Short-term dynamics and interactions of marine protist communities during the spring-summer transition

Abstract

We examined the short-term variability, by daily to weekly sampling, of protist assemblages from March to July in surface water of the San Pedro Ocean Time-series station (eastern North Pacific), by V4 Illumina sequencing of the 18S rRNA gene. The sampling period encompassed a spring bloom followed by progression to summer conditions. Several protistan taxa displayed sharp increases and declines, with whole community Bray-Curtis dissimilarities of adjacent days being 66% in March and 40% in May. High initial abundance of parasitic Cercozoa Cryothecomonas longipes and Protaspis grandis coincided with a precipitous decline of blooming Pseudo-nitzschia diatoms, possibly suggesting their massive infection by these parasites; these cercozoans were hardly detectable afterwards. Canonical correspondence analysis indicated a limited predictability of community variability from environmental factors. This indicates that other factors are relevant in explaining changes in protist community composition at short temporal scales, such as interspecific relationships, stochastic processes, mixing with adjacent water, or advection of patches with different protist communities. Association network analysis revealed that interactions between the many parasitic OTUs and other taxa were overwhelmingly positive and suggest that although sometimes parasites may cause a crash of host populations, they may often follow their hosts and do not regularly cause enough mortality to potentially create negative correlations at the daily to weekly time scales we studied.

Fig. 1

Time series dot plot showing the most abundant eukaryote OTUs, each displaying a relative abundance ≥2% in at least one sample. Note uneven sampling intervals, and symbols along the top showing community clustering (Fig. 2). Letters between quotation marks correspond to inferred trophic properties of selected OTUs based on closest known relatives: A autotroph, H heterotroph, P parasite, C with chloroplast (relatives include mixotrophs), U unassigned. The area in gray color corresponds to the temporal dynamics of chlorophyll a concentration during the studied period. To show the relative differences in sequence abundance for the displayed OTUs, the circle size is on a 0-1 scale representing proportions. SG stands for “Syndiniales Group”

Fig. 2

Protist communities clustered into five broad community types, at >20% Bray–Curtis similarity. Dendrogram from UPGMA clustering of Bray–Curtis dissimilarity based on the 300 most abundant OTUs. * duplicate DNA extract. Pie chart plots indicate high-ranking taxonomy distribution of the clustered protist community groups. The classification is based on Adl et al. [83]

Fig. 3

Canonical correspondence analysis of protist community composition and physicochemical parameters. Temp: Temperature (°C), PO₄³⁻: Orthophosphate (uM), SiO₃²⁻: Silicate, DPD: dominant wave period (s), MWD: direction from which DPD is coming (degree), MVHT: significant wave height (m), APD: average wave period (seconds), WPS: Wrigley precipitation sum (cm), WSPD: Wrigley (station) wind speed average (km hr⁻¹). Wrigley refers to the weather station at Wrigley Marine Science Center on Santa Catalina Island

Fig. 4

Association networks derived from the 300 most abundant protists and showing only statistically significant correlations between OTUs (Spearman correlation ≥│0.6│; p < 0.01; false discovery q < 0.05). Sub-network (a) included a total of 103 unique nodes and 348 edges (correlations) among nodes, and Sub-structural network (b) included a total of 125 unique nodes and 643 edges. Node sizes correspond to the number of neighbors (i.e., degree). Nodes colors and shapes correspond to main protist groups and their ecological behavior, respectively. Solid lines represent positive correlations and dashed lines, negative correlations