CHEN Jie , GUO Kangli , Daniel C. O. Thornton, and WU Yi

1) Fourth Institute of Oceanography, Ministry of Natural Resources, Beihai 536007, China

2) Department of Oceanography, Texas A & M University, College Station, Texas 77843, USA

Abstract The presence of diatoms is accompanied by the production of a large amount of extracellular polymeric substances,which are mainly composed of carbohydrates. Transparent exopolymer particles (TEP) are a large class of extracellular polymeric substances with high stickiness that promotes the formation of aggregates and marine snow, which affects marine bio-carbon pump efficiency. The purpose of this research was to determine how temperature increases affect the allocation of cellular carbohydrates and the formation and aggregation of TEP. The results showed that the responses of two different diatom species (Thalassiosira weissflogii and Skeletonema marinoi) differed according to temperature. The cell density and chlorophyll a concentration of the former were not significantly correlated with temperature, while those of the latter were significantly decreased with increasing temperature. This indicates that the two species of diatom may have different heat tolerance ranges. A temperature increase will promote significant formation of TEP by both types of diatoms, including aggregation of S. marinoi as the temperature rises, meaning that the high temperature will produce an aggregate with a larger particle size and thus may increase the sedimentation rate of organic carbon.Moreover, the TEP aggregation of T. weissflogii did not increase; therefore, its particle size was smaller, and so it may remain on the sea surface at high temperatures for longer periods. These influences have a profound impact on the biogeochemical cycling of carbon.

Key words transparent exopolymer particles; diatom; Thalassiosira weissflogii; Skeletonema marinoi; aggregation

1 Introduction

Diatoms in the ocean play an important ecological role,not only because of their high productivity but also because they can excrete large amounts of extracellular polymeric substances (EPS) (Hoaglandet al., 1993; Thornton, 2002;Underwood and Paterson, 2003). EPS constitute 10% of the carbon in the oceanic dissolved organic carbon (DOC)pool (Chinet al., 1998; Verdugoet al., 2004). Most EPS are acidic polysaccharides, which can coagulate into transparent exopolymer particles (TEP) (Alldredgeet al., 1993;Passow, 2002a). TEP are sticky, gel-like particles (Alldredgeet al., 1993; Engel, 2000; Passow, 2002b) that affect the formation of aggregates, as they collide with diatoms and other particulate organic carbon (POC) to form larger particles that sink rapidly in the water column (Ki?rboeet al., 1998; Thornton, 2002; Verdugoet al., 2004). Aggregates of diatoms sink as marine snow, exporting a rapid flux of particulate organic matter from the surface to the ocean’s interior (Passow and Alldredge, 1995; Passow,2002c). This process strongly affects the biological carbon pump and biogeochemical cycling of carbon in the ocean(Thornton and Thake, 1998; Passow, 2002b; Mariet al.,2017). Field investigations have found thatThalassiosira weissflogiiandSkeletonema marinoiare both widely distributed and representative species of marine diatoms, and the spring bloom of phytoplankton often occurs in the coastal waters from the temperate zone to the subarctic,accompanied by the appearance of a large number of aggregates (Ki?rboeet al., 1994; Bresnanet al., 2009; Degerlund and Eilertsen, 2010; Fukaoet al., 2010).T. weissflogiiis a centric solitary cell with low stickiness. Its aggregation is facilitated by TEP adhering to cells to become TEP-cell aggregates (Crocker and Passow, 1995).S.marinoiis a centric, chain-forming cell. It is sticky, and cells can adhere to one another to form of cell-cell aggregates (Ki?rboe and Hansen, 1993; Crocker and Passow,1995). Because these species are widespread in the ocean,they are used as laboratory model species in many microalgae studies (Crocker and Passow, 1995; Smith and Underwood, 1998; Claquinet al., 2008; Fukaoet al., 2010).Therefore, this study selectedT. weissflogiiandS. marinoias research subjects.

Since the industrial revolution, the Earth’s temperature has increased faster than at any other time in the past 420000 years (IPCC, 2013). Global warming has led to a rising temperature on the ocean’s surface. Models project that the ocean’s surface temperature will increase by 1 to 1.5℃ by the end of the 21st century (IPCC, 2013). An increasing water temperature on the surface of the ocean affects planktonic community structure (Lewandowska and Sommer, 2010; Sarmentoet al., 2010). As Kirbyet al.(2007) noted, warming the ocean contributes to an early onset of spring phytoplankton blooms. Increasing temperatures in the North Sea have changed the plankton community composition and seasonality. Metabolic processes, such as phytoplankton growth and microbial respiration, generally increase with elevated temperatures(Sarmentoet al., 2010; Keyset al., 2018). Zlotnik and Dubinsk (1989) reported that increased temperature triggers DOC excretion by phytoplankton. Thornton and Thake(1998) reported that more aggregates ofSkeletonema costatumformed in laboratory cultures grown at higher temperatures. In the open ocean, where the nutrient supply generally limits growth, temperature is a factor that influences the growth of organisms and may influence TEP production by phytoplankton (Wolfstein and Stal,2002; Engelet al., 2011). Laboratory studies indicated that temperature can affect cell metabolism imbalance and lead to increased excretion of primary photosynthetic products (Wolfstein and Stal, 2002; Claquinet al., 2008). For example, Claquinet al. (2008) found that the TEP concentration ofThalassiosira pseudonana,Pseudo-nitzschia,andIsochrysis galbanaincreased with increasing temperature in the range of 5–35℃. When diatom cells are under stress, especially at high temperatures, the cell’s metabolic rate and fixation of carbon by photosynthesis will be promoted (Claquinet al., 2008; Passow and Laws,2015). Therefore, in this study, a semi-continuous cultivation method was used, and the two diatoms were cultivated under three temperature conditions (20, 24 and 28℃).The initial temperature (20℃) was the usual ocean temperature where the species occur naturally. The highest temperature (28℃) was the extreme high temperature that may occur in the future ocean as a result of global warming. Cultures arrived in steady state after four or more generations during the acclimation period. In the open ocean, the cellular metabolic rate during this period was stable compared with the exponential growth and decay periods, and the concentration of TEP in the stable periods would be significantly species-specific (Passowet al., 1994;Corzoet al., 2000; Grossartet al., 2006; Fukaoet al.,2010), so the measured results could be explained by the temperature factor. The aim of this experiment was to test the hypothesis that temperature increases induce increased TEP production and aggregation in diatoms.

2 Materials and Methods

2.1 Materials

Diatoms used in the experiments were obtained from the National Center for Culture of Marine Algae and Microbiota (NCMA). Diatom species wereThalassiosira weiss-flogii(strain CCMP 1051) andSkeletonema marinoi(CCMP 2092). T. weissflogiiis a dominant planktonic species in euryhaline, warm water in the North Pacific. The optimal growth temperature range forT. weissflogiiis between 11℃ and 16℃.S. marinoiis another dominant planktonic species in the North Atlantic. The optimal growth temperature range forS. marinoiis between 11℃ and 16℃.

2.2 Methods

2.2.1 Culture conditions

The temperature experiment was designed to determine the effect of temperature on the release of TEP and aggregation by diatoms. Four replicate cultures ofT. weissflogii and S. marinoi(500 mL) were grown in semi-continuous cultures with autoclaved artificial seawater at three temperature treatments (20, 24, and 28℃). The initial temperature was the usual temperature in the ocean where the species occurred naturally. Under modified nutrient batch culture conditions, cultures in bottles were placed in a glass water bath filled with water. The culture temperature was controlled by using a thermocirculator (VWR model 1196D) to manipulate the water in a glass water bath. A photon flux density of 150 μmol m?2s?1on the surface of cultures with a 14 h light:10 h dark cycle was provided for the cultures. Cell concentrations in the cultures were determined every day, and additional samples were taken from the cultures only when it was established that they were in steady state. All samples were taken at the time of daily dilution from the volume of culture that was discarded each day. After arriving at steady state, cultures were left to acclimate to the new temperature for at least four generations before sampling. Cultures were sampled three times at each temperature and were maintained for more than 3 days between sampling times.

2.2.2 Cell concentration

Samples (1 mL) were transferred to a small glass vial,and a drop of Lugol’s iodine was added to fix the cells.Four hundred cells were counted by light microscopy at 10× magnification using a hemocytometer (Fuchs-Rosenthal ruling, Hauser Scientific) (Guillard and Sieracki, 2005).

2.2.3 Chlorophyll a concentration

Five milliliters from each replicate culture was filtered through 25 mm GF/C filters. Chlorophyllawas extracted from the filters in 15 mL sterile polypropylene centrifuge tubes (VWR Scientific) containing 5 mL of cold (4℃)90% acetone. A pretest proved that acetone did not dissolve these tubes. Cells were disrupted on the filters using a sonicator (Qsonica, 125 Watts, 20 kHz) for 10 min with the amplitude at 40% in 5 s pulses with 5 s pauses between pulses to prevent heat buildup. The tubes were kept on ice during sonication. After sonication, the filters were extracted in the dark overnight at 4℃. The extractions were centrifuged at 1000gat 4℃ for 20 min, and the chlorophyllaconcentration in the supernatant was measured using a Turner Designs 700 fluorometer (Arar and Collins, 1997), which had been calibrated using chlorophyllastandards (Sigma).

2.2.4 Carbohydrate measurement

There are different fractions of carbohydrates in microalgae cultures, including cell-associated carbohydrate and dissolved extracellular carbohydrate. Total carbohydrate is the total amount of carbohydrate in a volume of culture and includes both the particulate (e.g., cells) and dissolved carbohydrate components. A 1 mL sample from each replicate was stored in an autoclaved 1.5 mL microcentrifuge tube, and the samples were kept frozen (?20℃) until analysis. After centrifugation of 50 mL of the culture at 5000g(30 min, 4℃), 1 mL of the supernatant was placed in a sterile 1.5 mL microcentrifuge tube and stored frozen(?20℃) until analysis. Lastly, the rest of the supernatant in the centrifuge tube was removed, leaving only the cell pellet in the bottom of the centrifuge tube. The different carbohydrate fractions were extracted following the method of Underwood (Underwoodet al., 1995, 2004), and their concentrations were measured using the phenol-sulfuric acid (PSA) method (Duboiset al., 1956).

2.2.5 TEP staining and analysis

TEP were stained using a modification of the method of Alldredgeet al. (1993) and Passowet al. (1994), and Passow and Alldrege (1995) and the microscopic analysis of TEP proceeded according to Loganet al. (1994). One 1 mL sample of each replicate was diluted with 1 mL of 0.2 μm filtered artificial seawater and then filtered onto a 0.4 μm pore size polycarbonate filter (Whatman) under low pressure (＜ 150 mmHg). Samples were diluted to 2 mL to produce a random distribution of particles on the filters(Hobbieet al., 1977). TEP particles on the membrane were stained with 1 mL of Alcian Blue (0.02% in 0.06% acetic acid at pH 2.5). After filtration, the membrane was washed twice with 1 mL of 0.2 μm filtered UHP water (ultrapure water) and mounted on a Cytoclear slide (GE Water & Process Technologies) using immersion oil. These slides enable TEP to be observed on top of the filters using a light microscope with illumination from below the slide. Ten images of TEP on each slide were taken using a microscope (Axioplan 2, Carl Zeiss MicroImaging) at 100× magnification. The TEP concentration and area were analyzed from light micrographs using either Axio Vision 4.8 software (Carl Zeiss Micro-Imaging) or, later in the project, a method using Image J software (National Institutes of Health) (Engel, 2009). Using Axio vision 4.8 to perform image analysis on TEP, it is necessary to perform the best ellipse fit for each TEP particle to determine the size and abundance of TEP. The image analysis produced a set of data that included the total number of particles, total area, and mean individual particle area. The TEP concentration and area in the cultures were calculated using these data. The reason why the TEP Xanthan gum method was not used in this study is that it is only representative of a symbolic component, and it is not TEP. It can only semi-quantitatively determine the concentration of TEP and cannot directly show the aggregate size of TEP (Alldredgeet al.,1993; Bittaret al., 2018); therefore, it is unable to analyze the effect of temperature on the TEP aggregate size. The TEP area method, with the help of a microscope, can quantify the degree of aggregation of TEP at different temperatures (Engel, 2009). It can express the TEP concentration and the total change in area of each TEP particle size with a temperature change more accurately and directly.

2.2.6 Aggregation

The particle size distribution (PSD) and volume concentration of particles in the cultures was measured using laser scattering following the method of Rzadkowolski and Thornton (2012). A laserin situscattering and transmissometry instrument (LISST-100X, Type C; Sequoia Scientific, Bellevue, WA, USA) was used to measure the volume concentration of particles in 32 logarithmically placed size bins over the size range 2.5 to 500 μm. In each size bin, the largest particle diameter was 1.18× the smallest diameter (LISST-100X User’s Manual, Sequoia Scientific).Particle size was not referred to as absolute size, rather the median size in a size range (Rzadkowolski and Thornton, 2012). Samples of culture (approximately 150 mL) were added to the LISST chamber. Particles in the light path attenuated and diffracted the laser light, and the scattered light struck a detector. The detector included 32 concentric rings, which indicated a series of size ranges from small to large. The sizes of particles were determine based on which rings the diffracted light hit. The PSD and volume concentration of particles in the samples were estimated by LISST SOP software (Sequoia Scientific, Bellevue,WA, USA). PSDs were blank corrected by subtracting the PSD of 0.2 μm filtered artificial seawater.

2.2.7 Cell permeability

SYTOX Green (Invitrogen S7020) is a plasma membrane-impermeable nucleic acid stain that is used to test cell permeability (Veldhuiset al., 2001; Franklinet al.,2012). In cells with a compromised plasma membrane,the nucleus inside the cells becomes stained with SYTOX Green and fluoresces with an emission peak of 523 nm when excited by a 450 to 490 nm source. Culture samples were stained by the SYTOX Green method, which is adapted from (Veldhuiset al., 2001). One milliliter of culture was stained with 40 μL of working SYTOX Green stock solution (50 μmol L?1solution) for one hour in the dark (Franklinet al., 2012). The stained sample (0.5 mL) was mixed with 1.5 mL of filtered (0.2 μm) artificial seawater and the mixture filtered through a 0.4 μm polycarbonate filter. Filters were rinsed twice using artificial seawater and mounted on glass slides using immersion oil. Slides were stored in the dark at ?20℃. The proportion of SYTOX Greenlabeled cells was counted by examining 400 cells from each slide at 400× magnification using a fluorescence microscope (Axioplan 2, Carl Zeiss MicroImaging).

2.2.8 Statistical analysis

Data were analyzed using SigmaPlot 10.0 and SYSTAT 11 (Systat software). One-way analysis of variance (ANO VA) was conducted on all data except TEP data. The data were checked to ensure that they met the assumptions of normality and equality of variance. If the data did not meet these assumptions, they were log(x+ 1) transformed before analysis, or a non-parametric ANOVA was carried out on ranks (Kruskal-Wallis ANOVA). Correlation analysis was conducted using the Pearson product moment correlation.

3 Results

3.1 Cell Concentration

Cultures arrived at steady state after four or more generations during the acclimation period. Cell concentrations in the steady state were less than 10% of the deviation of average cell abundance. Cell concentrations in the steady state cultures were not significantly different from those in the cultures ofT. weissflogiigrown at different temperatures, with an average cell abundance of 6.67×104± 0.36×104cells mL?1(mean ± SD) (Fig.1). However,there was a significant negative correlation between cell concentration and temperature (R= ?0.973,P＜ 0.05,n= 36)in the cultures ofS. marinoi, decreasing from 1.82×105±0.08×105cells mL?1(mean ± SD) at 20℃ to 0.70×105±0.04×105cells mL?1(mean ± SD) at 28℃ (Fig.1). Therefore, temperature caused a decrease in cell abundance in cultures ofS.marinoi, but not in cultures ofT. weissflogii.

3.2 Chlorophyll a Concentration

The chlorophyllaconcentration was determined at each temperature (Fig.2). There was no relationship between chlorophyllacontent and temperature inT. weissflogii.The chlorophyllaconcentration decreased as the temperature increased from 20 to 24℃ and then increased to a concentration of 74.67 ± 6.62 μg L?1at higher temperature of 28℃ (Fig.2). Unlike inT. weissflogii, there was a negative relationship between chlorophyllaconcentration and temperature (R= ?0.883,P＜ 0.05,n= 36) in the cultures ofS. marinoi(Fig.2). Their chlorophyllaconcentrations decreased at 20℃ to 13.77 ± 1.37 μg L?1(mean ± SD)at 28℃. Chlorophyllaconcentrations were significantly different at different temperatures in both species (T. weissflogii,F2,35= 34.085,P＜ 0.05;S. marinoi,F2,35= 214.473,P＜ 0.05).

Fig.1 Cell abundance with time in T. weissflogii and S. marinoi grown in semi-continuous cultures at 20℃, 24℃, and 28℃.Error bars ± SD (n = 4 replicate cultures).

The chlorophyllacontent per cell presented the same trend for chlorophyllaconcentration with temperature in two species. The highest chlorophyllacontent per cell in the cultures ofT. weissflogiiwas 1.06 ± 0.11 pg cell?1(mean± SD), which was ten times higher than that in theS. marinoicultures. The maximum chlorophyllacontent per cell was only 0.19 ± 0.02 pg cell?1(mean ± SD) in the cultures ofS. marinoi.

3.3 Carbohydrate Concentration

In the cultures ofT. weissflogii, total carbohydrate concentration per cell had a positive correlation with temperature, increasing from 0.31 ± 0.05 ng cell?1(mean ± SD) at 20℃ to 0.45 ± 0.08 ng cell?1(mean ± SD) at 28℃ (Fig.3A).There was a significant difference in total carbohydrate content per cell at different temperatures (F2,35= 41.399,P＜ 0.05). However, there was no relationship between dissolved extracellular carbohydrate concentrations per cell or cell-associated carbohydrate concentration per cell and temperature.

In cultures ofS. marinoi, there were significant positive correlations between total carbohydrate per cell (R=0.391,P＜ 0.01,n= 36), dissolved extracellular carbohydrate per cell (R= 0.336,P＜ 0.01,n= 36), cell-associated carbohydrate per cell (R= 0.792,P＜ 0.01,n= 36) and temperature (Fig.3B). There was a significant difference in cell-associated carbohydrate at different temperatures in the cultures ofS. marinoi(F2,35= 28.366,P＜ 0.05). The increase in total carbohydrate was associated with more extracellular carbohydrate and more carbohydrate stored in the cells.

Fig.2 Chlorophyll a concentration and chlorophyll a content per cell in semi-continuous cultures of T. weissflogii and S. marinoi grown at 20, 24, and 28℃. A and C, chlorophyll a concentration; B and D, chlorophyll a content per cell. Black circles represent the chlorophyll a concentrations (n = 12). Green trianglesrepresent chlorophyll a per cell (n = 12). Solid lines represent the mean value of chlorophyll a content in cultures at 20, 24 and 28℃ (n = 36). *means statistically significant by SPSS.

Fig.3 Carbohydrate allocation in semi-continuous cultures grown at 20, 24 and 28℃. Error bars show the mean ± SE(n = 12). A, in the cultures of T. weissflogii; B, in the cultures of S. marinoi. Green bars represent total carbohydrate concentration per cell. Purple bars represent dissolved extracellular carbohydrate per cell. Grey bars represent cell-associated carbohydrate per cell.

3.4 TEP Formation

TEP concentration and TEP area were analyzed in Image J. The production of TEP responded differently to increasing temperature in the two species. The images of TEP associated with the two species are shown below (Figs.4A,B).

Fig.4 Image of TEP in semi-continuous cultures at the stable phase of T. weissflogii (A) and S. marinoi (B). TEP were stained by Alcian Blue and are shown as blue particles. The relationship between TEP content and temperature in semi-continuous cultures of T. weissflogii and S.marinoi. C, TEP concentration; D, mean TEP size; E, total TEP area; F, total TEP area per cell. Error bars show mean± SD (n = 120).

Several measures of TEP dynamics are shown in Fig.4.There was an increase in TEP concentration in the cultures ofT. weissflogiias temperature increased (Fig.4C).In contrast, the TEP concentration in the cultures ofS. marinoidecreased with rising temperature (Fig.4C). There were significant differences between TEP concentration and temperature in both cultures (T. weissflogii,F2,35=19.482,P＜ 0.05;S. marinoi,F2,35= 19.482,P＜ 0.05). The mean size of individual TEP did not change significantly with temperature in the cultures ofT. weissflogii, whereas TEP size became larger in higher temperature cultures ofS. marinoi(Fig.4D). Because TEP particles with different sizes occurred in the cultures, TEP production was determined by total TEP area (TEP concentration × mean TEP size). The total area of TEP in the cultures ofT. weissflogiiwas greater at higher temperature (Fig.4E). However, in cultures ofS. marinoi, the total area of TEP was not correlated with temperature (Fig.4E). When total TEP area was normalized to the cell number, total TEP area per cell increased with temperature in both cultures (T. weiss-flogii,R= 0.683,P＜ 0.01,n= 36;S. marinoi:R= 0.530,P＜0.01,n= 36), (Fig.4F), indicating that greater TEP production occurred in cultures growing at higher temperatures.

3.5 Particle Size Distribution and Aggregation

The particle size distributions (PSD) of the two species grown at different temperatures are shown below (Fig. 5).All particles with an ESD ≥ 63 μm, which was larger than the individual cell size, were designated as aggregates. The ratio of the volume concentration of aggregates to the volume concentration of unaggregated cells in the cultures ofT. weissflogiidecreased with elevated temperature, from 2.36 at 20℃ to 1.59 at 28℃ (Fig.5G), indicating that aggregation was greater at lower temperatures. In contrast,the ratio of the total volume of aggregates to the total volume of unaggregated particles in cultures ofS. marinoiincreased with temperature, from 1.37 at 20℃ to 3.91 at 28℃(mean ± SD) (Fig.5H).Thus,aggregation increased with rising temperature in cultures ofS. marinoiand decreased with temperature in cultures ofT. weissflogii.There were significant differences in the ratio of aggregated to unaggregated diatoms at different temperatures (T. weissflogii,F2,35= 4.377,P＜ 0.05;S. marinoi,F2,35= 27.689,P＜ 0.05).

3.6 Cell Permeability

Fig.5 Particles size distributions (PSD) and the ratio of aggregated to unaggregated diatoms in semi-continuous cultures grown at 20, 24, and 28℃. Bars represent the volume concentration in each size bin in the cultures grown at different temperatures. Bar shows mean ± SD (n = 1200). Solid circles represent the ratio of aggregates to unaggregated volume in cultures at different temperatures. Dashed lines represent the mean value of the total volume concentration. Solid lines represent the mean value of the ratio. * means statistically significant by SPSS.

Fig.6 Images of permeable T. weissflogii and S. marinoi cells. Damaged cells fluoresce green, and chlorophyll a fluorescence is shown in red. A, T. weissflogii; B, S. marinoi. Relationship between the proportion of SYTOX Green-stained cells out of 400 cells in the different cultures and at the different temperatures (n = 12). C, in the cultures of T. weissflogii; D, in the cultures of S. marinoi. * indicates statistically significant by SPSS.

Permeable cells ofT. weissflogiiandS. marinoiare shown in the images below (Figs.6A, B). There was a positive relationships between the proportion of SYTOX Green-labeled cells and temperature in both species (T. weissflogii,R=0.636,P＜ 0.05,n= 36;S. marinoi,R= 0.827,P＜ 0.05,n=36) (Figs.6C, D), indicating that cells grown at a higher temperature were more permeable. The proportion of SYTOX Green-labeled cells increased from 2.6% (mean)at 20℃ to 4.1% (mean) at 28℃ in cultures ofT. weissflogiiand elevated from 2.6% at 20℃ to 5.5% at 28℃ in cultures ofS. marinoi. There were significant differences in SYTOX Green-stained cell numbers at different temperatures in both cultures (T. weissflogii,F2,35= 11.49,P＜ 0.001;S. marinoi,:F2,35= 46.733,P＜ 0.001). Hence, the hypothesis of a greater proportion of the population with compromised cell membranes at higher temperatures was accepted for both species.

4 Discussion

4.1 Temperature Affects Cell Growth

Cell growth was monitored in cultures ofT. weissflogiiandS. marinoigrown at temperatures between 20 and 28℃.Our results showed that cell abundances in cultures ofT.weissflogiiwere consistent, irrespective of temperature(Fig.1). The chlorophyllaconcentration per cell also showed no significant difference in the cultures ofT. weissflogiigrown at different temperatures (Fig.2), indicating that the thermal range between 20 and 28℃ did not affect the growth ofT. weissflogii. UnlikeT. weissflogii, a decrease in cell abundance ofS. marinoiwas observed as the temperature rose (Fig.1). The cell concentration ofS. marinoiat the higher temperature of 28℃ was one of third of that at the lower temperature of 20℃. In addition, a lower chlorophyllaconcentration per cell occurred at higher temperatures in the cultures ofS. marinoi(Fig.2). Thus, these two organisms could have different widths of thermal tolerance and/or different temperature ranges. In the field,T.weissflogiiis distributed worldwide in the oceans and contributes to spring blooms in coastal surface water with temperatures between 3 and 24℃; for example, in the Atlantic Ocean, Pacific Ocean, Hawaiian Seas, and Indonesian Seas during spring and autumn at temperatures of 15 to 24℃ (Armbrust and Galindo, 2001; Sorhannuset al.,2010). Compared withT. weissflogii, many observations of bloomsin situhave confirmed thatS. marinoidominates in coastal waters with a cooler temperature range between 2 and 17℃. For instance, anS. marinoibloom occurred on the surface of the Baltic Sea where the temperature was approximately 4℃ (Kaeriyamaet al., 2011).Barofskyet al. (2010) reported anS. marinoibloom in Raunefjord, Western Norway, at temperatures of 7 to 8℃.In the North Atlantic,S. marinoiis most abundant during the spring bloom, with a temperature of 2 to 7℃ (Sarnoet al., 2005). Kentet al. (1995) reported that mixed blooms ofThalassiosiraspp. andSkeletonemaspp. occurred in the coastal ocean off of British Columbia, Canada, indicating that these two diatoms have overlapping distributions. From these observations, we propose that the range of thermal tolerance ofT. weissflogiiis wider than that ofS. marinoi, indicating thatT. weissflogiimay have an advantage in competing withS. marinoias ocean surface temperatures increases in the next several decades.

4.2 Temperature Affects TEP Production

Many studies have indicated that temperature is an important factor that influences photosynthesis in microalgae and could also indirectly influence TEP production.The effect of temperature on TEP production might be associated with carbohydrate allocation. Wolfstein and Stal(2002) found that temperature affected cellular metabolic imbalances, leading to increased excretion of primary photosynthetic products. Our results indicated that, although the temperature increase promoted the two diatoms to produce more carbohydrates (Fig.3), the impact on their carbohydrate allocation was different. The carbohydrate allocation ofT. weissflogiichanged with increasing temperature (Fig.3A), and the dissolved extracellular carbohydrate it released would form more TEP (Fig.4C). The carbohydrate allocation ofS. marinoidid not change significantly with temperature. It released more carbohydrate in a dissolved state (Fig.3B), but the TEP concentration decreased (Fig.4C). This is consistent with the findings of Zlotnik and Dubinsy (1989) that more dissolved primary production is excreted at higher temperatures. This may be because the TEP production per cell fit the temperature model of carbohydrate allocation in two species (Fig.4F).In studies of the relationship between TEP production and temperature in microalgae, Fukaoet al. (2012) indicated that TEP production decreased with increasing temperature in the diatomsCoscinodiscus granii. Nevertheless,Claquinet al. (2008) showed an increase in TEP production as temperature increased to an optimal temperature for diatoms ofThalassiosira pseudonana,S. marinoi, andPseudo-nitzschia fraudulenta. Our results are in agreement with the observations of Claquinet al. (2008) and might be the result of elevated enzyme activity at higher temperatures, which is related to TEP production mechanisms.Previous studies proposed that TEP precursors could be attributed to cell exudation or cell lysis (Bhaskar and Bhosle, 2005; Thornton and Chen, 2017). Our results showed that cell permeability increased with temperature (Fig.6).However, more permeable cells did not directly lead to more dissolved carbohydrate outside the cells because dissolved extracellular carbohydrate did not increase with temperature. Thus, whether cell permeability contributes to TEP precursors awaits future verification.

The composition of carbohydrates varied among the different carbohydrate fractions associated with the cell wall,cell storage, and extracellular carbohydrates. For example,carbohydrates associated with storage are rich in glucan,whereas cellulose is a major component of carbohydrates associated with the cell wall. The variation in carbohydrate allocation in the cells at different temperatures may be associated with different quantities or compositions of cell surface carbohydrates. In addition, the different compositions may be attributed to diverse chemical and/or physical properties. Thus, TEP with different compositions might have different characteristics, such as variable stickiness. Some studies have reported that different compositions of organic matter and different types of TEP were formed under different growth conditions (Underwoodet al., 2004; Engelet al., 2011). Therefore, further studies are necessary to investigate TEP and TEP compositions related to distinct characteristics of TEP, such as stickiness and structure.

4.3 Temperature Affects Aggregation

Our results indicated that aggregation inS. marinoiwas enhanced with temperature elevation, whereas aggregation inT. weissflogiiincreased at a lower temperature (Fig.5).In a study of the effect of temperature on aggregation,Thornton and Thake (1998) found a positive correlation between aggregate concentration and temperature in cultures ofSkeletonema costatum, which was consistent with what we found in cultures ofS. marinoi. The difference in aggregate formation with elevated temperature between the two species is probably due to their distinct stickiness.Previous experiments suggested that cells have different aggregation patterns based on their stickiness. Cell-TEP aggregation occurred in cells with low stickiness, and cellcell aggregates formed when cells were very sticky (Ki?rboe and Hansen, 1993; Crocker and Passow, 1995). Several studies have shown thatT. weissflogiiare not sticky,and aggregation must be facilitated by TEP in the form of cell-TEP aggregates (Ki?rboe and Hansen, 1993; Crocker and Passow, 1995). However,Skeletonema costatumare sticky cells, and their aggregation proceeds because of their high stickiness (Ki?rboe and Hansen, 1993). In observations of diatom aggregation, variation in stickiness has been recorded (Ki?rboeet al., 1998). Many studies have shown that cell stickiness can vary with temperature(Ki?rboe and Hansen, 1993; Thornton and Thake, 1998;Cisternaset al., 2019). Thus, we propose thatS. marinoicells became stickier and produced more aggregation at high temperatures, whereasT. weissflogiihad lower stickiness and produced less aggregation at higher temperatures.Aggregation is a source of marine snow; therefore, the response of aggregate formation to temperature also influences the vertical flux of carbon in the ocean. Sticky cells, such asS. costatum, will tend to aggregate during blooms and result in a fast-sinking flux of organic carbon.In contrast, if the cells have low stickiness, such asT.weissflogii, they will remain in surface waters for a long time during the bloom. Many observations of aggregates ofS. costatumin the coastal ocean and their subsequent sedimentation have been documented (Crocker, 1993).The sinking of diatom aggregates and marine snow plays a critical role in the rapid transfer of primary production from the euphotic zone to deeper ones (Spunginet al.,2018). TEP production and aggregation have a significant impact on the biogeochemistry of organic matter and the ecology of marine snow. However, there have been few studies of the effect of temperature on TEP production and the attendant effect on marine biogeochemistry.Hence, the effect of temperature on the stickiness of different species, and their distinct aggregation mechanisms are a potential avenue for further investigation.

5 Conclusions

In conclusion, we showed that temperature affects TEP production and aggregation in the diatom speciesT. weissflogiiandS. marinoi. However, the response to changes in temperature differed between the two species. As the temperature increased, cell abundance decreased in the cultures ofS. marinoibut not in the cultures ofT. weissflogii, indicating that these two organisms could have different widths of thermal tolerance and/or different temperature ranges. Our results suggest that TEP production can be enhanced at higher temperatures, which were associated with greater TEP production and greater permeability in the cultures. However, there was not sufficient evidence to demonstrate a relationship between TEP formation and cell lysis. It remains unclear if TEP precursors were created by leakage from permeable cells or exudation by cells. Such a question would be an interesting topic of future study. In our results, the effect of a temperature increase on TEP aggregation differed between the two diatom species. More aggregates ofS. marinoiformed in higher-temperature cultures, indicating thatS. marinoicells became stickier at higher temperatures. In contrast,fewer aggregates occurred in cultures ofT. weissflogiiat higher temperatures, despite enhancement of TEP production. This indicates that temperature may affect TEP composition and consequently influence their chemical or physical properties and aggregate formation. TEP production and aggregate formation have important roles in the transport of carbon in the ocean. Therefore, accumulation ofS. marinoiin the form of marine snow in response to elevated temperatures will enhance the transportation of carbon to the deep ocean.

Acknowledgements

This work is supported by the National Natural Science Foundation of China (No. 31500411), the Guangxi Zhuang Autonomous Region International Platform Project (No.2019AC17008), the Guangxi Beihai Science and Technology Research Focus (Nos. 201995048 202082021 and 20 19D05), the U. S. National Science Foundation (No. OCE 0726369), the Special Fund for Asian Regional Cooperation ‘2019 China-ASEAN Marine Science and Technology Cooperation Seminar Project’, the China Asia-Pacific Economic Cooperation (APEC) Cooperation Fund Project ‘APEC Typical Regional Coral Reef Ecosystem Comprehensive Assessment Technology and Management Cooperation Research’ and the ‘Bilateral and Multilateral International Cooperation’ Project of the Central Financial Allocation Program in 2019 and 2020.