.9- and 1.5-fold higher, respectively, than those of non-inoculated controls. These findings highlight that soil physicochemical conditions modulate root-associated bacterial assemblages and that a unique and dominant root-associated genus in favorable habitats may include strains capable of promoting S. europaea growth. Our results provide new ecological insights into plant–microbe–soil interactions and offer a promising framework for in situ conservation and restoration of rare halophytic species in brackish ecosystems.

Halophytes are critical for maintaining coastal ecosystem biodiversity (1). As outlined in technical notes from the Convention on Biological Diversity (2), biodiversity is essential for the sustainable development of various economic activities, particularly those related to agriculture, forestry, and fisheries. Many halophytic species, including Salicornia veneta, Limonium legrandii, and Atriplex lanfrancoi, are presently classified as vulnerable or endangered on the IUCN Red List (3). In Japan, the halophytes Salicornia europaea, Suaeda japonica, and Triglochin palustris are listed as vulnerable in the fifth edition of the Japanese Red List published in 2025 by the Ministry of Environment (4).

The decline in number of halophytic species is largely driven by anthropogenic pressures. For example, Limonium daveaui, endemic to Lusitania, was once widespread along the west coast of Portugal but is now restricted to the Tagus Estuary (5). Conservation of these endangered halophytes is vital for preserving coastal biodiversity. Enhancing survival through reintroduction and promotion of population and genetic diversity among native halophytes is essential (6). Reintroduction of local species has gained recognition as a critical conservation issue in international treaties and legislation (7–9). However, conserving halophytes on clifftops, rocky shores, sandy beaches, and salt marshes poses significant challenges because of their habitat specialization and vulnerability to disturbances from invasive species and human activities (10, 11).

Key breakthroughs in successful reintroduction of rare species could lie in detailed understanding of their microenvironmental requirements, including soil characteristics (12, 13). For instance, research on the halophyte Limonium multiflorum, which is endemic to the west coast of Portugal and listed on the IUCN Red List, suggests that in situ conservation in local “hotspots” is crucial for sustaining biodiversity (10). Studies focusing on soil and host-associated microbes have recently gained attention in conservation biology. Walling et al. (14) reported that integrating microbial research on both host-associated and free-living microbes into conservation biology represents an essential step forward for conserving charismatic organisms and maintaining ecosystem functions under ongoing global environmental change. They also emphasized that identifying the minimal set of functional microbial groups required to support key processes in compromised hosts or ecosystems should be considered a critical first step in guiding conservation efforts. Understanding key features of microenvironments within specific habitats is thus vital for the conservation of rare and vulnerable species.

Several studies have characterized microenvironments in native habitats of halophytes, focusing on soil physicochemical properties and rhizosphere and root endosphere microbiota compositions. For example, the diversity and composition of the root endophyte community associated with Suaeda salsa are significantly influenced by soil salinity (15). Another study found that the bacterial communities associated with the stems and roots of Arthrocaulon macrostachyum differed markedly between natural salt marshes and artificial sea salt farms, with differences attributed to variations in soil salinity (16). The distinct origins of salinity at the two sites, one naturally saline and the other anthropogenic, impacted bacterial composition associated with S. europaea (17, 18). To date, these studies have typically focused on two to three saline environments rather than on a single local habitat, which may not fully capture the complexities introduced by the sum of environmental factors.

In this study, Salicornia europaea was selected as an endangered plant species that has been confirmed to grow only in Hokkaido, Kagawa and Okayama Prefectures in Japan according to the fifth edition of the Japanese Red List (4) published by the Ministry of the Environment. To minimize variations caused by climatic conditions, salinity sources, and human activities, we established an environmental model by selecting three distinct zones within Lake Notoro, a brackish lake, each characterized by unique soil textures and distinct S. europaea growth patterns. The objective of this study was to elucidate the relationships among soil bacterial communities, edaphic factors, and S. europaea plant growth. We hypothesized that (i) the three habitats differing in the growth performance of S. europaea harbor distinct soil physicochemical characteristics, (ii) these edaphic variations drive differences in bacterial diversity and community structure associated with S. europaea, and (iii) a unique and dominant root-associated genus in favorable habitats may include strains capable of enhancing the growth of S. europaea.

By linking edaphic properties and soil bacterial community structure with plant growth and evaluating the growth-promoting potential of a candidate bacterial genus, this study provides insights that may contribute to future efforts in soil microbiome conservation, restoration, and management for endangered halophytic plant species.

Plant and soil sample collection

Naturally growing S. europaea plants were procured from three distinct locations: Misaki (44°6′5″N/144°11′37″E), Lakeside Park (44°2′51″N/144°11′20″E), and Ubaranai (44°0′50″N/144°6′59″E) on the shore of Lake Notoro in Hokkaido, Japan (Figure 1). To account for variations in samples from native habitats for Salicornia plants, we collected samples from the above-described areas within Lake Notoro over two consecutive years: August 23, 2021, and August 23, 2022. This research was conducted with the approval of Hokkaido Prefecture. Sampling was conducted according to methodology described by Yamamoto et al. (19). For 16S rRNA metagenomic analysis, whole plants were excavated with their surrounding soil in blocks measuring approximately 20 cm in length, 20 cm in width, and 10–20 cm in depth. Bulk soil samples were collected as described above for soil physicochemical and 16S metagenomic analyses. The plants and their soil cores were transported to the laboratory and the root systems were sampled within 12 h of removing the plants from their natural habitat. Twenty individual plants were pooled into four samples, each containing five plants. Rhizosphere and root compartments were separated from the samples for bacterial community profiling.

S. europaea growth parameters

To measure plant density, quadrats measuring 50 cm × 50 cm were randomly positioned within the habitat of S. europaea at each sampling site, as described by Shipley et al. (20), with four replicates conducted each year. Randomly selected plants were harvested from the soil blocks to measure shoot length, fresh weight, and dry weight. Shoot length was measured using Vernier calipers. The aboveground organs were dried at 80 °C for three days to obtain dry weights.

16S metagenomic analysis

Sample preparation

Rhizospheres and root endospheres were isolated as described previously (19). Roots were collected and segmented into 5.0 cm lengths, with an initial incision made 0.5 cm below the root base. Segments were placed in 50 mL sterile tubes containing 40 mL of PBS-S buffer (130 mM NaCl, 7 mM Na 2 HPO 4, 3 mM NaH 2 PO 4, pH 7.0, 0.02% Silwet L-77), then washed with shaking at 180 rpm for 20 min. Roots were subsequently transferred to new sterile tubes, and soil suspensions were centrifuged at 4,000 × g for 20 min. The resultant pellets, representing rhizospheres (soil-root interfaces), were immediately frozen in liquid nitrogen and stored at −80 °C.

To enrich for endophytic bacteria, the roots underwent an additional washing procedure, followed by transfer to fresh sterile tubes containing 40 mL of PBS-S buffer. Samples were sonicated for 10 min in a water bath sonicator (Model 5510, Branson Ultrasonics Corporation, Danbury, CT, USA) at 40 kHz. Roots were then washed again in fresh PBS-S buffer, and surface moisture was removed using Kimtowel ® (Nippon Paper Crecia Co., Ltd., Tokyo). The cleaned roots were frozen in liquid nitrogen and stored at −80 °C. These tissues were defined as root endosphere in accordance with the method of Yamamoto et al. (19).

Following the harvest of all plant tissue from the soil block, bulk soil samples were collected at a depth of 0.5 to 5.5 cm beneath the soil surface, corresponding to a root length of 5 cm. These samples were also frozen in liquid nitrogen and stored at −80 °C. DNA was extracted from the rhizosphere (Rh), root endosphere (R), and bulk control soil (Bl) samples for further analyses.

DNA extraction, PCR amplification, and sequencing

Root samples were finely ground using a sterilized mortar and pestle. The resulting powders were subsequently cooled in liquid nitrogen. DNA was extracted from rhizosphere (Rh), bulk soil (Bl), and powdered root samples (0.5 g). Total DNA was extracted using a NucleoSpin Soil Kit (Macherey-Nagel, Düren, Germany), which included buffer SL1 and enhancer SX, which are known for their efficacy in obtaining high-quality and high-quantity DNA from paddy soils (21). Extracted DNA concentrations were assessed using a Qubit dsDNA HS assay kit (Thermo Fisher Scientific, USA), following the manufacturer’s protocol. To minimize amplification of chloroplast and mitochondrial 16S rRNA genes during PCR, we targeted and amplified the V5–V7 region using specific primer pairs: 799F (5′- TCGTCGGCAGCGTCAGATGTGTATAAGAGACAG AACMGGATTAGATACCCKG-3’) and 1193R (5’- GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAG ACGTCATCCCCACCTTCC-3’) (19). These primers include overhang adapters for next-generation Illumina sequencing (underlined regions). Our PCR amplicon library was prepared according to the Illumina 16S Sample Preparation Guide (15044223-b; Illumina Inc., San Diego, CA, USA) excluding the index assignment step. A unique index assignment was subsequently performed using IDT for Illumina DNA/RNA UD indices (Illumina, Inc.). Library quality was assessed using an Agilent 4200 TapeStation (Agilent Technologies, Inc.) and sequenced as paired-end reads of 300 bp on an Illumina NextSeq 1000 sequencing platform at the NODAI Genome Research Center at Tokyo University of Agriculture.

Processing and analysis of obtained sequences

Obtained sequences were analyzed following the previous methods (19).

Bioinformatics and statistical analyses

Phylogenetic community structures were analyzed using R version 3.4.3, employing phyloseq (22) and pheatmap (23) packages, with subsequent normalization after filtering chloroplast and mitochondrial ASVs. Differences in diversity indices, including alpha diversity measures (richness, Shannon indices, and Simpson indices), were evaluated using one-way analysis of variance (ANOVA). Non-metric multidimensional scaling (NMDS), based on phylogenetic β-diversity metrics (unweighted and weighted UniFrac distances), was applied to assess interactions within and between bacterial communities. The LEfSe submodule within MicrobiomeAnalyst (https://www.microbiomeanalyst.ca) was utilized to determine effect sizes, using a false discovery rate (FDR)-adjusted p-value threshold of 0.05 and a log LDA score cutoff of 2.0. Statistical analyses of plant density, growth parameters, and alpha diversity data were performed using one-way ANOVA with the Tukey-Kramer test (α = 0.05).

Bacterial inoculation analysis

S. europaea seeds and bacterial type strains for inoculation assays

Seeds of Salicornia europaea were collected from naturally occurring populations in Lake Notoro, Hokkaido, Japan. Two strains, Luteibacter aegosomaticola JCM 35472 and Luteibacter aegosomatis JCM 35471, were obtained from the Japan Collection of Microorganisms (JCM), RIKEN BioResource Research Center, Japan. Three additional strains, Luteibacter jiangsuensis DSM 22396, Luteibacter yeojuensis DSM 17673, and Luteibacter rhizovicinus DSM 16549, were acquired from the Deutsche Sammlung von Mikroorganismen und Zellkulturen (DSMZ), Germany.

Preparation of five Luteibacter type strain inocula

For inoculum preparation, five Luteibacter type strains were streaked onto Luria–Bertani (LB) agar plates and incubated at 30 °C for three days. Individual colonies were then cultured in LB broth at 30 °C with shaking at 180 rpm for three days. Prior to inoculation of S. europaea seeds, each culture was adjusted to an optical density at 600 nm (OD 600) of 0.6 using LB broth and subsequently re-cultured at 30 °C with shaking at 180 rpm for 3 h. To remove residual LB components, the bacterial cells were harvested by centrifugation at 6,000 × g for 10 min, supernatants were discarded, and pellets were resuspended in 10 mM MgSO 4. This washing step was repeated thrice. The resulting bacterial suspensions were used for S. europaea seed inoculations.

Seed surface sterilization, germination, and bacterial inoculation of seeds and roots

S. europaea seeds were soaked in 70% (v/v) ethanol for 2 min, surface-sterilized with 3% (v/v) sodium hypochlorite for 7 min, then rinsed ten times with sterile Elix water. After surface sterilization, seeds were subjected to vernalization at 4 °C for one week. Vernalized seeds were transferred to tissue culture containers (72 mm × 72 mm × 100 mm; SPL Life Sciences, Korea) containing Murashige and Skoog (MS) medium supplemented with 2% sucrose and 0.8% agar. Each container was added to 1 mL of a prepared Luteibacter type strain suspension for bacterial inoculation and incubated for 14 days at 22 °C under a 16 h light/8 h dark photoperiod. Seedlings were subsequently transferred to MS medium supplemented with 1% NaCl, 2% sucrose, and 0.8% agar; re-inoculated with 1 mL of their respective Luteibacter suspension, and incubated for an additional 28 d under the same conditions. We used MS medium supplemented with 1% NaCl rather than 3% NaCl, which was identified as optimal for S. europaea growth in our previous study (28). Under the aseptic MS medium conditions used in the present study, 1% NaCl was considered suitable for evaluating the effects of bacterial inoculation, whereas 3% NaCl was optimal under the soil conditions used in our previous study (28).

Forty-two days after seeding, inoculated plants were collected to measure fresh weight, dry weight, shoot length, and root length. Photographs of the collected plants were taken prior to fresh weight measurements. Samples were then dried at 80 °C for three days to determine dry weights. Shoot and root lengths were quantified using the ImageJ software (https://imagej.net/ij/). Root samples were used for bacterial re-isolation and observation of bacterial colonization on root surfaces using confocal laser scanning microscopy, as described in the following section.

Re-isolation of inoculated bacteria

Root samples from ten inoculated individual plants were pooled and washed with 35 mL of sterile PBS (130 mM NaCl, 7 mM Na 2 HPO 4, 3 mM NaH 2 PO 4, pH 7.0). Pooled roots were then homogenized using a Biomasher II (Nippi, Japan) with 250 µL of sterile PBS, and tenfold serial dilutions of homogenate were spread onto LB agar plates. The plates were incubated at 25 °C for seven days. Individual colonies were selected and re-streaked onto LB agar plates for plants inoculated with L. aegosomatis or L. jiangsuensis. Genomic DNA was extracted using a NucleoSpin Microbial DNA Kit (MACHEREY–NAGEL, Germany) after cultivation on LB agar plates. The 16S rRNA gene was amplified using the universal primer pair 16SrRNA9F and 16SrRNA1541R using Q5 DNA polymerase (New England Biolabs, USA). PCR products were purified using a NucleoSpin Gel and PCR Clean-up kit (MACHEREY–NAGEL, Germany). Purified PCR products were sequenced by FASMAC Co., Ltd. (Kanagawa, Japan) using universal primers (Supplementary Table S1).

Confocal scanning laser microscopy analyses

L. aegosomatis and L. jiangsuensis colonization patterns in 42-day-old inoculated S. europaea roots were examined using a confocal laser scanning microscope (A1R 1 Ni-E with NIS Elements AR software; Nikon, Tokyo, Japan) equipped with a LU-N4/N4S 4-laser unit and a Nikon CFI Apo Lambda S 40XC WI objective. Freshly collected roots were washed twice with sterile PBS and stained using a LIVE/DEAD BacLight Bacterial Viability Kit (Invitrogen, USA), according to the manufacturer’s instructions, to visualize localization of inoculated bacteria on root surfaces. Stained root samples were mounted in PBS under glass coverslips and excited with a 408.9-nm laser for root autofluorescence, a 489.1-nm laser for SYTO 9 fluorescence, and a 561.6-nm laser for propidium iodide fluorescence. A DM405/488/561/640 dichroic mirror and a BA570–620 emission filter were used for observation. Pinhole diameter was set at 20.4 µm.

Physicochemical analyses of soils

For each site and sampling year, four independent soil samples were analyzed. Soil samples for physicochemical analyses were air-dried (≤2 mm particle size) after collection from each plot. Particle size distribution was determined using the pipette method (24). Air-dried samples were treated with hydrogen peroxide to remove organic matter, followed by ultrasonic dispersion. A suspension was then prepared by adding sodium hexametaphosphate solution. Particle mass in each particle size fraction was measured, and fraction proportion was calculated accordingly. Soil pH (H 2 O) and electrical conductivity (EC) were measured by adding distilled water to air-dried soil samples at respective solid-to-liquid ratios of 1:2.5 and 1:5, followed by shaking for 1 h. Total organic carbon (TOC) and total organic nitrogen (TON) were determined by dry combustion using a CN analyzer (SUMIGRAPH NC-220F, SCAS, Japan) after finely grinding the air-dried soil to pass through a 0.5 mm sieve. Water-soluble base cations were quantified via sequential extraction. Briefly, 2 g of air-dried soil was mixed with 30 mL of distilled water and shaken for 1 h. Supernatants were then collected. This extraction was repeated thrice. The combined extract was diluted to a final volume of 100 mL, and Ca 2+, Mg 2+, K +, and Na + concentrations were determined by inductively coupled plasma–atomic emission spectrometry (ICP-AES; ICPS-8100, Shimadzu, Japan).

Growth parameters of S. europaea in three areas within Lake Notoro

The growth parameters of S. europaea in the three areas within Lake Notoro showed similar patterns in 2021 and 2022 (Figure 2; Supplementary Tables S2, S3). Shoot length, fresh and dry weight, and number of primary branches were highest in S. europaea from Lakeside Park, followed by Ubaranai and Misaki. Plant densities of S. europaea from both Ubaranai and Misaki were significantly higher than those from Lakeside Park. No significant plant density differences were observed between S. europaea specimens from Ubaranai and Misaki. In this study, among the three areas selected within Lake Notoro, S. europaea exhibited optimal growth in the Lakeside Park region.

Physicochemical soil parameters associated with S. europaea from each area within Lake Notoro

Soil physicochemical parameters associated with Salicornia europaea in each region of Lake Notoro exhibited consistent patterns in 2021 and 2022 (Table 1). According to USDA soil classification, the textural classes of soils in the three regions of Lake Notoro were sand in Misaki, sandy loam in Ubaranai, and sandy clay in Lakeside Park, indicating that the soil in Lakeside Park had the highest fertility, followed by that in Ubaranai. Total organic carbon (TOC), total organic nitrogen (TON), and carbon-to-nitrogen ratio (C/N) in the soils from Lakeside Park and Ubaranai were higher than those from Misaki. Although the mean TON concentration in Lakeside Park was approximately twofold higher than those in Ubaranai and Misaki in 2022, the relatively large variability among Lakeside Park samples may explain why no statistically significant differences were detected among the areas. Water-soluble sodium (Na +) and magnesium (Mg 2+) ion concentrations were highest in soil from Lakeside Park, compared with those from Ubaranai and Misaki, consistent with their electrical conductivity (EC) values. Soil pH measurements from the three sampling regions indicated that soils from Lakeside Park and Ubaranai were slightly acidic, whereas those from Misaki are alkaline. These results suggest differences in edaphic factors among the three sampling sites.

General analyses of sequencing data

A total of 32,069,328 high-quality non-chimeric sequence reads were obtained following filtration and normalization of 61,534,594 raw reads from NextSeq 1000 sequencing of 72 samples (Supplementary Table S4). In total, 331,110 ASVs were identified, with ASVs per sample ranging from 1,537 to 9,953 (Supplementary Table S4). As shown in Supplementary Figure S1, the alpha rarefaction curves for the ASVs in each sample approached saturation, indicating that most bacterial groups were captured.

Comparative analysis of bacterial diversity and community structure in S. europaea in three areas within Lake Notoro

Alpha diversity metrics indicated higher bacterial diversity in bulk soil (Bl) and rhizosphere (Rh) than in roots (R) across all sampling sites within Lake Notoro. Species richness and Shannon indices specifically revealed the highest bacterial diversity in Bl at Misaki, whereas Simpson indices demonstrated that the lowest bacterial diversity occurred in R at the same location (Figure 3; Supplementary Table S5).

NMDS and PERMANOVA, calculated using the unweighted UniFrac distance, indicated a clear separation between bacterial communities among sampling locations (R 2 = 0.163, Pr(>F) < 0.001), sample source (R 2 = 0.086, Pr(>F) < 0.001), and sampling year (R 2 = 0.081, Pr(>F) < 0.001) (Figure 4A; Supplementary Table S6). Similar results were observed for NMDS based on weighted UniFrac distances, with sampling location (R 2 = 0.245, Pr(>F) < 0.001), sample source (R 2 = 0.299, Pr(>F) < 0.001), and sampling year (R 2 = 0.007, Pr(>F) < 0.001) being significant factors distinguishing bacterial communities across the three sampling areas (Figure 4B; Supplementary Table S6). PERMANOVA indicated that sampling location was the most significant factor in separating bacterial communities.

Physicochemical variables significantly impacting community structure included soil particle size, EC, pH, TOC, TON, C/N, and water-soluble Ca 2+ concentration (Figures 4A, B; Supplementary Table S7). The bacterial community in Misaki was influenced by sandy and slightly alkaline soil and elevated Ca 2+ concentrations, whereas the communities at Ubaranai and Lakeside Park were shaped by high TOC and TON, elevated EC, and high clay content, indicating that the higher fertility and salinity of soils at Ubaranai and Lakeside Park contributed to the clear separation of their bacterial communities from those at Misaki.

Structural classification of genus-level bacterial communities

Observational analysis of the top 50 identified genera revealed that the bacterial genera were classified and clustered under 13 bacterial phyla based on a heatmap analysis (Figure 5; Supplementary Table S8). Ten genera (Bacteroidetes BD2-2, SEEP-SRB1, Bacteroidetes VC2.1 Bac22, Candidatus Thiobios, BD2–11 terrestrial group, Calditrichaceae, Truepera, Luteibacter, Winogradskyella, and Stappiaceae) were highly abundant (>2% relative abundance) in the samples from Lakeside Park. Among these ten genera, Luteibacter, Winogradskyella, and Stappiaceae were specifically enriched in R samples from Lakeside Park. In contrast, B2M28, Desulfuromonas, and Gillisia were more abundant in the Ubaranai samples, with Gillisia increasing in Rh samples. In Misaki, Bacillus, BIrii41, Woeseia, Acanthopleuribacter, Defluviicoccus, Ilumatobacter, and Phaeodactylibacter were highly abundant, with BIrii41, Acanthopleuribacter, Defluviicoccus, and Ilumatobacter specifically attracting Rh, R, or both.

Statistically significant microbial communities at genus level among three sampling areas within a sample type

Based on LEfSe analysis, the top 10 unique genera (LDA score > 2.0, and FDR-adjusted P -value < 0.05) were identified as specialized communities among the three sampling areas within each sample type (Figure 6). LEfSe analysis identified several genera unique to the Lakeside Park site, where S. europaea exhibited its highest growth parameters in Lake Notoro; among these unique genera, we focused on the genera shown in Figure 5. Five genera (Bacteroidetes BD2-2, Candidatus Thiobios, SEEP-SRB1, Calditrichaceae, and Bacteroidetes VC2.1 Bac22) were significantly enriched in the Bl samples from Lakeside Park. Luteibacter and Bacteroidetes BD2–2 were also notably elevated in R samples from Lakeside Park. Bacteroidetes BD2–2 was consistently identified as a unique genus in the Bl and R samples. Similar results were obtained for the relative abundances of these six genera, as shown in Supplementary Figure S2.

Effect of inoculation with Luteibacter type strains on S. europaea growth

For inoculation assays, we focused on Luteibacter, a unique and dominant root-associated genus in favorable habitats (Lakeside Park) because its type strains are readily accessible. Inoculation with L. aegosomatis achieved the greatest promotion of S. europaea growth across all measured parameters (Figures 7A–D). Plants inoculated with L. jiangsuensis exhibited significantly enhanced growth as assessed by fresh weight, dry weight, and shoot length (Figures 7A–C). Roots inoculated with L. aegosomatis exhibited extended primary root elongation and increased lateral root emergence (Figure 7E). To assess bacterial colonization persistence, we re-isolated the inoculated strains from S. europaea roots. Both L. aegosomatis and L. jiangsuensis were successfully re-isolated from roots inoculated with their respective type strains (Figure 7E, right panels) with 16S rDNA sequences identical to the original strains. To determine whether the inoculated bacteria established viable populations on root surfaces, we visualized the presence of L. aegosomatis and L. jiangsuensis on inoculated roots. A small number of living cells of both L. aegosomatis and L. jiangsuensis were observed on root surfaces, despite a predominance of dead cells. (Figure 7F).

Among the three areas examined within Lake Notoro, Lakeside Park was characterized by the most favorable overall growth performance of S. europaea, with the highest shoot length, fresh and dry weights, and number of primary branches, whereas plant density was lowest at this site. This inverse density–size pattern may partly reflect differences in intraspecific competition, with lower plant density potentially allowing greater access to available resources and promoting individual plant growth. Density-dependent self-thinning may also contribute to this pattern, although the present observational data do not allow these mechanisms to be distinguished.

S. europaea exhibited highest growth at Lakeside Park, where the soil was classified as sandy clay, compared to Ubaranai and Misaki, which respectively possess sandy loam and sandy soils. Sandy soils, characterized by large pore spaces, facilitate water drainage but are typically low in nutrients. Clay-rich soils retain more water and provide essential nutrients for plant growth (25). The fine particles, extensive surface area, and negative charge of clay minerals are particularly important for enhancing soil fertility and influencing soil physicochemical properties (26), corroborated by research in which sandy clay loam was shown to provide better growth conditions for maize than loamy sand (27). These findings suggest that S. europaea thrives better in clay-rich soils such as those at Lakeside Park, consistent with the growth preferences of crop species. EC values and concentrations of water-soluble Na + and Mg 2+ were highest in Lakeside Park among the three sampling areas. The highest amounts of carbon and nitrogen were also found in the soil at Lakeside Park. A previous study showed that the poorest S. europaea growth was observed in the absence of salt, whereas the best growth was observed in the presence of 3% NaCl (28). Similar results were also obtained for S. brachiata: plant growth and biomass yield clearly increased with increasing salinity (29). Ramasubramanian and Uyyavanthan (30) reported that S. brachiata action on supplemental fertilizers played a vital role in restoring soil fertility, particularly levels of available nitrogen and organic carbon in saline soils. Taken together, soil texture, soil salinity, and inherent soil nutrient levels significantly impact S. europaea performance under natural saline conditions. Several studies have explored the relationship between soil pH and Salicornia plants growth. Favorable productivity has been observed across various pH ranges in nutrient solutions, including 5.5–5.8 (31, 32), and above 8.0 (33, 34). S. europaea exhibited no significant changes in growth and development spanning a broad pH spectrum (5.29–8.52) (35). Differences in soil pH among the three sampling sites in Lake Notoro are thus unlikely to fully account for the observed variation in plant growth in this study.

Across the three sampling sites in Lake Notoro, bacterial diversity was higher in Bl and Rh than in R. Previous studies have shown that microbial density is typically greater in rhizospheres than in roots, with bacterial diversity and richness progressively decreasing from soil to root endosphere (19, 36, 37). This decline is driven by host-specific factors, including genotype and immune responses, which filter the microbial community from a diverse, soil-derived microbiome to yield a more specialized, host-adapted population (37). In the root endosphere, bacterial diversity at Lakeside Park and Ubaranai was higher than that at Misaki. Plant-associated microbial diversity is a crucial factor for plant health. The diversity of soil microbial communities plays a pivotal role in maintaining plant health as it is closely associated with microbial balance and resistance to pathogen invasion (38). Accordingly, the higher bacterial diversity observed in root endospheres at Lakeside Park and Ubaranai may contribute to improved S. europaea root health, thereby promoting overall plant growth.

The bacterial community structure differed significantly among the three sampling areas, reflecting variations in soil textural classes, EC, pH, TOC, TON, C/N ratio, and water-soluble Ca 2+ concentrations. Strong correlations between microbial communities and nutrients, including soil carbon or nitrogen content and phosphorus content, and other characteristics, including soil EC, have been identified in studies across various scales (39–41). Szymańska et al. (17) reported similar findings, demonstrating that the distinct endophytic bacterial communities of Salicornia europaea between a natural habitat and an anthropogenic site (meadows near a soda factory) correlated with soil properties including CaCO 3, saturation percentage, EC, Na +, K +, Mg 2+, Cl −, SO 4 2−, HCO 3−, TOC, TON, Ca 2+, and C/N ratio. In this study, we selected three sampling sites within a natural habitat that shared a common saline origin. Na +, K +, and Mg 2+ concentrations were not significantly associated with differences in bacterial community structure (Supplementary Table S7). Instead, our data highlight edaphic soil textural classes and pH as additional critical factors influencing the relationship between the S. europaea bacterial community and edaphic conditions in native habitats. Seaton et al. (42) demonstrated that soil textural heterogeneity promotes bacterial diversity, and that textural composition significantly influences bacterial community structure, with specific microbial taxa showing associations with clay and fine silt particles. In our study, the three sampling areas share silt and clay but in significantly different proportions. Differences in bacterial community structure among the three sites is primarily associated with the ratio of clay and silt particles.

Soil pH plays an important role in shaping bacterial communities in saline soils that provide habitats for halophytes, including S. europaea, in Ebinur Lake, China (43), consistent with the findings of this study. Our study also reveals that soil pH influences bacterial communities in bulk soil and those associated with S. europaea rhizospheres and roots. Although several studies have reported that soil pH does not affect Salicornia plant growth, as described above, pH-induced shifts in bacterial communities within saline soils possibly influence plant growth. Based on heatmap and LEfSe analyses at the genus level, unique genera were identified at each sampling site. Focusing on the unique genera in the Lakeside Park samples, where S. europaea exhibited the highest growth parameters within Lake Notoro, significant enrichment of Bacteroidetes BD2-2, Candidatus Thiobios, SEEP-SRB1, Calditrichaceae, and Bacteroidetes VC2.1 Bac22 was observed in Bl samples. Bacteroidetes BD2–2 has been identified as an important degrader of complex organic materials such as rhizodeposits and algal cell wall polysaccharides (44). Bacteroidetes VC2.1 Bac22 has demonstrated the ability to utilize polysaccharides, protein polymers, and fatty acids, while flexibly acquiring energy through NO/N 2 O reduction and polysulfide reduction. These metabolic capabilities suggest a role in mineralization of complex organic carbons and N 2 O sink function in deep-sea hydrothermal vent ecosystems (45). Furthermore, the SEEP-SRB1 group, a lineage of uncultured bacteria, has been identified as a major sulfate-reducing group in Lake Harutori, in a residential area in Kushiro, Japan (46). Sulfate reduction, a key pathway mediated by dissimilatory sulfate-reducing bacteria, plays a critical role in organic carbon mineralization in estuarine and coastal sediments (47). Bacteroidetes BD2-2, Bacteroidetes VC2.1, Bac22, and SEEP-SRB1 may thus contribute to organic carbon mineralization in the bulk soil region of Lakeside Park, in keeping with the data from Table 1, wherein Bl in Lakeside Park had higher TOC and TON than those observed in Misaki. This could indirectly support Salicornia plant growth by enhancing nutrient availability in this environment.

Candidatus Thiobios zoothamnicoli, within the genus Candidatus Thiobios, is known to form symbiotic relationships with sessile ciliates such as Zoothamnium niveum in sulfur-rich environments. This bacterium fixes carbon under both oxic and sulfidic conditions, transferring it to host cells for support (48). The phylum Calditrichota, within the family Calditrichaceae, is an anaerobic organoheterotroph that is abundant in marine sediments (49). Candidatus Thiobios and Calditrichaceae may thus be of little importance for S. europaea growth in Lakeside Park.

Root samples collected from Lakeside Park showed significant Luteibacter enrichment. Previous research by Guglielmetti et al. (50) identified Luteibacter rhizovicinus MIMR1 as a promising candidate for biofertilizer applications because this strain can chelate ferric ions, solubilize monocalcium phosphate, and produce indole-3-acetic acid. L. rhizovicinus MIMR1 has also been shown to localize to root tips and along the rhizoplane of barley, where it markedly enhances aerial biomass, root weight, and root length. In our study, members of the genus Luteibacter were abundant in the root endosphere and rhizosphere of S. europaea collected from the Lakeside Park area of Lake Notoro and from Lake Tofutsu (43°57′04″N/144°21′40″E) (19, 51). These findings suggest that Luteibacter enriched in the Lakeside Park area, may similarly contribute to S. europaea growth promotion.

To test this hypothesis, five Luteibacter type strains were inoculated into S. europaea grown on MS medium under aseptic conditions, and the relationship between each Luteibacter strain and the host plant was evaluated on a one-to-one basis. Inoculation with L. aegosomatis or L. jiangsuensis most strongly promoted S. europaea growth. Both strains were successfully re-isolated from the roots of inoculated plants, although dead cells were more prominently detected on root surfaces. L. aegosomatis and L. jiangsuensis were originally isolated from the intestines of Aegosoma sinicum larvae and soil samples collected at the Dongjing Pesticide Factory in China (52, 53). Notably, neither species has previously been reported as a plant growth-promoting bacterium. In contrast, inoculation with L. rhizovicinus, previously shown to promote barley growth (50), exerted no significant effect on S. europaea growth. These findings suggest that the unique and dominant root-associated genus Luteibacter associated with S. europaea in the Lakeside Park area of Lake Notoro may include strains capable of promoting the growth of the host plant. Future studies should isolate and characterize the S. europaea –associated Luteibacter strains responsible for plant growth promotion and evaluate their effects under field conditions.

From the perspective of soil microbiome engineering, the present study goes beyond a descriptive comparison of bacterial communities among habitats by linking site-specific edaphic properties and root-associated bacterial community structures with plant growth and subsequently examining the growth-promoting potential of a candidate bacterial genus. These findings provide ecological insights that may help inform the future development of microbiome-based conservation strategies. Such approaches may involve not only the selection and inoculation of beneficial bacterial strains but also the management of soil physicochemical conditions that support their establishment and function, providing a potential framework for microbiome-based conservation, restoration, and management of halophytic habitats.

We report here that distinctive soil physicochemical properties, characteristic bacterial community structures, and the presence of a unique and dominant root-associated bacterium in areas supporting the vigorous growth of rare and vulnerable halophytic species may represent critical environmental determinants for in situ conservation of these plants.

Walling et al. (14) emphasized that microbial genomics should move beyond the hypothesis-generating stage of descriptive surveys toward hypothesis-testing approaches that directly evaluate microbial functions, thereby providing clearer insights into their ecological roles and practical implications for conservation. In line with this concept, we have proposed three hypotheses: (i) habitats differing in S. europaea growth performance possess distinct soil physicochemical properties; (ii) these edaphic variations shape the bacterial diversity and community structures associated with S. europaea; and (iii) a unique and dominant root-associated genus in favorable habitats contributes to enhanced plant growth. Hypotheses (i) and (ii) were supported by soil physicochemical analyses and 16S rRNA gene-based community profiling. 16S rRNA gene-based community profiling identified the unique and dominant root-associated genus Luteibacter. Hypothesis (iii) was partially supported by inoculation assays using five Luteibacter type strains under aseptic conditions in bench-scale laboratory experiments. This study represents a next-generation experimental approach that bridges descriptive microbiome research with functional verification, thereby offering new ecological insights and practical directions for conservation of rare halophytic species. Nevertheless, the field survey was conducted at only three sites over two years, which represents a limitation of this study, and therefore provides correlative rather than direct causal evidence for the relationships among edaphic factors, bacterial communities, and plant growth. In addition, the inoculation assays were performed under axenic MS-agar conditions rather than in soil. Therefore, it remains unclear whether similar growth-promoting effects would be observed when S. europaea is cultivated in sterile sand or soil, where the physical and chemical conditions differ substantially from those of MS-agar. Further studies are needed to determine whether the growth-promoting effects observed under axenic MS-agar conditions can also be reproduced in soil and under field conditions.