Particle-associated bacteria differentially influence the aggregation of the marine diatom Minutocellus polymorphus

The aggregation of phytoplankton leads to the settling of particulate organic carbon in the form of marine snow, making it an important process in marine biogeochemical cycles. Diatoms >20 µm in size are considered to contribute appreciably to sinking particle fluxes due to aggregation and the production of transparent exopolymeric particles (TEP), the matrix for marine snow aggregates; however, it is not known whether nano-sized (2–20 µm) diatoms are able to aggregate and produce TEP. Here, we tested the aggregation and production of TEP by the nano-diatom Minutocellus polymorphus and investigated if interactions with bacteria influence aggregation by comparing axenic M. polymorphus cultures with co-cultures of the diatom with bacterial taxa known to colonize marine snow particles. We found that M. polymorphus form sinking aggregates and produce TEP comparably to other phytoplankton groups and that aggregation and TEP production were influenced depending on the species of bacteria added. Aggregation was enhanced in the presence of Marinobacter adhaerens HP15, but not in the presence of Pseudoalteromonas carrageenovora or Vibrio thalassae. Cell aggregation mediated by interactions with specific bacterial species are possible mechanisms behind the export of nano-sized diatoms, such as M. polymorphus, especially in oligotrophic open ocean regions where small phytoplankton dominate.


INTRODUCTION
Diatoms are an abundant and ubiquitous class of phytoplankton responsible for contributing up to 40% of the ocean's primary productivity [1]. Their blooms are often followed by aggregation and subsequent sedimentation events that account for~40% of the total export of particulate organic carbon (POC) [2], thus, they are considered significant contributors to export fluxes and global biogeochemical cycles. In biogeochemical models, diatoms are frequently considered to be micro-phytoplankton (>20 µm), but, many ubiquitous species overlap with the nano-(2-20 µm) and picoplanktonic (≤2 µm) size-classes [3]. Due to their minute size, nano-and pico-sized diatoms are typically considered to contribute negligibly to particle fluxes [4]; however, there is increasing evidence against this notion [5][6][7]. Observations of abundant frustules belonging to Minidiscus trioculatus (ca. 2 µm) in sediment traps deployed in the Northwestern Mediterranean at 2400 m [8] as well as the consistent representation of M. trioculatus and Minutocellus polymorphus (ca. 4-6 µm) in sizefractionated water samples (≥180 µm) collected at 700 m depth [9] suggest high sinking velocities as a result of their incorporation into sinking particles and/or aggregation mechanisms. Furthermore, DNA-based analyses of particle trap material from the Sargasso Sea have demonstrated significant overrepresentation [5] and specificity [10] of these small diatom genera in sinking particles compared to the ambient seawater, further suggesting an underestimation of their role in particle fluxes, especially in open ocean regions, where small phytoplankton are most abundant [11].
Hypothesized mechanisms of export of pico-and nanophytoplankton include aggregation and subsequent increase of their effective size [7,12] which is dependent on the abundance, size, sticking efficiency, and relative sinking velocities of surrounding particles [13]. Particle sticking efficiency can be a result of cell surface properties [13] and/or cellular exudates [14,15] including exopolysaccharides such as the transparent exopolymeric particles (TEP). TEP further facilitate aggregation and subsequent settling of cells during phytoplankton blooms and are a significant component of the matrix of marine snow aggregates [16,17]. Phytoplankton are considered the main producers of TEP [18], and the presence of bacteria has been shown to increase this production [19][20][21][22], suggesting interactions that may have significant implications in aggregation and enhanced sinking rates of senescent phytoplankton. In addition to enhancing the production of TEP by phytoplankton, some bacteria can produce TEP [23], as well as consume and/or modify its precursors [24,25]. As a result, bacteria-mediated aggregation in phytoplankton can vary with the ambient bacterial species.
To explain the observations suggesting a significant contribution of nano-sized diatom cells to sinking particles and their export fluxes, controlled laboratory experiments are necessary to test the potential mechanisms behind the export of these abundant, yet often-overlooked, diatom genera. In this study, our objectives were: (1) to investigate whether the marine nanodiatom Minutocellus polymorphus forms aggregates and produces TEP and (2) to determine if the presence of distinct marine bacteria distinct marine bacteria enhance the aggregation of M. polymorphus. We chose a two-phase approach with a first set of experiments focusing on the growth phase of diatoms and the second set targeting aggregation and sedimentation at the end of a bloom (Table 1). Experiments investigated the diatom M. polymorphus in combination with one of three known particleassociated bacteria [10,26]. We hypothesized that (1) M. polymorphus forms fast-sinking aggregates within a matrix of TEP, and (2) that aggregation and TEP production by M. polymorphus is enhanced in the presence of bacteria.

Diatoms and bacterial isolates
Stock cultures of marine Minutocellus polymorphus [27] (CCMP497, National Center for Marine Algae and Microbiota, NCMA) were maintained in L1 medium [28] prepared in artificial seawater [29]  TEP and micro-aggregate formation during growth Experimental conditions. Treatments consisted of triplicate cultures of axenic M. polymorphus as a bacteria-free negative control treatment and M. polymorphus in co-culture with each of the three bacterial species (Table 1). The co-culture experiment with M. adhaerens (growth experiment 1) was performed independently from that with V. thalassae and P. carrageenovora (growth experiment 2), each with their own triplicate negative control treatments of axenic M. polymorphus (Table 1). Exponentially growing M. polymorphus cells were used to inoculate all treatments (ca. 10 5 cells mL −1 ). Bacterial cell concentrations were about an order of magnitude higher than diatom concentrations. Experimental cultures were grown in the same conditions as the stock cultures. 100 mL samples were taken every other day and immediately used to determine the volume of micro-aggregates in cultures as described below. Samples were then preserved in glutaraldehyde (1% [v/v] final concentration, Sigma-Aldrich) for TEP measurements and for the quantification of singlecell abundances in cultures. All axenic treatments were tested for bacterial and fungal contamination prior to experiment inoculation and at the end of all experimental periods by inoculation in L1 medium with added peptone and methylamine-HCl, as suggested by the NCMA. Axenic M. polymorphus treatments were bacteria-free throughout the experiments. Co-culture treatments were tested for contamination of non-cocultured bacteria at every sampling period by epifluorescence microscopic observation of cell morphologies, as well as colony morphologies by streaking onto Marine Agar (BD Difco 2216, Becton Dickinson, NJ; ZoBell 1941) plates at days 1, 10, and following the end of the experimental periods.
Volume concentrations of micro-aggregates "suspended" in cultures (i.e., non-sinking particles with an equivalent spherical diameter [ESD] of 5-60 µm) were determined at every sampling period using a Multisizer 3 Particle Counter (Beckman Coulter, CA). Prior to fixation with glutaraldehyde, duplicate samples were diluted to a 1-10% final particle concentration with Isoton II diluent (Beckman Coulter, CA) and aggregates were Minutocellus polymorphus were incubated with or without the addition of bacteria in flasks and sampled throughout their growth (representative of bloom conditions), to determine the production of transparent exopolymeric particles (TEP) and the formation of micro-aggregates, and in roller tanks to investigate the formation of sinking aggregates (representative of end of bloom conditions). The growth experiment with the addition of Marinobacter adhaerens (growth experiment 1) was carried out independently from that with the addition of Pseudoalteromonas carrageenovora and Vibrio thalassae (growth experiment 2). Roller tank experiments were performed with total cell numbers of 10 3 cells mL −1 (lower abundance) and 10 5 cells mL −1 (higher abundance).
measured and quantified with a 100 µm aperture tube. The volume concentration of aggregates was calculated in µm 3 per mL.
TEP. TEP concentrations in the co-cultures and axenic cultures were determined as described by Passow and Alldredge [30]. 10 mL of glutaraldehyde-fixed culture samples were filtered through duplicate 0.4 µm pore-size polycarbonate membranes (GVS Life Technologies, ME) at a low and constant vacuum pressure (100 mm Hg). The retained TEP was subsequently stained with 0.5 mL of the acidic polysaccharide-specific Alcian Blue (AB) dye (8GX, Sigma-Aldrich), followed by a 0.5 mL rinse with MilliQ water for the removal of excess stain and stored at −40°C until analysis. Prior to staining, the pre-calibrated [31] 0.02% (w/v) AB working solution pH-adjusted with 0.06% (v/v) acetic acid (final pH 2.5) was passed through a 0.2 µm Acrodisc syringe filter (Pall Corporation, NY) to remove the undissolved dye. Membranes were soaked in 6 mL of 80% (v/v) sulfuric acid for 3 h to extract the AB-stained TEP and absorption was then measured using a spectrophotometer (Shimadzu UV-1900i, Shimadzu, Kyoto, JP) at 787 nm. Duplicate stained filters with sterile media functioned as blanks. TEP concentrations were calculated using a calibration factor of the AB dye determined with xanthan gum (f-factors: 81.70 for experiments with M. adhaerens and 83.83 for experiments with V. thalassae and P. carrageenovora) and expressed in µg of xanthan gum equivalent units (µg XG eq.) as described by Bittar et al. [31]. To compare TEP concentrations between treatments and with other phytoplankton groups, concentrations were normalized to diatom cell abundances and biovolumes, respectively. TEP production rates were calculated as previously described [32].

Aggregation in roller tanks
To investigate the formation of visible aggregates by M. polymorphus with and without the presence of bacteria, diatom and bacteria co-cultures were incubated in roller tanks that simulate the natural collision of particles as they would occur in situ [33]. M. polymorphus pre-cultures were incubated until the late exponential phase of growth was achieved (ca. 10 days). Bacteria were grown overnight in Marine Broth prior to being washed three times in artificial seawater. Bacteria were then incubated in artificial seawater for 24 hours prior to inoculation within roller tanks to minimize nutrient carry-over. M. polymorphus pre-cultures were diluted to cell abundances simulating spring conditions in the Sargasso Sea as observed by Cruz and Neuer [20] (ca. 10 3 cells mL −1 , hereafter termed "lower cell abundance experiments"; Table 1) and incubated in cylindrical 1.25 L Plexiglass roller tanks [33] with artificial seawater [29]. Axenic M. polymorphus were inoculated at a final cell concentration of 10 3 cells mL −1 and co-culture treatments were inoculated with M. polymorphus cells at 10 2 cells mL −1 and bacteria at 10 3 cells mL −1 . Axenic bacterial treatments were inoculated at a final concentration of 10 3 cells mL −1 (Table 1). Control treatments consisted of the following: axenic M. polymorphus, bacteria-only (V. thalassae, P. carrageenovora, or M. adhaerens), and the addition of sterile silica microspheres (2.5 µm, 1.4 × 10 9 beads mL −1 stock concentration; Bangs Laboratories, IN) to axenic M. polymorphus cell suspensions (final concentrations: 10 2 M. polymorphus cells mL −1 ; 10 3 beads mL −1 ) and to bacteria-only cell suspensions (final concentrations: 10 2 beads mL −1 and 10 3 bacterial cells mL −1 ). An experiment with higher total cell abundances (ca. 10 5 cells mL −1 , hereafter termed "higher cell abundance experiments"; Table 1) was performed with the addition of M. adhaerens, composed of axenic treatments with M. polymorphus or bacteria and treatments with M. polymorphus and the addition of M. adhaerens (Table 1). Duplicate tanks for each treatment were rotated on a rolling platform at 3.5 rotations per minute in the dark at 24°C for 7 days and checked daily for aggregate formation. At the end of the incubation period, the number of visible aggregates (ca. > 0.5 mm) formed in each tank was counted and photographs of aggregates were taken within the roller tanks. The ESD of each imaged particle was then determined using ImageJ image analysis software (http://rsb.info.nih.gov/ij/). To determine sinking velocities [34], aggregates from each roller tank were gently transferred with a wide-bore pipette into a 1 L graduated cylinder and released at 1 cm under the airwater interface. The cylinder was filled with artificial seawater at the same salinity and temperature as roller tanks, and the settling time of each aggregate was determined through a vertical distance of 32.6 cm. Settling times were subsequently converted to velocities in meters per day.

Statistical analyses
Data are presented as the means of triplicate cultures with standard deviations. Two-sample Student's t-tests were performed to check for statistically significant differences in the formation of micro-aggregates and the production of TEP between axenic and co-culture treatments in growth experiments 1 and 2. Kruskal-Wallis corrected for multiple comparisons with a post-hoc Dunn's test were used to test for significant differences in the formation of sinking aggregates between treatments in the roller tank experiments. Statistical analyses were performed using R [35] software.

Data deposition
All experimental data are as well as their associated metadata are available through the Biological and Chemical Oceanography Data Management Office (https://bco-dmo.org/dataset/876461).

RESULTS
TEP and micro-aggregate formation during growth Cell growth. M. polymorphus cells grew for the first 9-11 days of incubation in axenic and co-culture treatments, with maximum growth rates of 0.6 ± 0.  (Fig. 3). P. carrageenovora-only aggregates were observed in co-cultures throughout the experimental period (Fig. 3h).
TEP. TEP accumulated in co-cultures and axenic treatments as the abundance of M. polymorphus cells increased (Figs. 1d and 2e).
In axenic M. polymorphus cultures, most of the production of TEP occurred during periods of exponential growth (Table 2), with concentrations increasing from 2.1 ± 0.8 µg XG eq. mL −1 on day 1 to 4.1 ± 1 µg XG eq. mL −1 on day 9 for experiments with M. adhaerens (Fig. 1d). In experiments with V. thalassae and P. carrageenovora, TEP concentrations in axenic M. polymorphus cultures increased from 1.3 ± 0.8 µg XG eq. mL −1 on day 1 to 3.6 ± 0.6 µg XG eq. mL −1 on day 11 (Fig. 2e). To compare between treatments with differing cell abundances, we normalized TEP concentrations to diatom cell numbers. Maximum cell-normalized TEP concentrations in axenic M. polymorphus and in co-cultures occurred during the exponential phase ( Table 2). The addition of V. thalassae led to higher cell-normalized TEP concentrations during the decline phase of the co-cultured M. polymorphus (t-test, p = 0.01; Table 2, but not in the exponential (p = 0.4) or stationary (p = 0.07) phases. There were no differences in TEP production rates between axenic treatments and co-cultures with V. thalassae (t-test, p > 0.05 for all growth phases; Table 2). We also found no significant differences in the TEP production rates or cellnormalized TEP concentrations between axenic cultures and cocultures with M. adhaerens (t-test, p > 0.05 for all growth phases; Table 2). Compared to axenic treatments, cell-normalized TEP concentrations were significantly higher in co-cultures with P. carrageenovora during the exponential growth period of the co-cultured M. polymorphus (t-test, p = 0.04; Table 2), but significantly lower during the decline periods of the co-cultured M. polymorphus (t-test, p = 0.04; Table 2). TEP production rates were also significantly lower in co-cultures with P. carrageenovora during the stationary phase of M. polymorphus (t-test, p = 0.04; Table 2). These lower TEP production rates and cell-normalized TEP concentrations occurred in parallel to increases in P. carrageenovora abundances (Fig. 2b).

Aggregation in roller tanks
In addition to quantifying the development of suspended micro-aggregates in batch cultures, we performed roller tank experiments to enhance the formation of visible, sinking aggregates. M. polymorphus formed visible aggregates without the addition of bacteria in all experiments, with fewer aggregates (0-2 aggregates) formed in experiments with a lower total cell abundance (10 3 cells mL −1 ) compared to the experiments containing a higher total cell abundance (0-5 aggregates; 10 5 cells mL −1 ; Fig. 4; Tables 3 and 4). Axenic M. polymorphus aggregates in the higher cell abundance experiments were brown in color, compact, and resilient to handling for the determination of sinking velocities, in contrast to aggregates formed with the addition of M. adhaerens and those formed in the lower cell abundance experiments, which were white, amorphous, and much more fragile. Aggregates in the higher cell abundance experiments had ESD's of 1.7-2.2 mm (n = 3 aggregates; Table 4) and sank at velocities of 225-485 m d −1 (n = 5 aggregates; Table 4). The addition of M. adhaerens led to the formation of a significantly higher number of aggregates compared to axenic M. polymorphus in both experiments-for instance, in the higher cell abundance experiments, 0-5 aggregates formed in axenic treatments, while 9-16 aggregates formed with the addition of M. adhaerens (Kruskal-Wallis and post-hoc Dunn's test, p = 0.04, n = 3 tanks; Fig. 4; Tables 3 and 4). No aggregation was observed with the addition of V. thalassae or P. carrageenovora ( Fig. 4a; Table 3). Compared to treatments with axenic M. polymorphus and with axenic bacteria, there was no significantly enhanced aggregation with the addition of 2.5 µm silica beads (Kruskal-Wallis and post-hoc Dunn's test, p > 0.05 for all; Fig. 4; Table 3).

DISCUSSION
Our results support the hypothesis that the marine nano-diatom Minutocellus polymorphus forms fast-sinking (up to 485 m d −1 ) aggregates within a matrix of TEP. The sinking rates of these aggregates were comparable to those formed by bloom-forming diatom genera that aggregate readily in laboratory conditions and are associated with events of high POC flux, such as Chaetoceros spp., Navicula spp., and Skeletonema spp., with sinking rates ranging between ca. 100 and 400 m d −1 (17,36,37). Our findings support that cell aggregation is the likely mechanism leading to the observations of nano-and pico-sized diatoms, such as M. polymorphus, in sediment traps deployed in the Sargasso Sea [5,10] and in the Mediterranean [8].
Despite the lower abundance of cells in roller tanks compared to other aggregation studies (10 3 in our lower cell abundance experiments vs. up to 10 6 cells mL −1 ; 20, 36-38), M. polymorphus formed visible, sinking aggregates. As aggregation is highly dependent on the number of surrounding particles and the probability with which they adhere after collision (stickiness; α) [13], a likely explanation for the formation of aggregates despite relatively low particle numbers is the high α of M. polymorphus due to the production of TEP. Cell volume-normalized TEP concentrations in M. polymorphus cultures were two times higher than in cultures of other diatom species known to readily produce TEP and sticky particulate mucus [15] (Table S1). These diatoms are known to contribute appreciably to phytoplankton blooms and therefore contribute to events of high POC flux [39,40]. Likewise, M. polymorphus bloom at cell abundances of up to 10 3 cells mL −1 and predominate in sediment trap material relative to the populations in the euphotic zone [5,10], the latter possibly being a result of the formation of TEP-rich, yet fast-sinking aggregates as seen in our study. The sinking rates for M. polymorphus aggregates shown herein would allow for cells within aggregates to escape grazing and remineralization through the microbial loop in the water column, contrary to the common assertion that they are entirely remineralized because of their small size and low sinking rates. In nature, the formation of TEP-rich, sticky aggregates by M. polymorphus would likely scavenge additional ballasting material, such as biogenic or lithogenic minerals [41], further enhancing the sinking velocities of their aggregates.
We also found that the aggregation and TEP production of M. polymorphus is differentially enhanced by the species of bacteria present. V. thalassae and P. carrageenovora were pre-selected based on indicator species analyses [42,43] performed on 16S rRNA libraries of bacterial communities on sinking aggregates collected from the mesopelagic of the Sargasso Sea, whereby Vibrio spp. and Pseudoalteromonas spp. were specific to sinking particles [10]. A third species, Marinobacter adhaerens HP15 [26,44], was selected based on their propensity to attach to and enhance the aggregation of the diatom Thalassiosira weissflogii [19]. When co-cultured with M. polymorphus, M. adhaerens HP15 caused a significant increase in the formation of suspended micro-aggregates (5-60 µm; Table 2) and of  Tables 3 and 4), similar to its enhancement of aggregation in the diatom Thalassiosira weissflogii found by Gärdes et al. [19]. However, co-cultures with M. adhaerens did not produce more TEP than axenic treatments, in contrast with the increase of TEP in co-cultures with T. weissflogii as reported by Gärdes et al. [45]. It is possible that TEP and/or its precursors were simultaneously being consumed and exuded by M. adhaerensultimately leading to a pool of bacteria-derived exopolymeric substances and/or TEP that are stickier than those derived from the diatoms [46][47][48], and to the observed enhanced formation of suspended micro-aggregates without increased TEP concentrations (Fig. 1c, d; Table 2). Elucidating the compositional differences of exopolymeric substances and investigating the production of other exopolymers that could also play a role in aggregation would help explain the mechanisms for the aggregation observed between M. polymorphus and M. adhaerens. Contrastingly, the addition of P. carrageenovora (a member of the Gammaproteobacteria) did not lead to enhanced aggregation in roller tanks ( Fig. 4; Table 3), but to   Table 4). Lowercase letters indicate significant differences between treatments.   an enhanced volume concentration of suspended micro-aggregates (Fig. 2d, Table 2) along with lower cell-normalized TEP concentrations compared to axenic diatom cultures ( Table 2). Members of the Gammaproteobacteria are able to hydrolyze complex polysaccharides [49]. Specifically, Pseudoalteromonas spp. readily attach to [50] and degrade a wide-range of phytoplankton-derived polysaccharides, including the relatively-recalcitrant fucose-containing sulphated polysaccharides [51,52], by utilizing many different carbohydrate-active enzymes [51]. Therefore, M. polymorphusderived TEP was likely being consumed by P. carrageenovora which is also supported by the decrease in TEP concentrations that occurred parallel to increases in P. carrageenovora cell abundances (Fig. 2b, e). As known colonizers and degraders of marine particles and phytoplankton-derived organic matter [51,53,54], the greater volume concentration of suspended micro-aggregates in P. carrageenovora treatments during the stationary and decline phases of the co-cultured M. polymorphus (Table 2) may be a result of aggregation of P. carrageenovora cells around free TEP and/or around M. polymorphus aggregates in order to consume the diatomderived TEP (Figs. 2d and 3h). Lastly, no enhanced aggregation was observed with the addition of V. thalassae to M. polymorphus in roller tanks or cultures; however, TEP was significantly higher, coinciding with an increase in bacterial abundances. V. thalassae may have produced TEP or enhanced the production of TEP by M. polymorphus within co-cultures, but the lack of increased suspended micro-aggregates indicates a decoupling with TEP production. This finding also suggests other mechanisms of bacteria-mediated aggregation besides influencing TEP production and/or modifying its precursors, as was found with M. adhaerens. These results show that various TEP and aggregation-influencing interactions may occur within a natural bacterial community. In the environment, these interactions would be more complex as a result of interactions between diverse bacterial and phytoplankton communities, their grazers, as well as the influence of nutrient availability and other environmental factors [45] on these interactions. Cell biovolume-normalized TEP concentrations of M. polymorphus cultures were an order of magnitude lower than those of Synechococcus sp. as found by Cruz and Neuer [20] (Table S1). Despite these lower TEP concentrations, M. polymorphus formed sinking aggregates in roller tanks, in contrast to Synechococcus, which do not form aggregates in axenic conditions unless lithogenic material is added [20]. One possible explanation for this is M. polymorphus' larger effective size due to the formation of long chains, in contrast to the unicellular lifestyle of Synechococcus. The presence of larger particles with higher excess densities within M. polymorphus roller tanks allow for greater collision rates and thus facilitate aggregation. The greater aggregation in M. polymorphus is also seen in cultures, whereby suspended micro-aggregate volume concentrations were approximately five times higher (39 ± 19 × 10 6 µm 3 mL −1 in axenic cultures during the exponential phase) than in Synechococcus [20] (7.6 ± 0.41 × 10 6 µm 3 mL −1 in axenic cultures), likely due to the presence of aggregated diatom chains (Fig. 3). Additionally, TEP produced by M. polymorphus may be stickier [47,48] than Synechococcus-derived TEP, further increasing the chances of aggregation of M. polymorphus chains.
An alternative mechanism explaining the presence of M. polymorphus in 18S rRNA datasets from large size-fractionated seawater [9,55] and sediment trap samples [5,10] is their endosymbiotic affiliation with foraminifera (usually > 100 µm) [56]. Foraminifera are common components of sinking material collected within sediment traps and are therefore significant contributors to particle fluxes [57,58], as their large size and CaCO 3 -derived shells allow for high sinking rates. The assimilation of M. polymorphus as endosymbionts by foraminifera could serve as a mechanism transporting them to depth, similar to the export of Phaeocystis spp. as endosymbionts of Acantharia [59] and the diazotrophic cyanobacteria Richelia spp. as endosymbionts of rhizosolenid diatoms [40]. Additionally, it is possible that aggregation, as seen in our study, allows for M. polymorphus cells and chains to become readily available as prey for zooplankton, with subsequent export in the form of fecal pellets [60,61]. Nevertheless, our observations give further evidence to the underestimation of the contribution of these diatom genera to particle fluxes.
Our study is the first to show the potential for nano-sized eukaryotic phytoplankton to form sinking aggregates and produce TEP and provides evidence that this aggregation is differentially influenced by the ambient bacterial community. We also show that bacteria-mediated aggregation of phytoplankton is likely influenced by other exudates other than the TEP, which calls for the investigation of other potential aggregation-enhancing organic compounds. Our results contribute to the understanding of the mechanisms by which eukaryotic pico-and nanophytoplankton escape remineralization in the euphotic zone and are exported to the deep ocean.