Spatial Patterns and Overlap of Sedimentary and Rhizosphere Microbiomes of the Seagrass Zostera capensis

TL;DR

This study explores the microbial communities in the sediment and roots of a South African seagrass species to understand their roles in ecosystem health and function.

Contribution

The first characterization of rhizosphere microbiomes of Zostera capensis using 16S rRNA metabarcoding across multiple estuaries.

Findings

Sediment and rhizosphere microbiomes of Zostera capensis share 34 genera but also have distinct core genera.

Both sediment and rhizosphere microbiomes show significant spatial variability influenced by local and estuary-specific factors.

The rhizobiome is enriched for nutrient cycling pathways potentially beneficial to Zostera capensis.

Abstract

Seagrasses are important nature‐based solutions for climate change mitigation and adaptation due to their carbon stocks and ecosystem service co‐benefits. Characterising microbial communities in seagrass sediments and rhizospheres is essential for understanding their roles in biogeochemical cycling, seagrass health, and potential contributions to ecosystem functioning. However, the extent to which seagrass microbiomes are shared at different spatial scales is not well understood. We utilised 16S rRNA metabarcoding to characterise prokaryotic communities in the sediments of the seagrass Zostera capensis at three estuaries spanning the environmental gradient of South Africa. In addition, we characterised the rhizosphere microbiome (rhizobiome) to better understand rhizosphere and sediment community dynamics. Overall, after accounting for community in adjacent seawater, we found that
Z. capensis
sediment and rhizosphere microbiomes largely overlap, sharing 34 genera but also harbour core genera. The sediment microbiome and rhizobiome showed significant spatial variability, suggesting that both local‐scale and broader estuary‐specific factors shape site‐level microbial signatures. Further, predictive functional analysis showed that the rhizobiome was enriched for nutrient cycling pathways potentially beneficial to
Z. capensis
. Our findings support the exploration of sediment and rhizosphere microbial communities for monitoring ecosystem health and assessing impacts from threats such as pollution and climate change.

Sedimentary microbiomes and, for the first time, rhizosphere microbiomes of Zostera capensis, South Africa's most abundant seagrass, were characterised using 16S rRNA metabarcoding across three estuaries. After accounting for seawater‐derived microbes, we found that sediment and rhizosphere communities largely overlapped but also harbor distinct core genera.

Figures

(a) Location of estuaries sampled in this study. (b) Sampling design for characterising the microbial diversity of the water column, sediments and seagrass rhizosphere. Note that for sediment sampling, the top 1 cm of sediment was discarded, with the 1–5 cm retained for analysis. Sediment and seawater samples were collected from the three estuaries, while rhizosphere samples were only collected from the Olifants River Estuary.

Microbial community structure for (a), top phyla (> 0.001 relative abundance) and (b), top genera (> 0.0001 relative abundance) in the sediment, rhizosphere sediments (Olifants only), and water samples at the three sampled estuaries.

Observed ASV and Shannon alpha diversity indices. For the sediments at (a), the estuary level; (b), sampling site level and (c), between sources at Olifants River Estuary.

Principal co‐ordinate analysis (PCoA) for beta diversity under the UniFrac distance matrix (a), for the sediments at the estuary level and (b), for the sediment and rhizosphere microbial communities at Olifants River Estuary.

Overview of shared genera and core microbiomes and their taxonomy for (a) the sediment microbial communities at the three estuaries; (b) the taxonomic ranks of the core genera; (c) the rhizosphere core microbiome; and (d) the taxonomic ranks of the rhizosphere core microbiome.

Full text

Introduction

1

Seagrasses are a polyphyletic group of flowering plants fully adapted to marine environments, inhabiting intertidal and subtidal zones of estuarine and coastal habitats on all continents except Antarctica (Les et al. 1997; Short et al. 2016; UNEP‐WCMC and Short 2021). Throughout their distribution, seagrasses are highly productive ecosystem engineers that modify surrounding sediments and water currents, providing a wide range of ecosystem services (Nordlund et al. 2016). They are also recognised as nature‐based solutions for climate change mitigation through their carbon sequestration potential with regionally and globally significant carbon stocks (Macreadie et al. 2021; Raw et al. 2023; Ndhlovu et al. 2024). However, ecosystem functioning and the capacity of seagrasses to deliver ecosystem services are closely linked to the structure and function of microbial communities associated with them (Hurtado‐McCormick et al. 2022; Corinaldesi et al. 2023; Crump and Bowen 2024). Further, seagrass sediments and associated microenvironments, including plant tissue (endosphere), leaves (phyllosphere) and roots together with rhizomes (rhizosphere), have been identified as ecological micro‐niches that can harbour microbial taxa important for seagrass ecosystem health and associated function (Uku et al. 2007; de la Garza Varela et al. 2023; Tasdemir et al. 2024). As seagrasses continue to experience population declines globally due to anthropogenic activities (Unsworth et al. 2018; Dunic et al. 2021), this also negatively impacts the provision of ecosystem services (Nordlund et al. 2016). As such, seagrass‐associated microbial communities, inextricably linked to plant and soil health, are emerging as important indicators of coastal ecosystem health (Conte et al. 2021) and may also support seagrass restoration (Corinaldesi et al. 2023; Sun, Zhao, et al. 2024).

Numerous studies have highlighted the importance of symbiotic interactions between microbial taxa associated with seagrass microenvironments and their seagrass hosts in promoting seagrass growth and survival (Tarquinio et al. 2019; Mohr et al. 2021). This has led to the consideration of seagrass hosts and their associated microbial communities as a superorganism, termed the seagrass holobiont (Ugarelli et al. 2017; Conte et al. 2021). For example, the nitrogen‐fixing symbiosis between the bacterium Candidatus Celerinatantimonas neptuna and the Mediterranean seagrass Posidonia oceanica (Mohr et al. 2021) provides robust evidence of a specific, mutually beneficial symbiosis that serves as a well‐documented exemplar of holobiont functionality. Inoculation experiments in the seagrass Thalassia hemprichii with plant growth‐promoting rhizobacteria (PGPR) resulted in enrichment of sulphur‐cycling bacterial communities and enhanced photosynthetic performance (Zhou et al. 2024). There is also evidence suggesting that PGPR strains enhancing plant growth indicators in Zostera marina demonstrate their potential for helping restore degraded seagrass meadows (Sun, Zhao, et al. 2024). The holobiont framework suggests that due to the intricate relationships between the seagrass host and its associated microbial taxa, the health of one may indicate the health of the other (Ugarelli et al. 2017; Tarquinio et al. 2019). Consequently, the microbiome may serve as an ecological indicator for seagrass health (Mejia et al. 2016; Conte et al. 2021; Hurtado‐McCormick et al. 2022). However, there remain numerous questions about operational definitions, and critical empirical evidence supporting the seagrass holobiont concept is still lacking. These gaps include uncertainties about the degree of host specificity within the microbial community, the extent of functional and genomic redundancy among microbial taxa, and the levels at which selection operates (Conte et al. 2021; Vadillo Gonzalez et al. 2025).

A key question to understanding host‐microbiome associations, holobionts and host‐microbial taxa specificity is whether a signature community or set of taxa or functions is shared between hosts (Neu et al. 2021). In seagrasses, the sampled microenvironment also influences the composition of microbial communities; for example, microscale heterogeneity has been detected between epiphyte communities and rhizosphere microbiome (rhizobiome) (Crump and Koch 2008; Cúcio et al. 2016; Mejia et al. 2016; Ettinger et al. 2017), with distinct leaf, root and sediment microbiomes attributed to local environmental conditions (Fahimipour et al. 2017). Further, there is evidence for ‘core taxa’ that are shared between different seagrass structures or microenvironments, which are hypothesized to play key roles in plant survival and sediment biogeochemistry (Hurtado‐McCormick et al. 2019). In microbial community studies, core taxa are often referred to as the core microbiome with the terms used interchangeably in the literature (Neu et al. 2021). Identifying such a core microbiome provides a framework for characterising microbial taxa that may be involved in promoting host survival (Mohr et al. 2021) or performing functions beneficial to the seagrass host (Hurtado‐McCormick et al. 2022). Associations between seagrass and the core taxa are influenced by host plant ecology, physiology, environmental conditions and metabolic properties of the microbes (Ugarelli et al. 2017), although not all microbial taxa associated with seagrass microenvironments are beneficial to seagrass hosts (Tarquinio et al. 2019; Conte et al. 2021). While a core microbiome may exist, its composition is not static and can vary across space and time. The composition of the core microbiome is also influenced geographically and temporally, determined by the distinct microenvironment on the seagrass, for example, phyllosphere versus rhizospheres (Hurtado‐McCormick et al. 2019), and the associated environmental conditions, such as oxygen levels (Martin et al. 2018), organic substrates, root or rhizome exudates (Zhang et al. 2024) and light availability (Martin et al. 2018).

Among the seagrass microenvironments, the rhizosphere, at the interface of seagrass and sediment, is likely an important micro‐niche for seagrass plant growth and survival as well as investigating the holobiont framework (Cúcio et al. 2016; de la Garza Varela et al. 2023; Sun, Liu, et al. 2024; Zhou et al. 2024). The rhizosphere is involved in plant processes that facilitate the exchange of nutrients between the seagrass and its associated microbial community, for example, nitrogen (Mohr et al. 2021; Pfister et al. 2023). Several studies show that rhizosphere microbial communities and those of surrounding vegetated sediments differ significantly (Jensen et al. 2007; Bourque et al. 2015; Cúcio et al. 2016), as well as between the rhizosphere and the phyllosphere (Fahimipour et al. 2017). For example, Jensen et al. (2007) found that sulfate‐reducing (SRB) and sulfate‐oxidising bacteria (SOB) inhabit seagrass rhizospheres, influenced by the available oxygen produced by the seagrass and also suggested that SOB may benefit seagrasses by removing toxic hydrogen sulphide (H_2_S). While the environment has been shown to shape the microbiomes associated with the seagrass rhizosphere (Cúcio et al. 2016; Zhang et al. 2020). Other studies suggest that host identity is more important (Sun, Liu, et al. 2024), underscoring the need for additional studies to disentangle how species‐specific traits and environmental factors interact to shape the rhizosphere microbial communities.

The Cape‐Dwarf eelgrass, Zostera capensis, is endemic to Africa, extending from cool‐temperate to tropical environments from South Africa to Kenya, where it inhabits intertidal and shallow subtidal habitats (von der Heyden et al. 2024). Populations are highly fragmented throughout its range, with sustained population declines that have led to an “Endangered” listing (Adams and van der Colff 2018). As with all seagrasses, Z. capensis is a recognised ecosystem engineer providing a wide range of ecosystem services (Raw et al. 2023; von der Heyden et al. 2024) and due to its recognition as an important component of nature‐based solutions (NbS) for climate change mitigation, Z. capensis has been the focus of several recent studies with efforts to quantify its carbon stocks (Raw et al. 2023; Ndhlovu et al. 2024) and investigations to restore its populations (Amone‐Mabuto et al. 2023; Mokumo et al. 2023; Watson et al. 2023).

In South Africa, Z. capensis habitats are distributed in sheltered estuaries along the ~3000 km coastline (Raw et al. 2023; von der Heyden et al. 2024), spanning a cool‐to‐warm environmental gradient across four biogeographical regions (Van Niekerk et al. 2019). It has also been sugested that several climatic factors, including nutrient composition, temperature, and precipitation, explain the biogeography of soil microbiomes in sub‐Saharan Africa (Cowan et al. 2022), and a metabarcoding study along the heterogeneous South African coastline revealed that biogeographic boundaries, shaped by environmental factors, consistently influenced the distribution of taxa across marine metazoans, protists, and bacteria (Holman et al. 2021). Given the strong link between microbial communities and environmental conditions and their rapid response to change (Hurtado‐McCormick et al. 2019; Valverde et al. 2021; Banister et al. 2022; de la Garza Varela et al. 2023), understanding how microbial communities in seagrass ecosystems are shaped by these climatic gradients could provide insights into their responses to environmental dynamics (Conte et al. 2021). However, for Z. capensis, little is known about seagrass‐associated microbial diversity, with only one study focused on characterising the microbial communities in the sediments of coastal vegetated ecosystems, including Z. capensis (Ndhlovu et al. 2026). As more studies support using microbial communities as ecological indicators (Liu et al. 2017; Hurtado‐McCormick et al. 2019; Tarquinio et al. 2019; Conte et al. 2021), baseline data on the microbiomes of the sediment and rhizosphere of Z. capensis may prove useful for monitoring ecosystem health and potential restoration efforts. Here, we characterised the prokaryotic microbial communities in the sediments of the endangered seagrass, Z. capensis, across three South African estuaries, distributed across ~1800 km of coastline. More specifically, the aim of this study was to characterise the spatial variation of sediment and rhizosphere microbiomes in Z. capensis . By examining the inter‐ and intra‐estuarine variability of microbial communities, while accounting for communities in the water, we determined the genera that may represent a potential core microbiome associated with sediments in the beds of Z. capensis . We also identified the rhizosphere core microbiome within seagrass meadows in one of the estuaries. This study provides baseline microbial ecological data for an endangered seagrass species in a geographically underrepresented region. Our work provides a foundational understanding of seagrass‐microbiome interactions and highlights the necessity of acknowledging the role of microbiomes for seagrass persistence and resilience to increasing anthropogenic and climate change pressures within an African context.

Experimental Procedures

2

Study Sites and Sampling

2.1

Samples were collected at three estuaries in South Africa: Olifants, Breede, and Mngazana (Figures 1a and S1 for satellite images) between February and June 2023. The Olifants River Estuary supports extensive intertidal Z. capensis meadows (47.47 ha) and represents the westernmost extent of the species' distribution (Bornman et al. 2008). The Breede River Estuary, located approximately 800 km southeast of the Olifants, supports Z. capensis at a smaller (6 ha) areal extent. Further, in the east coast, 1000 km along the coastline from the Breede, the Mngazana River Estuary supports the smallest (2 ha) Z. capensis areal cover of the three sites (von der Heyden et al. 2024). All three estuaries are permanently open systems, with Z. capensis occurring along salt‐marsh margins; however, at Mngazana, seagrass co‐occurs with both salt marsh and mangrove habitats (Rajkaran and Adams 2010; Adams et al. 2016).

(a) Location of estuaries sampled in this study. (b) Sampling design for characterising the microbial diversity of the water column, sediments and seagrass rhizosphere. Note that for sediment sampling, the top 1 cm of sediment was discarded, with the 1–5 cm retained for analysis. Sediment and seawater samples were collected from the three estuaries, while rhizosphere samples were only collected from the Olifants River Estuary.

Sediment and water samples were collected at the three estuaries at three sampling sites. In addition to sediment and seawater samples, Z. capensis rhizosphere samples were also collected from the Olifants River Estuary. To account for intra‐estuarine variability in carbon (Ndhlovu et al. 2024) and microbial communities (Ndhlovu et al. 2026), three sampling sites were identified in each estuary: one at the estuary mouth herein referred to as the lower, a middle site, and an upper site further in the upper reaches of the estuary based on accessibility (sites ranged from ~0.3 to ~4 km apart depending on the configuration of the estuary and accessibility of sites; Figure S1 for satellite images).

Sediment samples were collected from Z. capensis meadows at spring low tide using a sterile 20 mL syringe with the top cut off, which was inserted into the sediment to a depth of 5 cm, where the 1–5 cm depth was immediately placed on ice in separate 50 mL tubes and the 0–1 discarded (Figure 1b). We targeted the 1–5 cm soil depth to minimise the influence of transient microbial taxa and to better represent taxa present in the more stable subsurface. Sediment samples were collected from within seagrass meadows, in between plants, avoiding roots or rhizomes. Each sediment sample was a composite sample pooled from sediment samples collected in the same general area (1–2 m radius) with three replicates collected at a distance of ~20 m apart, resulting in 27 samples (three estuaries × three sites × three replicates) across the three estuaries.

To account for microbial taxa shared between the water column and seawater, water samples were collected adjacent at points to sediment sampling sites by filtering ~250 mL through a 0.22 μm Sterivex filter, which was preserved in 2 mL of ATL buffer (Qiagen, Hilden, Germany). Three replicates of the water samples, including one field blank consisting of distilled water, were collected at each site, resulting in 27 water samples (three estuaries × three sites × three water replicates) and nine field blanks.

At the Olifants River Estuary, seagrass plants adjacent to the sediment collection sites were sampled to obtain the rhizosphere. Rhizomes, along with the roots, were collected from multiple seagrass plants in the same general area (1–2 m radius) by digging into the sediment to expose and retrieve them. After removing the seagrass leaves, the rhizomes were shaken to remove loose sediment, placed into 50 mL tubes, immediately stored on ice, and frozen at −20°C at the laboratory. Replicates were collected following the protocol for the sediment, resulting in nine (three sites × three replicates) rhizosphere samples.

Microbial DNA Extraction and Sequencing

2.2

For sediment samples, DNA was extracted using the DNeasy PowerSoil Kit (Qiagen, Hilden, Germany), but in order to maximise the detection of microbiomes, four extractions were carried out per replicate to a maximum of ~1 g of sediment. To collect the sediment from the rhizosphere, rhizomes were rinsed with 5–10 mL of phosphate‐buffered saline (PBS) solution, which was shaken for 5 min, and the sediment was collected by centrifugation following the methods of Fan et al. (2022). The DNA extraction was then performed on the rhizosphere sediment following the same procedure used for the sediment samples. To extract DNA from microbial communities filtered onto Sterivex cartridges, the DNeasy Blood & Tissue Kit (Qiagen, Hilden, Germany) was used following the protocol described by Rossouw et al. (2024). Extracted DNA samples (108 sediment samples, 36 rhizosphere, 27 water, 10 lab blanks and 9 field blanks) were sent to BGI Genomics, Hong Kong, for PCR amplification and sequencing. Briefly, the hypervariable V4 region of the 16S rRNA gene was amplified using the primer set 515F (5′‐GTGYCAGCMGCCGCGGTAA‐3′) and 806R (5′‐GGACTACNVGGGTWTCTAAT‐3′) (Hurtado‐McCormick et al. 2022) and the PCR products sequenced on the DNBSEQ platform.

Sequence Data Processing

2.3

Raw sequence reads were processed using QIIME2 v2023.71 (Bolyen et al. 2019), where sequence reads were denoised and filtered for singletons and chimaeras with amplicon sequence variants (ASVs) generated using the DADA2 pipeline (Callahan et al. 2016). Taxonomy was assigned to ASVs using a classifier trained on sequences generated by the 515F/806R primer set, using the SILVA ribosomal RNA gene database (release 138; 99% OTUs) (Quast et al. 2013) as a reference. Those ASVs with taxonomic ranks assigned to “Chloroplast” and “Mitochondria” together with ASVs assigned to the “Eukaryota” kingdom were removed. We also corrected for contaminants in the field and extraction blanks using the R package decontam v1.22.0 under a prevalence threshold of 0.1.

Identification of Sedimentary and Rhizosphere Core Microbiome

2.4

To identify the sediment core genera, sediment samples from the three estuaries were used, and Olifants estuary samples only for rhizosphere core genera using the core_members function in the microbiome R package v1.24.0 with detection set to 0 and prevalence set to 0.89 (i.e., any genera in at least 89% of samples with a relative abundance above 0). For the identification of core genera from seawater, detection was set to 0 and prevalence set to 0.5. While prevalence cutoff values used here are arbitrary, a higher threshold was chosen for sediment samples to minimise the inclusion of rare or spurious taxa. In contrast, a relatively lower threshold was applied to water samples to account for differences in extraction methods, capture a broader community, and reduce false positives in the sediment core microbiome. This was conducted at the estuary level for the different sample sources, with the comparison of sources conducted at the genus level. Core genera for sediment and rhizosphere were identified as those genera not detected from other sources, that is, restricted to one source only based on the detection and prevalence cutoff.

Functional Prediction of Sediment and Rhizosphere Microbial Communities

2.5

To predict functions and pathways of the microbial community, the phylogenetic investigation of communities by reconstruction of unobserved states (PICRUSt2) software v2.5.2 (Douglas et al. 2020) was used. This analysis was performed using ASVs where pathways were assigned using the MetaCyc Metabolic Pathway Database (Caspi et al. 2020). Predicted pathways were then tested for differential abundance between sediment and rhizosphere for samples collected from Olifants River Estuary.

Statistical Analysis

2.6

Alpha diversity metrics (observed ASVs and Shannon) were calculated using the estimate_richness function in the phyloseq R package v1.46.0. Comparisons of alpha diversity at the estuary level and between estuaries were conducted using the Kruskal‐Wallis Test (rstatix v0.7.2), with post hoc analysis performed using the Dunn Test (dunn_test function in the rstatix package). To examine the microbial community composition within estuaries (between sampling sites) and between estuaries, a permutational multivariate analysis of variance (PERMANOVA) was conducted using the adonis2 function implemented in the vegan package v2.6.6.1 under a Bray‐Curtis distance matrix as well as unweighted and weighted UniFrac distance matrices.

To identify ASVs that were differentially abundant between the rhizosphere and adjacent surrounding sediment at Olifants River Estuary, a differential abundance analysis using the DESeq2 package v1.42.1 was performed. Pairwise comparisons were conducted between sample types using the phyloseq_to_deseq2 function from phyloseq, followed by the DESeq2 function with the test set to ‘Wald’ and fitType set to ‘parametric’ with the differently abundant identified as those meeting the adjusted p‐value threshold of 0.01.

To determine the differentially abundant pathways between sources, the pathway_daa function in the ggpicrust2 package v1.7.4 was utilised, using the linear models for differential abundance analysis of microbiome compositional data (LinDA) method under a Benjamini‐Hochberg correction for multiple testing with an adjusted p‐value cut‐off of 0.01.

Results

3

Microbial Sequence Data and Taxonomic Assignment

3.1

A total of 5,620,320 raw sequence reads were obtained from 190 samples (with little to no reads in field and laboratory blanks) and assigned to 37,149 ASVs after quality filtering and taxonomical assignment. Following the merging of extraction replicates, rarefaction analysis revealed sufficient sequencing depth and sampling effort (Figure S2). The ASVs were filtered for non‐prokaryotic phyla and decontaminated using a prevalence‐based approach, resulting in a total of 35,901 ASVs. At the domain level, 94.9% of these were assigned to bacteria, while 5% were assigned to archaea, with 88.1% assigned to the genus level across both domains (Figure S3 and Table S1).

Microbial Community Structure of Sedimentary and Rhizosphere Zostera capensis

3.2

The relative abundance and prevalence of the microbial communities varied between water, sediment and rhizosphere, as well as between sampling sites (Figures 2a and S4), with the microbial community in the rhizosphere of Z. capensis characterised by 70 phyla compared to 57 and 76 phyla found in the water and sediment, respectively. The most abundant sediment phyla were Proteobacteria (25.8%), Desulfobacterota (16.2%), and Chloroflexi (10.8%), while the Olifants River Estuary rhizobiome was dominated by Proteobacteria (27.6%), Desulfobacterota (25.2%), and Bacteroidota (11.3%). In the adjacent seawater across all estuaries, the most abundant phyla were Proteobacteria (51.2%), Desulfobacterota (27.6%), and Bacteroidota (17.8%). While at the genus level, the most frequently detected genera across sediments and rhizosphere included Woeseia, Rhodopirellula, Desulfosarcina, and Halioglobus, as well as several candidate and uncultured lineages (Figure 2b).

Microbial community structure for (a), top phyla (> 0.001 relative abundance) and (b), top genera (> 0.0001 relative abundance) in the sediment, rhizosphere sediments (Olifants only), and water samples at the three sampled estuaries.

Alpha Diversity

3.3

Comparisons of alpha diversity metrics in sediment samples revealed significant differences among the three estuaries for Observed ASVs (H(2) = 7.20, p < 0.05) while the Shannon index did not show significant results (H(2) = 1.53, p > 0.05; Figure 3a). Furthermore, post hoc pairwise comparisons for the observed ASV alpha diversity index showed significant differences for the Breede and Mngazana estuaries (p < 0.05) and between the Mngazana and Olifants estuaries (p < 0.01), while the Shannon index did not show significant results. At the site level, only the observed ASV for Breede River Estuary showed significant differences between sites (H(2) = 7.20, p < 0.05; Figure 3b and Table S2) and the Shannon alpha index showed significant intra‐estuarine variability for both Breede (H(2) = 6.49, p < 0.05) and Mngazana (H(2) = 6.49, p < 0.05).

Observed ASV and Shannon alpha diversity indices. For the sediments at (a), the estuary level; (b), sampling site level and (c), between sources at Olifants River Estuary.

For the Olifants estuary, where rhizosphere samples were also collected, we found significant differences between the sources (i.e., water, sediment and rhizosphere) for Observed ASV (H(2) = 21.56, p < 0.01; Figure 3c) and Shannon index (H(2) = 18.16, p < 0.01). Furthermore, pairwise comparisons showed significant differences between the rhizosphere and sediment for the observed ASV (p < 0.05), while the Shannon index did not (Table S3).

Beta Diversity

3.4

Beta diversity analyses using PERMANOVA (adonis) showed significant differences in sediment microbial communities between sampled estuaries under three distance matrices (Bray‐Curtis: R^2^ = 0.50, Pseudo‐F = 12.6, p < 0.01, Figure 4a and Table S4 for UniFrac and weighted UniFrac distance matrices). Within estuaries, the PERMANOVA showed significant intra‐estuarine spatial variability in the sediment microbial community composition for all estuaries (Table S5). Pairwise comparisons of beta diversity between estuaries by PERMANOVA (adonis2) also showed significant variability in the microbial composition under the distance matrices (Table S6). Sediment and rhizosphere microbial communities at Olifants River Estuary showed significant variability only in pairwise comparisons under the UniFrac distance matrix (R^2^ = 0.47, Pseudo‐F = 10.7, p < 0.01, Figure 4b and Table S6 for other distance matrices).

Principal co‐ordinate analysis (PCoA) for beta diversity under the UniFrac distance matrix (a), for the sediments at the estuary level and (b), for the sediment and rhizosphere microbial communities at Olifants River Estuary.

Core Microbiome and Differentially Abundant Taxa

3.5

We identified Z. capensis sediment microbial taxa unique to each of the estuaries, where the Breede River Estuary showed the highest number of unique genera, Further, we detected a core sediment microbiome consisting of 18 genera that was shared among the three estuaries (Figure 5a,b). After accounting for the core genera in sediment and water, we identified a potential rhizosphere core microbiome of 15 genera (Figure 5c,d). The differential abundance analysis between the sediment and rhizosphere at Olifants River Estuary found that 20 ASVs across eight phyla were found to be differentially abundant between the sediment and rhizosphere at Olifants River Estuary (Table S7). Differentially abundant taxa were dominated by uncultured lineages belonging to families Flavobacteriaceae, Cyanobiaceae, and Pseudoalteromonadaceae in members where genera Synechococcus_CC9902 and uncultured were differentially abundant in the sediment, while genera Pseudoalteromonas, Pseudomonas, Candidatus Thiodiazotropha, and uncultured bacterium genera were differentially abundant in the rhizobiome.

Overview of shared genera and core microbiomes and their taxonomy for (a) the sediment microbial communities at the three estuaries; (b) the taxonomic ranks of the core genera; (c) the rhizosphere core microbiome; and (d) the taxonomic ranks of the rhizosphere core microbiome.

Predicted Functions of the Microbial Community

3.6

The predictive functional analysis showed there were 110 significantly differentially abundant MetaCyc pathways between the Z. capensis rhizosphere and sediment microbial communities at Olifants River Estuary. Pathways in the rhizosphere included sulfur‐related pathways (sulfoglycolysis, superpathway of sulfate assimilation and cysteine biosynthesis, sulfate reduction I), methane‐related pathways (superpathway of methanogenesis and methanogenesis from acetate), nitrogen‐related pathways (nitrifier denitrification, nitrate reduction I (denitrification) and urea cycle) and organic carbon degradation‐related pathways, among others (Figure S5).

Discussion

4

South African Estuarine Microbial Communities Differ at Large and Small Spatial Scales

4.1

Characterising the diversity and community structure of seagrass‐associated microbial communities across different environments is an important first step towards understanding their functions in seagrass health and resilience (Ugarelli et al. 2017; Tarquinio et al. 2019; Conte et al. 2021), supporting restoration efforts (Bourque et al. 2015; Sun, Zhao, et al. 2024) and providing a more holistic understanding of the role of microbiomes in blue carbon sequestration and storage (Hurtado‐McCormick et al. 2022; Corinaldesi et al. 2023). How microbial communities vary at different spatial scales or to what extent they are shared, particularly across environmentally heterogeneous systems, can support efforts to identify functionally important bacterial groups, that is, those involved in, for example, nutrient cycling and bioremediation (Sun, Liu, et al. 2024; Sun, Zhao, et al. 2024; Zhou et al. 2024). In our study that captures one of the world's pronounced coastal environmental gradients, we show that sediment microbial communities vary by site, likely driven by local environmental conditions, with each estuary exhibiting unique microbial signatures. Significant intra‐ and inter‐estuarine variability suggests that both local conditions and estuary influence microbial composition. However, our work also identified Z. capensis potential sedimentary and rhizosphere core microbiomes, providing novel insights into genera that may be important for seagrass ecosystem functioning and health in the region.

In the sediments of Z. capensis, we detected highly abundant phyla, including the Proteobacteria and Bacteroidota, known to be very successful across diverse biomes and which are consistently retrieved in high abundances globally (Bahram et al. 2018). Additionally, the high abundance of Desulfobacterota in seagrasses is well established, where they are associated with sulphur metabolism, including sulfate‐reducing prokaryotes as a result of anoxic conditions (Smith et al. 2004; Jensen et al. 2007; Banister et al. 2022; Markovski et al. 2022). Linked to anoxic conditions, the physicochemical properties of seagrass sediments, including their salinity, continuous inundation, and elevated H_2_S levels, shape their associated microbial communities (Smith et al. 2004; Jensen et al. 2007; Banister et al. 2022). These communities can differ across seagrass species and from adjacent bare sediment (Bourque et al. 2015; but see Ugarelli et al. (2018)). Microbial communities also vary by sampling locations (Ugarelli et al. 2018; Garcias‐Bonet et al. 2021), with differences between marine and terrestrial systems along a diversity continuum rather than discrete groups (Ruff et al. 2024). For comparison with terrestrial microbiomes from the region, including sites close to where we sampled estuarine systems, a recent study by Cowan et al. (2022) of soil microbiomes across sub‐Saharan Africa identified the Actinobacteriota as the most dominant components of terrestrial microbial communities. Given the differences between terrestrial and estuarine ecosystems, the difference in most abundant taxa is not surprising, but highlights the wide diversity of microbial communities throughout southern Africa.

Our analysis of Z. capensis sediment microbiomes revealed that intra‐ and inter‐estuarine microbial diversity differed between the three estuaries along a pronounced cool‐warm environmental gradient of its coastline. Although showing mixed results, alpha diversity analyses suggest that taxonomic richness and evenness may be driving the significant differences between and within estuaries, respectively. Sediment microbial communities have been shown to vary at small spatial scales, for example, in relation to distance from the edge of a meadow in Z. marina (Ettinger et al. 2017), with local environmental conditions, such as nutrient availability and temperature, partly explaining the spatial heterogeneity of microbiomes (Szitenberg et al. 2022). Our results suggest that local environmental conditions and potentially large‐scale factors such as estuary‐specific conditions contribute to shaping unique microbial communities. This is underscored by significant community structure and clustering of samples in the PCoA analysis of sampling sites and estuaries, suggesting that each sampling location exhibits a unique environmental microbial signature. This mirrors previous work that biogeographic boundaries along the South African coastline shape the diversity of metazoans, protists, and bacteria (Holman et al. 2021) and more broadly underscores our knowledge of the biological and genetic diversity of the South African coastal marine systems (Dalongeville et al. 2022). As such, microbial communities may well add unique elements to our understanding of coastal diversity in the region. These findings are also consistent with recent proposals of the environment shaping microbial communities across pronounced biogeographic barriers. For example, distinct seagrass‐associated microbial communities were detected on both sides of Wallace's Line, a historic biogeographic boundary in Indonesia (Wainwright et al. 2024). We note that inter‐estuarine comparisons may be influenced by differences in sampling times as microbial community structure varies temporally (Yi et al. 2020); therefore, our analysis offers only a snapshot of microbial communities at the time of sampling. Furthermore, while we did not investigate direct links between microbial taxa and local environmental conditions, microbial communities have been shown to respond quickly to environmental changes (Bengtsson et al. 2017). As such, the significant spatial variability observed here supports the use of microbial communities as indicators of local environments, with strong potential for biomonitoring (Conte et al. 2021).

The Rhizobiome of

Z. capensis Differs From Microbial Communities in Sediments

4.2

Fine‐scale differences between rhizosphere, sediment, and water microbiomes were detected at Olifants River Estuary, as demonstrated by significant variability in both alpha and beta diversity metrics. Although the surrounding water and sediment influence the seagrass rhizosphere, it has been shown to maintain distinct microbial communities in other species (Jensen et al. 2007). This is supported by our results, where the two top phyla by abundance, Proteobacteria and Desulfobacterota, were consistent for rhizobiome and sediment; however, the third‐most abundant phylum in the rhizosphere was Bacteroidota, compared to Chloroflexi in the sediment. Community structure differences between the rhizobiomes and sediment microbiomes are further underscored by the detection of differentially abundant taxa. These differences are indicative of different conditions, reflecting the oxic and high organic content in the rhizosphere and root exudates, favouring the Bacteroidota (Zhou et al. 2021), while the Chloroflexi are generally anaerobic and thrive on the high organic matter in seagrass sediments (Ling et al. 2021). In addition to the differences in the abundance of phyla, our results showed significantly higher microbial richness in the sediments than in the rhizosphere, which has been demonstrated in other studies where rhizospheres harbour fewer taxa than the surrounding sediment, leading to a hypothesis of rhizospheres selecting for microbial diversity (Cúcio et al. 2016; Zhang et al. 2020; de la Garza Varela et al. 2023). This is also consistent with seagrasses producing antimicrobial products and eliminating certain taxa within their vicinity (Gono et al. 2022; Tasdemir et al. 2024). Moreover, further support for the differences between the sediment and rhizosphere communities were also observed at the genus level, where the genus Desulfosarcina, an SOB involved in organic degradation (Cúcio et al. 2018), was most dominant in the rhizosphere as seen in three European seagrass species (Cúcio et al. 2016). In contrast, in the sediment, the genus Woeseia, involved in sulphur oxidation in high organic content marine sediments (Sun et al. 2020), was the most prevalent. Our findings suggest that, like other seagrasses, Z. capensis can shape its associated microbial community; however, further investigations are required on the role of habitat and the mechanisms behind the interplay of the rhizosphere community with its seagrass host.

A Potential Core Microbiome for

Z. capensis Associated Sediments and Rhizosphere

4.3

The determination of a core microbiome depends on several factors, including, for example, the cut‐offs used to determine taxonomic delineation, as well as the incorporation of abundance and prevalence (Shade and Stopnisek 2019). Further, there is no strong consensus on which taxonomic ranks or thresholds best resolve operational definitions of core microbiomes (Neu et al. 2021). Methods may vary as well, for example, Sanders‐Smith et al. (2020) used indicator species analysis to identify core taxa of the Z. marina microbiome. In this study, our determination of a potential sediment core microbiome was based on an abundance and prevalence approach. Although the prevalence cutoffs used were arbitrary, they are more stringent compared to other studies, see for example, Ugarelli et al. (2024) used a lower cut‐off of 80% compared to the 89% used here.

After accounting for microbial taxa that may be recruited from the surrounding water and removing these from the sediment communities, our conservative approach identified a core microbiome, consisting of 18 genera, which are restricted to the sediment of Z. capensis and found across the sampling sites in the three estuaries. Among these was the Bathyarchaeia, a ubiquitously occurring archaeal lineage associated with biogeochemical cycles of carbon and nitrogen (Bates et al. 2011). We also identified the genus Bacillus, important for nitrogen cycling and soil health, which has been associated with the epiphytic and endophytic microbiome of Z. marina (Tasdemir et al. 2024). The core microbiome also included Candidatus Thiobios, an SOB, which is consistent with other seagrass sediment microbiome studies (Banister et al. 2022). In addition to these, we also identified several uncultured and uncharacterised genera; for example, members of the Acidobacteria included uncharacterised subgroups: 10, 17, 21, 22 and 23, some of which have been implicated in the degradation of organic matter (de Chaves et al. 2019).

In response to the anoxic, high organic content and sulphidic conditions of their sediments, seagrasses can alter their rhizosphere microenvironments, for example, by increasing oxygen which can influence the structure of associated microbial communities (Brodersen et al. 2018). To identify those taxa influenced by these conditions and consistently associated with the Z. capensis rhizosphere, similar to other studies (Cúcio et al. 2016; Brodersen et al. 2018; Zhang et al. 2020), we determined the core microbiome of the rhizosphere. Among the genera identified was Desulfonema, a group of SRBs that thrive in high organic content and anoxic conditions that have been identified in the rhizosphere of other seagrass species, for example, Halophila ovalis , Cymodocea nodos, T. hemprichii , and Syringodium isoetifolium (Ling et al. 2021). However, it is important to note that the high abundance of Desulfonema has also been associated with stressed seagrasses (Martin et al. 2020) and proven to be useful as an ecological indicator for early detection of stress in seagrass ecosystems (Martin et al. 2020; Conte et al. 2021). Due to the importance of sulphur metabolism to coastal vegetated ecosystems, several studies have investigated the activities of SRBs and SOB associated with seagrasses, including in Zostera noltii (Cifuentes et al. 2003) and Z. muelleri (Brodersen et al. 2018). We also identified the genus Thiogranum, associated with sulphur oxidation, a significant process in the global sulphur cycle (Mori et al. 2015). Sulphur oxidation is important for seagrass as it lowers the concentration of the toxic H_2_S in the sediment, thereby promoting the growth and survival of seagrasses (Fraser et al. 2023). Additionally, we detected several taxa involved in the degradation of organic matter, such as Ornatilinea (Podosokorskaya et al. 2013), which are important for their roles in the carbon cycle. Taxa involved in carbon degradation pathways are important for investigating climate change and carbon storage potential (Hurtado‐McCormick et al. 2022; Corinaldesi et al. 2023). Other genera identified in this as members of the rhizosphere were uncultured or uncharacterised with limited taxonomic information, suggesting that with our work we have scraped the tip of the iceberg for southern African coastal microbiomes and that additional work is needed to update the reference databases to be able to characterise microbial communities in regional seagrass, and other coastal vegetated, ecosystems.

Potential Functions of Seagrass‐Associated Microbial Communities

4.4

To understand how the microbial communities may influence seagrass ecology and function, we predicted the potential functions of the detected microbial communities. Overall, we found that the rhizosphere microbiome was enriched for, among others, nutrient (nitrogen and sulphur) and organic carbon degradation‐related pathways. The presence of nitrogen‐related pathways is consistent with other studies that have identified diazotrophic bacteria and archaea in seagrass rhizobiomes, for example, in Z. marina (Caffrey and Kemp 1990) and in P. oceanica (Mohr et al. 2021). The nitrogen pathways identified here are important for the biogeochemical cycle of nitrogen, transforming it into various forms, some of which have been studied in seagrasses elsewhere (Caffrey and Kemp 1990; Nielsen et al. 2001; Mohr et al. 2021). These nitrogen transformations are important for seagrasses, as certain forms can be assimilated by the plant for growth and survival, making nitrogen processes a key factor in the interactions between seagrasses and bacteria (Ugarelli et al. 2017; Tarquinio et al. 2019). Similarly, sulphur metabolism is also important for the seagrass due to the potentially toxic levels of H_2_S found in their sediments (Holmer et al. 2005; Fraser et al. 2023). Seagrasses also require sulphur for amino acid production, which is important for growth, with transformations of sulphur by, for example, SRB an important process for seagrass growth and health (Nielsen et al. 2001; Cifuentes et al. 2003). The anoxic conditions that promote the production of H_2_S and the growth of SOBs also favour the proliferation of methanogenic microorganisms (Tan et al. 2025). It was therefore unsurprising to find pathways related to methanogenesis within the rhizosphere (Tan et al. 2025). With increasing concerns about methane in seagrass meadows (Eyre et al. 2023), identifying the taxa involved in its methanogenesis is crucial in developing ecological indicators for gaining insight into the factors that increase methane production in seagrass ecosystems (George et al. 2020). We also identified numerous pathways involved in organic matter degradation, which may contribute to the carbon cycle in these ecosystems through enhancing the preservation of organic carbon in sediments, reinforcing their role as blue carbon sinks and contributing to climate change mitigation (Liu et al. 2017; Macreadie et al. 2021). Pathways identified in the rhizosphere microbiome in this study are consistent with nutrient exchange and interactions found in previously described symbiotic relationships that characterise the seagrass holobiont (Ugarelli et al. 2017; Tarquinio et al. 2019; Conte et al. 2021). However, as the PICRUSt2 analysis used here is limited to inferring metabolic function, only providing predicted functions, whole genome metagenomic sequencing, as well as experimental validation of these predicted functions, are required. Further investigations are also required into how these functions may form the basis for symbiotic relationships between the seagrass host plant and its rhizosphere microbiome to provide support for the holobiont of Z. capensis .

Conclusion

5

Consistent with other studies of seagrass ecosystems conducted elsewhere, we found that the endangered Z. capensis sediments and rhizosphere harbour unique taxa. We found significant differences in microbial composition between the sampled estuaries, suggesting that local‐scale environmental conditions and broader estuary‐specific factors help shape the Z. capensis sediment microbial community structure. Additionally, this study provides the first characterisation of the microbial community in the rhizosphere of Z. capensis and its potential core microbiome. These core taxa represent potential members of the Z. capensis holobiont and may be candidates for aiding in the restoration of this ecologically significant yet highly threatened seagrass species. However, confirming a true core microbiome requires sampling across the species' full geographic and environmental range. As our findings rely exclusively on 16S rRNA gene amplicon sequencing, the functional inferences remain limited. Complementary methods in future studies, such as metagenomics, metatranscriptomics, or culture‐based investigations, could provide deeper functional resolution of these microbial communities. Moreover, because our study lacks environmental parameter data (which is difficult to capture given the dynamic nature of estuarine environments), interpretations regarding microbial functions or ecological roles should be considered preliminary and hypothesis‐generating rather than indicative of causal relationships. Therefore, although further empirical work is required, our data support the increasingly recognised view that microbial communities in the seagrass sediments and rhizospheres can be utilised to monitor these increasingly important ecosystems.

Author Contributions

Andrew Ndhlovu: conceptualization, methodology, software, data curation, investigation, formal analysis, writing – review and editing, writing – original draft, visualization. Sophie von der Heyden: conceptualization, funding acquisition, writing – original draft, writing – review and editing, project administration, resources.

Funding

AN was supported through a grant from the African Research Universities Alliance (ARUA: Africa‐Europe CoRE Nature‐based Solutions for Climate Change Adaptation and Mitigation) and a Grantholder‐linked Postdoctoral Fellowship from the National Research Foundation (NRF) awarded to SvdH.

Ethics Statement

Seagrass and sediments from three estuaries (Olifants, Breede, and Mngazana) were sampled under a permit from the Department of Forestry, Fisheries, and the Environment (DFFE: RES2023‐49), with an additional permit from Cape Nature (CN35‐87‐16,936) for sample collection in the Olifants River Estuary.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Data S1: Supporting Information.