Dinophysis is an obligate mixotroph that relies on consumption of the ciliate, Mesodinium rubrum, to grow and form harmful algal blooms (HABs). In this study, blooms of Dinophysis acuminata in two NY, USA, estuaries were studied over the course of three years (2019-2021) using discrete samples and an Imaging FlowCytobot (IFCB) to capture images of plankton 20-150 mm. The darkness of Dinophysis images on the IFCB was used to quantify the “fullness” or feeding state of Dinophysis cells. Culture experiments performed to ground truth this approach revealed a highly significant correlation (R=0.98; p<0.001) between the darkness of Dinophysis cells and the abundance of Mesodinium. With a quantitative scale developed to track the fullness of Dinophysis cells, ecosystem observations revealed the percentage of “full” Dinophysis cells increased during blooms of a large-morphotype Mesodinium that preceded the initiation of Dinophysis blooms. A smaller morphotype Mesodinium appeared during Dinophysis bloom peaks suggesting they supported bloom maintenance. While the relative abundance of diatoms was elevated before Dinophysis blooms, other dinoflagellates and tintinnids increased in abundance during these HABs indicating they emerged within a consortium of heterotrophs and mixotrophs that may have collectively filled the same open niche as Dinophysis. This study reveals the manner in which different Mesodinium populations co-bloom with Dinophysis to support plastid acquisition, bloom initiation and bloom maintenance and contextualizes these changes within the larger plankton community succession associated with these HABs.

Water from Cold Spring Harbor (40.863374° N, -73.46334° W) and Northport Harbor (40.899593° N, -73.35302° W; Figure S1) was collected weekly prior to the onset of the Dinophysis blooms (May) and continued through the demise of the blooms (July-August) in 2019, 2020, and 2021.

Temperature, salinity, and dissolved oxygen were measured on site with a YSI 556 handheld sonde.

To prepare samples for IFCB analyses, 1 L of water was gently concentrated on a 20 mm sieve to ensure cells were never dried and resuspended into 15 mL or 40 mL of filtered seawater. Concentrated, 3 mL live samples were analyzed with IFCB configured to capture images of targets that exceeded trigger thresholds on light scattering (side angle) or chlorophyll a fluorescence.

Images were analyzed with the ifcb-analysis MATLAB code package (Sosik and Olson 2007, Sosik et al. 2016).

Taxa were identified to lowest possible taxonomic level and were broadly grouped into diatoms, dinoflagellates (excluding Dinophysis), aloricate ciliates (excluding Mesodinium), tintinnids, larval zooplankton, and rotifers. Relative abundances of these taxa as well as Mesodinium and Dinophysis were calculated relative to the total plankton count. Two types of Mesodinium were identified and classified based on size (small: ~10 mm, large: ~25mm).

Used the feature "texture_average_gray_level" (TAGL) as a proxy for Dinophysis feeding. The TAGL can rangefrom 0-255 where 0 is black and 255 is white. All Dinophysis images during this study ranged from 51 to 204 on this scale and were, therefore, binned into three levels that provided near equal observations across the bins; 51-100 (full), 100-110 (partially full), 110-204 (partially empty; Figure 1). The average TAGL for each Dinophysis imaged on each date was then sorted into one of the three levels and the percentage of the total images that were within each level was determined.

A Dinophysis acuminata strain isolated from Meetinghouse Creek, NY, (DANY; Hattenrath-Lehmann and Gobler, 2015; Hattenrath-Lehmann et al., 2021; Ayache et al., 2023) was fed Mesodinium rubrum (strain JAMR) that was fed Teleaulax amphioxia (strain JATA), both isolated from Inokushi Bay, Oita Prefecture, Japan, in 2007 (Nishitani et al., 2008). All cultures were kept in f/2 media at salinity 25 and18 °C with ~80 µEin m^-2 s^-1 on a 12:12 light:dark cycle. DANY was fed JAMR ad libitum prior to the first day of IFCB sampling. After one week, all Mesodinium were removed from the culture with a 20-mm sieve. IFCB samples were obtained and analyzed in triplicate using side scattering and chlorophyll a triggering modes concurrently the day following the last feeding and weekly for five weeks. The TAGL for each Dinophysis image was then analyzed for each week following the procedure described above.