The trees of the cultivar Arbequina grown at the three geographically distinct locations (Carrizo Springs, Moulton, Berclair) were selected to analyze the microbiome profile of olive rhizosphere soil and endosphere of roots over three seasons (Spring, summer, and fall). Amplicon sequencing of 16S rRNA and ITS regions on the Illumina Miseq platform generated 8,971,015 bacterial and 8,049,726 fungal raw sequence reads. After filtering, 8,879,079 bacterial and 6,868,796 fungal high-quality sequence reads were obtained.
Effects of location and seasons on OTUs
The number of OTUs was used to provide a comprehensive overview of the microbial structure and distribution, where 14190 bacterial OTUs were identified in the rhizosphere (RS) and root endosphere (RE) samples. A Venn diagram representing OTU distribution in the rhizosphere and root endosphere of all samples revealed that 73.57% (10,441 OTUs) were only found in the rhizosphere soil as opposed to the roots (26.12%, 3707 OTUs), and both niche samples shared 0.03% (5 OTUs) of the total bacterial OTUs. Of all locations, rhizosphere soils shared 43.81% (4575 OTUs), while roots endosphere shared 43.78% (1623 OTUs of total bacterial OTUs). The most unique bacterial OTUs were found in rhizosphere soils (14.13%, 1476 OTUs) in the Carrizo Springs and Moulton endosphere roots (17.56%, 651 OTUs). Across seasons, 5277 OTUs (50.54%) and 1895 OTUs (51.11%) of rhizosphere and root endosphere OTUs were shared between seasons, where1176 OTUs in spring (11.26%) and 639 OTUs in fall (17.23%) exhibited the unique rhizosphere and root endosphere OTUs (Fig. 1; Fig.S1). In the case of fungi, 11068 OTUs were generated. While both niches shared 7.17% (794 OTUs) of the total fungal OTUs, 94.71% (10483 OTUs) were exclusively detected in rhizosphere soil compared to roots (5.29%, 585 OTUs). In Berclair, rhizosphere soils (30.5%, 2817 OTUs) and endosphere (22.57%,195 OTUs) exhibited the most unique fungal OTUs. The rhizosphere soils of all locations shared 21.29% (1966 OTUs), whereas the endosphere roots shared 33.62% (2786 OTUs). However, for seasons, spring (27.81%,2304 OTUs) and summer (29.37%, 220 OTUs) displayed the unique OTUs for rhizosphere soils and root endosphere, respectively. Similarly, the rhizosphere and root endosphere shared 25.16% (2,786 OTUs) and 41.35% (528 OTUs) OTUs between seasons (Fig.S2).
Location and seasons structure the microbial community composition
The relative abundance of dominant bacterial phyla varied across all locations and seasons for both sample categories. In the analysis, RS is an acronym for rhizosphere soil, RE for root endosphere, CS for Carrizo Springs, M for Moulton, B for Berclair, Sp for spring, Su for summer, and F for fall. Phylum Proteobacteria (RS, 16.42%; RE, 15.498%) and Actinobacteria (RS, 12.63%; RE, 16.47%) dominated the bacterial communities in both the rhizosphere and endosphere (Fig. 1, Table S1) followed by Proteobacteria (RS, 16.42%; RE, 15.498%) and Firmicutes (RS, 4.877%; RE, 11.1124.25%). The archaeal phylum Crenarcheota was found only in the rhizosphere (4.92%), while it showed zero abundance in the endosphere. At the genus level, Bacillus was identified as the dominant taxon across niches (RS, 3.41%; RE, 5.78%), followed by Rubrobacter (5.27%) in the rhizosphere. Actinophytocola (3.49%), Unidentified Streptosporangiales (2.38%), and Pseudonocardia (1.6%) were identified only in the root endosphere. Intriguingly, a greater proportion of "Others" under phyla and genus was discovered in both niches, indicating the possibility of a diverse bacterial composition (Fig. 1A, Fig.S3A).
Considering the influence of locations within seasons for given rhizosphere soil niches for taxonomic composition, The top known predominant bacterial phyla were Actinobacteriota in Moulton in all seasons (Sp,15.38%; Su,14.19%; F,14.06%), Proteobacteria in Carrizo Springs summer and fall (Su,10.52%; F,7.62%) and Moulton summer (7.01%), and Firmicutes (Sp,6.45%; F,6.84%), archaeal phyla Crenarcheota (Sp,8.65%; Su,5.66%) were higher in Berclair location. In the root endosphere, there was an abundance of Actinobacteriota in Carrizo Springs in all seasons (Sp,17.66%; Su,19.88%; F,33.54%) while Proteobacteria was greater in Carrizo Springs summer (23.88%) and Berclair in all seasons respectively (Sp,23.02%; Su,16.96%; F,16.38%). The summer and fall of Berclair were also enriched with other phyla like Bacteroidota (Su,4.52%; F,6.81%), while Firmicutes was found in higher numbers in the spring and summer seasons of Moulton (Sp,17.87%; Su,18.82%) (Fig.S3B-C). At the genus level, Rubrobacter was dominant in the rhizosphere of Berclair in the spring and fall (8.8%) seasons, followed by Bacillus in the spring and fall of Berclair (4.64%) and Carrizo Springs (4.86%). The Berclair and Carrizo Springs rhizosphere was also enriched with Candidatus_Nitrosophaera (3.40%) in spring and fall. In contrast, the root endosphere of Carrizo Springs was enriched with Actinophytocola (F,11.19%), Cronobacter (Su,9.65%), and Unidentified Streptosporangiales (Sp,3.04%; Su,7.08%; F,10.01%). Berclair endosphere was abundant in Corynebacterium in spring (8.39%) (Fig. 1B-C).
Like bacteria, the relative abundance of fungal phyla varied across all geographic regions and climatic seasons in both niches. Phylum Ascomycota (RS,71.09%; RE,41.37%) dominated the fungi communities, which was the most abundant phylum in both the rhizosphere and endosphere (Fig. 1, Table S1), followed by Basidiomycota (RS, 10.96%; RE, 30.95%). While Mortierellomycota was exclusively identified in rhizosphere soil (3.60%), Glomeromycota was found in abundance in the root endosphere (7.72%). At the genus level, in addition to "Others" found across both niches, members of Chaetomiaceae were identified as the dominant taxon across the rhizosphere (RS,17.61%; RE,1.52%), while unidentified Agaricales (17.41%) and Xylariales (15.97%) were dominant in the root endosphere. Furthermore, a higher percentage of "Others" under the order, family, and genus was found in both niches, suggesting a potential for a diverse fungal composition (Fig. 2A, Fig. S4A). The top four abundant fungal phyla in all rhizosphere niches of locations within seasons were Ascomycota, with the highest in Berclair during fall (79.25%), in spring (79.71%) at Carrizo Springs and in fall (78.5% at Moulton. Basidiomycota came next during summer in Moulton (23.11%) and Berclair (14.89%) and during spring at Carrizo Springs (15.4%). Mortierellomycota was more prevalent in Berclair during spring (7.47%) and summer (7.19%). The phylum Glomeromycota was equally abundant in Carrizo Spring's summer and fall seasons (Su,6.10; F,3.52%). Phylum Ascomycota dominated the root endosphere and was particularly abundant in all locations in springtime (CZ,70.90%; M,63.2%; B,63.54%), followed by Basidiomycota in the Berclair (58.91%) and at Carrizo Springs in summer (35.58%) and fall (38.96%) seasons. Glomeromycota showed the highest abundance in summer of all locations (CZ,25.81%; M,11.12%; B,10.86%) (Fig. S4B-C).
At the family level, members of Chaetomiaceae were enriched in all seasons of Berclair (Sp,32.87; Su,11.83%, F,24.63%), Hypocrealeas were dominant in the rhizosphere of Moulton summer (24.32%), Pleosporaceae (Sp,18.02%) and Cucurbitariaceae (Su,21.34%) in Carrizo Springs, while Berclair rhizosphere was enriched with Mortierellaceae during spring and summer (Sp,7.47%; Su,7.19%). In contrast, in the root endosphere, Unidentified Xylariales dominated the spring of Berclair (49.85%) and Moulton (48.21%), Moulton root endosphere was found to be abundant in Herpotrichilleaceae during spring and summer (Sp,16.29%; Su,13.53%) and Glomeraceae was enriched in Carrizo Springs root in summer (25.77%) (Fig. 2B-C). Using linear discriminant analysis of effect size (LEfSe), we aimed to identify high dimensional biomarkers that most likely explained observed differences between locations or seasons. The linear discriminant analysis effect size (LEfSe) with logarithmic LDA > 2 was used to identify biomarker taxa. Overall, rhizosphere soil contained a more considerable number of bacterial biomarkers than endosphere soil (Fig.S5A). LEfSe analysis indicated that bacterial genera Rubrobacter, Candidatus_Nitrosophaera, and unidentified Nitrosophaeraceae were enriched in the rhizosphere. Root endosphere showed a significant presence of many genera, such as Bacillus, Actinophytocola, Streptosprangiales, and Pseudonocardia. Between locations, Candidatus _Nitrosophaera and members of Nitrosophaeraceae were the predominant rhizosphere soil genus in Berclair. Carrizo Springs was enriched with Sphingomonas, unidentified Geminicoccaceae, and Actinomarinales, while the Moulton rhizosphere contained Microtrichiales and Actinobacteria. In contrast, the root endosphere of Berclair was enriched with abundant genera such as Corynebacterium, Niastella, unidentified Micromonosporaceae, Rhizobium, and Promicromonospora. In comparison, Carrizo Springs and Moulton were enriched with only a few genera, such as Streptosporangiales Clostridium_senso_stricto_6 in Carrizo Springs and Exiguobacteium in Moulton (Fig. 3A, C).
Concerning fungal biomarkers, the rhizosphere was enriched with genera such as Mortierella, Aspergillus, Pyrenochaeta, Trichoderma, Acrophialophora, Humicola, Dothiora and unidentified Chaetomiaceae, Hypocreales and Agaricales, root endosphere was enriched with genera such as Penicillium, Subulicystidium, Pocescoma and unidentified members of Agaricomycetes, Xylanases, Auriculariales, and Cerataobasidicae (Fig.S6A). In rhizosphere soils of locations, besides members of Chaetomiaceae, Penicillium, Mortierella, Leucocoprinus, Humicola, Wilcoxina, and Rhexothecium were abundant in Berclair. Moulton carried members of Hypocreales, Pleosporales, Aspergillus, Trichoderma, Acrophailophora, and Curvularia. While no unique distinguishable biomarker fungi existed in Carrizo Springs, it contained moderate numbers of Hypocreales, Aspergillus, Curvularia, and Herpotrichellaceae. In contrast, the genus Penicillium was a prominent biomarker, followed by Mortierella and unidentified members of Ceratobasidiaceae, Micrascaceae, and Sordariomycetes in the root endosphere of Carrizo springs. On the other hand, the Moulton and Berclair root endospheres were frequently abundant with species from the genera Malassezia and Serendipitaceae, respectively (Fig. 3B, D).
No significant features were identified that distinguished the biomarkers among the three seasons in both the bacteria and fungi. We also examined the existence of distinct biomarkers between niches for each season. Spring showed enriched bacterial biomarkers in the rhizosphere soil, while summer showed more dominant biomarkers in the root endosphere than in other seasons. Genera like Rubrobacter, Nitrosophaeraceae, and Gaiellales genera were detected in all seasons in the rhizosphere during the spring and fall. While Microvirga was identified as a significant biomarker in soils in the summer and fall, Candidatus _Nitrosophaera was predominant in the spring and fall, and class α-Proteobacteria members were found in summer rhizosphere soil. The root endosphere of all seasons contained the genus Actinophytocola. Staphylococcus, Pseudonocardia, and Kibdelosporangium were exclusively identified in spring, Bacillus, Cronobacter, and Streptomyces in summer, and Saccharimonadales in fall. The root endosphere also included Rhizobium and Niastella during spring and summer and Streptosporangiales in the summer and fall (Fig.S5 B-D). Seasonally, fall demonstrated a greater concentration of enriched fungal biomarkers in the rhizosphere soil than in other seasons. While Mortierella, Aspergillus, Penicillium, Trichoderma, and Acrophialophora were found in the rhizosphere throughout all seasons. Wilcoxina, Clonostachys, and Acremonium were exclusively seen in the fall, and Pyrenocheta was only detected in the summer. While the genus Pyrenechaeota was only detected in the summer, the fall showed biomarkers like Trichoderma and Humicola. The predominant biomarkers in the root endosphere genera include Malassezia and members of Auriculariales in the spring, respectively, whereas Xylariales were found across all seasons. (Fig.S6B-D).
Bacterial and fungal taxonomic richness and diversity
Alpha diversity
Rarefaction curves demonstrated distinct differences in bacterial and fungal community composition as sampling depth increased for each sample from three locations in three seasons, indicating adequate sequencing effort and the total diversity within the sample was captured. With the increase in sample size, the Specaccum (species cumulative curve), like the rarefaction curves, showed the rate of increase of new species of bacteria and fungi (Fig. S7).
To verify its validity, alpha diversity metrics like Shannon (H') and Inverse Simpson (1/D) indices were used to assess the diversity of bacterial and fungal communities within samples across niches, locations, and seasons. Bacterial communities differed significantly (Kruskal-Wallis, P < 0.001) in diversity and species richness among overall sample categories, with rhizosphere soil exhibiting the highest diversity based on all metrics (H' = 6.27; 1/D = 162.61) (Fig. 4A, Fig.S8A). Season and site significantly affect the diversity of bacterial communities. For example, the site explained 83.91%, and the season explained 78.45% of the significant variation in bacterial diversity in the rhizosphere and root endosphere, respectively. The indices estimated a significant difference (Kruskal-Wallis, P < 0.05) in niches between locations for bacterial diversity and richness in rhizosphere soil. Moulton rhizosphere soil exhibited much higher richness and diversity than other locations (H’=6.45). In contrast, none of the used metrics revealed a significant difference in the α-diversity of the root endosphere niche in locations (Table S1; Fig. 4C). For rhizosphere soils, seasonal differences in bacterial diversity and richness were not statistically significant, whereas in root endosphere, seasonal differences in α -a significant variation in observed richness can explain diversity. At the same time, the evenness index comparison (InvSimpson) was not statistically significant (Kruskal-Wallis, P > 0.05). In particular, the root endosphere had a greater bacterial diversity in the fall season (H' = 6.27) than in the other seasons based on the diversity metrics used in the study (Fig. 4E, 4h; Table S1; Fig.S8A, C, E). Fungi communities differed considerably (Kruskal-Wallis, P < 0.001) in diversity and species richness between niches, with rhizosphere soil showing the highest diversity across all locations and seasons (H’= 3.44; 1/D = 13.06) (Table S2; Fig. 4B; S8B). However, no statistically significant variation in fungal richness and evenness in the rhizosphere and root endosphere was observed according to the metrics used (Kruskal-Wallis, P > 0.05) (Fig. 4D, F; S8D, F).
Beta diversity: Permutational multivariate analyses of variance (PERMANOVA) of the Bray-Curtis distance matrix showed that compartments (R2; Bacteria:0.451; Fungi:0.123) significantly influenced microbial community differences (P = 0.001). Furthermore, locations in the rhizosphere (R2; Bacteria:0.305, Fungi:0.216) and endosphere (R2; Bacteria:0.146, Fungi:0.156) influenced microbial diversity. Seasons did not affect bacterial diversity in either niche; only fungal diversity was found in the rhizosphere (R2 = 0.084, P = 0.001). The principal coordinate analysis (PCoA) based on Bray-Curtis's distance revealed that the 54 samples of the rhizosphere and endosphere bacterial communities from the same plant species exhibited a clear tendency to group based on sample type, with two principal component scores accounting for 45.7% and 4.3% of the total variations for bacteria and 14.8% and 7.0% of the total variations for fungi, respectively (Fig. 5). Rhizosphere bacterial samples from Berclair and Moulton tend to have separate clusters and Carrizo Springs formed separate clusters from other sites. There was an overlapping clustering for the root endosphere of Berclair and Moulton, while samples from Carrizo Springs within seasons were scattered (Fig. 5C, E). Carrizo Springs' rhizosphere soil samples for fungi were dispersed for all seasons, while Moulton and Berclair samples overlapped. For root endosphere samples, samples did not cluster clearly between Berclair and Moulton locations, while Carrizo Springs clustered from other locations, although not season-wise, because they overlapped (Fig. 5D, F). These differences were also shown by NMDS in which clear separation of samples according to the locations was observed.
In contrast, samples based on seasons displayed higher overlapping for bacteria and fungi (Fig.S9). Comparable results were observed in the Bray-Curtis dissimilarity distance-based hierarchical cluster analysis, demonstrating that the samples clustered into distinct groups based on the species composition of each sample. Rhizosphere samples were grouped separately from root endosphere samples. Based on locations and seasons, the two groups were even further subdivided. Based on locations, for bacteria, it was evident that the rhizosphere soil samples from Carrizo Springs and Moulton clustered with those from Berclair, whereas the root endosphere samples from Berclair and Moulton clustered within Carrizo Springs. Likewise, root endosphere samples of fall and spring were located within summer, which agreed with PCoA and NMDS analyses. As for fungi, the rhizosphere soil and root endosphere samples from Moulton and Berclair strongly clustered separately with those from Carrizo Springs based on their geographic locations. The rhizosphere samples from the fall and spring were grouped within the summer, whereas the root endosphere samples from the seasons were intermingled (Fig.S10).