Samples were collected in the northeast of Shantou, Guangdong, China (longitude: 23°55′9″N, latitude: 117°25′9″E; altitude: 0 m) at the coastal zone of Yunxiao. The rhizosphere samples were collected by selecting three zones, labeled R1, R2, and R3, extending from the intertidal zone to the deeper areas of the mangrove forest. Rhizosphere sediments were defined as those directly adhering to the roots of mangrove plants, while non-rhizospheres were those collected 3 m away from the mangrove roots (Li et al., 2016). Polyvinyl chloride pipes of 4.2 cm in diameter and 50 cm in length were used to collect sediments at depths ranging from 0 to 20 cm, as described by Allaouia et al. (2022). Depths were designated as follows: Ni-1 (0–5 cm), Ni-2 (5–10 cm), Ni-3 (10–15 cm), and Ni-4 (15–20 cm), where N represents either R (rhizosphere) or NR (non-rhizosphere). After sediments were collected, basic treatments, such as removing small pebbles, stones, and roots, were performed on-site. Samples were then transported to the laboratory for biological and molecular analyses. Each sample was divided into two groups: one stored at 4 °C for determining the physical and chemical characteristics of the sediments, and the other stored at −20 °C for DNA analysis before DNA extraction.

Once the samples were collected, a hand-held thermometer (model no. LBT-4S, Xi’an Lonn M&E Equipment Co., Ltd.), a pH meter (Benchtop pH meter inoLab® pH 7,110 Set 4), and an oxidation–reduction potential (ORP) meter (PINPOINT® ORP Monitor) were used to measure temperature, pH, and ORP values at different depths. To determine the physical and chemical characteristics of rhizosphere and non-rhizosphere sediments, they were air-dried, crushed, and sieved through a 2-mm mesh. Next, 0.5 g of the crushed sediment sample was added to an Erlenmeyer flask and digested using the aqua regia extraction method (Victório et al., 2020). After digestion, the sample was diluted to 50 mL with deionized water and filtered through filter papers. The microelements and nutrients such as iron (Fe), manganese (Mn), zinc (Zn), magnesium (Mg), potassium (K), and calcium (Ca) were analyzed by inductively coupled plasma optical emission spectrometry (ICPOES) (Allaouia et al., 2022; Asnat et al., 2022; Elyamine et al., 2023). In contrast, other nutrients, such as total nitrogen (TN), nitrate (NO₃-), and nitrite (NO₂-), were determined using the Kjeldahl method as described in Konopiński (2020). A double digestion with H₂SO₄/HClO4 was used to analyze the phosphorus content. Carbon and sulfur were determined by dry combustion using a high-temperature induction furnace as described in Lavkulich et al. (1970).

The method used to design MSM and to determine the effect of pyrene stress during the acclimation was inspired by Hui et al. (2018), with a slight personal modification. The details of the method and the composition of the mineral salt culture medium can be found in our previous study (Elyamine et al., 2023). Briefly, 5 g of sediment was added to 50 mL of MSM containing 50 mg/L of pyrene, and the mixture was incubated in a shaker at 25 °C and 150 rpm for 30 days. Each treatment was performed in three replicates to ensure statistical validity and scientific rigor. Every 30 days, 5 mL of the old sample was transferred to a new MSM, and genomic DNA from each old individual was extracted and stored at −80 °C for optimal preservation until further processing.

The PowerSoil® DNA isolation kit (MoBio Laboratories, Carlsbad, CA, USA) was used to extract total genomic DNA from the different samples before and after acclimation. PCR amplification of 16S rRNA genes from the V3-V4 region of the samples was performed using the same method described in our previous study (Allaouia et al., 2022; Asnat et al., 2022; Elyamine et al., 2023) using the 338F universal primers (5’-ACTCCTACGGGAGGCAGCAG-3′) and 806R (5’-GGACTACHVGGGTWTCTAAT-3′). The extracted DNA was sent to the Sangon Biotec Institute (SBI) platform in Shanghai, China, for sequencing. DNA concentrations and purity were measured using a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, USA). Computational analysis was performed as described in Elyamine et al. (2023). Indeed, the duplication and filter-qualification of the raw fastq files, sequence classification, annotation, and alpha diversity distance calculations were performed using Quantitative Insights Into Microbial Ecology (QIIME version 1.9). UPARSE software (version 7.0.1001) was used to group the filtered sequences, with Operational Taxonomic Unit (OTU) clustered with a 97% similarity cutoff. At a 97% confidence threshold, the taxonomy of each 16S rRNA gene sequence was analyzed using the 16S rRNA database and the Ribosomal Database Project (RDP) classifier (version 2.12). The functional gene composition of the bacterial community was determined using PICRUST.

To assess the extent of pyrene degradation by bacteria present in MSM, at each generation, the concentration of pyrene was extracted using dichloromethane as an extracting solvent and analyzed using high-performance liquid chromatography (HPLC) as described in previous studies (Willis et al., 1996; Wang et al., 2016a; Allaouia et al., 2022; Asnat et al., 2022). After six successive transfers, the pyrene concentration was found to be stable. Along with pyrene concentration, the microbial community in each transfer was also determined.

Data on sediment characteristics and pyrene degradation concentration were processed using analysis of variance (ANOVA) in SPSS software (version 20). The Duncan post-hoc test with a 95% confidence level was used to identify differences among the multiple stepwise means. For taxonomic analysis, similarity analysis (ANOSIM) was used to assess similarities between different experimental groups. The alpha diversity and species richness of the different communities in in situ and ex situ environments were assessed using the Chao1 and Shannon indices, calculated in the R language software (version 3.6.0) (Konopiński, 2020; Konopiński, 2020). The Shannon index was interpreted as follows: the higher the Shannon value, the higher the community diversity. At a 97% confidence threshold, the taxonomy of each 16S rRNA gene sequence was analyzed using the 16S rRNA database and the RDP classifier (version 2.12). The distance matrix and similarity or difference in sample community composition were performed using Canonical Correlation Analysis (CCA) and UniFrac distances, respectively, in QIIME version 1.9.01. The graphs for relative abundances and taxonomies were generated using the R language (version 3.6.0) and OriginPro 8. SigmaPlot 12 software was used to plot the pyrene degradation curves in the different experimental groups.

Rhizosphere and non-rhizosphere physical and chemical properties are shown in Table 1. The pH values for the rhizosphere and non-rhizosphere showed no significant difference. However, in the rhizosphere (R), the pH was slightly lower than that of the non-rhizosphere (NR).

In total, 121,774, 121,521, 122,158, and 121,198 quality-filtered sequences were identified, respectively, in non-rhizosphere (NR) and rhizosphere samples R1, R2, and R3 through 16S rRNA gene sequencing (Table 2). No significant difference was noted in the number of identified OTUs between NR and R samples (ANOVA, p < 0.05). However, when the samples collected in the different rhizospheres are compared, the deeper we move into the mangrove forest, the higher the bacterial abundance becomes. Moreover, when considering “depth” as a factor, in both non-rhizosphere and rhizosphere samples, the greater the depth, the fewer OTUs were identified. This result indicated that bacterial abundance depends on the canopy, which is influenced by light availability and the depth factor. The Shannon and Chao1 indices, which were used to estimate the specific richness, showed that bacterial diversity and abundance were mostly and significantly higher in the top layer (0–10) than in the underlying samples (Table 2). No significant difference was noted in the specific richness between rhizosphere and non-rhizosphere samples.

The second-generation sequencing technology was used to analyze the mangrove rhizosphere and non-rhizosphere microflora in situ based on OTUs annotation results. In order to explore the influence of in situ environmental properties on the flora in the mangrove sediments, CCA analysis was performed based on 16S rRNA gene high-throughput sequencing (Figure 1).

The impact of ORP was not as important as that of K, P, pH, and Fe on the microflora of NR, R3-2, and R2-3. Ammonia nitrogen (NH₃-N) was strongly correlated with the microflora present in the rhizosphere (R1, R1-4, and R3-1). Mn, NO₂-, and Ca were positively correlated with the microflora in non-rhizosphere (NR3 and NR4) and rhizosphere (R1-1 and R2-1). Carbon, nitrogen, and sulfur were correlated with the microflora present in R3-3 and R2-4. Nitrate (NO3-) was strongly correlated with the microflora in R2, R3-4, and NR2. The microelement Mg was slightly correlated with the microflora in R1-2. Consequently, even though each element impacted the different experimental depth layers differently, it was revealed that the abundance of both rhizosphere and non-rhizosphere microflora clearly depends on environmental and local abiotic factors.

The single-sample multi-level species composition diagram showed the mangrove rhizosphere/non-rhizosphere single-sample results, moving from the inside to the outside through multiple concentric circles (Supplementary Figure S1). From the single-sample results, it can be observed that the distribution and richness of bacteria in the mangrove ecosystem vary. However, naturally, the distribution of the bacterial community in the rhizosphere and the non-rhizosphere seems to be different. This is why the diversity and relative abundance at the phylum and class levels were evaluated and plotted in Supplementary Figure S2 and Figure 2, respectively. At the class level (Figure 2), generally, the bacterial diversity in both rhizosphere and non-rhizosphere did not show major differences. The distribution of taxonomic classes in the two experimental groups and across the different depth levels was almost the same. However, a slight difference was noted in the relative abundance between the experimental groups. To be more focused and objective, only results where the percentage was greater than 1% were considered, and those ≤1% were classified as other. Except Sphingobacteriia (2.54% vs. 1.06%), uncultured Chloroflexi (2.21% vs. 1.26%), Bacteroidia (4% vs. 1.64%), and Phycisphaerae (2.71% vs. 1.3%), which were slightly (p > 0.05) more abundant in the non-rhizosphere samples, all other classes were slightly more abundant in the rhizosphere samples (p > 0.05). The most abundant identified classes were Desulfuromonadia, Anaeroline, Desulfobulbia, Alphaproteobacteria, and Gammaproteobacteria, with 10.81, 12.57, 13.4, 12 and 14.18%, respectively, in rhizosphere sediments, compared to 9.10 9.18 11.63, 11.86, and 13.28% in the non-rhizosphere. In addition, when the “depth factor” was considered, the deeper the layer, the lower the bacterial abundance. Clearly, in the top layer (0–10 cm), microbial abundance was significantly higher, whereas at greater depths, only bacteria with a relative abundance ≤1% increased. This result supports existing evidence that, in soil, bacteria are more abundant at the surface than in the deeper compartments.

Bacterial relative abundance in both mangrove rhizosphere and non-rhizosphere was further evaluated at the phylum level (Supplementary Figure S2). The relative abundance of Planctomycetota was higher in the non-rhizosphere samples than in the rhizosphere, with 10.74 and 6.42%, respectively. Proteobacteria were the most dominant phylum, with 48.84% in the rhizosphere and 44.77% in the non-rhizosphere, followed by Chloroflexi, Bacteroidota, and Desulfobacterota, with 12.67% vs. 10, 11.88% vs. 9.79, and 8% vs. 5%, respectively. Taking the depth factor into account, in the non-rhizosphere, the upper layer (0–5 cm) harbored more bacteria than the deeper zone (10–20 cm).

The mangrove rhizosphere and non-rhizosphere samples were used in an acclimation process using pyrene as an external stress factor and a carbon source, through a succession of culture incubation for 30 days in MSM. At the end of the process, the concentration of pyrene and the bacterial community present were evaluated. For the process of acclimation, since the upper layer (0–10 cm) is the most frequented by soil microbes and also since the degradation process was conducted in aerobic conditions, we only considered the results of the two top layers (0–5 and 5–10 cm).

To assess the different bacterial communities after the different cultural transfers (first, before the last, and the last transfer, denoted G1, G5, and Gf, respectively) under pyrene stress, Illumina pairwise sequencing of the 16S rRNA gene was performed. The analysis of bacterial abundance indicated that the in situ community had the greatest abundance among the acclimated communities. Indeed, a significant difference was noted when we compared bacterial community abundance during the acclimation period with that in situ; this is probably due to the pollutant stress in both the rhizosphere and the non-rhizosphere. The bacterial communities in mangrove rhizosphere and non-rhizosphere sediments after acclimation were classified at different taxonomic levels, including class (Supplementary Figure S3), family (Figure 4), and genus levels (Figure 5). As in the in situ samples, OTUs whose relative abundance was ≤1% were classified as other. In the first transfer (G1), the most identified OTUs were divided into 19 families (Figure 4), dominated by Burkholderiaceae (25% vs. 11.43% in non-rhizosphere vs. rhizosphere), Rhodobacteraceae (11% vs. 20%), Bacillaceae (11% vs. 10.59%), Xanthobacteraceae (7.84% vs. 8.97%), Pseudomonadaceae (6% vs. 7.70%), and Moraxellaceae (5% vs. 8.63%). However, in both before the last and the last transfers (G5 and Gf), these dominant families increased slightly. A few OTUs belonging to Burkholderiaceae, Rhodobacteraceae, and Bacillaceae increased in G5, with 28.37% vs. 12, 14.36% vs. 20 and 13% vs. 9.77% in non-rhizosphere vs. rhizosphere samples, respectively. In Gf, these families increased to 28.59% vs. 14, 15.67% vs. 21.36, and 13.96% vs. 15.84%, respectively, in non-rhizosphere vs. rhizosphere samples.

The bacterial classification at the genus level (Figure 5) also revealed clear changes. In total, 23 genera were identified in the first transfer (G1), dominated by Burkholderia (23.9% vs. 9.23%), Rhodobacter (4% vs. 10.95%), Bacillus (11% vs. 10%), Xanthobacter (6.82% vs. 7.62%), Dyella (4.9% vs. 7%), Pseudomonas (6% vs. 7.70%), and Acinetobacter (5% vs. 8.63%), respectively, in non-rhizosphere vs. rhizosphere samples. In the other two transfers (G5 and Gf), Thiomonas, Ochrobactrum, Ancylobacter, Defluviimonas, and Hyphomicrobium were almost absent, whereas the dominant genera increased slightly. This was the case for Burkholderia, Rhodobacter, Bacillus, and Pseudomonas, which accounted for approximately 25% vs. 9.65, 8% vs. 11, 13% vs. 9.77, and 7% vs. 8% in G5 and 26% vs. 12, 9% vs. 12, 13% vs. 15, and 7% vs. 8% in Gf.

The graphic in Supplementary Figure S4 shows the heatmap correlation matrix representing the correlation between different OTUs identified in the rhizosphere and non-rhizosphere samples. This correlation helps to understand the linear relationship between the different identified OTUs in the different experimental groups. OTUs identified in the rhizosphere sample R2-2 were positively correlated with those of the non-rhizosphere sample NR2 (r = 0.93, p = 0.02), the rhizosphere samples R3-2 (r = 0.72, p = 0.03) and R1-1 (r = 0.81, p = 0.03), the non-rhizosphere sample NR1 (r = 0.74, p = 0.04), and the rhizosphere samples R1-2 (r = 0.69, p = 0.03) and R3-1 (r = 0.62, p = 0.03). Interestingly, the identified OTUs in the non-rhizosphere sample NR1 were highly correlated with the non-rhizosphere sample NR2 (r = 0.96, p = 0.04).

The scatterplot matrix presented in Supplementary Figure S5 highlights the correlation between different genera identified in the experimental groups during the acclimation process under pyrene stress. The genus Bacillus was strongly correlated with Pseudomonas and Martelella (r = 0.83, p = 0.04). Similarly, Xanthobacter was correlated with Pseudomonas, Martelella, Altererythrobacter, and Sphingobium (r = 0.86, p = 0.02). The genus Dyella was correlated with Martelella, Altererythrobacter, and Sphingobium (r = 0.88, p = 0.03). Altererythrobacter was positively correlated with Pseudomonas and Sphingobium (r = 0.81, p = 0.02), and finally, the genus Sphingobacterium was correlated with Altererythrobacter and Kordiimonas (r = 0.77, p = 0.04).

The bacterial community structure of the different experimental groups during the acclimation process under pyrene stress was similar in the top layer. Therefore, it was crucial to identify and determine common bacteria among the different experimental groups. A core microbiome, conventionally considered as taxa present in all samples, was thereafter identified. In Figure 6, the identified core microbiome after the acclimation process of rhizosphere and non-rhizosphere samples is shown. In total, 310 OTUs, defined as the “core microbiome,” were identified, and few or no OTUs were specific to individual experimental groups. It is important to note that when we talk about genomes, we are referring to “genes,” which are putative protein-coding sequences. As we are talking about the acclimation process under pyrene stress, it can be speculated that these 310 core OTUs may have roles in tolerance to, or in the metabolism of, pyrene.

The variation of pH in this study is consistent with numerous scientific reports and studies. Numerous chemical reactions in soil and sediment are linked to the pH value. Indeed, it is reported by several studies that nutrient availability in general, and its accessible form, depends largely on the soil pH (Schneider et al., 2013; Tripathi et al., 2019). The soil pH is known to be influenced by various factors in soil and sediment, especially root respiration and soil microorganisms, which are a source of proton (H+) production, thus lowering the pH in soil and sediment (Hinsinger et al., 2003), and in the present case, the rhizosphere compartment. This can explain why a slightly lower pH was observed in the rhizosphere than in the non-rhizosphere. Nitrogen forms were found to vary from the rhizosphere to the non-rhizosphere and from one depth to another. This result is consistent with the existing knowledge on nutrients in rhizosphere and non-rhizosphere sediment. Indeed, mineralization is one of the main processes by which soil sediment enriches itself in different nutrients (Vicente et al., 2018). In the rhizosphere, both mineralization and immobilization of different nutrients by sediments can occur simultaneously and reach equilibrium (Liu et al., 2020). This could explain the difference found in different nitrogen forms in one compartment and the other.

The concentration of ammonia nitrogen (NH₃-N) was generally higher in the non-rhizosphere than in the rhizosphere. This observed variation could be explained by rhizosphere priming. Indeed, research has shown that priming, defined as the short-term change in the mineralization of organic matter originating in the soil due to the release of C deposited by fresh roots via root exudates, cells sloughing, and epidermal cell death, can influence soil N cycling (Zhang et al., 2013; Pathan et al., 2015; Wanapaisan et al., 2018)). On one hand, this system is considered a kind of evolutionary mutualism between plants and rhizosphere microbes that benefit from plant-derived C and, in turn, mutually benefit plants through enhanced microbial mineralization of organic N from the soil. On the other hand, studies demonstrate that the effects of priming on soil N cycling and stimulation of extracellular enzymes involved in N mineralization are a key mechanism in mobilizing soil nutrients (Dash et al., 2013; Díaz et al., 2013). It is clear that the effects of priming on soil nitrogen cycling can vary considerably depending on many factors such as the identity and phenology of plant species, microbial status, and soil type (Hinsinger et al., 2003). In this specific case, the study was carried out in a complex ecosystem, namely the mangrove area, where many factors involved could, in one way or another, modify this evidence. However, an in-depth study on priming in mangroves could well shed light on this established scientific reality.

The significant variation in carbon content in the rhizosphere would be linked to the products of the photosynthesis process released by plants. Indeed, it is known that plants, through the phenomenon of photosynthesis, fix atmospheric CO₂, using it and producing organic matter (Zhi et al., 2024). The products of photosynthesis are released by root exudation and are discharged underground (Yang et al., 2015). This is used as a key mechanism to reduce CO₂ in the air and sequester it in the ground. Here, we have been discussing carbon sequestration for years (Shukla et al., 2022).

Phosphorus in its phosphate form is highly insoluble in soils, and it can be bound strongly to oxides of Ca and Fe and to soil organic matter (Chiu et al., 2002). This could explain the reduced concentration of P in non-rhizospheres. However, in rhizospheres, the roots can exude organic acids such as malic and citric acids into the rhizosphere, effectively reducing the pH of the rhizosphere and thus solubilizing the P bound to soil minerals (Sokolova, 2015). Indeed, the presence of calcium, silicon, magnesium, and potassium can improve the mobilization of phosphorus not only by decreasing its sorption capacity but also by promoting the dissolution of phosphate precipitates (Lei et al., 2024). Studies have demonstrated that the joint action of precipitation-dissolution, sorption–desorption, and microbial immobilization-mineralization is generally considered to be the key process that regulates phosphorus mobilization in soil (Cho et al., 2012). Thus, by modifying the soil pH, the phosphorus content and the transformation between the phosphorus bound to different soil elements, such as Fe, Al, and calcium (Ca), can be varied.

Metal speciation in sediment depends on pH and the chemical composition of the plant root exudates (Chiu et al., 2002; Jitendra et al., 2017). Research reported that Mn, Fe oxides, and clays can adsorb cationic heavy metals or bind with them to form co-precipitates. No difference was observed in the concentration of Zn availability in both the rhizosphere and the non-rhizosphere. Additionally, its concentration was extremely low. This can be explained by the fact that this metal can interact easily with other metals such as Fe and Mn. It has been reported that oxides of Fe and Mn can specifically sorb Zn in sediment (Worms et al., 2007). Indeed, the change in redox potential permits the formation of Fe or Mn oxides, which increases the adsorption of Zn. Furthermore, the low concentration of carbon in the non-rhizosphere limited the formation of soluble complexes with Zn. The high availability of K, Mg, and Ca in the mangrove rhizosphere sediment was due to competition for exchange sites. These microelements are known to be highly absorptive and competitive on exchangeable sites (Worms et al., 2007).

The main reason for the bacterial variation in different experimental treatments was attributed to the different nutrients available, which in turn are conditioned by the physical and chemical properties of the sediments (Allaouia et al., 2022; Yang et al., 2024). Although it is evident that different soil properties interfere and influence soil microorganisms’ composition and structure, when considering the “depth” factor, the variation is even more considerable between rhizospheres and non-rhizospheres. It was found in this study that in the superficial layers (0–10 cm), the abundance of the bacterial community was much greater. This could be explained by the fact that in the upper layers, the available nutrient concentrations are higher than in the deeper layers. It is obvious that in the rhizosphere vicinity, due to root exudates, nutrients are more abundant (González-López and Ruano-Rosa, 2020; Trevisan et al., 2024), which would result in a considerable abundance of microorganisms in these surroundings. Unlike the rhizospheres, in the non-rhizosphere compartments, and especially in the coastal zone, mineralization, tidal variations, and water runoff upstream of the mangroves would be the origin of the nutrients and their variations (Liu et al., 2020). The unequal distribution of the nutritional resources of bacteria on either side of the rhizosphere and the non-rhizosphere would, in fact, be further evidence supporting a fact already demonstrated, that nutrient abundance is synonymous with microorganism abundance in soil. A positive correlation is therefore established between the various factors and the abundance of bacteria, and this demonstrates that a cause–effect relationship exists between nutrient availability and microorganisms’ biomass.

The distribution of soil bacteria supports the reality that microbial distribution in soil depends on where different nutrients are available. Generally, following the notion of the “rhizosphere effect” defined exclusively as the stimulation and creation of favorable conditions for the multiplication and development of soil microorganisms in the vicinity of roots through root exudates, studies tend unanimously to believe that the microbial biomass is more considerable in the rhizospheres than in the bulk sediment (Balyan and Pandey, 2024; Trevisan et al., 2024). It would be reasonable to believe in an abundance of microbes in the rhizosphere rather than on the other side, since it is known that root exudates are an excellent source of nutrients for the development of microbes. However, in this study, although a slight abundance of bacteria was noted in the rhizosphere, there was no significant difference in the abundance of bacteria between the rhizosphere and the non-rhizosphere. This is contradictory to the various scientific reports, which support the fact that the trend is always in favor of the rhizosphere. Nonetheless, three distinct points explain the present observation in order to illuminate our results: (i) selective effects on microorganisms in the rhizosphere zones, (ii) strong competition between rhizosphere bacteria and plants for nutrients, and (iii) competition between bacteria and other microorganisms, particularly fungi present in the rhizosphere area.

Indeed, for the first point, it is important to know that, although root exudates generally increase the biomass of soil microorganisms in the rhizosphere, this may not always be the case under different circumstances. Other scientific research supports a rather flexible opinion that there may be a selective effect on microorganisms in the rhizosphere zones due to various reasons, such as variations in root exudates depending on plant species, soil type, and microbial composition and diversity. For example, Allaouia et al. (2022), DeAngelis et al. (2009), and Mukerji et al. (2006) have reported that the composition of rhizosphere microbes is a function of root quality and composition (DeAngelis et al., 2009; Allaouia et al., 2022). This could therefore lead to a large variation in microbial biomass in the rhizosphere. In this specific case, the study was conducted in mangroves, which are quite special plants, where the excretion of root exudates and the presence of seawater could be significant causes of the variation of nutrients on both sides of the rhizosphere and non-rhizosphere, which would result in variations in microbial abundance between the two experimental groups.

For the second point, it is important to recognize that among the bacteria present in the rhizosphere, there can be a strong competition for nutrients with the plants, which would lead to considerable variation in the bacterial biomass in the rhizosphere (Rengel and Marschner, 2010; Balyan and Pandey, 2024). Studies have shown that in the event of nutrient deficiency, plants adapt in different ways, including morphological modification of the roots, changes in nutrient absorption and transport mechanisms, modification of soil pH, and exudation in general—particularly of organic compounds—to maximize the availability of nutrients (Marschner and Rengel, 2023). These changes in the rhizosphere could therefore affect and cause the bacterial variation observed. Other studies show that most available nutrients are absorbed by the roots in the apical regions, even before the rhizosphere is colonized by microorganisms (Nannipieri et al., 2008). Furthermore, it has also been reported that the mangroves could compete with anammox bacteria for available nitrogen and thus reduce the abundance of anammox bacteria in the deep zone (20–21 cm) (Li et al., 2011). Therefore, the low nutrient concentration in deep areas and the lack of difference between rhizosphere and non-rhizosphere could be attributed to the variation and distribution of root exudates in the mangrove ecosystem.

For the last point, it is clear that another competition between bacteria and other microorganisms, particularly fungi present in the rhizosphere area, could also explain the bacterial variation observed. A study supporting the competition between bacteria and fungi was conducted by Blum et al. (2004). The authors revealed an inverse relationship between the two types of microorganisms and concluded that competition occurs between them.

The concept of evolution and adaptation states that when environmental conditions experience extreme disruptions, it alters the structure of the community of living beings in this ecosystem. However, those with the necessary traits adapt themselves and manage to survive (Rastogi et al., 2012). In this study, pyrene, which is a toxic pollutant, was used as a carbon source for some microorganisms, but at the same time, it can also be toxic for others. Based on the evolutionary theory, only bacteria with the necessary genes conferring tolerance to pyrene stress were present in acclimation samples. This adaptation can be morphological, physiological, and genetic. The appearance of new phylotypes would confirm hypotheses regarding the adaptation response of the bacterial community. The bacterial community in the culture succession was studied using Illumina sequencing of the 16S rRNA gene (Figures 4, 5). We can deduce that the OTUs identified in the majority of the different samples after 30 days of incubation under pyrene stress are functional with respect to pyrene metabolism. This result is similar to those obtained by Wang et al. (2016a) and Hui et al. (2018), showing that changes in the bacterial community in an external stress culture occur over time. The diversity of the bacterial community in the G5 and Gf transfers changed in profile; some taxa that were present in the first generation G1 transfer, such as Thiomonas, Ochrobactrum, Ancylobacter, Defluviimonas, and Hyphomicrobium, progressively disappeared. This can be explained on one hand simply by natural selection, due to the toxicity of pyrene or its metabolites, or on the other hand by the reduction of redundancy during the process of co-acclimation. The isolated OTUs were mainly identified as Burkholderiaceae, Rhodobacteraceae, Bacillaceae, Xanthobacteraceae, Pseudomonadaceae, and Moraxellaceae. These families join the list of families of bacteria involved in, or that may be involved in, the degradation of pyrene. In the final transfer generation (Gf), the results indicated that certain sets of bacteria, “core” OTUs abundant and ubiquitous in the different experimental groups (R, NR, and different depth layers) were identified (Figure 6) and were mainly dominated by the genera Burkholderia, Rhodobacter, Bacillus, and Pseudomonas. These bacteria, which were enriched during the acclimation process, are believed to be in one way or another largely responsible for the bacterial responses to the degradation of pyrene. Many of the taxa have been classified as “uncultured”; however, we believe that these bacteria undoubtedly have potential pyrene degradation functions. The fact that they cannot currently be cultivated would simply be evidence that the majority (>99%) of microorganisms in the environment cannot be cultivated in the laboratory.