Water mass-driven multiple ecological effects determine the biodiversity and assembly processes of microbial flagellates’ communities in subtropic-tropic marginal seas of China

Running head: Microbial flagellate community assembly

Xin Guo¹, Qiang Liu¹, Xiaoqing Lin¹, Xinyi Zheng¹, Cheng Huang¹, Mengwen Pang², Lingfeng Huang^{1 *}

¹Key Laboratory of the Ministry of Education for Coastal and Wetland Ecosystems, College of the Environment and Ecology, Xiamen University, Xiamen, Fujian, China

²Department of Ocean Science, Hong Kong University of Science and Technology, Hong Kong, China

^*Corresponding Author:Lingfeng Huang: Phone (Office): +86 592-2188455, Fax (Office): +86 592-2188455, Phone (Home): +8613906053195, iD: 0000-0002-2428-8147 , E-mail:huanglf@xmu.edu.cn

Abstract:

Marine microbial flagellates form an important part of marine ecosystems, and play an essential role in maintaining ecosystem functions. However, the underlying biogeographic processes and ecological effects that shape marine microbial flagellate communities (MFCs) on the geographical scale (~ 2,000 km) remain unclear, especially how their composition is related to movements of water masses. In this study, high-throughput sequencing of 18S rRNA genes was conducted to survey two size-fractioned groups (0.8–2.0 μm for pico-sized and 2.0–20 μm for nano-sized groups) of MFCs in three subtropic-tropic marginal seas of China. Furthermore, the impacts of environmental factors, spatial factors, and water masses on MFCs were explored and compared across different spatiotemporal conditions. The results demonstrate non-random biogeographic distributions of MFCs in the studied area. These distributions were affected by several ecological processes, such as environmental selection, dispersal limitation, neutral process, and interactions within communities. These processes were driven by complex water masses that formed on a geographical scale. Notably, environmental heterogeneity was identified as the principal determinant of MFCs in each sea area. However, the importance of spatial factors increased with the spatial scale, which weakened biotic interactions within the community on a geographical scale. This effect was more apparent in nano-sized MFCs, indicating stronger dispersal limitation because of their larger cells and weaker dispersal ability. In summary, this study expands the available knowledge on the dynamic biogeographic patterns of MFCs associated with water masses on a geographical scale where strong spatial and environmental gradients exist.

Keywords: Microbial flagellate, community assembly, water mass, biodiversity, co-occurrence network, geographical scale

Introduction

The marine microbiol food web plays a key role in maintaining both the structure and function of the marine ecosystem and is as important as the classic food chain (Sherr & Sherr, 2009). Microbial flagellates, with a cell size less than 20 μm, have become a crucial part of microbial food web due to their large biomass, high diversity, multiple trophic modes, and broad ecological niches in the global marine ecosystem (Throndsen, 1997). With differing trophic modes, e.g., autotrophy (Haeder & Lebert, 2009), mixotrophy (Caron, 2016), heterotrophy (Jeuck et al., 2017), and parasitic lifestyle (Lukes, Skalicky, Tyc, Votypka, & Yurchenko, 2014), microbial flagellates greatly contribute to biogeochemical cycling. They conduct photosynthesis, phagotrophy, osmotrophy, and parasitism, thus maintaining carbon balance and forming an essential trophic link for the transfer of energy from the microbial food web to the classic food web (Worden et al., 2015). The term flagellates is not a taxonomic definition, but rather represents an assembly of single-celled protists with similar ecological characteristics—they share one or more flagella, which they use for locomotion or feeding (Throndsen, 1997).

In recent years, the rapid development of DNA sequencing technology and in particular, the application of high-throughput sequencing (HTS) combined with metabarcoding approaches, has revealed incredible diversity of marine microbial flagellates, far exceeding the recognition from traditional microscopic identification methods (Stoeck et al., 2010). However, few studies distinguished flagellates from other protists or microbial eukaryotes, but rather pooled them into a community (Guo, Wu, & Huang, 2020; Z. M. Xu et al., 2020). The reason for this approach was likely associated with a lack of ecological knowledge on most eukaryotic lineages, especially those mainly known by environmental sequences, which makes their division into flagellates and not-flagellates challenging. However, it is necessary to study microbial flagellates separately from other protists, such as ciliates (the main predators of microbial flagellates mostly with a size greater than 20 μm (Sherr & Sherr, 2009)) and fungi (which usually attach to substrates via hyphae or rhizoids and obtain nutrients exclusively by feeding osmotrophically (Richards, Jones, Leonard, & Bass, 2012)). Thus, the unique ecological status and adaption mechanisms to the environment of microbial flagellates in the microbial food web must be explored.

For more than one century, ecological research has predominantly focused on the key scientific issue of the formation, distribution, and maintenance of biodiversity, i.e., biological community assembly (Konopka, 2009). Community assembly embodies the response and adaptation of organisms to short-term environmental disturbance (Logue, Findlay, & Comte, 2015) and long-term climatic changes (Caron & Hutchins, 2012). Therefore, understanding community composition and assemblage will help to understand their relationship with the environment, and further explore how the community structure affects microbially-mediated ecological functions (Logue et al., 2015). Niche theory (Keddy, 1992) and neutral theory (Hubbell, 2006) are two major classic ecological theories that focus on explaining the community assembly process. Niche theory emphasizes niche differentiation of species under the action of environmental filtering, leading to dynamic community distributions in response to environment gradients (Caron & Hutchins, 2012). In contrast, neutral theory suggests that community dynamics are stochastic processes subject to the random effects of dispersal, ecological drift, and historical events (Martiny et al., 2006). An increasing number of studies have suggested that both processes play essential roles in the assembly of microbial communities (Xiong et al., 2017; Zhang, Zhao, Dai, Jiao, & Herndl, 2014). However, many factors may affect the relative importance of these two processes, examples for which are environmental changes, trophic interactions, interference events, dispersal effects, community connectivity, spatial scales, and various biological attributes such as niche width and species abundance (Hanson, Fuhrman, Horner-Devine, & Martiny, 2012; Langenheder & Lindström, 2019; Q. Xu et al., 2022). Exploring the relative contributions of both processes on microbial community assembly is the key for understanding the ecology and biogeography of microbial communities (Martiny et al., 2006).

The subtropic-tropic marginal seas of China have sloping topography and clear monsoon climate characteristics, and form complex seasonally changing water masses. Mixing occurs between these water masses, which results in mixtures of biological communities with different properties at various distances from their formation areas (Liu & Tanhua, 2021). Therefore, seasonal changes of microbial communities not only reflect community internal dynamics, but also biological flow between different communities in adjacent waters. This should be considered with the movement and mixing of water masses to better understand temporal dynamics and spatial variations (Fuhrman, Cram, & Needham, 2015). Studying the biogeography of microbial eukaryotic communities on a geographical scale is vital because it is the scale at which different water masses create contrasting hydrological conditions and spatial heterogeneity (Hong et al., 2011; Lee & Chao, 2003) that shapes spatiotemporal distribution patterns of microbial communities (W. Li et al., 2018). The transformations of these influences over time and space can be observed in the framework of water masses (Liu & Tanhua, 2021). However, previous studies on the biogeographic distribution of microbial eukaryotic communities were mostly restricted to regional sea areas at small spatial scales. For example, for the Xiamen coast, it has been shown that the distribution of microbial eukaryotic communities was affected by complex mechanisms such as environmental factors, dispersal limitation, neutral process, and biotic interactions (W. D. Chen, Pan, Yu, Yang, & Zhang, 2017). The same was shown for the Yangtze River plume and its adjacent waters (Guo et al., 2020). Recently, the geographical or global distributions of various microbial eukaryotes were explored for the Tara Oceans (De Vargas et al., 2015; Sommeria-Klein et al., 2021), Baltic Sea (Hu, Karlson, Charvet, & Andersson, 2016), and Mediterranean Sea (Zhou et al., 2018). Multiple ecological factors that shape spatiotemporal microbial patterns were identified. However, the effects of water masses on microbial eukaryotic communities at a geographical scale (~2,000 km) have not been explored, especially for cases where these effects are associated with strong environmental gradients and spatial factors. Moreover, such research has rarely been reported along a large scale of marginal seas in China and more attention is warranted (Z. M. Xu et al., 2020). As a result, notable knowledge gaps exist regarding how environmental conditions, spatial factors, and movement of water masses jointly determine the assembly of microbial eukaryotic communities on a geographical scale.

By using high-throughput sequencing of the 18S rRNA gene, this study described the spatiotemporal distribution of microbial flagellate communities (MFCs) in subtropic-tropic marginal seas of China. These data were used to explore how their distributions were related to different ecological processes driven by multiple water masses. For this study, MFCs were divided into two size-fractioned groups: pico-sized (0.8–2.0 μm) and nano-sized (2–20 μm) MFCs. This size fractionation was employed to assess the potential impact of cell size on diversity and ecological processes because cell size is a prime factor in many physiological and metabolic processes (e.g., picoeukaryotes have a larger cell surface to volume ratio to grow and disperse easier compared to nanoplankton) (Elferink et al., 2017; Martiny et al., 2006; Worden et al., 2015). For both size groups, the diversity, community composition, and biogeographic distribution were studied and compared. The effects of multiple ecological processes (including environmental selection, dispersal limitation, neutral process, and biological interactions) on MFCs were also assessed. It was hypothesized that the biogeographic distribution of MFCs depends on a variety of ecological processes, the degree of importance of which changes under different spatial scales, for different ecological MFC groups, and in different seasons. Testing the validity of these hypotheses helped to assess the diversity, distribution patterns, and assembly mechanisms of marine MFCs.

Materials and methods

Study area and sampling

Environmental and biotic samples were collected from three subtropic-tropic marginal seas of China, spanning 15.16–32.34°N and 114.78–126.6°E (Fig. 1a, Fig. 2a). These seas include the East China Sea (ECS), the Taiwan Strait (TS), and the northeastern South China Sea (SCS). Sampling in the ECS was conducted in spring (13–29 May, 2019) and autumn (9–28 September, 2019), sampling in the TS was conducted in spring (4–15 April, 2019) and summer (21–30 July, 2019), and sampling in the SCS was conducted in summer (11 July to 11 August, 2017). All summer and autumn samples were considered as in transitional phases from summer to autumn, as these phases showed similar environmental variations and community compositions (Fig. S1 and Supplementary Methods A). Thus, to better compare spatiotemporal differences among samples, the sampling seasons were simplified to spring and summer-autumn (Supplementary Methods A). The spatial scales of each sea area (i.e., ECS, TS, and SCS) were referred to as “regional scales”, whereas the spatial scales of two sea areas in spring (ECS and TS) or three in summer-autumn (ECS, TS, and SCS) connected by water masses were referred to as “geographical scales”. Samples from ECS and TS were collected at two water depths (2 m beneath the sea surface representing the surface layer and 3–4 m above the seabed representing the bottom layer). Samples from the SCS were collected at three layers (surface, deep chlorophyll maximum (DCM), and the bottom of the photic zone, referred to as Bottom, at a depth of ~200 m) (Fig. 1a, Fig. 2a). Maps of sampling stations were made by Ocean Data View v.5.2.1.

Totals of 5–10 L of seawater were pre-filtered sequentially through 200-μm and 20-μm nylon meshes to remove meso- and micro-plankton. The water was then sequentially filtered through 2-μm and a 0.8-μm polycarbonate membranes (47 mm in diameter, Millipore, USA) to obtain nano-sized (2.0–20 μm) and pico-sized (0.8–2.0 μm) microbial plankton samples. In total, 330 samples (167 pico-sized and 163 nano-sized samples) were collected (Table S1). Filters were immediately frozen in liquid nitrogen in the field and stored at -20 ℃ in the laboratory. DNA was extracted within one month after sampling. Details of the sampling of abiotic and biotic environmental factors are presented in Supplementary Methods B, including temperature, salinity, fluorescence, dissolved oxygen (DO), dissolved inorganic nutrients, chlorophyll a (Chl a ), and abundance of heterotrophic bacteria (HB), cyanobacteria (Cyan), pico-sized eukaryotic algae (pico-Euk), and nano-sized flagellate (NF).

DNA extraction, high-throughput sequencing, and sequence analyses

DNA was extracted using the PowerWater^® DNA Isolation Kit (Qiagen, USA) following the manufacturer’s instructions. PCR amplification was conducted using primers with barcodes to get the V4 region of eukaryotic 18S rRNA gene: 3NDF (Cavalier-Smith, Lewis, Chao, Oates, & Bass, 2009) with the reverse primer V4_euk_R2 (Brate et al., 2010). Purified PCR products were sequenced using V3 chemistry with a paired-end 2 × 250 bp run on an Illumina Hiseq 2500 platform (Illumina, Inc., San Diego, CA, USA). Sequencing data were processed on Mothur v.1.40.5 (Schloss et al., 2009) following data quality filtering, operational taxonomic unit (OTU) clustering, and taxonomic classification (Guo et al., 2020). Sequences were clustered into OTUs applying a 97% similarity threshold and were taxonomically classified based on the PR² protist database (Guillou et al., 2013) at an 80% confidence level. Normalization was conducted to enable comparison between samples in different sequencing coverage depths (Fig. S2a (i, ii)), based on the lowest sequence count (1,083 sequences for pico-sized samples and 1,107 sequences for nano-sized samples) from each dataset (Table S2, Fig. S2a (iii, iv)). To focus on our target organism, we then only selected OTUs affiliated to the microbial flagellates according to the literature (Table S3) (Adl et al., 2019; Archibald et al.; Throndsen, 1997). Taxa that are not affiliated with flagellates were removed from the datasets. Finally, two OTU tables of pico-sized and nano-sized MFCs were formed and used for further analyses and comparisons (see Supplementary Methods part C for details on this method).

Statistical analyses

Determination of water masses and environmental properties

The typical water masses in spring and summer-autumn were determined based on temperature and salinity data using fuzzy cluster and temperature-salinity water similarity methods as proposed by Zhu et al. (2019) (see details under Supplementary methods D). Temperature and salinity data used for water mass classification were collected in situ with the SeaBird conductivity-temperature-depth (CTD) profiler (SBE917plus, SeaBird Corp., WA, USA), covering all stations throughout the entire regions of cruises to better identify the distribution of water masses (Figs. S3, S4, and S5). After classifying the water masses in the whole study area in spring and summer-autumn, only the stations with the environmental and biotic samples were chosen for the further analysis and matched with corresponding water masses (Fig. 1a, Fig. 2a). All environmental factors in the selected stations were log(x+1) transformed before analyses. Principal component analysis (PCA) was conducted to test the explanatory power of significant environmental variables (p < 0.001) on the variation of samples. The significance of environmental differences among groups was further calculated using permutational multivariate analysis of variance (Adonis).

Diversity analysis and community compositions

Rarefaction curves, species accumulation curves, and Goods’ coverage index were calculated to test the sequencing depth. The two alpha-diversity indices Chao1 and Shannon were calculated to evaluate the richness and diversity of MFCs. OTU data were Hellinger-transformed prior to analyses to provide unbiased estimates (Legendre, 2008). Analyses of similarities (ANOSIM) were conducted to assess the difference in MFCs among groups based on Bray-Curtis similarity matrixes. Community dissimilarity was visualized by principal coordinate analysis (PCoA). Relative richness (i.e., number of OTUs) and relative abundance (i.e., number of reads) of lineages at different taxonomic levels were plotted using OriginPro 2019.

The effects of water masses, environmental factors, and spatial factors on MFC assembly

Correlation analyses and Mantel tests based on Spearman’s rank correlation coefficients were conducted to assess the relationships among MFCs’ alpha diversity and compositions with environmental parameters. Water masses were represented by temperature and salinity, whereas environmental factors were represented by all biotic and abiotic factors except for temperature and salinity. A set of spatial factors was generated using the principal coordinates of neighbor matrices (PCNMs) approach based on the longitude and latitude coordinates of each sampling site (Dray, Legendre, & Peres-Neto, 2006). Variations of water masses, environmental factors, and spatial factors among samples were analyzed by calculating the Euclidean distances between sampling points. The relative contributions of these three factors to MFC variances were explored using variance partitioning analysis (VPA) with adjusted R² coefficients based on redundancy analysis (RDA) (Legendre, 2008). Simple and partial Mantel tests was conducted to assess the effects of three factors on alpha diversity (i.e., Euclidean distance of samples based on Chao1 and Shannon indices) and beta diversity (Bray-Curtis distance) of MFCs (Sun, Wang, Laws, & Huang, 2020).

Community neutral model

A community neutral model (NM) was used to assess the potential importance of neutral processes on MFC assembly by fitting the relationship between OTU occurrence frequency and relative abundance (Sloan et al., 2006). In this model,R² indicates the fitness to the NM andNm values represent the metacommunity size (N) times immigration rate (m).

Co-occurrence network analyses

To assess the biotic interactions in MFCs, Spearman’s rank correlations (r ) between pairs of OTUs were calculated using the WGCNA package (Pei, Chen, & Zhang, 2017).P -values were adjusted by “q-values” using a modified version of the false discovery rate (FDR) test(Storey, 2003). Only robust (|r | > 0.6) and statistically significant (p < 0.05) correlations were incorporated into the network and visualized using the Gephi 0.9.2 interactive platform (https://www.gephi.org/, 6^th Sept., 2020 ). Several network-level topological properties were computed in the igraph R package to explore the complexity and stability of the network structure in different spatiotemporal conditions (Deng et al., 2012). All statistics analyses were performed in R (v.3.6.1) using the named packages (https://www.r-project.org/, 30^th Aug., 2020) unless otherwise stated.

Results

Water masses in the study area and their environmental heterogeneity

There were 5 and 10 water masses recognized in spring (Fig. 1a, b) and summer-autumn (Fig. 2a, b), respectively, across the area of the selected stations. Differences of environmental characteristics among these water masses were assessed with PCA (Fig. 1c, Fig. 2c). The Minzhe coastal water (Mc) along the ECS had low salinity and temperature, and high Chl a concentration, nutrients, and photosynthetic plankton in spring (Fig. 1b, c). In summer-autumn, it had low salinity and higher water temperature (Fig. 2b, c). The ECS was largely affected by the East China Sea surface water (Es) in both seasons (Fig. 1a, Fig. 2a). The branch of the Kuroshio surface water (Ks) intruded the ECS from the bottom in spring (Fig. 1a), while Kuroshio subsurface water (Ku) rose along the continental slope of the ECS in summer-autumn and mixed with the Ks, thus forming Kuroshio surface-subsurface mixed water (Km) (Fig. S4b, Fig. 2a). Both water masses had high salinity (Fig. 1b, c; Fig. 2b, c). The northern TS was mainly affected by the Coastal mixed water (Cm) in both seasons, and had similar environmental conditions with the Es (Fig. 1, Fig. 2). However, the southern TS had different distribution of water masses in spring and summer-autumn. In spring, the surface layer of the southern TS was mainly affected by the South China Sea surface water (Ss) (Fig. 1a), which might have originated from the branch of the Ks that crossed into the SCS from the Luzon Strait, thus the environmental conditions of the Ss and the Ks were similar (Fig. 1b, c). In summer-autumn, the Cm and the Zhujiang River diluted water (Zd) affected the surface layer of the southern TS, while the bottom layer was affected by the South China Sea surface-subsurface mixed water (Sm) (Fig. 2a). Samples collected at three layers in the SCS in summer-autumn in 2017 were affected by South China Sea surface water (Ss-2017), South China Sea subsurface water (Su-2017), and Pacific Ocean subsurface water (Pu-2017) (Fig. 2a). In these abbreviations, the year 2017 was added after the name as a distinction from the SCS water mass in 2019 (Supplementary methods D). Furthermore, water masses were sorted into three types according to their salinity, source, and distance from the shore. Specifically, the Mc and Zd were considered coastal waters, the Cm, Es, and Ss were considered mixed waters, while the Ks, Km, Sm, Ss-2017, Su-2017, and Pu-2017 were considered oceanic waters (Fig. 1, Fig. 2). The general environmental parameters of the above-mentioned water masses are listed in Table S5.

Throughout the studied spatiotemporal scale, the environmental properties presented differences in several dimensions: seasonal, water masses, latitudinally from north to south area, longitudinally from coastal to oceanic waters, and vertically from surface to bottom layers (Table S6). Overall, the greatest heterogeneity of environmental factors was associated with water masses, followed by latitudinal and vertical, seasonal, and longitudinal comparisons (Adonis, Table S6). This indicated a significant influence of water masses on environmental conditions on a geographical scale, which shaped the biogeographic patterns of MFCs (Table S7).

Community diversity, composition, and non-random biogeographic distribution at the geographical scale

After subsampling 1,083 and 1,107 sequences per sample for the datasets of pico- and nano-sized protists, totals of 180,861 sequences (3,227 OTUs) and 180,441 sequences (3,019 OTUs) were retained, respectively (Table S4). The slightly increasing trends of rarefaction curves (Fig. S2a), the species accumulation curves for OTUs (Fig. S2b), and the high values of Good’s Coverage index (82.48–99.82%, Fig. S2c) indicated good recovery of sequenced protist taxa. The relative richness and abundance of microbial flagellates accounted for about 70% of protists (Table S4), indicating the dominant role of microbial flagellates in protist communities.

Alpha diversity of MFCs displayed distinct spatiotemporal distribution patterns. Summer-autumn samples possessed higher alpha diversity than spring samples, although the difference was not significant (Fig. 3). Alpha diversity followed significant increasing trends with distance of water masses from the shore (Fig. 3). The bottom layers of the ECS and the TS had significantly higher alpha diversity than surface layers (Fig. S7). Correlation analyses demonstrated that alpha diversity was significantly associated with water chemistry (i.e., salinity and dissolved oxygen), abundance of food sources (i.e., heterotrophic bacteria and cyanobacteria), and abundance of nano-sized flagellates (Fig. S8).

For taxonomic compositions of pico-sized and nano-sized MFCs, the assigned 1,489 and 1,340 OTUs were classified into 32 and 34 deep-branching lineages among seven supergroups, respectively Fig. S9). Alveolata was the most dominant and diverse supergroup in both size-fractioned communities, accounting for about 60% of OTUs and more than 80% of DNA reads. It was followed by Archaeplastida, which contributed 2.66% and 3.70% of the total OTUs number and 10.11% and 6.25% of the total reads in pico- and nano-sized MFCs, respectively (Fig. S9). Stramenopiles, Rhizaria, and Hacrobia had high relative richness values, but low relative abundance. The sums of the relative richness of these three supergroups were 32.05% and 25.67% in pico- and nano-sized MFCs, respectively, while the sums of the relative abundance only accounted for 7.56% and 5.52%, respectively (Fig. S9).

The biogeographic distribution of MFCs was reflected by the results of PCoA (Fig. S10) and ANOSIM (Table S7). Overall, the strongest differences in community composition were found among water masses, followed by latitudinal comparisons of different sea areas, while seasonal, longitudinal, and vertical differences were relatively weaker (Table S7). Mantel tests showed strong associations of MFCs composition with water temperature, nutrients, and microbial abundance (Fig. S8). In addition, the dominant taxa responded differently to environmental conditions (Fig. S11c, d), leading to non-random spatiotemporal distributions of MFCs in the water masses across the studied area (Fig. 4). Dinophyceae and the marine alveolates (MALVs) were the two most abundant groups in Alveolata, dominating in all water masses (Fig. 4). However, MALVs displayed lower relative abundance in spring (P< 0.05), while Dinophyceae had high abundance in both seasons (Fig. 4). Trebouxiophyceae, a major class of Archaeplastida, was found to be widely distributed from coastal to oceanic waters in two seasons (Fig. 4). Another two key Archaeplastida lineages were Micromonas and Prasino-Clade, with an average relative abundance up to 15.04% in Mc and 8.48% in Sm in summer-autumn, respectively (Fig. 4b). Cryptophyta and Haptophyta, the two most abundant phyla of Hacrobia, were widely distributed in all water masses (Fig. 4), despite their average relative abundances of only ~2% (Fig. S9). Cercozoa, a major class of Rhizaria, with the average relative abundance of only ~1%, was more distributed in coastal and mixed currents (Fig. 4b, d). We also found only about 1% relative abundance of Labyrinthulea, a class in the Stramenopiles, occurred in the Km in summer-autumn, whereas acccounted for less than 0.1% on average in other water masses (Fig. 4b, d). Apusozoa was discovered with relative abundances up to 5.08% and 13.80% in the Ss-2017 and Pu-2017, respectively, which was less than 0.01% on average in other water masses (Fig. 4b).

Water mass-driven spatial effects and environmental heterogeneity on microbial flagellates’ communities

Simple and partial Mantel tests showed that MFCs covaried significantly with water mass, environmental heterogeneity, and spatial variability (Table 1). Among these factors, water mass was most important in most cases (Table 1). However, the contributions of the environmental and spatial factors to the MFCs’ variation differed under different spatiotemporal conditions. In terms of the spatial scale, the impacts of environmental factors were dominant in MFC assembly in each sea area (i.e., ECS, TS, and SCS), where the impact of spatial factors was weak and non-significant (Table 1). These sea areas were considered as regional scale in this study. However, at the geographical scale of two or three sea areas, the importance of spatial factors was significantly enhanced, especially for nano-sized communities (Table 1). This effect was more apparent when samples from the SCS in summer were added to analyses. Moreover, alpha diversity variation also showed stronger correlations with spatial variations on the geographical scale (Table S8). To further clarify the potential impact of spatial factors on structuring MFCs’ distributions at the geographical scale, we applied a neutral model. The NM indicated that neutral process explained more than 50% (R² > 0.50) of the variation in the occurrence frequency of OTUs in MFCs (Fig. 5). In addition, seasonal variations of the importance of environmental and spatial factors on MFCs were also observed (Table 1). To eliminate the influence of the spatial scale, samples from the SCS in summer were excluded from comparisons. The results demonstrated greater spatial effects and weaker environmental effects on both sizes of MFCs in spring than in summer-autumn (Table 1).

MFCs co-occurrence networks

The present study explored the co-occurrence network that generated from strongly significant pairwise correlations (|r | > 0.6, p < 0.05) among OTUs, to search for how biotic interactions could be important to structure the MFCs. The community on a geographical scale had more than eight sub-modules, composed of multiple taxonomic lineages in two size-fractions with different species abundance (Fig. S12). According to various topological properties, networks on the geographical scale were loosely and weakly connected compared with the three regional networks (Table 2). Specifically, the network at the geographical scale held lower indices of average degree (19.499), betweenness centralization (0.014), degree centralization (0.020), and connectedness (0.007), while the values of average clustering coefficient (0.642), average path length (5.034), diameter (12), modularity (0.766), and number of sub-modules (276) were larger than those of regional networks (Table 2). Correspondingly, we also observed a clear seasonal pattern of co-occurrence networks. On the geographical scale, community displayed more closely connected networks in spring and relatively looser networks in summer-autumn according to corresponding topological features (Table 2). This is particularly pronounced considering the samples from the SCS (Table 2).

Discussion

In this study, the spatial scale ranged between 100–2,000 km with a substantial scope that led to the clear spatial distributions of different water masses. On the one hand, the distribution of water masses forms certain connections that drive marine physicochemical conditions and microbial communities to display gradient distributions on the geographical scale. On the other hand, water masses from different directions and with differences in speed and hydrological characteristics may impose a certain dispersal limitation on microbial communities and therefore cause complex distribution patterns.

Geographical alpha diversity

The spatiotemporal patterns of alpha diversity of MFCs were largely influenced by environmental conditions. The present study revealed associations of alpha diversity with water chemistry and microbial abundance (Fig. S8). Previous studies found an increasing trend of microeukaryote diversity with salinity (Z. M. Xu et al., 2020), which is consistent with the results of the present research (Fig. S8), explaining the increase of diversity from coastal waters to oceanic waters (Fig. 1b; Fig. 2b; Fig. 3; Table S5). It has been predicted that the expansion of hypoxic areas in the oceans could favor trophic relationships dominated by protistan micro-grazers with the ability to tolerate low-oxygen environments (Caron & Hutchins, 2012). In this study, alpha diversity showed negative relationships with the dissolved oxygen and abundance of food sources (Fig. S8). This is one of the possible explanations of the higher diversity of MFCs in the bottom layer of the ECS and TS (Fig. S7), where an annual hypoxia cycle was observed because of decomposition and stratification of organic matter (Z.-Y. Zhu et al., 2011). As the main food sources of flagellates, higher abundance of heterotrophic bacteria and cyanobacteria (or other suitable environmental conditions) may facilitate the rapid growth and reproduction of certain species that thus become abundant and suppress the growth of others in the community; consequently, the diversity decreases (Caron & Hutchins, 2012). In this study, a higher abundance of nano-sized flagellates was observed in coastal waters with sufficient food abundance (Table S5), which may be the cause for their lower diversity compared with oceanic waters (Fig. 3).

In addition to the presented environmental factors, several other unmeasured environmental factors may affect alpha diversity. For example, the lower diversity of surface layer waters may be related to the strong ultraviolet radiation the surface layer is subjected to, which restricts the growth and behavior of certain UV sensitive protists (Caron & Hutchins, 2012; Rodríguez‐Martínez et al., 2009). pH is also an important factor that impacts calcification and motility in some protists (Caron and Hutchins 2012). The higher diversity of the bottom layer of the ECS and the TS could also be a result of environmental heterogeneity because of poor patch connectivity at the bottom (Yeh et al., 2015) causing competition and niche differentiation.

Geographical assemblages

In this study, the geographical assemblages of MFCs were strongly characterized by water masses, while the latitudinal, longitudinal, vertical, and seasonal differences were also displayed (Table S7). These were related to the environmental heterogeneity such as temperature, nutrients, and microbial abundance according to the Mantel tests (Fig. S8). Temperature and nutrients are the main driving factors that facilitate the growth of phytoplankton and bacteria (Chang et al., 2022) and increase grazing pressure of nanoflagellates (Choi et al., 2012), resulting in “bottom-up” controls on the MFCs (Caron & Hutchins, 2012). The influence of temperature on MFCs may reflect the community response to the variation of water masses, because temperature is considered one of the two parameters to identify the water mass in this study. It was reported that the plankton community composition from oligotrophic to productive areas in the tropical and subtropical regions were highly variable (Armengol, Calbet, Franchy, Rodriguez-Santos, & Hernandez-Leon, 2019). The nutrients could reflect the anthropogenic impacts on MFCs’ variation from the coastal to oceanic waters due to the human activities such as fishing, shipping, etc.

More specifically, our discovery of non-random spatiotemporal distributions of MFCs among water masses revealed strong niche differentiation in the community, especially for the dominant taxa, which played important role in community assembly (Fig. 4, Fig. S11c, d). For example, two dominant groups Dinophyceae and MALVs were prevalent in all the water masses, but their ecological niches were different. The sizes of both groups were clearly different, with larger cells of Dinophyceae and smaller MALV sizes (Fig. S9). The size difference reveals one of the intrinsic characteristics of two groups that may cause nich seperation through many physiological and metabolic processes such as nutrient uptake, growth and reproduction, dispersal, etc (Martiny et al., 2006). A recent study of the seasonality of microbial eukaryotes has reported that spring blooms of specific taxa could reduce the advantages of MALV hosts or affect the cell viability of MALVs themselves, thereby decreasing the abundance of MALVs (Marquardt, Vader, Stubner, Reigstad, & Gabrielsen, 2016).It also reported the predominance of Dinophyceae throughout the year (Marquardt et al., 2016). This was probably due to their mixotrophic nutrient mode that could either feed on bacteria, cyanobacteria or other smaller protists or compete for nutrients with phytoplankton (Stoecker, 1999), thus showing the relationships and potential interactions with other organisms (Fig. S11c, d). Dinophyceae is also one of the common parasitic hosts of MALVs (Guillou et al., 2008), so their distribution was correlated to a certain extent (Fig. S11c, d). In addition, MALV-II have been found to have a higher richness and diversity than MALV-I (Guillou et al., 2008), which was consistent with the findings on the present study (Fig. S9).

Archaeplastida was the second dominant supergroups in this study.Micromonas and Prasino-Clade were mainly found in Mc and Sm in summer-autumn respectively, displaying very different distributions in water masses (Fig. 4b). Another study also observed the highest abundances of Micromonas and Prasino-Clade in the west coast of the English Channel in summer (Not et al., 2004). Both studies reflected the dominance of photosynthetic picoeukaryotes in the euphotic zone of coastal areas (Not et al., 2004). The prevalence of Micromonaswas also related to the highest abundance of bacteria in Mc in summer-autumn (3.36 ×10⁶ mL^-1, Table S5), because it is a well-known bacterivorous protist (Marquardt et al., 2016) which showed positive correlations with bacteria (Fig. S11c). Another dominant taxa Trebouxiophyceae has been found to distributed widely in the studied area, especially in coastal waters in spring (Fig. 4a, c). The most abundant taxon (OTU2) of Trebouxiophyceae belonged to the genus Picochlorum , according to the BLAST results in the NCBI database (https://blast.ncbi.nlm.nih.gov/Blast.cgi, 30^th Apr., 2020). Picochlorum is a microbial green alga with wide ecological niche and a diameter of 2–3 μm. It is considered an extremophile with tolerance to highly variable environments such as largely fluctuating salinity and temperature, which includes water bodies from fresh water to 3-fold salinity of seawater and within a temperature range of 16–33 ℃ (Wang, Lambert, Giang, Goericke, & Palenik, 2014). Picochlorum was also reported to grow better under low temperature and low illuminance in winter and spring, thus were prone to algal blooms (Somogyi et al., 2009).

Lineages in Hacrobia, Rhizaria, and Stramenopiles were highly diverse with very low relative abundance for each taxon (Fig. S9). Taxa belonging to Cryptophyta and Haptophyta are usually mixotrophic types, which offer survival advantages over autotrophic phytoplankton or heterotrophic protists (Lepere, Masquelier, Mangot, Debroas, & Domaizon, 2010). This facilitated them to distribute widely in all water masses (Fig. 4) and played active roles in network relationships (Fig. S12b). Cercozoa was mainly found in coastal and mixed water masses, which may be affected by high productivity and high microbial activities of the microbial food loop in these water masses (Huang et al., 2011). Cercozoa feed on bacteria,Synechococcus , small algea, and even other protists (Lepere et al., 2010), which yielded positive correlations with these biotic factors (Fig. S11c, d). Similarly, Cercozoa was also a key member in networks with low relative abundance (Fig. S12b). Labyrinthulea was mostly discovered in the Km in summer-autumn in this study (Fig. 4b, d). A study on the seasonal distribution of Labyrinthulea in estuaries and coastal waters during six years found that Labyrinthulea was generally more abundant in summer than in other seasons (Ueda, Nomura, Doi, Nakajima, & Honda, 2015). This was related closely with the warmer water and higher concentration of dissolved organic matter in summer (Ueda et al., 2015). The input of the Kuroshio water mass brings a lot of nutrients and organic matters to the shelf area of the ECS (H. M. Li, Shi, Wang, & Han, 2014), which promotes microbial activities and the decomposition effect of Labyrinthulea (Ueda et al., 2015).

Here we also observed a high relative abundance of Apusozoa in the SCS in summer-autumn compared with other water masses (Fig. 4b). A previous study by Xu et al. (2017) at a similar location in the SCS basin showed that Apusozoa were mainly distributed in deep waters below 1000 m, where they had higher metabolic activity than in shallow waters (D. P. Xu et al., 2017). It has also been previously reported that Apusozoa were commonly detected in freshwater or sediments, with a high concentration of organic matter. They were identified as euryhaline organisms, which distributed across a wide range of salinity levels from freshwater to seawater (Torruella, Moreira, & Lopez-Garcia, 2017). The above relevant research showed that Apusozoa may inhabit a very wide ecological niche and should be further explored.

These results indicated that there was a non-random spatiotemporal distribution of microbial flagellates in the water masses across the studied area. This was related to the specific environmental conditions and the connected or isolated spatial distributions of different water masses. As the carrier of the biological community, water mass had environmental heterogeneity and spatial distribution characteristics, both of which had certain correlations (Fig. S11). This indicated that the role of water mass on the MFCs may be reflected in the regulation of environmental factors and spatial factors. Thus, multiple interactive factors jointly determined the MFCs’ assembly process in different water masses or areas.

Controlling mechanisms shaping MFCs diversity and assembly process

Effect of spatial scale

In this study, we studied microbial flagellates’ communities across the subtropic-tropic marginal seas in China on the geographical scale., of which the assembly process was shaped by environmental gradients and spatial distances under complex movements of water masses (Table 1). However, their contributions to the MFCs variation differed when the spatial scales, target organism groups, and seasons were different. The most apparent effect was related to the spatial scale. The correlations between spatial factors and MFCs can be equivalent to the distance-based dispersal limitation effect on biological communities (Hanson et al., 2012). It has been reported that the spatial scale of research areas determines which assembly process will be more important (Langenheder & Lindström, 2019). Specifically, at the smallest spatial scale, taxa can spread quickly among fully connected communities, leading to a high degree of similarity (Leibold et al., 2004). In this case, the interaction between organisms is strong (Genitsaris, Monchy, Breton, Lecuyer, & Christaki, 2016). With increasing spatial scale, species slowly and sufficiently diffuse into their respective habitats through the environmental filtering effect (Moreno-Pino et al., 2016). This effect is the dominant factor of community assembly for three regional sea areas in this work. Finally, on a larger spatial scale, for example the geographical scale from subtropic to tropic seas as addressed in the present study, dispersal limitation plays the leading role in determining community assembly (Langenheder & Lindström, 2019). In addition, spatial effects also provide evidence of the importance of historical processes for microbial composition, as certain historical environmental conditions have been preserved until today because of limitations imposed on dispersal (Hanson et al., 2012).

However, the variation caused by geographic distance may also originate from unobserved environmental parameters, such as irradiance or pH, which could lead to overestimating the importance of spatial factors (Sun et al., 2020). Another factor that may cause an overestimation of role of the spatial factors in community assembly could originate from the repeated calculation of the spatial variables because of repeated station settings in different seasons or water layers. However, when the seasons and layers were calculated separately, the spatial factors still played leading roles in most circumstances (Table S9). Thus, we speculated that the importance of spatial factors increased and exceeded that of environmental factors in determining MFC composition from the regional to the geographical scale.

The importance of spatial factors on the geographical scale was further corroborated through NM. According to neutral theory, the effect of drift interacting with dispersal could create a distance-decay pattern, reflecting purely spatial effect and unexplained variation even in the absence of environmental selection (Hanson et al., 2012). In this study, more than 50% explanation of neutral process was a relatively intermediate value of fitness to the NM compared to other studies (Fig. 5). For example, neutral model fitted 25% to the biogeographic pattern of microbial plankton in Xiamen Bay (W. D. Chen et al., 2017), a smaller area compared to the current study. Xu et al. (2020) found around 74% explanation of neutral process for the variation in microbial eukaryotic communities along the southern coastal areas of China. They attributed this to the high and random dispersal rate driven by the Zhejiang-Fujian (Zhe-Min) Coastal Current (Z. M. Xu et al., 2020). In this study, multiple water masses existed and their movement may facilitate high dispersal rates of microorganisms that inhabited the same water mass, but also limited exchanges of MFCs across water masses, thus resulting in a relatively intermediate explanation level of neutral process on communities. Therefore, water masses had the greatest relationships with MFCs in most cases independent of whether the spatial scale was regional or geographical (Table 1). Accordingly, this suggested that water mass was a key factor that influence MFC assembly through environmental conditions and spatial variability (Sun et al., 2020).

Mechanisms in shaping different size groups of MFCs

According to Mantel test results, the impact of environmental factors on pico-sized MFCs exceeded spatial factors for both regional or geographical scales (Table 1). However, for nano-sized MFCs, spatial factors exerted stronger effects on the geographical scale (Table 1). This result is mirrored by VPA, which showed greatest explanations of environmental factors (8.3%) on pico-sized MFCs and spatial factors (7.9%) on nano-sized MFCs (Fig. S11a, b). This was further supported by the NM that showed larger Nm values for pico-sized MFCs, reflecting their strong dispersal ability (Fig. 5).

Studies on west Greenland (Elferink et al., 2017) and the Southern Ocean (Clarke & Deagle, 2020) reported the increase of dispersal limitation with the cell size of microbial plankton. This may be related to intrinsic characteristics such as cell size, population size, and diffusion mode (i.e., active or passive) (Martiny et al., 2006). With passive diffusion modes of both size fractions, the larger cells of nano-sized MFCs are more susceptible to dispersal limitation because of their weaker dispersal ability (Hanson et al., 2012; Martiny et al., 2006). Moreover, the smaller pico-sized MFCs require less resources from the environment to grow and reproduce, and are thus more likely to reach higher population density, making long-distance dispersal easier to achieve (Martiny et al., 2006).

Seasonal variations of the MFCs assembly mechanisms

Mantel tests demonstrated that the spatial factors had greater impacts on MFCs than environmental factors in spring, while the situation was reverse in summer-autumn (Table 1). This may be related to the monsoon and currents exchange in different seasons (Fig. S6). In spring, monsoons exchange frequently, alternating from the northeast to the southwest. The movement of water masses is also complex and multi-directed, with the Yangtze River diluting water from the west, the warming-up of Jiangsu-Zhejiang-Fujian (Su-Zhe-Min) Coastal Currents, Kuroshio branch waters entering the SCS across the Luzon Strait and extending northward (D. Y. Chen et al., 2020), and the branch of the Kuroshio surface water extending downwards and wedge-shaping into the ECS shelf area (Qi, Yin, Zhang, Yang, & Xu, 2014) (Fig. S6a). In summer-autumn, because of the prevalence of the southwesterly monsoon, the Zhe-Min Coastal Currents weaken (Hong et al., 2011), while the Yangtze River mediated water dilution strengthens (Lee & Chao, 2003). The SCS surface currents run through the TS and extend to the southern ECS (Lee & Chao, 2003; Yeh et al., 2015), while the Kuroshio branch waters retreats southward without entering the SCS from the northern Luzon Strait (Hong et al., 2011; J. Zhu et al., 2019). Instead, the branch of the Kuroshio subsurface water rises along the continental slope of the ECS in summer and retreats in autumn (Qi et al., 2014) (Fig. S6b). The Zhujiang River diluted water considerably extends northeastward in summer and autumn (J. Zhu et al., 2019).

Overall, the water masses in spring exhibit complex and diverse characteristics with each type being relatively weak, and hydrographic boundaries may impose strong dispersal limitation on MFCs (Yeh et al., 2015). This may also explain the weak correlations between MFCs and environmental factors in spring (Table 1), as highly selective environmental conditions may weaken the response of biological communities to the environment and produce a lag-back effect (Guo et al., 2020). In contrast, the water masses in summer-autumn are relatively simple but have higher speed and flux than those in spring (Lee & Chao, 2003). This led to weaker dispersal limitation on MFCs, as well as relatively stable and homogenized environmental conditions (Table S6), showing strong correlations with MFCs (Table 1) which can quickly respond and adapt to the environment. This was further reflected in the NM with a higher degree of fitness R² in spring than in summer-autumn for both sizes, although the difference was weak (Fig. 5). This indicated a relatively stronger spatial effect of dispersal interacting with drift or other neutral processes on MFCs in spring(Hanson et al., 2012).

Other possible factors in shaping MFCs diversity and assemblages

Notably, the tested variables explained only a little of the variability in MFCs, both in Mantel tests and VPA (Table 1, Fig. S11). This indicated more complex mechanisms underlying the determination of MFCs (e.g., other unmeasured environmental components or undetected neutral process such as ecological or genetic drift, or accidental or historical events) because of the limited sampling efforts. Furthermore, many studies have shown that the co-occurrence network is a supplement to the unexplainable part of environmental factors and spatial factors in the determination of MFC assemblage, reflecting the importance of inter-species relationships in shaping the community structure (Genitsaris et al., 2016; Guo et al., 2020). In this study, the complex and strong associations in co-occurrence networks showed putative non-random coexistence patterns of MFCs under different spatiotemporal conditions, indicating multiple biological interactions such as mutualisms through cross-feeding or metabolic exchange, antagonisms through competition, predation or parasitism, and functional redundancy (Hunt & Ward, 2015).

The large number of sub-modules in the network at the geographical scale reflected numerous micro-habitats that provide a wide range of niches for organisms to maintain a relatively stable community structure under environmental heterogeneity (Deng et al., 2012). Organisms having high co-occurrence in the same sub-modules indicate similar or overlayed ecological niches with highly correlated relationships, while organisms with different sub-modules display different niches and poor correlations with each other (Deng et al., 2012). However, an increasing spatial scale increase the number of sub-modules and decreases the degree of network connectivity (Table 2), leading to weaker inter-species correlations within the community. This may be related to the distance-decay effect of community similarity suggesting a positive relationship between spatial distance and community distance (Nekola & White, 1999). In terms of seasonal patterns, the co-occurrence networks were more closely connected in spring (Table 2). As stated above, complex, multi-directional but weak movement of multiple water masses in spring could cause considerable differences in environmental conditions in different water masses (Table S6), i.e., high habitat heterogeneity. Complex connections between species usually form under high environmental pressure (Genitsaris et al., 2016). However, communities with closer connections between species tend to exhibit poor stability because of their susceptibility and vulnerability to environment disturbance, which in turn leads to lower biodiversity (Konopka, 2009). Overall, these results showed a more closely connected network (Table 2) and lower diversity (Fig. 3) in spring MFCs. Correspondingly, under the action of simple but strong water masses in summer-autumn, the connections within MFCs were relatively looser (Table 2), leading to better stability and higher diversity of MFCs in summer-autumn (Fig. 3).

Conclusion

The results of this study demonstrate that the non-random biogeographic distribution of MFCs in the subtropic-tropic marginal seas of China depends on the balance of multiple ecological processes. These are environmental selection, dispersal limitation, neutral process, and interaction within MFC communities. These effects are associated with water masses on the geographical scale and the relative importance of each effect changed with spatial scales, target organism groups, and seasons. In this study, the importance of spatial factors increased with the spatial scale, especially for nano-sized MFCs. This mainly represented the dispersal limitation effect, which also weakened interactions within the community at the geographical scale. Community network interactions were more closely connected in spring because of the complexity and multi-directionality of water masses, which enhanced environmental heterogeneity. Overall, to obtain a comprehensive understanding of MFC biogeography, future studies must determine the relative importance of different ecological processes and how they interact to influence MFC assembly under specific spatiotemporal conditions.

Acknowledgments

We sincerely thank Dr. Yonggang Huang for providing CTD data, nutrient data, and Chl a data of the ECS (2019) cruises. We thank Dr. Bangqin Huang, Dr. Xin Liu, and Fujian Institute of Oceanography to provide CTD, Chl a , and nutrient data of the TS (2019) cruises respectively. We also thank Dr. Wei Zhao and Minhan Dai for providing CTD and nutrient data of the SCS (2017) cruise. We are also grateful to the captains and crews of R/V “Xiangyanghong 18”, “Yan Ping II” and R/V “Dongfanghong 2” for their helps in the field sampling from the ECS, TS, and SCS, respectively. We also appreciate Dr. Jia Zhu and Dr. Jianyu Hu for the guidance on the water mass classification and constructive comments on the article. This work was supported by the National Natural Science Foundation of China (no. 41676131 and no. 41876155) and the National Basic Research Program of China (973 Program, no. 2011CB409804).

References

Adl, S. M., Bass, D., Lane, C. E., Lukes, J., Schoch, C. L., Smirnov, A., . . . Zhang, Q. (2019). Revisions to the Classification, Nomenclature, and Diversity of Eukaryotes. Journal of Eukaryotic Microbiology, 66 (1), 4-119.https://doi.org/10.1111/jeu.12691

Archibald, J. M., Simpson, A. G. B., Slamovits, C. H., Margulis, L., Melkonian, M., Chapman, D. J., & Corliss, J. O. (Eds.). (2017).Handbook of the Protists : Springer, Cham.

Armengol, L., Calbet, A., Franchy, G., Rodriguez-Santos, A., & Hernandez-Leon, S. (2019). Planktonic food web structure and trophic transfer efficiency along a productivity gradient in the tropical and subtropical Atlantic Ocean. Scientific Reports, 9 (1), 2044.https://doi.org/10.1038/s41598-019-38507-9

Brate, J., Logares, R., Berney, C., Ree, D. K., Klaveness, D., Jakobsen, K. S., & Shalchian-Tabrizi, K. (2010). Freshwater Perkinsea and marine-freshwater colonizations revealed by pyrosequencing and phylogeny of environmental rDNA. The ISME journal, 4 (9), 1144-1153.https://doi.org/10.1038/ismej.2010.39

Caron, D. A. (2016). Mixotrophy stirs up our understanding of marine food webs. PNAS, 113 (11), 2806-2808.https://doi.org/10.1073/pnas.1600718113

Caron, D. A., & Hutchins, D. A. (2012). The effects of changing climate on microzooplankton grazing and community structure: drivers, predictions and knowledge gaps. Journal of Plankton Research, 35 (2), 235-252.https://doi.org/10.1093/plankt/fbs091

Cavalier-Smith, T., Lewis, R., Chao, E. E., Oates, B., & Bass, D. (2009). Helkesimastix marina n. sp.(Cercozoa: Sainouroidea superfam. n.) a gliding zooflagellate of novel ultrastructure and unusual ciliary behaviour. Protist, 160 (3), 452-479.https://doi.org/10.1016/j.protis.2009.03.003

Chang, C. W., Miki, T., Ye, H., Souissi, S., Adrian, R., Anneville, O., . . . Hsieh, C. H. (2022). Causal networks of phytoplankton diversity and biomass are modulated by environmental context. Nature Communications, 13 (1), 1140.https://doi.org/10.1038/s41467-022-28761-3

Chen, D. Y., Lian, E. G., Shu, Y. Q., Yang, S. Y., Li, Y. L., Li, C., . . . Su, N. (2020). Origin of the springtime South China Sea Warm Current in the southwestern Taiwan Strait: Evidence from seawater oxygen isotope. Science China-Earth Sciences, 63 (10), 1564-1576.https://doi.org/10.1007/s11430-019-9642-8

Chen, W. D., Pan, Y. B., Yu, L. Y., Yang, J., & Zhang, W. J. (2017). Patterns and processes in marine microeukaryotic community biogeography from Xiamen coastal waters and intertidal sediments, southeast China.Frontiers in microbiology, 8 , 1912.https://doi.org/10.3389/fmicb.2017.01912

Choi, K. H., Yang, E. J., Kim, D., Kang, H. K., Noh, J. H., & Kim, C. H. (2012). The influence of coastal waters on distributions of heterotrophic protists in the northern East China Sea, and the impact of protist grazing on phytoplankton. Journal of Plankton Research, 34 (10), 886-904.https://doi.org/10.1093/plankt/fbs046

Clarke, L. J., & Deagle, B. E. (2020). Eukaryote plankton assemblages in the southern Kerguelen Axis region: Ecological drivers differ between size fractions. Deep-Sea Research II, 174 , 104538.https://doi.org/10.1016/j.dsr2.2018.12.003

De Vargas, C., Audic, S., Henry, N., Decelle, J., Mahé, F., Logares, R., . . . Probert, I. (2015). Eukaryotic plankton diversity in the sunlit ocean. Science, 348 (6237), 1261605.https://doi.org/10.1126/science.1261605

Deng, Y., Jiang, Y. H., Yang, Y. F., He, Z. L., Luo, F., & Zhou, J. Z. (2012). Molecular ecological network analyses. Bmc Bioinformatics, 13 (113), 1-20.https://doi.org/10.1186/1471-2105-13-113

Dray, S., Legendre, P., & Peres-Neto, P. R. (2006). Spatial modelling: a comprehensive framework for principal coordinate analysis of neighbour matrices (PCNM). Ecological Modelling, 196 (3-4), 483-493.https://doi.org/10.1016/j.ecolmodel.2006.02.015

Elferink, S., Neuhaus, S., Wohlrab, S., Toebe, K., Voss, D., Gottschling, M., . . . John, U. (2017). Molecular diversity patterns among various phytoplankton size-fractions in West Greenland in late summer. Deep-Sea Research II, 121 , 54-69.https://doi.org/10.1016/j.dsr.2016.11.002

Fuhrman, J. A., Cram, J. A., & Needham, D. M. (2015). Marine microbial community dynamics and their ecological interpretation. Nature Reviews Microbiology, 13 (3), 133-146.https://doi.org/10.1038/nrmicro3417

Genitsaris, S., Monchy, S., Breton, E., Lecuyer, E., & Christaki, U. (2016). Small-scale variability of protistan planktonic communities relative to environmental pressures and biotic interactions at two adjacent coastal stations. Marine Ecology Progress Series, 548 , 61-75.https://doi.org/10.3354/meps11647

Guillou, L., Bachar, D., Audic, S., Bass, D., Berney, C., Bittner, L., . . . Christen, R. (2013). The Protist Ribosomal Reference database (PR2): a catalog of unicellular eukaryote Small Sub-Unit rRNA sequences with curated taxonomy. Nucleic Acids Research, 41 (D1), D597-D604.https://doi.org/10.1093/nar/gks1160

Guillou, L., Viprey, M., Chambouvet, A., Welsh, R. M., Kirkham, A. R., Massana, R., . . . Worden, A. Z. (2008). Widespread occurrence and genetic diversity of marine parasitoids belonging to Syndiniales (Alveolata). Environmental microbiology, 10 (12), 3349-3365.https://doi.org/10.1111/j.1462-2920.2008.01731.x

Guo, X., Wu, L. N., & Huang, L. F. (2020). Spatiotemporal patterns in diversity and assembly process of marine protist communities of the Changjiang (Yangtze River) plume and its adjacent waters.Frontiers in microbiology, 11 , 579290.https://doi.org/10.3389/fmicb.2020.579290

Haeder, D.-P., & Lebert, M. (2009). Photoorientation in Photosynthetic Flagellates. In T. Jin & D. Hereld (Eds.), Chemotaxis: Methods and Protocols (Vol. 571, pp. 51-65).

Hanson, C. A., Fuhrman, J. A., Horner-Devine, M. C., & Martiny, J. B. H. (2012). Beyond biogeographic patterns: processes shaping the microbial landscape. Nature Reviews Microbiology, 10 (7), 497-506.https://doi.org/10.1038/nrmicro2795

Hong, H. S., Chai, F., Zhang, C. Y., Huang, B. Q., Jiang, Y. W., & Hu, J. Y. (2011). An overview of physical and biogeochemical processes and ecosystem dynamics in the Taiwan Strait. Continental Shelf Research, 31 (6), S3-S12.https://doi.org/10.1016/j.csr.2011.02.002

Hu, Y. O. O., Karlson, B., Charvet, S., & Andersson, A. F. (2016). Diversity of Pico- to Mesoplankton along the 2000 km Salinity Gradient of the Baltic Sea. Frontiers in microbiology, 7 (679).https://doi.org/10.5589/fmicb.2016.00679

Huang, B. Q., Xiang, W. G., Zeng, X. B., Chiang, K. P., Tian, H. J., Hu, J., . . . Hong, H. S. (2011). Phytoplankton growth and microzooplankton grazing in a subtropical coastal upwelling system in the Taiwan Strait.Continental Shelf Research, 31 (6), S48-S56.https://doi.org/10.1016/j.csr.2011.02.005

Hubbell, S. P. (2006). Neutral theory and the evolution of ecological equivalence. Ecology, 87 (6), 1387-1398.https://doi.org/10.1890/0012-9658(2006)87[1387:Ntateo]2.0.Co;2

Hunt, D. E., & Ward, C. S. (2015). A network-based approach to disturbance transmission through microbial interactions. Frontiers in microbiology, 6 , 1182.https://doi.org/10.3389/fmicb.2015.01182

Jeuck, A., Nitsche, F., Wylezich, C., Wirth, O., Bergfeld, T., Brutscher, F., . . . Arndt, H. (2017). A Comparison of Methods to Analyze Aquatic Heterotrophic Flagellates of Different Taxonomic Groups.Protist, 168 (4), 375-391.https://doi.org/10.1016/j.protis.2017.04.003

Keddy, P. A. (1992). Assembly and response rules: two goals for predictive community ecology. Journal of Vegetation Science, 3 (2), 157-164.https://doi.org/10.2307/3235676

Konopka, A. (2009). What is microbial community ecology? The ISME journal, 3 (11), 1223-1230.https://doi.org/10.1038/ismej.2009.88

Langenheder, S., & Lindström, E. S. (2019). Factors influencing aquatic and terrestrial bacterial community assembly. Environmental Microbiology Reports, 11 (3), 306-315.https://doi.org/10.1111/1758-2229.12731

Lee, H.-J., & Chao, S.-Y. (2003). A climatological description of circulation in and around the East China Sea. Deep-Sea Research II, 50 (6-7), 1065-1084.https://doi.org/10.1016/s0967-0645(03)00010-9

Legendre, P. (2008). Studying beta diversity: ecological variation partitioning by multiple regression and canonical analysis.Journal of Plant Ecology, 1 (1), 3-8.https://doi.org/10.1093/jpe/rtm001

Leibold, M. A., Holyoak, M., Mouquet, N., Amarasekare, P., Chase, J. M., Hoopes, M. F., . . . Gonzalez, A. (2004). The metacommunity concept: a framework for multi-scale community ecology. Ecology letters, 7 (7), 601-613.https://doi.org/10.1111/j.1461-0248.2004.00608.x

Lepere, C., Masquelier, S., Mangot, J. F., Debroas, D., & Domaizon, I. (2010). Vertical structure of small eukaryotes in three lakes that differ by their trophic status: a quantitative approach. The ISME journal, 4 (12), 1509-1519.https://doi.org/10.1038/ismej.2010.83

Li, H. M., Shi, X. Y., Wang, H., & Han, X. R. (2014). An estimation of nutrient fluxes to the East China Sea continental shelf from the Taiwan Strait and Kuroshio subsurface waters in summer. Acta Oceanologica Sinica, 33 (11), 1-10.https://doi.org/10.1007/s13131-014-0550-2

Li, W., Wang, M. M., Pan, H. Q., Burgaud, G., Liang, S. K., Guo, J. J., . . . Cai, L. (2018). Highlighting patterns of fungal diversity and composition shaped by ocean currents using the East China Sea as a model. Molecular Ecology, 27 (2), 564-576.https://doi.org/10.1111/mec.14440

Liu, M., & Tanhua, T. (2021). Water masses in the Atlantic Ocean: characteristics and distributions. Ocean Science, 17 (2), 463-486.https://doi.org/10.5194/os-17-463-2021

Logue, J. B., Findlay, S. E., & Comte, J. (2015). Editorial: Microbial Responses to Environmental Changes. Frontiers in microbiology, 6 , 1364.https://doi.org/10.3389/fmicb.2015.01364

Lukes, J., Skalicky, T., Tyc, J., Votypka, J., & Yurchenko, V. (2014). Evolution of parasitism in kinetoplastid flagellates. Molecular and biochemical parasitology, 195 (2), 115-122.https://doi.org/10.1016/j.molbiopara.2014.05.007

Marquardt, M., Vader, A., Stubner, E. I., Reigstad, M., & Gabrielsen, T. M. (2016). Strong Seasonality of Marine Microbial Eukaryotes in a High-Arctic Fjord (Isfjorden, in West Spitsbergen, Norway).Applied and Environmental Microbiology, 82 (6), 1868-1880.https://doi.org/10.1128/aem.03208-15

Martiny, J. B. H., Bohannan, B. J. M., Brown, J. H., Colwell, R. K., Fuhrman, J. A., Green, J. L., . . . Staley, J. T. (2006). Microbial biogeography: putting microorganisms on the map. Nature Reviews Microbiology, 4 (2), 102-112.https://doi.org/10.1038/nrmicro1341

Moreno-Pino, M., De la Iglesia, R., Valdivia, N., Henriquez-Castilo, C., Galan, A., Diez, B., & Trefault, N. (2016). Variation in coastal Antarctic microbial community composition at sub-mesoscale: spatial distance or environmental filtering? FEMS microbiology ecology, 92 (7), fiw088.https://doi.org/10.1093/femsec/fiw088

Nekola, J. C., & White, P. S. (1999). The distance decay of similarity in biogeography and ecology. Journal of Biogeography, 26 (4), 867-878.https://doi.org/10.1046/j.1365-2699.1999.00305.x

Not, F., Latasa, M., Marie, D., Cariou, T., Vaulot, D., & Simon, N. (2004). A single species, Micromonas pusilla (Prasinophyceae), dominates the eukaryotic picoplankton in the western English Channel.Applied and Environmental Microbiology, 70 (7), 4064-4072.https://doi.org/10.1128/aem.70.7.4064-4072.2004

Pei, G., Chen, L., & Zhang, W. (2017). WGCNA Application to Proteomic and Metabolomic Data Analysis. Methods in Enzymology, 585 , 135-158.https://doi.org/10.1016/bs.mie.2016.09.016

Qi, J., Yin, B., Zhang, Q., Yang, D., & Xu, Z. (2014). Analysis of seasonal variation of water masses in East China Sea. Chinese Journal of Oceanology and Limnology, 32 (4), 958-971.https://doi.org/10.1007/s00343-014-3269-1

Richards, T. A., Jones, M. D. M., Leonard, G., & Bass, D. (2012). Marine Fungi: Their Ecology and Molecular Diversity. In C. A. Carlson & S. J. Giovannoni (Eds.), Annual Review of Marine Science, Vol 4(Vol. 4, pp. 495-522).

Rodríguez‐Martínez, R., Labrenz, M., Del Campo, J., Forn, I., Jürgens, K., & Massana, R. (2009). Distribution of the uncultured protist MAST‐4 in the Indian Ocean, Drake Passage and Mediterranean Sea assessed by real‐time quantitative PCR. Environmental microbiology, 11 (2), 397-408.

Schloss, P. D., Westcott, S. L., Ryabin, T., Hall, J. R., Hartmann, M., Hollister, E. B., . . . Weber, C. F. (2009). Introducing mothur: open-Source, platform-Independent, community-supported software for describing and comparing microbial communities. Applied and Environmental Microbiology

75 (23), 7537-7541.https://doi.org/10.1128/aem.01541-09

Sherr, E. B., & Sherr, B. F. (2009). Food Webs, Microbial. In M. Schaechter (Ed.), Encyclopedia of Microbiology (Third Edition)(3rd ed., pp. 174-189). Oxford: Academic Press.

Sloan, W. T., Lunn, M., Woodcock, S., Head, I. M., Nee, S., & Curtis, T. P. (2006). Quantifying the roles of immigration and chance in shaping prokaryote community structure. Environmental microbiology, 8 (4), 732-740.https://doi.org/10.1111/j.1462-2920.2005.00956.x

Sommeria-Klein, G., Watteaux, R., Ibarbalz, F. M., Karlusich, J. J. P., Iudicone, D., Bowler, C., & Morlon, H. (2021). Global drivers of eukaryotic plankton biogeography in the sunlit ocean. Science, 374 (6567), 594-599.https://doi.org/doi:10.1126/science.abb3717

Somogyi, B., Felföldi, T., Vanyovszki, J., Ágyi, Á., Márialigeti, K., & Vörös, L. (2009). Winter bloom of picoeukaryotes in Hungarian shallow turbid soda pans and the role of light and temperature. Aquatic Ecology, 43 (3), 735-744.https://doi.org/10.1007/s10452-009-9269-0

Stoeck, T., Bass, D., Nebel, M., Christen, R., Jones, M. D., BREINER, H. W., & Richards, T. A. (2010). Multiple marker parallel tag environmental DNA sequencing reveals a highly complex eukaryotic community in marine anoxic water. Molecular Ecology, 19 (s1), 21-31.

Storey, J. D. (2003). The positive false discovery rate: A Bayesian interpretation and the q-value. Annals of Statistics, 31 (6), 2013-2035.https://doi.org/10.1214/aos/1074290335

Sun, P., Wang, Y., Laws, E., & Huang, B. Q. (2020). Water mass-driven spatial effects and environmental heterogeneity shape microeukaryote biogeography in a subtropical, hydrographically complex ocean system - A case study of ciliates. Science of the Total Environment, 706 , 135753.https://doi.org/10.1016/j.scitotenv.2019.135753

Throndsen, J. (1997). The Planktonic Marine Flagellates. In C. R. Tomas (Ed.), Identifying Marine Phytoplankton (pp. 591-729): Academic Press.

Torruella, G., Moreira, D., & Lopez-Garcia, P. (2017). Phylogenetic and ecological diversity of apusomonads, a lineage of deep-branching eukaryotes. Environmental Microbiology Reports, 9 (2), 113-119.https://doi.org/10.1111/1758-2229.12507

Ueda, M., Nomura, Y., Doi, K., Nakajima, M., & Honda, D. (2015). Seasonal dynamics of culturable thraustochytrids (Labyrinthulomycetes, Stramenopiles) in estuarine and coastal waters. Aquatic Microbial Ecology, 74 (3), 187-204.https://doi.org/10.3354/ame01736

Wang, S., Lambert, W., Giang, S., Goericke, R., & Palenik, B. (2014). Microalgal assemblages in a poikilohaline pond. Journal of Phycology, 50 (2), 303-309.https://doi.org/10.1111/jpy.12158

Worden, A. Z., Follows, M. J., Giovannoni, S. J., Wilken, S., Zimmerman, A. E., & Keeling, P. J. (2015). Rethinking the marine carbon cycle: Factoring in the multifarious lifestyles of microbes. Science, 347 (6223), 735-735.https://doi.org/10.1126/science.1257594

Xiong, W., Ni, P., Chen, Y. Y., Gao, Y. C., Shan, B. Q., & Zhan, A. B. (2017). Zooplankton community structure along a pollution gradient at fine geographical scales in river ecosystems: The importance of species sorting over dispersal. Molecular Ecology, 26 (16), 4351-4360.https://doi.org/10.1111/mec.14199

Xu, D. P., Li, R., Hu, C., Sun, P., Jiao, N. Z., & Warren, A. (2017). Microbial Eukaryote Diversity and Activity in the Water Column of the South China Sea Based on DNA and RNA High Throughput Sequencing.Frontiers in microbiology, 8 , 1121.https://doi.org/10.3389/fmicb.2017.01121

Xu, Q., Luo, G., Guo, J., Xiao, Y., Zhang, F., Guo, S., . . . Shen, Q. (2022). Microbial generalist or specialist: Intraspecific variation and dormancy potential matter. Molecular Ecology, 31 (1), 161-173.https://doi.org/10.1111/mec.16217

Xu, Z. M., Li, Y. F., Lu, Y. H., Li, Y. D., Yuan, Z. W., Dai, M. H., & Liu, H. B. (2020). Impacts of the Zhe-Min coastal current on the biogeographic pattern of microbial eukaryotic communities.Progress in Oceanography, 183 , 102309.https://doi.org/10.1016/j.pocean.2020.102309

Yeh, Y. C., Peres-Neto, P. R., Huang, S. W., Lai, Y. C., Tu, C. Y., Shiah, F. K., . . . Hsieh, C. H. (2015). Determinism of bacterial metacommunity dynamics in the southern East China Sea varies depending on hydrography. Ecography, 38 (2), 198-212.https://doi.org/10.1111/ecog.00986

Zhang, Y., Zhao, Z. H., Dai, M. H., Jiao, N. Z., & Herndl, G. J. (2014). Drivers shaping the diversity and biogeography of total and active bacterial communities in the South China Sea. Molecular Ecology, 23 (9), 2260-2274.https://doi.org/10.1111/mec.12739

Zhou, J., Song, X., Zhang, C. Y., Chen, G. F., Lao, Y. M., Jin, H., & Cai, Z. H. (2018). Distribution Patterns of Microbial Community Structure Along a 7000-Mile Latitudinal Transect from the Mediterranean Sea Across the Atlantic Ocean to the Brazilian Coastal Sea.Microbial Ecology, 76 (3), 592-609.https://doi.org/10.1007/s00248-018-1150-z

Zhu, J., Zheng, Q. A., Hu, J. Y., Lin, H. Y., Chen, D. W., Chen, Z. Z., . . . Kong, H. (2019). Classification and 3-D distribution of upper layer water masses in the northern South China Sea. Acta Oceanologica Sinica, 38 (4), 126-135.https://doi.org/10.1007/s13131-019-1418-2

Zhu, Z.-Y., Zhang, J., Wu, Y., Zhang, Y.-Y., Lin, J., & Liu, S.-M. (2011). Hypoxia off the Changjiang (Yangtze River) Estuary: Oxygen depletion and organic matter decomposition. Marine Chemistry, 125 (1-4), 108-116.https://doi.org/10.1016/j.marchem.2011.03.005

Data Accessibility and Benefit-Sharing

The raw sequence data have been deposited in the NCBI Sequence Read Archive (SRA) under the BioProject accession number PRJNA673365 (https://www.ncbi.nlm.nih.gov/bioproject/PRJNA673365).

Benefits Generated: Benefits from this research accrue from the sharing of our data and results on public databases as described above.

Author Contributions

The contributions of each author are listed as follows. Xin Guo: Study design, Investigation, Samples processing, Data analysis, Writing the manuscript, Review and editing. Qiang Liu, Xiaoqing Lin, Xinyi Zheng, Cheng Huang, and Mengwen Pang: Investigation, Samples processing, Review and editing. Lingfeng Huang: Conceptualization, Review and editing, Funding support, Project administration, Supervision.

Tables and Figures

Table 1 Simple and partial Mantel tests were performed to show the effects of water mass (WM), environmental factors (E), and spatial factors (S) on the beta diversity of pico- and nano-sized microbial flagellate communities based on Bray-Curtis distance.