Seasonal and spatial patterns of eukaryotic phytoplankton communities in an urban river based on marker gene

Effects of seasonality on physiochemical properties of water

The characteristics of the physical and chemical factors of the river are shown in Table 1. Variance analysis showed that the environmental conditions during the four seasons were significantly different (P < 0.01), indicating the seasonal changes can affect the water quality in the Fenhe River. Although the spatial changes were not significantly different, the samples from downstream sites were separated from those from the upstream and midstream sites during the four seasons (Fig. 1). In addition, the levels of nitrite, phosphate, and DOC in the upstream sites were lower than those in the downstream sites, suggesting that the water quality in the former was better than that in the latter.

Phytoplankton diversity and richness analysis

The total raw reads of all samples ranged from 782,712 (autumn) to 1,547,327 (summer). After removing chimera, the number of reads in each sample ranged from 41,526 to 85,363, with a mean value of 61,629. The temporal and spatial richness estimates and diversity indices are shown in Fig. 2. The Simpson diversity index ranged between 0.08 and 0.35 with an average value of 0.15 ± 0.06, and the Shannon diversity index varied from 1.86 to 3.17 with a mean value of 2.65 ± 0.37. The Shannon and Simpson diversity indices of winter were significantly different (P < 0.001) from those of other seasons, whereas the Chao and ACE indices of all the four seasons were significantly different (P < 0.001). Overall, the diversity value in summer and autumn was equivalent, while that in winter was lowest due to its lower temperature, which is not suitable for the growth of algae. Although there were no obvious differences between the sampling sites, the upstream sites had high Shannon index but low Simpson index in spring, indicating that phytoplankton communities in these sites had high diversity.

Differences in phytoplankton communities analyzed based on LEfSe

The seasonal and spatial differentially abundant taxa (i.e., the biomarkers) from the phylum to the genus level identified by LEfSe analysis are shown in Fig. 3. The linear discriminant analysis (LDA) score describes the degree to which the relative abundance of various microbial groups in given microbial communities consistently changes between seasons²⁴. Our results emphasized that family and genus were main the potential biomarkers in different seasons were identified and found four to seven phytoplankton taxa unique to each season. The phytoplankton groups that were enriched in spring included Mediophyceae, Stephanodiscales, Scenedesmaceae, Desmodesmus, and Cyclotella, and their LDA scores were higher than 4.5. The order Chlamydomonadales, the family Characiaceae, and the genus Pseudoschroederia were the biomarkers that preferred summer; their LDA scores were > 4.5. Fewer phytoplankton groups were enriched in autumn, Discostella and Scenedesmus; their LDA scores were higher than 4.5. The taxa enriched in winter samples mainly included Eustigmatales at the order level, Neochloridaceae and Monodopsidaceae at the family level, Monoraphidium and Nannochloropsis at the genus level. In spatial analysis, the biomarkers enriched at the upstream sites included the Thalassiophysales at the order level, Catenulaceae at the family level, Amphora, Neochloris, and Hindakia at the genus level. Trebouxiophyceae ordo incertae sedis (order), Coccomyxaceae and Pseudomuriellaceae (families), Pseudomuriella, Coccomyxa, and Chloroidium (genus) were enriched in the midstream sites. The downstream sites were mainly enriched with the members of the phyla Chlorophyta, the order Scotinosphaerales, the families Chromochloridaceae and Scotinosphaeraceae, and the genera Scherffelia, Chromochloris, and Scotinosphaera.

Phylogenetic analysis and distribution characteristics of phytoplankton communities

In this study, the first 50 representative sequences of samples collected in winter were different from those of samples collected in other seasons. Therefore, two phylogenetic evolution trees were constructed (Figs. 4 & 5). Firstly, we used the 50 most abundant OTUs to blast against the NCBI database, downloaded each of the most relevant sequences with the clearest taxonomic annotations, and then clustered them with our 50 OTUs. As shown in Fig. 4, 36 OTUs in Chlorophyta formed a monophyletic clade with sequences belonging to 27 genera. In Bacillariophyta, 7 OTUs (OTU1268, OTU708, OTU938, OTU941, OTU1252, OTU1170, and OTU1096) formed a monophyletic clade with sequences belonging to 2 genera, Cyclotella and Discostella, indicating that the species in these two genera were highly abundant. At the same time, we found that OTU1170 had the closest relationship with the Peridiniopsis jiulongensis H.Gu (now is accepted taxonomically as Unruhdinium jiulongense (H.Gu) Gottschling) diatom endosymbiont, which suggests that OTU1170 may be a Peridiniopsis diatom endosymbiont sequence. Besides, OTU1170 was clustered with OTU1252 and OTU1096, which belonged to Discostella. Previous studies have shown that the endosymbionts of Peridiniopsis may originate from the Discostella-like species²⁵, indicating the two are closely related. For the phylogenetic analysis of Ochrophyta, 7 OTUs were closely related to the genera such as Nannochloropsis, Eustigmatos, Tetraëdriella, and Goniochloris, which have not or hardly been reported to be present in the Fenhe River, reflecting that there are substantially more species of Ochrophyta in this river.

As illustrated in Fig. 5, the abundance of some genera such as Neochloris, Botryosphaerella, Schroederia, and Chromochloris was higher in winter than in other seasons. For Ochrophyta, 6 OTUs formed the monophyletic clade with sequences belonging to 4 genera, including Vischeria, Nannochloropsis, Pseudotetraëdriella, and Vacuoliviride. Two OTUs from Cryptophyta were clustered with Cryptomonas obovoidea Pascher and Hemiselmis cryptochromatica C.E.Lane & J.M.Archibald, respectively, with a support rate of 100%.

To find the most widely distributed genera in the Fenhe River, a heatmap showing the relative abundance of 50 genera was plotted (Fig. 6). The heatmap clearly showed that the genus Desmodesmus, Cyclotella, and Fragilaria had the highest relative abundance and were most widely distributed in spring, which is in agreement with the occupancy-abundance relationship. In summer, the average proportions of Pseudomuriella, Pseudopediastrum, Rotundella, and Oocystis were higher than those of other genera at the same sites. Mychonastes, Scenedesmus, Neodesmus, Follicularia, and Schroederia were widely distributed in autumn. The Cryptomonas genus at site S2 reached its peak during winter.

Relationships between phytoplankton communities and physicochemical parameters

The dynamic pattern of phytoplankton communities along the Fenhe River can be influenced by the fluctuations of physical and chemical variables. In general, specific phytoplankton at the different taxonomic levels (mainly phylum, class, order, family, and genus) were related to certain environmental variables (see Supplementary Table S1 online). For instance, at the phylum level, the variations of most members of the Bacillariophyta and Ochrophyta were significantly associated with both physical and chemical properties (P < 0.01 and P < 0.05). Although the Chlorophyta had no significant correlation with these parameters, at the class level, Chlorophyceae was positively associated with all studied environmental variables, both the physical and chemical properties. Taken together, the phytoplankton communities in the Fenhe River were influenced by the seasonal changes of physical and chemical factors.

To further determine crucial environmental parameters causing the seasonal changes of phytoplankton, we employed Pearson correlation analysis to analyze the relationship between the relative abundance of representative taxa and water quality. Among the effects of environmental variables upon phylotypes at different classification levels, the influence of temperature, phosphate, and DOC on phytoplankton was extensive. For example, the relative abundance of the phyla Bacillariophyta, the class Chlorophyceae, the order Sphaeropleales, and the families Mychonastaceae, Radiococcaceae, Scenedesmaceae, and Selenastraceae was significantly correlated with temperature, phosphate, and DOC concentration (P < 0.05). Among these taxa, the family Selenastraceae was significantly negatively correlated with temperature, phosphate, and DOC concentration (P < 0.05). Additionally, temperature and phosphate concentration had negative correlations with the relative abundance of the phylum Ochrophyta (including the class Eustigmatophyceae and the order Eustigmatales). Overall, phytoplankton phylotypes associated with the temperature, phosphate, and DOC covered a wider range of taxonomic resolutions than nitrate and nitrite, suggesting that carbon and phosphorus are more likely to be the factors affecting the formation of phytoplankton communities in the Fenhe River than nitrogen.

Phytoplankton functional groups and driving factors

The proportion of phytoplankton functional groups divided according to the references ^19,36 is shown in Fig. 7. Group J was dominant in spring, summer, and autumn, accounting for 54.97%, 50.25%, and 48.29%, respectively (Fig. 7a). Besides, group F also presented high proportions, its average proportion accounted for 21.54% in spring, 29.18% in summer, and 46.24% in autumn. Group X1 occurred during all seasons, however, the proportion was particularly high in winter. Groups B and X2 were also dominant in winter compared to other seasons.

The RDA ordination diagram of environmental variables and phytoplankton functional groups is shown in Fig. 7b. The first two axes explained 82.89% and 3.87% of the cumulative variance of the relationship of species-environmental variables, and the eigenvalues were 0.8289 and 0.0387, respectively (Table 2). The species-environment correlation values of axis 1 and axis 2 were 0.9807 and 0.6262, respectively. Monte Carlo permutation tests showed that environmental factors such as water temperature (F = 46.3, P = 0.002), nitrite (F = 13.5, P = 0.002), and phosphate (F = 3.8, P = 0.014) were the main factors regulating the phytoplankton functional groups. In Fig. 7b, axis 1 was positively correlated to pH while negatively correlated to water temperature. Axis 2 was negatively correlated with nitrite, phosphate, and DOC. Moreover, groups X1 were greatly affected by pH, and group J showed a strong positive correlation with temperature. Groups C and Xph showed different degrees of positive correlation with phosphate.

Cluster analysis based on environmental variables (Fig. 1) showed that the upstream and midstream samples were clustered into a branch, however, they were separated from the downstream samples, indicating that the upstream and midstream environments were similar, and showed certain environmental differences from the downstream. The variance partitioning analysis also showed the same spatial difference (Table 3). The pH could explain 41.8% (P < 0.01) of the changes in phytoplankton functional groups of upstream and midstream, followed by phosphate (15.6%) and nitrite (10.1%). Water temperature had the largest explanation rate (39.4%) (P < 0.05) to phytoplankton functional groups at the downstream of river, followed by phosphate (36.5%), and that the explanation rate of pH was relatively low (12.2%).

Source: Ecology - nature.com