The influence of extracellular polysaccharides, growth rate, and free coccoliths on the coagulation efficiency of Emiliania huxleyi

    The influence of extracellular polysaccharides, growth rate, and free coccoliths on the coagulation efficiency of Emiliania huxleyi Jennifer Szlosek Chow, Cindy Lee, Anja Engel PII: DOI: Reference:

S0304-4203(15)00104-8 doi: 10.1016/j.marchem.2015.04.010 MARCHE 3243

To appear in:

Marine Chemistry

Received date: Revised date: Accepted date:

21 August 2014 30 April 2015 30 April 2015

Please cite this article as: Chow, Jennifer Szlosek, Lee, Cindy, Engel, Anja, The influence of extracellular polysaccharides, growth rate, and free coccoliths on the coagulation efficiency of Emiliania huxleyi, Marine Chemistry (2015), doi: 10.1016/j.marchem.2015.04.010

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

PT

The influence of extracellular polysaccharides, growth rate, and free

SC

RI

coccoliths on the coagulation efficiency of Emiliania huxleyi

a

MA

NU

Jennifer Szlosek Chowa,b, Cindy Leea , and Anja Engela,b,c*

School of Marine and Atmospheric Sciences, Stony Brook University, Stony Brook, New York

TE

D

11794-5000

Alfred Wegener Institute for Polar and Marine Research, 27515 Bremerhaven, Germany

c

Current Address: GEOMAR Helmholtz-Centre for Ocean Research Kiel, Düsternbrooker Weg

AC CE P

b

20, 24105 Kiel, Germany

*Corresponding author: Anja Engel ([email protected])

Running head: Coagulation efficiency of E. huxleyi cells

First Author Last Name Disambiguation: Last Name Chow, Middle Name Szlosek

1

PT

ACCEPTED MANUSCRIPT

RI

Acknowledgements

We thank Nicole Händel for technical assistance and Corinna Borchard for sharing and

SC

discussing chemostat data. This work was supported by the U.S. National Science Foundation Chemical Oceanography Program, the Helmholtz Association (HZ-NG-102), and the German

NU

Academic Exchange Service (DAAD). This research is a contribution to the German Research Foundation Collaborative Research Center 754 (DFG SFB754) Programme on Climate-

MA

Biogeochemistry Interactions in the Tropical Ocean. We gratefully acknowledge the anonymous

AC CE P

TE

D

reviewers for their careful reading and valuable comments.

2

RI

PT

ACCEPTED MANUSCRIPT

SC

Abstract

Coagulation of small particles results in the formation of larger aggregates that play an

NU

important role in the biological pump, moving carbon and other elements from the surface to the deep ocean and seafloor. In this study, we estimated the efficiency of particle coagulation of the

MA

coccolithophore Emiliania huxleyi at different growth rates using Couette flow devices at a natural shear rate. To determine the impacts of chemical and biological factors involved in aggregate formation, we investigated how variance in organic matter composition, and in

D

particular the presence of extracellular polysaccharides (EP), including transparent exopolymer

TE

particles (TEP) and acidic polysaccharides attached to the coccolith surface, affect the coagulation efficiency (α). When E. huxleyi was grown in a chemostat at different growth rates,

AC CE P

coagulation efficiency increased from ~0.40 to 1 as cell growth rates declined and nutrients became more limited. With declining growth rate the concentration of EP and the number of detached coccoliths increased. Overall a close correlation between coagulation efficiency of E. huxleyi and the ratio of EP to total particle volume was observed. The minimum value of α of ~0.4 determined during this study is higher than estimates published for other phytoplankton cells, and may be related to the presence of EP attached to coccoliths. Based on our findings, we suggest that E. huxleyi is more prone to form aggregates, particularly during the decline of blooms, when increased production of EP and enhanced shedding of coccoliths coincide. This may be one explanation for why blooms of E. huxleyi play an important role in the biological carbon pump, efficiently enhancing the vertical flux of particles, as has been suggested by sediment trap studies.

3

RI

PT

ACCEPTED MANUSCRIPT

SC

1. Introduction

Phytoplankton-derived aggregates are a major source of the organic matter transported

NU

from the surface ocean to the deep sea (Honjo, 1982). Aggregates in seawater can form through physical coagulation processes, i.e., the collision and subsequent adhesion of individual particles

MA

that can continuously increase aggregate size from a few µm to mm size, or by biological activity, i.e., mucus net feeding, typically producing larger sized aggregates (Alldredge, 1979, McCave, 1984, Kiorboe et al., 1990, see also Burd and Jackson, 2009, for review). The onset

D

and magnitude of phytoplankton aggregate sedimentation events in the upper ocean are

TE

dependent on the stages of phytoplankton blooms as well as the phytoplankton functional groups present (Stemmann et al., 2002; Jin et al., 2006). Coccolithophores are an important group of

AC CE P

phytoplankton that produce large blooms (Balch et al., 2005) and produce calcite tests that affect aggregate and fecal pellet settling rates (De La Rocha and Passow, 2007; Engel et al. 2009, Iversen and Ploug, 2010). In the past decade increasing attention has been given to the role of mineral ballast on the efficiency of carbon export (Armstrong et al., 2002; Klaas and Archer, 2002; Passow and De La Rocha, 2006). Data and model results indicate that CaCO3 plays a dominant role in exporting organic matter (Klaas and Archer, 2002; Jin et al., 2006), yet relatively little is known about what controls coccolithophore aggregation (De La Rocha and Passow, 2007; Biermann and Engel, 2010). The rate at which particles coagulate into aggregates depends on the size and concentration of the colliding particles. Jackson and Kiørboe (2008) suggested that coagulation rate increases with the square of the particle concentration, and that particle loss due to coagulation will eventually equal the algal growth rate and thus be an upper limit on cell concentration. For a bloom, the maximum algal concentration is hence a function of cell net growth rate, turbulent shear rate, algal diameter, and coagulation efficiency. The coagulation efficiency (α) is defined as the fraction of total particle collisions that result in particles attaching 4

ACCEPTED MANUSCRIPT to one another and has also been referred to as stickiness (Kiørboe et al., 1990; Engel, 2000), stickiness efficiency (Kahl et al., 2008), sticking coefficient (Logan et al., 1995), coalescence efficiency (McCave, 1984), and collision efficiency factor (Gibbs, 1982). This value has been

PT

used in aggregation models to interpret relative changes in carbon export over the duration of a bloom (Jackson and Lochmann, 1993; Kahl et al., 2008; Karakas et al., 2009). In a study of

RI

diatoms, Kahl et al. (2008) found that the coagulation efficiency was dependent on the

SC

physiological state of the cells, and that this could have a significant effect on the critical concentration of algal cells, and in turn the subsequent export flux. An approximation of the

NU

coagulation efficiency can be derived experimentally using Couette flow devices (van Duuren, 1968; Drapeau et al., 1994; Engel, 2000; Vieira et al., 2008) or annular flumes (Kahl et al.,

MA

2008).

One of the physiological factors that appears to affect coagulation efficiency and promote phytoplankton aggregation is the production of extracellular polymeric substances, particularly

D

extracellular polysaccharides (EP) (Logan et al., 1995; Passow and Alldredge, 1995a; Engel,

TE

2000). EP are the major fraction of all extracellular polymeric substances produced by algal cells (Hoagland et al 1993). For the purpose of this study we will define EP as any polysaccharides

AC CE P

released external to the plasmalemma of the cell, including dissolved polysaccharides, discrete polysaccharidic exopolymer particles such as transparent exopolymer particles (TEP) as well as polysaccharidic surface coatings.

Passow and Alldredge (1995a) found the ratio TEP:cells to be a crucial factor in determining when aggregation and sedimentation of a diatom-dominated mesocosm bloom occurred, and the peak concentration of TEP during diatom blooms has been shown to be directly related to the sedimentary loss of total particulate organic carbon (POC) or nitrogen (PN) (Engel et al., 2014). Coccoliths and intact coccolithophores have long been observed in association with mucus material (Honjo, 1982; Cadée, 1985; de Wilde et al., 1998). Recent studies have documented the association of coccolithophores with TEP, and the formation of coccolithophore aggregates within laboratory roller tanks (Engel et al., 2009; Biermann and Engel, 2010). The amount of mineral ballast, TEP abundance and size, chemical composition of dissolved organic matter exudates from coccolithophores, and the abundance of free coccoliths are all contributing factors to coccolithophore aggregation (De La Rocha and Passow, 2007; 5

ACCEPTED MANUSCRIPT Engel et al., 2009; Biermann and Engel, 2010). Surface-active polysaccharides, such as acidic sugars like uronic acids and sulfonic sugars, have been shown to correlate with coagulation efficiency and TEP concentration (Mopper et al., 1995; Passow and Alldredge, 1995a). Of

PT

particular relevance to the coagulation of the coccolithophore Emiliania huxleyi is the production of polysaccharides known to be important for the formation and attachment of coccoliths to the

RI

coccosphere during cell growth (de Jong et al., 1976, 1979). Engel et al. (2009) proposed that

SC

these TEP precursors might catalyze formation of aggregates from calcifying coccolithophore cells at a rate faster than that observed for non-calcifying strains. Thus, a motivation of our study

NU

was to investigate the effect that TEP composition and abundance has on coagulation efficiency and rate of aggregate formation.

MA

Here, we report on coagulation studies performed with the cosmopolitan species E. huxleyi and the relationship of coagulation to cell growth rate, TEP, and calcite ballast. We investigated the processes that influence coagulation efficiency for E. huxleyi using Couette flow

D

devices as cells undergo nutrient depletion. We sought a more complete understanding of the

TE

processes that yield coccolithophore aggregates by characterizing TEP, the sugar composition of

2. Methods

AC CE P

the colloidal pool, and the abundance of coccoliths.

2.1. Experimental set-up

Four experiments using Couette flow devices were carried out to study coagulation of the coccolithophorid Emiliana huxleyi; this study was part of a larger 28-day study using flowthrough chemostats (see Borchard et al., 2011; Borchard and Engel, 2012). The Couette flow device experiments were carried out on days 10, 14, 22 and 25 of the larger study (Fig. 1). Hereafter we refer to the experiment days on which the Couette flow device experiments were run as day 10, day 14, day 22 and day 25 (D10, D14, D22, D25). Material used for the Couette flow device experiments came from one chemostat (T, pCO2) of the larger study. Experiments were started at the same time on each experiment day (3 h after the light cycle started) and run in semi-darkness at 15°C for 4 h. Each experiment was carried out in duplicate using two Couette flow devices (referred to as CC1 and CC2) that were identical to those used in Engel (2000) and similar in design to that of Drapeau et al. (1994) (see section 2.1.2). 6

ACCEPTED MANUSCRIPT

2.1.1. Chemostat set-up

PT

The work presented here used material from the overflow of a single chemostat incubator (constant temperature of 14˚C and aeration of 300 µatm CO2) that was part of the larger

RI

chemostat study mentioned above. The set-up of the chemostats is described in more detail in Borchard et al. (2011) and Borchard and Engel (2012). In brief, 4 chemostats, each with a

SC

volume of 9.2 L, were maintained at either a constant temperature of 14˚C (constant aeration of 300, 550, and 900 µatm CO2) or 18˚C (constant aeration of 900 µatm CO2), with a light:dark

NU

cycle of 16h:8h with a photon flux density of 300 μmol photons m-2 s-1 (TL-D Delux Pro, Philips; QSL 100, Biospherical Instruments, Inc.), and with constant stirring of ~50 rotations

MA

min-1. Cell abundances of E. huxleyi (strain PML B92/11; Plymouth Culture Collection of Marine Microalgae) were maintained at ~105 cells mL-1 at dilution rates (D) of 0.30 d-1 for 12 days (days 5-16) followed by 12 days at D = 0.10 d-1 (days 17-28). In chemostats, cell abundance

D

is controlled by the nutrient concentration of the inflow media, while growth rate µ (d-1) is

TE

controlled by the rate of dilution D (d-1) with the nutrient media. The dilution rate D is defined as D=F/V with F (mL d-1) for the rate of inflow of nutrient media, and V (mL) for total volume of

AC CE P

the chemostat incubator. During steady state, the growth rate equals the dilution rate, and cell abundances remain constant over time. Steady state growth of cells inside the chemostats was reached after day 14. The growth medium was 0.2-µm filtered North Sea water sterilized using UV-light for 24 h and enriched with trace metals and vitamins according to f/2 medium (Guillard and Ryther, 1962); nutrients were added to reach final concentrations of 30 µmol L-1 NO3-1 and 1 µmol L-1 PO4-3. Hence, cells experienced P-limitation, and P cell quotas fell below a critical value at the lower growth rate (Borchard et al., 2011). Coccolithophore cell inoculates were in the diploid stage and >99% calcifying, as determined by microscopic observation (Borchard pers. comm).

2.1.2. Couette flow devices Each Couette flow device consisted of a horizontally-mounted Couette chamber with a fill port and a sampling port and a moveable cap at one end, a crank to operate the end cap, and a motor and drive shaft at the opposite end. The Couette chamber consists of a fixed inner cylinder 7

ACCEPTED MANUSCRIPT and an outer rotating cylinder that creates a two-dimensional flow that promotes coagulation (Donnelly, 1991). The mean shear rate, , of the laminar flow depends on the angular velocity and the inner and outer radii (van Duuren, 1968). The annular space between the two cylinders , during each Couette flow device experiment was

PT

holds ~1.2 L of medium. Mean shear rate,

0.86 s-1. This shear rate is within the range typical for the ocean surface (10-2 to 10 s-1) (Grant et

RI

al., 1962; Soloviev et al., 1988) and has been used by Engel (2000).

SC

Prior to each experiment day, culture material for filling the Couette flow devices was first collected in a single sterile bottle (5 L) from the overflow of the chemostat. The duration of

NU

the overflow collection was dependent on the chemostat dilution rate (D); collection of ~3 L of material required a collection time of 1.5 days at dilution rate 0.3 d-1 and 3 days at dilution rate

MA

0.1 d-1 (see sampling scheme in Fig. 1). Overflow was directed onto the collection vessel walls to minimize disturbance of particles. The collection vessel was sealed using parafilm and there was no stirring of the material within the vessel during the duration of the collection.

D

Relative exponential growth rates (µ) for each of the four experiment days were

TE

determined from changes in cell abundance (N) in the chemostat over the duration of overflow

AC CE P

collection (t2-t1) at constant dilution rate (D) according to:

(1)

where t2 are days 10, 14, 22 and 25 and t1 are days 8.5, 12.5, 19 and 22 for each experiment, respectively. Growth rates are reported in Table 1. Cell growth during the 4-h duration of the Couette flow device experiments was also monitored in a parallel 1-L bottled sample; no detectable growth was observed during the 4-h experiments.

2.2. Analytical methods Samples representing material “before coagulation” were removed from the chemostat on each of the four experiment days (10, 14, 22, 25). These samples were analyzed for biogeochemical parameters (carbon, nitrogen, phosphorous, total alkalinity), carbohydrates, and TEP. At the conclusion of the 4-h Couette flow device experiment, the Couette chambers were

8

ACCEPTED MANUSCRIPT emptied by gentle outflow via tubing, and samples were taken to represent material “after coagulation” for carbohydrate analysis and TEP measurements.

PT

2.2.1. Chemical analysis

A detailed description of biogeochemical methods used during this study is given in

RI

Borchard et al. (2011) and Borchard and Engel (2012). Biogeochemical samples came from the

SC

chemostat. In brief, samples for nutrient analysis were filtered through 0.2-µm syringe filters, frozen at -20ºC and analyzed spectrophotometrically. Detection limits were 0.3 µmol L-1 for N

NU

and 0.1 µmol L-1 for P. For total particulate carbon (TPC), particulate organic carbon (POC), particulate nitrogen (PN) and particulate organic phosphorous (POP), samples were filtered

MA

through GF/F filters in duplicate. POC filters were acidified to remove particulate inorganic carbon (PIC). C and N concentrations were determined by elemental analyzer. PIC was calculated as the difference between TPC and POC. Total alkalinity (TA) samples were filtered

D

through GF/F filters and analyzed by titration, with a precision of ±3 µmol kg-1 seawater (SW).

TE

Data were compared with the Certified Reference Material for oceanic CO2 analyses of A.

AC CE P

Dickson (Scripps Institution of Oceanography) with an accuracy of <3 µmol kg-1 SW.

2.2.2. Carbohydrate analysis

For total combined carbohydrates (tCCHO), duplicate 20-mL samples taken from the chemostat (before coagulation) and from each Couette chamber (after coagulation) were placed in combusted (8 h at 500°C) glass vials. For high molecular weight (>1kDa) dissolved combined carbohydrates (HMW-dCCHO), duplicate 20-mL samples were filtered through 0.45-µm syringe-filters (GHP membrane, Acrodisk, Pall Corporation) into combusted glass vials. Samples were stored at -20˚C prior to analysis. Determination of tCCHO and HMW-dCCHO was done by high-performance anion-exchange chromatography with pulse amperometric detection (HPAECPAD) after acid hydrolysis according to Engel and Händel (2011) on a Dionex 3000 with an AS50 autosampler. Samples were desalted using 1 kDa MWCO membranes before hydrolysis. The method permits detection of neutral sugars [fucose (Fuc), rhamnose (Rha), arabinose (Ara), galactose (Gal), glucose (Glc), mannose (Man), and xylose (Xyl)], amino sugars [galactosamine (GalN) and glucosamine (GlcN)], and uronic acids (URA) including galacturonic acid (GalURA) and glucuronic acid (Glc-URA). Ara+GalN and Man+Xyl were quantified together due to 9

ACCEPTED MANUSCRIPT co-elution. In our data analysis we did not distinguish between the two uronic acids and refer to Gal-URA+Glc-URA as URA. The detection limit was 10 nmol kg-1. Samples were either unfiltered, 0.45-µm syringe-filtered, or 1000-kDa ultrafiltered via centrifugation (1000 kDa

PT

MWCO, Macrosep™, Pall Corporation); we refer to these size fractions as total combined carbohydrates (tCCHO) , <0.45 µm-HMW-dCCHO, and <1000 kDa-HMW-dCCHO.

RI

Mass balance of sugars was not always achieved among the two HMW dissolved pools;

SC

half of the <0.45 µm-HMW-dCCHO samples had lower concentrations than the <1000 kDaHMW-dCCHO samples, contrary to expectation given that the <0.45 µm fraction theoretically

NU

should include macromolecules >1000 kDa. Higher concentrations of the <1000 kDa-HMWdCCHO may have been an artifact, because the ultrafiltration was not conducted in a fully

MA

quantitative manner. For this reason, we focus on the trends in <1000 kDa-HMW-dCCHO composition between the four experiment days and not on absolute concentrations.

D

2.2.3. Transparent exopolymeric particles (TEP) analysis

TE

For analysis of TEP concentration, between 10-20 mL were taken from the chemostat (before coagulation) and from the Couette chamber after conclusion of the 4-h experiment (after

AC CE P

coagulation). These samples were filtered in triplicate onto 0.4 µm polycarbonate filters (Nuclepore), stained with 1mL of Alcian Blue (AB) for 4 seconds and rinsed with several mL of ultrapure water. Filters were stored in polypropylene tubes at -20°C until analysis using the colorimetric technique of Passow and Alldredge (1995b). TEP concentrations are expressed in Xanthan Gum equivalents per liter (µmg XG eq L-1). The detection limit of TEP abundance was ~50 µg Xanthan Gum equivalents (Xeq) L-1. The amount of AB-stainable material on cell surfaces was calculated from cell abundances assuming 2.59±0.4 pg Xeq. cell-1 after Engel et al. (2004). Colorimetric TEP data (TEPcolor) represent discrete TEP, i.e., were corrected for the presence of the AB-stainable material on cell surfaces. Area and volume of TEP > 0.2 µm2 were determined by microscopy as described in Engel (2009) for samples taken from the chemostat (before coagulation) and from the Couette chamber (after coagulation). Duplicate slides were prepared using volumes of 5-10 mL. Each filter mounted for microscopy was photographed with a Zeiss AxioCam MC5 at 200 magnification on a Zeiss AxioSkop plus 2 compound microscope. AxioCam software was used to record 30-35 fields of view at 4000999 pixels for each filter, with 0.33 µm  0.33 µm pixel 10

ACCEPTED MANUSCRIPT size. Image, and correspondingly, particle analysis was carried out using Image J software (ImageJ 1.42h by W.S. Rasband, NIH, Bethesda, MD); we determined area, perimeter, and fit ellipse (major and minor axes). Additionally, we used ImageJ to distinguish coccoliths, cells, and

PT

circular aggregates from other particles in the field of view by using the circularity tool (Engel, 2009). The parameter circularity is calculated by ImageJ using the equation: circularity = 4π ×

RI

(area/ perimeter2) yielding values between 0 (oblong, non-circular) and 1 (perfect circle).

SC

One effect of TEP is to increase total particle volume, which leads to higher collision rates and potentially higher coagulation efficiencies (Kiørboe et al., 1994; Jackson, 1995). The

NU

volume fraction of TEP (ΦTEP; in ppm) can be measured microscopically, but as mentioned above, Alcian Blue also stains the EP surface coating of E. huxleyi, so the total volume of AB-

MA

stainable material must be corrected for the presence of AB-stained solid particles already detected by the Coulter Counter Multisizer (see section 2.3.1) such as free E. huxleyi cells and cells bound in aggregates; we correct the “after coagulation” ΦTEP,micro samples for this affect. visible AB-stained particles, Φmicro (ppm). For this, the equivalent spherical volume (ESV)

TE

micro)

D

This can be accomplished by first calculating the volume fraction of all microscopically (denoted

of each particle was calculated from the area measurements (Engel, 2009) and summed up to

AC CE P

total Φmicro (ppm). Second, the ‘cell’ volume fraction due to AB-stained cell surface coating, Φcell,micro, is determined by the sum of ESV of all particles that satisfy the criterion of a circularity value approximating a sphere (≥ 0.7) and ferret diameter between 3.5 and 7 µm. Third, the ‘aggregate’ volume fraction of AB-stained material, Φagg,micro is determined by the sum of ESV of all particles with a ferret diameter >7 µm. Thus, ΦTEP,micro = Φmicro − Φcell,micro – Φagg,micro.

2.3. Particle analysis 2.3.1. Particle analysis during experiments During each 4-h Couette flow device experiment, subsamples of ~ 20 mL were taken after 1, 15, 30, 60, 90, 120, 150, 180, and 240 min from each Couette flow device, and kept cold and dark for up to 30 min prior to particle size measurement with a Coulter Counter Multisizer (section 2.3.2). These subsamples were acquired from the Couette flow devices by stopping the motor for ~30 seconds to crank the end cap deeper into the Couette chamber until ~20 mL of material was pushed out of the sampling port. The motor was then restarted and the Couette flow 11

ACCEPTED MANUSCRIPT device experiment was resumed. Particle size measurements made on each of the experiment time points (e.g., 1, 15, 30 min, etc) by the Coulter Counter Multisizer included detection of the particle equivalent spherical diameter (ESD), enumeration of particles discretized into binned

PT

ranges of ESD, and measurement of total particle concentration. Particles detected by the Coulter Counter Multisizer are hereafter referred to as Coulter Counter particles (CCP). The decrease in

RI

total particle concentration that occured as the cells coagulated was determined from Coulter

SC

Counter Multisizer data and used to estimate the apparent coagulation efficiency, α (see section

NU

2.4).

2.3.2. Coulter Counter Multisizer set-up

MA

The Multisizer™ 3 COULTER COUNTER® (Beckman Coulter) was outfitted with a 100-µm aperture and detected particles in the size range 2.17 - 59.2 µm equivalent spherical diameter (ESD). Triplicate measurements were performed on 0.5-1 mL subsamples and

D

averaged. If necessary, the sample was diluted with 0.2-µm pre-filtered seawater to keep the

TE

coincidence of particles at the aperture < 5%. Coincidence counts (when the orifice of the aperture is occupied by more than one particle) reduces the accuracy of particle counts by

AC CE P

underestimating the true count. The total volume of solid particles per milliliter (µm3 mL-1) was converted to volume fraction (in ppm) by dividing by 106.

2.4. Determination of coagulation efficiency Following coagulation theory, the early stages of coagulation will reduce the numerical abundance of particles within each Couette chamber during the 4-h experiment according to: (2) where Φ is the initial volume fraction of particles (in ppm) and Ct and C0 are the particle concentrations (mL-1) at times t and 0, respectively (Kiørboe et al., 1990). The value α is the stickiness parameter describing the apparent coagulation efficiency, i.e., the probability that two particles stick together when they collide, with values ranging from 0 to 1. Eq. 2 uses the rectilinear coagulation coefficient (

, where

is the mean diameter of particles), which

reduces to the constant 8 Φ/π (Kriest and Evans, 1999) in the exponent rather than 7.824 Φ/π, as in other coagulation efficiency studies (Kiørboe et al., 1990; Dam and Drapeau, 1995; Engel

12

ACCEPTED MANUSCRIPT 2000). This rectilinear coagulation coefficient (Elimelech et al., 1995) is used for three reasons: 1) to compare results of this study to other Couette flow device studies that employed the rectilinear coagulation kernel; 2) to simplify the calculations because all particles are assumed to

PT

have the same diameter; and 3) to account for the predominance of coagulation by shear in the Couette flow device.

RI

In our application of Eq. 2, we determined C and Φ by summing the Coulter Counter

SC

detectable particle concentrations (mL-1) and volume concentrations (µm3 mL-1) for particles from the largest Coulter Counter bin (59.2 µm) down to particles having a diameter equal to the

NU

smallest cell (Elimelech et al., 1995). This lower limit on particle size is designated by the size spectrum cut-off value, xcut, and varied between 3.0 and 3.6 µm equivalent spherical diameter

MA

(ESD) for this study. The volume fraction, Φ, is given with a scaling factor of 1:106 ( ppm) as in other Couette flow device studies (Kiørboe et al., 1990; Drapeau et al., 1994; Dam and Drapeau, 1995). Coagulation theory predicts particle volume to be conserved over the course of a Couette

D

flow device experiment. Taking advantage of this, Φ used in Eq. 2 equals the initial particle

TE

volume fraction, determined after 1 minute of being in the Couette flow device. We use the 1 minute time point to represent the initial conditions of the particle size spectra (i.e., coagulation

AC CE P

start time). This allows for the startup of the flow within the annular space of the Couette chambers such that time points (e.g., 1, 15, 30 min, etc) are taken once a fully developed steady shear flow had been established.

The coagulation efficiency of particles inside the Couette flow device can be approximated by resolving Eq. 2 for α with some modifications to address sources of variability in α. Because α is highly sensitive to the boundary size of particles being included in C and Φ, we report α as an average of all individuals, i.e., size-class-resolved α (x) from xcut to ~7 µm. Additionally, we take into account the variation of size (ESD) of cells (S2) within one population of cells from xcut to ~7 µm according to Kiørboe et al. (1990):

(3)

where n is the number of Coulter Counter bins between xcut and the end of the E. huxleyi monomer peak, ~7 µm ESD, thus including all individual cells from the smallest to largest size.

13

ACCEPTED MANUSCRIPT The Coulter Counter Multisizer does not detect TEP directly because of its gel-like composition rendering it almost the same density as water. Because TEP is not accounted for explicitly in Φ, α may be overestimated and >1 (Engel, 2000). The value α is therefore

PT

considered as the apparent coagulation efficiency. Estimates of the coagulation efficiency value, α’ (Engel, 2000), taking into account the presence of TEP were made by adding ΦTEP,micro to the

SC

RI

volume fraction of particles determined using the Coulter Counter Multisizer:

NU

(4)

2.5. Estimates of detached coccoliths

MA

Material used for Couette flow device experiments contained E. huxleyi cells as well as free coccoliths that were shed from the cells. Coccolith detachment was estimated using Coulter Counter Multisizer ESD, image analysis of microscope samples, and calculations based on Fritz

D

and Balch (1996). E. huxleyi coccoliths have been reported as having diameters of 2.5 to 3.0 µm

TE

(van der Wal et al., 1983). Fritz (1999) found E. huxleyi coccolith total length and total width to be 2.99 ± 0.29 µm and 2.40 ± 0.26 µm for a growth rate of 0.37 d-1, and 3.10 ± 0.30 µm and 2.50

AC CE P

± 0.25 µm for a growth rate of 0.20 d-1, which are similar to growth rates used in this study (Table 1). Using the Coulter Counter Multisizer, we counted coccoliths, which we defined as particles with an ESD less than the size spectrum cut-off value, xcut (mean 3.57 µm). Using the microscope and image analysis, we counted coccoliths defined as all particles < 3.5 µm that satisfy the criterion of a circularity value approximating a sphere (≥ 0.7), and used the above mentioned coccolith dimensions from Fritz (1999) and the xcut values estimated from Coulter Counter size spectra to determine coccoliths cell-1. To compare the effect of the detached coccoliths on the coccosphere diameter we estimated the number of layers of coccoliths on cells according to Balch et al. (1993) and determined the change in cell diameter, Δd, for each additional coccolith using an experimentally derived power law function Δd /coccolith = 0.195×n−0.465 (Fritz and Balch, 1996) where n is the cumulative number of coccoliths attached to the cell. We assumed that non-calcifying cells had a diameter of 2.7 µm and each additional layer of coccoliths added 0.798 µm (Fritz and Balch, 1996) to the preexisting diameter. We calculated the surface area of coccospheres using the mean cell diameter from the Coulter Counter particle size spectrum. The surface area per coccolith was 14

ACCEPTED MANUSCRIPT taken from Balch et al. (1993) and used to estimate the cumulative number of attached coccoliths necessary for each layer and the change in coccosphere diameter for every added/lost plate. For example, from day 14 to day 22 the cumulative number of attached coccoliths decreased from 31

PT

to 24, and the decrease in diameter per coccolith lost, Δd/coccolith, was 0.04 µm. We then used the change in cell diameter per experimental growth rate to estimate detached coccoliths and the

RI

ratio of detached to attached coccoliths (e.g., from day 14 to day 22, ~3 coccoliths were lost per

SC

cell, decreasing the mean cell diameter by 0.17 µm).

NU

2.6. Principal components analysis

A correlation matrix was used to evaluate the relationships between measured variables.

MA

Measured variables included: total coccolith-associated polylsaccharides (tCAP); <0.45 µm high molecular weight dissolved coccolith-associated polysaccharides (<0.45 µm-HMW-dCAP); <1000 kDa high molecular weight dissolved coccolith-associated polysaccharides (<1000 kDa-

D

HMW-dCAP); growth rate; non-circular aggregates; circular aggregates; total combined

TE

carbohydrates (tCCHO); total uronic acids (tURA); <0.45 µm high molecular weight dissolved uronic acids (<0.45 µm-HMW-dURA); <1000 kDa high molecular weight dissolved uronic acids

AC CE P

(<1000 kDa-HMW-dURA); TEP to chlorophyll-a ratio (TEP:Chl a); detached coccoliths per cell (Coccoliths:Cell); total alkalinity; and α. The significance of correlation coefficients was determined by 2-tailed Student’s t-test at p<0.05 and p<0.01 significance. In addition, principal components analysis (PCA) with the correlation matrix was carried out using MATLAB to investigate data trends in multi-dimensional space. To address the different units of the measured variables, sample data was demeaned (the sample mean was subtracted from each observation) and standardized prior to PCA. PCA represents each sample with a sample site score and each variable with a different scale, a variable loading. The full PCA determines a set of sample site scores and associated variable loadings for every component (number of components is equal to the number of variables measured). Two PCAs were conducted, one using the sugar composition data and a second using the biogeochemical variables associated with aggregate formation (e.g., total alkalinity, TEP: Chl-a, aggregate circularity). Compositional variables plot positively or negatively with increasing absolute magnitude that is scalable to the influence the variable had on processes that explain the first and second orders of variation in samples. In environmental biogeochemical studies the sample site scores often spread out according to their source and 15

ACCEPTED MANUSCRIPT degree of degradation along PC1 and PC2 (e.g., Gõni et al., 2000; Abramson et al., 2010; Xue et

PT

al., 2011).

RI

3. Results 3.1. Chemical analyses

SC

Cell growth in the chemostats was controlled by the dilution rate and by the limiting nutrient, phosphate, which was not detectable in the chemostats at any time (Table 1). Total cell

NU

abundance reached steady state on day 14. On day 22 and the 2 days prior when the overflow was collected, we observed growth rates <0.1 d-1 of cells, which may be due to the cells still

MA

acclimating to changes in dilution rate from 0.3 to 0.1d-1 on day 17. The volume concentration of CCP showed little difference with growth rate (Table 2), indicating that biomass was controlled by nutrient concentration. Initial TA was 2440 µmol kg-1 SW; due to cell growth and

D

calcification, TA decreased to 1045-1815 µmol kg-1 SW, with the lower values being observed at

TE

lower growth rate, indicating continued calcification at reduced cell division. PIC and POC cell quotas increased with decreasing growth rate as the amount of calcite per cell increased. Further

AC CE P

information on biogeochemical results of the chemostat study are given in Borchard et al. (2011).

3.2. TEP and carbohydrate analyses TEPcolor concentrations are expressed in xanthan gum equivalents per liter (µg XG eq L-1) and were normalized to cell abundance (µg XG eq cell-1), ranged from below detection to ~830 µg XG eq L-1 (or b.d. to ~3.2 pg XG eq cell-1) and were higher at lower growth rate (Table 2). On day 22, TEPcolor per cell was ~25% lower for material sampled after the Couette flow device experiment than before. This may be due to TEP that was coagulated with cells (either preformed or formed during incubation) settling onto the inner cylinder of the Couette chamber. Another possible explanation is that material captured within the overflow vessel experienced conditions slightly different than cells that remained in the stirred chemostats. In accordance with chemostat and Couette flow device TEPcolor results, the TEP volume fraction (ΦTEP,micro) was higher at lower growth rates for the Couette chambers after coagulation (Table 2). When compared to higher growth rate results, the lower growth rate samples increased ΦTEP,micro by a factor of 3 and Φagg,micro by a factor of 7. As defined previously, Φagg,micro is the 16

ACCEPTED MANUSCRIPT volume fraction of aggregates. The volume fraction of AB-stained coccoliths (Φcocco,micro) taken from the Couette chambers, increased by nearly a factor of ~4 between days 14 and 25 (Table 2). In general, the concentrations of tCCHO and HMW-dCCHO increased at lower growth

PT

rates (Table 3). Composition (mol%) and concentration (µmol L-1) of amino sugars, neutral sugars and URA primarily varied with growth rate, but were also affected by sample source

RI

(chemostat study and Couette flow device experiments) (Table 3). PCA results showing variation

SC

of CCHO composition with growth rate are presented in section 3.5.

NU

3.3. Coagulation efficiencies

Particle coagulation efficiencies (α) in our study are given over the size range of E.

MA

huxleyi cells, i.e., 3-7 µm ESD, and by accounting additionally for the contribution of TEP to the total particle concentration (α’). Coagulation efficiencies ranged from α = 0.35 to 1.19 and α’ = 0.24 to 0.38 between the four experiments (Tables 4). Differences in coagulation efficiency

D

values determined in replicate Couette flow device experiments were below 20%, except for one

TE

experiment at growth rate µ=0.12 d-1, where the difference was 70% (day 25). Coagulation efficiencies decreased significantly with growth rate (p<0.01, n=8) (Fig. 2), and were more

AC CE P

variable at lower growth rates. Engel (2000) showed that accounting for enhancement of collisions due to the TEP particle abundance reduces the coagulation efficiency to less than or close to 1, the theoretical upper limit for the attachment rate/collision rate (Alldredge and McGillivary, 1991). We also found a reduction of coagulation efficiencies in this study, with maximum values of ΦTEP,micro−corrected coagulation efficiency (α’) well below 1. Apparent coagulation efficiency (α) was closely related to the ratio of EP (including both TEP and AB stainable material attached to the cell surface) to the total volume of CCPs (p<0.001, n=7) (Fig. 3). The lowest value determined for [EP]:[CCP] was 11 fg Xeq. µm³, equivalent to about 1 pg Xeq. cell-1 for a cell with 5.49 µm ESD as determined on day 10. For comparison, Engel et al. (2004) determined that adsorption of AB to the cell surface was equivalent to a concentration of 2.59±0.4 pg Xeq. cell-1 for E. huxleyi during a mesocosm experiment. Thus, EP concentration was comparatively low at growth rate 0.63 d-1, and no free TEP was detected (Table 2). An α value of ~0.4 determined on day 10 may therefore represent a minimum value for the E. huxleyi cell itself. It follows then that despite occurring on the day with the lowest growth rate (0.03 d-1 on day 22), the maximum α’ is 0.38. 17

ACCEPTED MANUSCRIPT On day 25, the apparent coagulation efficiency observed in one of the replicate Couette flow devices was again relatively low with α =0.35 (α’=0.09), while TEP concentration was high. We cannot explain this low coagulation efficiency value, particularly since the total volume

PT

of aggregates determined in this Couette chamber at the end of the Couette flow device experiment was high (Table 2), and thus treated it as an outlier; it was not included in the linear

SC

RI

regression analysis.

3.4.Coccoliths

NU

Despite differences in absolute numbers, coccolith detachment estimated by Coulter Counter ESD, image analysis of microscope samples, and calculations based on Fritz and Balch

MA

(1996) showed a similar trend. At coagulation start time (experiment time point of 1 min), the number of detached coccoliths per cell ranged from 1 to 17 when estimated from Coulter Counter size spectra, 1.5-7.4 based on image analysis, or 5.6-10.2 using calculations from Fritz

D

and Balch (1996). All estimates showed an increasing number of free coccoliths per cell with

TE

decreasing growth rate (Table 5). The loss of coccoliths from coccospheres may explain the reduction observed in the mean diameter of the cells.

AC CE P

Using biovolume results from Coulter Counter Multisizer data (output assumes particles have spherical morphology), we observed that coccolith-like particles (diameter < xcut) comprised 1–13% of the total volume of solid particles. This quantity increased during the Couette flow device experiments by factors of 1.2 ± 0.09, 1.4 ± 0.03, 1.9 ± 0.18, 3.1 ± 0.26 for days 10, 14, 22, and 25, respectively. This data indicated that coccoliths were lost from the cells within the Couette flow device.

3.5. Principal component analyses of chemical data and coagulation efficiencies PC1 site scores for the biogeochemical data set from the four Couette flow device experiments show clear differences between high and low growth rates and explain 53.7% of the variability (Fig. 4a). Site scores for the lower growth rate cells (days 22 and 25) are not significantly different from each other. Their negative site scores correspond to variable loadings known to be associated with aggregation (Fig. 4b). These include: all coccolith-associated polysaccharides (CAP), total and <1000 kDa-HMW-dCHHO of the acidic sugars (URA), total CCHO, coccoliths:cell, α, and circular and non-circular aggregates. In comparison, variable 18

ACCEPTED MANUSCRIPT loadings that correspond to experiments at high growth rates (days 10 and 14) are growth rate, total alkalinity, and to a lesser degree <0.45 µm-HMW-dCCHO fraction of URA. In general, the presence of CAP indicated cells with lower growth rates.

PT

The 19.5% of variability explained by the second principal component (PC2) illustrates the change in aggregate properties (e.g., aggregate shape and coverage of exopolymeric material

RI

on cells) between lower growth rate experiment days (days 22 and 25) (Fig. 4c). PC2 site scores

SC

alternate between negative and positive. Likely causes for the differences between the lower growth rate experiment days are best seen in the variable loadings (Fig. 4d). Day 25 is

NU

represented by positive loadings for parameters that are dominated by AB-stained cells, circular aggregates, total CCHO, are less affected by total CAP, TEP:Chl-a, α, and nearly unaffected by

MA

growth rate. Yet, day 22 is represented by negative loadings. These loading are affected most by the presence of <1000 kDa-HMW-dURA, <0.45 µm-HMW-dURA, tURA, 0.45 µm-HMWdCAP, coccoliths:cell, and non-circular aggregates (Fig. 4d). Overall, these results showed that

D

the abundance of AB-stainable EP on cell surfaces was correlated with circular aggregates, and

TE

thus AB-stained cells were a factor in determining aggregate shape. In a separate PCA (not shown), CCHO samples were differentiated by growth rate and

AC CE P

sample source (PC1, 61.4% variation explained), and abundance of Rha and URA versus Ara+GalN, Man+Xyl, and GlcN (PC2, 18.7% variation explained). Of the sugars separated in this study, all but the amino sugars GlcN and GalN are generally found in coccolithophores (Fichtinger-Schepman et al., 1979) and referred to as coccolithophore-associated polysaccharides (CAP). Those samples that shifted towards the Rha and URA end of the PC2 axis could be interpreted as having a relatively higher abundance of CAP. PCA results revealed a shift in Couette flow device samples towards variable loadings for Rha and URA. This shift was most pronounced for tCCHO material of days 22 and 25.

4. Discussion 4.1. Coagulation efficiency as a function of growth rate, EP concentration and free coccoliths Coagulation efficiencies reported in previous studies range between 0 and 1 (Table 6). An increase in coagulation efficiency with decreasing growth rate has been shown for diatoms in batch, mesocosm, and field cultures (Vieira et al., 2008; Kahl et al., 2008; and Engel, 2000; 19

ACCEPTED MANUSCRIPT respectively). In general, previous studies found coagulation efficiencies of ~0 to <0.2 when cells were nutrient replete and in exponential phase, ~0.25 as cells reached stationary phase, and then a more variable value close to 1 as the cells reached senescence. Our study showed that

PT

coagulation efficiency of the coccolithophore E. huxleyi was also related to changes in cell growth rate. The lowest α of exponentially growing E. huxleyi in this study was ~0.4 and hence

RI

in the upper range reported previously for other phytoplankton cells. Although there is

SC

considerable range in the absolute α values given in each of these studies, measurements using Couette flow device or laminar shear experiments allow for the detection of factors responsible

NU

for the initiation of phytoplankton aggregation.

One factor responsible for the relatively high lower boundary of coagulation efficiency values of E. huxleyi might be that calcified cells coagulate and form aggregates more readily

MA

because of the presence of both EP and Ca2+ at the coccolith surface. Ca2+ is important for promoting cationic bridge formation of acidic sugars included in mucus fibrils (Decho, 1990;

D

Leppard, 1995; Mopper et al., 1995). Engel et al. (2009) compared aggregate formation of

TE

calcified and non-calcified E. huxleyi and observed faster aggregation of calcified cells at similar cell abundance, suggesting a higher coagulation efficiency of calcified cells.

AC CE P

Crystal morphology of E. huxleyi during normal biomineralization is regulated by coccolith-associated polysaccharides (CAP), pH, and solution chemistry (de Jong et al., 1979; Kok et al., 1986; Henriksen and Stipp, 2009). Of the CAP monomers detected, mannose and the uronic acids are most prevalent within the CAP structure and relevant to the promotion of cationic bridge formation. The CAP backbone consists mainly of mannose (Kok et al., 1986), and each CAP contains ~60 uronic acid groups that preferentially bind Ca2+ relative to other cations (de Jong et al., 1976; Borman et al., 1982). In this study, mannose + xylose and the uronic acids, galacturonic and glucuronic acid, make up an average of 40 ± 28% and 34 ± 25% of the <0.45 µm-HMW- and <1000 kDa-HMW dCCHO pools, respectively. We found the ratio coccoliths:cell to be highly correlated to <1000 kDa- CAP (r2 = 0.88, p<0.1) possibly signifying the release of colloidal CAP from cell surfaces. Indeed, supporting the findings of Engel et al. (2009), we estimated α values for calcified E. huxleyi cells growing in exponential phase (growth rate ~0.6 d-1, day 10) that were relatively large compared to other coagulation experiments (Table 6). For E. huxleyi, our findings suggest that direct coverage of cells by EP, specifically CAP, 20

ACCEPTED MANUSCRIPT results in a relatively high attachment yield for a given number of cell-cell collisions compared to other phytoplankton species. Cell-cell binding can be directly observed in exponentially growing E. huxleyi cells (Fig. 5a) and may dominate aggregation processes during early bloom

PT

development. CAP are only one fraction of EP produced by coccolithophores; TEP forming from dissolved polysaccharide precursors are another major fraction. A larger fraction of

RI

photosynthate is often exuded by cells when growth rates decline, increasing the amount of

SC

extracellular polysaccharides and thus the rate of TEP formation (Borchard and Engel 2012), as also observed during this study. Our results suggest that coagulation efficiency of E. huxleyi is

NU

strongly related to the ratio of EP to total volume fraction of solid particles. Thus EP can be present either as TEP or directly attached to cells (Fig. 3). Increasing amounts of TEP thus

MA

further enhance aggregation of E. huxleyi, resulting in more TEP-cell attachment. This mechanism may explain the incorporation of E. huxleyi in large amorphous aggregates (Fig. 5b) that have also been reported in sediment trap material collected from the North Sea (Cadée

D

1985). However, we also observed a low coagulation efficiency in one of the replicate Couette

TE

flow device experiments despite high TEP concentration (day 25). This observation may be an indication that factors other than EP concentration, e.g., EP chemical composition, could impact

AC CE P

E. huxleyi coagulation efficiency at low growth rate, or that aggregate formation at high EP concentration does not follow the simplified particle coagulation dynamics assumed here. Another mechanism that could increase aggregation rates during the late stages of an E. huxleyi bloom is the larger number of coccoliths shed from the coccosphere at lower growth rates (Table 5). In that case, the total number of suspended particles increases resulting in a higher number of total collisions and thus enhances aggregation rate. Because detached coccoliths are usually covered by CAP, their coagulation efficiency is presumably as high as observed for CAP-covered coccospheres.

4.2. Implications for E. huxleyi aggregate settling and strength The mechanism of coagulation during an E. huxleyi bloom (e.g., cell-cell binding or cellTEP binding) influences the density and shape of aggregates, and therefore plays a crucial role in determining settling velocity and strength of aggregates. Thus, insights on the relationship between bloom conditions and specific cell attachment mechanism would improve our understanding of sedimentation events. Early thoughts about TEP involvement in phytoplankton 21

ACCEPTED MANUSCRIPT aggregation were that late in a bloom when cells are reaching senescence TEP is abundant and cell aggregation begins (Alldredge et al. 1993; Kiørboe and Hansen, 1993; Passow et al., 1994). However, results from several studies are in conflict with this prevailing model of TEP-

PT

facilitated aggregation. For example, there can be aggregation before cell senescence occurs (Hill, 1992; Kiørboe et al., 1994; Passow and Alldredge, 1995a) or, conversely, there can be little

RI

sedimentation despite high concentrations of TEP and cells (Pitcher et al., 1991; Kiørboe et al.,

SC

1998). Therefore, having TEP occurring late in the bloom when nutrients are depleted (Smetacek and Pollehne, 1986; Hoagland et al., 1993; Passow, 2002) is not a necessity or always sufficient

NU

for aggregation to occur (Passow and Alldredge, 1995a). Surface colloid and coagulation theory suggest that weak aggregates will result from coagulation with little involvement of a

MA

coagulation aid (e.g., organic polymers) much like aggregates formed by charge patch flocculation, such as at the turbidity maximum of an estuary (Edzwald et al., 1974; Somasundaran, 2006). Therefore, the capacity for the aggregate to resist disaggregation as it

D

settles through the surface ocean might be highly dependent on the quality of glue attachment to

TE

cells that developed during the initial stages of aggregation. Mass sedimentation of phytoplankton is most likely to occur when cells reach a

AC CE P

theoretical critical concentration (Jackson and Kiørboe, 2008), and the magnitude of export is enhanced by initial low coagulation efficiencies that allow a buildup of biomass in the euphotic zone (Jackson and Lochmann, 1993). Therefore shifts in EP abundance, or type (TEP vs. CAP) during a bloom has not only a controlling effect on the robustness of aggregates, but on the magnitude of export as well.

We propose a conceptual model that describes the influence of EP on coagulation and encompasses the above scenarios from past studies as well as data from this study by addressing concepts from coagulation theory (Somasundaran, 2006) and carrying capacity of organic matter (Passow 2004; Passow and De La Rocha, 2006). Our proposed model, represented in Fig. 6, relates the EP abundance and arrangement to aggregate formation mechanism and robustness. As such, Fig. 6 represents a theoretical model of how EP arrangement within an aggregate can affect the sinking velocity and fate of the aggregate as drag forces act on its sinking surface. For a given abundance of EP, coccolithophore cell density, size, and morphology, the association of EP is varied from dominating the aggregate material to being limited to the cell surface (Fig. 6 ac). 22

ACCEPTED MANUSCRIPT Cells embedded in large amorphous clumps of TEP (Fig. 6a) can either have negligible sinking rates, be neutral or even positively buoyant (Kiørboe et al., 1998; Azetsu-Scott and Passow, 2004; Azetsu-Scott and Niven, 2005; Mari, 2008). These follow most closely the type of

PT

coagulation referred to as network flocculation (Somasundaran, 2006). The commonly proposed configuration of TEP in phytoplankton aggregates is that free TEP binds together phytoplankton

RI

cells and other particle debris (Fig. 6b) (Krank and Milligan, 1988; Azam and Long, 2001;

SC

Passow, 2002). These aggregates have been described as delicate (Alldredge and Gotschalk, 1989; Alldredge et al., 1990) and require in situ sampling by SCUBA divers. Such aggregates

NU

often originate from diatom blooms and relate most closely to charge patch coagulation, which is similar in strength to salt-induced coagulation. The strength of these aggregates lies between that

MA

of aggregates created by charge neutralization and bridging coagulation (Somasundaran, 2006, and references therein). As these aggregates settle, shear acts on the downward facing side of the aggregates resulting in deformation, erosion and possible breakup (Hill, 1998).

D

Fig. 6c shows another possible configuration of phytoplankton aggregate where cells are

TE

coated with EP, such as CAP, forming tightly packed aggregates. In this instance, the surface coating of EP acts as a bridging coagulant that functions like a spring between particles reaching

AC CE P

beyond the repulsive barrier linking two or more particles together. If the chemical composition of the bridging coagulant allows some flexibility, these polymers can move to accommodate pressures due to shear (e.g., up to 10-3 s-1 at surface when binding particles of 10 µm radius), and allow particles to physically rearrange their configuration without disaggregating (Somasundaran, 2006 and references therein). Examples of bridging coagulation would include heavily ballasted aggregates that are typically small and spherical in shape (e.g., Passow, 2004; Passow and De La Rocha, 2006; Engel et al., 2009; Iversen and Ploug, 2010); such aggregates have been shown to achieve greater settling velocity rates in rotating tanks (Engel et al., 2009, Biermann and Engel, 2010; Iversen and Ploug, 2010). Thus, cell-cell binding driven by surface coating EP might yield small aggregates and rapid sedimentation of cells without aggregate break up and resuspension at the density gradient of the mixed layer.

5. Conclusions Coccolithophores have been shown to be important to carbon export (e.g., Honjo, 1982; Cadée, 1985; de Wilde et al., 1998), and their carbonate an important ballast mineral (Klaas and 23

ACCEPTED MANUSCRIPT Archer, 2002). We show here the first empirical values for coagulation efficiency of a calcifying coccolithophore, Emiliania huxleyi. We found coagulation efficiency to increase as cell growth rates decreased (i.e., as a bloom progresses), and to relate directly to the relative abundance of

PT

EP. The coagulation efficiencies determined for E. huxleyi were in the upper range of those reported in previous studies for diatoms at similar growth stages, despite different shear flow

RI

levels used to cause cell collisions.

SC

Using examples from coagulation theory, we argue that aggregates formed by coagulation of EP covered cells would resist disaggregation from shear better than aggregates

NU

formed by free TEP, gluing cells together tightly and settling at rates faster than cells embedded within large matrices of TEP. Hence, including data on the form of EP and its association with

MA

cells and aggregates along transects of coccolithophore blooms (from the center to the edge or old to new) would increase our ability to predict when and if there will be mass sedimentation of

AC CE P

TE

D

the bloom material.

24

ACCEPTED MANUSCRIPT References

PT

Abramson, L., Lee, C., Liu, Z., Wakeham, S.G., Szlosek, J., 2010. Exchange between suspended and sinking particles in the northwest Mediterranean as inferred from the organic

RI

composition of in situ pump and sediment trap samples. Limnol. Oceanogr. 55, 725-739.

SC

Alldredge, A.L., 1979. The chemical compositon of macroscopic aggregates in two neritic seas.

NU

Limnol. Oceanogr. 24:855-866.

Alldredge, A.L., and Gotschalk, C.C., 1989. Direct observations of the mass flocculation of

MA

diatom blooms: characteristics, settling velocities and formation of diatom aggregates. Deep-Sea Res. 36, 159-171.

D

Alldredge, A.L., Granata, T.C., Gotschalk, C.C., Dickey, T.D., 1990. The physical strength of

TE

marine snow and its implications for particle disaggregation in the ocean. Limnol.

AC CE P

Oceanogr. 35, 1415-1428.

Alldredge, A.L., McGillivary, P., 1991. The attachment probabilities of marine snow and their implications for particle coagulation in the ocean. Deep-Sea Res. 38, 431-443.

Alldredge, A.L., Passow, U., Logan, B.E., 1993. The abundance and significance of a class of large, transparent organic particles in the ocean. Deep-Sea Res. 40, 1131-1140.

Armstrong, R.A., Lee, C., Hedges, J.I., Honjo, S., Wakeham, S.G., 2002. A new, mechanistic model for organic carbon fluxes in the ocean based on the quantitative association of POC with ballast minerals. Deep-Sea Res. II 49, 219-236.

Azam, F., Long. R.A., 2001. Sea snow microcosms. Nature 414: 495-497.

Azetsu-Scott, K., Passow, U., 2004. Ascending marine particles: significance of transparent exopolymer particles (TEP) in the upper ocean. Limnol. Oceanogr. 49, 741-748. 25

ACCEPTED MANUSCRIPT

Azetsu-Scott, K., Nivent, S.E.H., 2005. The role of transparent exopolymer particles (TEP) in the transport of Th-234 in coastal water during a spring bloom. Cont. Shelf Res. 25, 1133-

PT

1141.

RI

Balch, W.M., Kilpatrick, K.A., Holligan, P.M., Cucci, T., 1993. Coccolith production and

SC

detachment by Emiliania huxleyi. J. Phycol. 29, 566-575.

NU

Balch, W.M., Gordon, H.R., Bowler, C., Drapeau, D.T., Booth, E.S., 2005. Calcium carbonate budgets in the surface global ocean based on MODIS data. J. Geophys Res. 110,

MA

doi:10.1029/2004JC002560.

Biermann, A., Engel, A., 2010. Effect of CO2 on the properties and sinking velocity of

TE

D

aggregates of the coccolithophore Emiliania huxleyi. Biogeosciences 7, 1017-1029.

Borchard, C., Borges, A.V., Händel, N., Engel, A., 2011. Biogeochemical response of Emiliania

AC CE P

huxleyi (PML B92/11) to elevated CO2 and temperature under phosphorous limitation: a chemostat study. J. Exp. Mar. Biol. Ecol. 410, 61-71.

Borchard, C., Engel, A., 2012. Organic matter exudation by Emiliania huxleyi under simulated future ocean conditions. Biogeosciences 9, 3405-3423, doi:10.5194/bg-9-3405-2012.

Borman, A.H., de Jong, E.W., Huizinga, M., Kok, D.J., Westbroek, P., Bosch, L., 1982. The role in CaCO3 crystallization of an acid Ca2+-binding polysaccharide associated with coccoliths of Emiliania huxleyi. Eur. J. Biochem. 129, 179-183.

Burd, A., Jackson G. A., 2009. Particle Aggregation. Annu. Rev. Marine. Sci..1:65-90.

Cadée, G.C., 1985. Macroaggregates of Emiliania huxleyi in sediment traps. Mar. Ecol. Progr. Ser. 24, 193-196.

26

ACCEPTED MANUSCRIPT Dam, H.G., Drapeau, D.T., 1995. Coagulation efficiency, organic-matter glues and the dynamics of particles during a phytoplankton bloom in a mesocosm study. Deep-Sea Res. II 42,

PT

111-123.

Decho, A.W., 1990. Microbial exopolymer secretions in ocean environments: Their role(s) in

SC

RI

food webs and marine processes. Oceanogr. Mar. Biol. Ann. Rev. 28, 73-153. de Jong, E.W., Bosch, L., Westbroek, P., 1976. Isolation and characterization of Ca2+-binding

NU

polysaccharide associated with coccoliths of Emiliania huxleyi (Lohmann) Kamptner.

MA

Eur. J. Biochem. 70, 611-621.

de Jong, E.W., van Rens, L., Westbroek, P., Bosch, L., 1979. Biocalcification by the marine alga

D

Emiliania huxleyi (Lohmann) Kamptner. Eur. J. Biochem. 99, 559-567.

TE

De La Rocha, C.L., Passow, U., 2007. Factors influencing the sinking of POC and the efficiency

AC CE P

of the biological carbon pump. Deep-Sea Res. II 54, 639-658.

de Wilde, P.A.W.J., Duineveld, G.C.A., Berghuis, E.M., Lavaleye, M.S.S., Kok, A., 1998. Latesummer mass deposition of gelatinous phytodetritus along the slope of the N.W. European continental margin. Progr. Oceanogr. 42, 165-187.

Donnelly, R.J., 1991. Taylor-Couette flow: The early days. Physics Today 44 (11), 32-39. Drapeau, D.T., Dam, H.G., Grenier, G., 1994. An improved flocculator design for use in particle aggregation experiments. Limnol. Oceanogr. 39, 723-729.

Edzwald, J.K., Upchurch, J.B.,O'Melia, C.R., 1974. Coagulation in estuaries. Environ. Sci. Techn. 8, 58-63.

Elimelech, M., Gregory, J., Jia, X., and Williams, R. (Eds.), 1995. Particle deposition and aggregation: measurement, modelling, and simulation. Elsevier, Oxford, pp. 441.

27

ACCEPTED MANUSCRIPT Engel, A., 2000. The role of transparent exopolymer particles (TEP) in the increase in apparent particle stickiness () during the decline of a diatom bloom. J. Plankton Res. 22, 485-

PT

497.

Engel, A., Delille, B., Jacquet, S., Riebesell, U., Rochelle-Newall, E., Terbrüggen, A.,

RI

Zondervan, I., 2004. Transparent exopolymer particles and dissolved organic carbon

NU

experiment. Aq. Microb. Ecol. 34: 93-104.

SC

production by Emiliania huxleyi exposed to different CO2 concentrations: a mesocosm

Engel, A., Szlosek, J., Abramson, L., Liu, Z., Lee, C., 2009. Investigating the effect of ballasting

MA

by CaCO3 in Emiliania huxleyi, I: Formation, settling velocities and physical properties of aggregates. Deep-Sea Res. II 56, 1396-1407.

D

Engel, A. 2009. Determination of Marine Gel Particles. In: Wurl, O. (Ed.), Practical Guidelines

TE

for the Analysis of Seawater. CRC Press, Boca Raton, FL, pp. 125-142.

AC CE P

Engel, A., Händel, N., 2011. A novel protocol for determining the concentration and composition of sugars in particulate and in high molecular weight dissolved organic matter (HMWDOM) in seawater. Mar. Chem. 127, 180-191.

Engel, A., Piontek, J., Grossart, H. P., Riebesell, U., Schulz, K. G., Sperling, M. (2014) Impact of CO2 enrichment on organic matter dynamics during nutrient induced coastal phytoplankton blooms. Journal of Plankton Research, 36 (3). pp. 641-657.

Fichtinger-Schepman, A.M.J., Kamerling, J.P., Vliegenthart, J.F.G., de Jong, E.W., Bosch, L., Westbroek, P., 1979. Composition of a methylated, acidic polysaccharide associated with coccoliths of Emiliania huxleyi (Lohmann) Kamptner. Carbohydrate Res. 69, 181-189.

Fritz, J.J., Balch, W.M., 1996. A light-limited continuous culture study of Emiliania huxleyi: determination of coccolith detachment and its relevance to cell sinking. J. Exp. Mar. Biol. Ecol. 207, 127-147. 28

ACCEPTED MANUSCRIPT

Fritz, J.J., 1999. Carbon fixation and coccolith detachment in the coccolithophore Emiliania

PT

huxleyi in nitrate-limited cyclostats. Mar. Biol. 133, 509-518.

RI

Gibbs, R.J., 1982. Floc stability during Coulter counter analysis. J. Sed. Pet. 52, 657-660.

SC

Grant, H.L., Stewart, R.W., Moilliet, A., 1962. Turbulence spectra from a tidal channel. J. Fluid

NU

Mech. 2, 241-268.

Guillard, R.R.L., Ryther, J.H., 1962. Studies of marine planktonic diatoms. Can. J. Microbiol. 8,

MA

229-239.

Hansen, J.L.S., Kiørboe, T., 1997. Quantifying interspecific coagulation efficiency of

TE

D

phytoplankton. Mar. Ecol. Prog. Ser. 159, 75–79.

Henriksen, K., Stipp, S.L.S., 2009. Controlling biomineralzation: the effect of solution

AC CE P

composition on coccolith polysaccharide functionality. Cryst. Growth Des. 9, 2088-2097.

Hill, P.S., 1992. Reconciling aggregation theory with observed vertical fluxes following phytoplankton blooms. J.Geophys. Res. 97, 2295-2308.

Hill, P.S., 1998. Controls on floc size in the sea. Oceanography 11, 13-18.

Hoagland, K., Rosowski, J., Gretz, M., Roemaer, S., 1993. Diatom extracellular polymeric substances: function, fine structure, chemistry, and physiology. J. Phycol. 29, 537-566.

Honjo, S., 1982. Seasonality and interaction of biogenic and lithogenic particulate flux at the Panama Basin. Science 218, 883-884.

Iversen, M.H., Ploug, H., 2010. Ballast minerals and the sinking of carbon flux in the ocean: carbon-specific respiration rates and sinking velocity of marine snow aggregates. 29

ACCEPTED MANUSCRIPT Biogeosciences 7, 2613-2624.

Jackson, G.A., Lochmann, S., 1993. Modeling coagulation of algae in marine ecosystems. In:

PT

Buffle, J. van Leeuwen H.P. (Eds.), Environmental Particles. Lewis Publisher, Ann

RI

Arbor, pp. 387-414.

SC

Jackson, G.A. 1995. TEP and coagulation during a mesocosm experiment. Deep-Sea Res. II 42,

NU

215-222.

Jackson, G.A., Kiørboe, T., 2008. Maximum phytoplankton concentrations in the sea. Limnol.

MA

Oceanogr. 53, 395-399.

Jin, X., Gruber, N., Dunne, J.P., Sarmiento, J.L., Armstrong, R.A., 2006. Diagnosing the

D

contribution of phytoplankton functional groups to the production and export of

TE

particulate organic carbon, CaCO3, and opal from global nutrient and alkalinity

AC CE P

distributions. Glob. Biogeochem. Cycl. 20, GB2015, doi:10.1029/2005GB002532.

Kahl, L.A., Vardi, A., Schofield, O., 2008. Effects of phytoplankton physiology on export flux. Mar. Ecol. Progr. Ser. 354, 3-19.

Karakas, G., Nowald, N., Schäfer-Neth, C., Iversen, M., Barkmann, W., Fischer, G., Marchesiello, P., Schlitzer, R. 2009. Impact of particle aggregation on fluxes of organic matter. Progr. Oceanogr. 83, 331-341.

Kiørboe, T., Andersen, K.P., Dam, H.G., 1990. Coagulation efficiency and aggregate formation in marine phytoplankton. Mar. Biol. 107, 235-245.

Kiørboe, T., Hansen, J.L.S., 1993. Phytoplankton aggregate formation: observations of patterns and mechanisms of cell sticking and the significance of exopolymeric material. J. Plankton Res. 15, 993-1018.

30

ACCEPTED MANUSCRIPT Kiørboe, T., Lundsgaard, C., Olsen, M., Hansen, J.L.S., 1994. Aggregation and sedimentation process during a spring phytoplankton bloom: a field experiment to test coagulation

PT

theory. J. Mar. Res. 52, 197-323.

Kiørboe, T., Tiselius, P., Mitchell-Innes, B., Hansen, J.L.S.,Visser, A.W., Mari, X., 1998.

RI

Intensive aggregate formation with low vertical flux during an upwelling-induced diatom

SC

bloom. Limnol. Oceanogr. 43, 104-116.

NU

Klaas, C., Archer, D.E., 2002. Association of sinking organic matter with various types of mineral ballast in the deep sea: Implications for the rain ratio. Glob. Biogeochem. Cycl.

MA

16, 1116, doi:10.1029/2001GB001765.

Kok, D.J., Blomen, L.J.M.J. Westbroek, P., Bijvoet, O.L.M., 1986. Polysaccharide from

D

coccoliths (CaCO3 biomineral): influence on cyrstallization of calcium oxalate

TE

monohydrate. Eur. J. Biochem. 158, 167-172.

AC CE P

Krank, K., Milligan, T.G., 1988. Macroflocs from diatoms: In-situ photography of particles in Bedford Basin, Nova Scotia. Mar. Ecol. Progr. Ser. 44, 183-189. Kriest, I., Evans, G.T., 1999. Representing phytoplankton aggregates in biogeochemical models. Deep-Sea Research II 46, 1841-1859. Leppard, G.G., 1995. The characterization of algal and microbial mucilage and their aggregates in aquatic ecosystems. Sci. Total Environ. 165, 103-131.

Logan, B.E., Passow, U., Alldredge, A.L., Grossart, H.-P., Simon, M., 1995. Rapid formation and sedimentation of large aggregates is predictable from coagulation rates (half-lives) of transparent exopolymer particles (TEP). Deep-Sea Res. II 42, 203-214. Mari, X. 2008. Does ocean acidification induce an upward flux of marine aggregates? Biogeosciences, 5, 1023-1031.

31

ACCEPTED MANUSCRIPT McCave, I.N., 1984. Size spectra and aggregation of suspended particles in the deep ocean. Deep-Sea Res. 31, 329-352.

PT

Mopper, K., Zhou, J., Ramana, K.S., Passow, U., Dam, H.G., Drapeau, D.T., 1995. The role of surface-active carbohydrates in the flocculation of a diatom bloom in a mesocosm. Deep-

SC

RI

Sea Res. II 41, 47-73.

Passow, U., Alldredge, A.L., 1995a. Aggregation of a diatom bloom in a mesocosm, The role of

NU

transparent exopolymer particles (TEP). Deep-Sea Res. II 42, 99-109. Passow, U., Alldredge, A.L., 1995b. A dye-binding assay for the spectrophotometric

MA

measurement of transparent exopolymer particles (TEP). Limnol. Oceanogr. 40, 1326-

D

1335.

TE

Passow, U., Alldredge, A.L., Logan, B.E., 1994. The role of particulate carbohydrate exudates

AC CE P

in the flocculation of diatom blooms. Deep-Sea Res. I 41: 335-357.

Passow, U., 2002. Transparent exopolymer particles (TEP) in aquatic environments. Progr. Oceanogr. 55, 287-333.

Passow, U., 2004. Switching perspectives, do mineral fluxes determine particulate organic carbon fluxes or vice versa? Geochem. Geophys. Geosyst. 5, Q04002.

Passow, U., De La Rocha, C.L., 2006. Accumulation of mineral ballast on organic aggregates. Glob. Biogeochem. Cycl. 20, GB1013, doi,10.1029/2005GB002579.

Pitcher, G.C., Walker, D.R., Mitchell-Innes, B.A., Moloney, C.L., 1991. Short-term variability during an Anchor station study in the southern Benguela upwelling system, phytoplankton dynamics. Progr. Oceanogr. 28, 39-64.

Smetacek, V., Pollehne. F., 1986. Nutrient cycling in pelagic systems, a reappraisal of the

32

ACCEPTED MANUSCRIPT conceptual framework. Ophelia 26, 401-428.

Soloviev, A.V., Vershinsky, N.V., Bezverchnii, V.A., 1988. Small scale turbulence

PT

measurements in the thin surface layer of the ocean. Deep-Sea Res. 35, 1859-1874.

RI

Somasundaran, P., 2006. Encyclopedia of Surface and Colloid Science. vol. 6, CRC, Taylor &

SC

Francis Group, New York, pp. 6675.

NU

Stemmann, L., G. Gorsky, J.-C.G., Marty, M.J.-C., Picheral, M., Miquel, J.-C. 2002. Four-year study of large-particle vertical distribution (0-1000m) in the NW Mediterranean in

MA

relation to hydrology, phytoplankton, and vertical flux. Deep-Sea Res. II 49, 2143-2162.

van der Wal, P., de Jong, E.W., Westbroek, P., de Bruijn, W.C., Mulder-Stapel, A.A., 1983.

D

Ultrastructural polysaccharide localization in calcifying and naked cells of the

TE

coccolithophorid Emiliania huxleyi. Protoplasma 118, 157-168.

AC CE P

van Duuren, F.A., 1968. Defined velocity gradient model flocculator. J. Sanit. Eng. Div., Proc. Am. Soc. Civil Eng. 94, 671-682.

Vieira, A.A.H., Ortolano, P.I.C., Giroldo, D., Oliveira, M.J.D., Bittar, T.B., Lombardi, A.T. Sartori, A.L., Paulsen, B.S. 2008. Role of hydrophobic extracellular polysaccharide of Aulacoseira granulata (Bacillariophyceae) on aggregate formation in a turbulent and hypereutrophic reservoir. Limnol. Oceanogr. 53, 1887-1899.

Waite, A.M., Gallager, S.M., Dam, H., 1997. New measurements of phytoplankton aggregation in a flocculator using videography and image analysis. Mar. Ecol. Prog. Ser. 155, 77–88.

Xue, J., Lee, C., Wakeham, S.G., Armstrong, R.A. 2011. Using principal components analysis (PCA) with cluster analysis to study the organic geochemistry of sinking particles in the ocean. Org. Geochem. 42, 356-367.

33

ACCEPTED MANUSCRIPT

Table 1. Biogeochemical properties of Emiliania huxleyi cultures for each of the four experiment days including total dissolved inorganic N (nitrate, nitrite, and ammonia) and

PT

chlorophyll-a (Chl-a) concentrations, particulate inorganic to organic carbon (PIC/POC) molar ratios, total alkalinity (TA), and cell abundance (mL-1). Growth rates determined

RI

for each experiment are also presented. Total P concentration was below detection limits. Growth (d-1)

Cells

Total DIN

Chl-a

PIC/POC

S=31.5

×105

(µmol L-1)

(µg L-1)

(µM/µM)

(µmol kg-1)

(mL-1)

0.15

1815

2.04

0.15 ± 0.04

1424

2.86

0.17 ± 0.11

1045

2.88

0.16 ± 0.20

1095

3.12

10

0.74

19.3 ± 0.2

0.35 ± 0.00

14

0.52

13.8 ± 1.3

0.03 ± 0.00

22

0.70

8.5 ± 0.7

0.12 ± 0.08

25

0.31

9.4 ± 0.5

AC CE P

TE

D

MA

0.63 ± 0.03

SC

Day

NU

Rate

TA at

34

PT

ACCEPTED MANUSCRIPT

Table 2. TEP and particle volume fraction results for samples taken from the chemostat as well as from the Couette flow devices at the end of the

RI

four experiments (4 h) as determined by colorimetric (TEPcolor, TEPcolor per cell) and microscope (Φmicro, ΦTEP,micro, Φcocco,micro, Φcell,micro, Φagg,micro) techniques (filtration cut-off of 0.4 µm). Concentration averages ± SD are given. TEPcolor, TEPcolor per cell, and ΦTEP,micro were corrected for Alcian

SC

Blue adsorption to cell surface. Φmicro represents total volume fraction of AB-stained material on microscope slides. Coulter Counter particle volume

TEPcolor (µg XG eq L-1)

TEPcolor per cell (pg XG eq cell-1)

Φmicro ΦTEP,micro (ESD>0.4 µm) (ESD>0.4 µm) (ppm) (ppm)

Φcocco,micro (ESD<3.5 µm) (ppm)

Φcell,micro (ESD 3.5–7 µm) (ppm)

Φagg,micro (ESD>7 µm) (ppm)

ΦCCP ESD≥xcut (ppm)

0.6 ± 0.0

5.1 ± 3.4

63.2 ± 34.8

18.7

11.3 ± 0.9

0.6 ± 0.0

6.6 ± 0.7

34.5 ± 18.4

25.4

12.3 ± 1.9

0.6 ± 0.1

6.8 ± 1.2

53.7 ± 0.9

29.8

−

−

−

−

23.9

MA

Sample ID

NU

fraction (ΦCCP) was determined from Coulter Counter size spectra (≥xcut, i.e., ≥cells) at 1 minute into the experiment.

Day 10, µ=0.63 d-1 b.d.

b.d.

77.4 ± 28.7

CC1

"

"

52.4 ± 16.8

"

"

72.7 ± 0.1

chemostat

126 ± 115

0.44

−

CC1

b.d.

b.d.

73.8 ± 9.6

20.2 ± 3.3

1.3 ± 0.0

24.4 ± 10.0

29.2 ± 3.8

28.0

"

"

62.8 ± 14.1

18.0 ± 3.1

1.3 ± 0.0

18.3 ± 11.6

26.5 ± 0.6

28.1

chemostat

556 ± 197

1.93

−

−

−

−

−

22.7

CC1

181 ± 167

0.63

174 ± 5.6

25.4 ± 1.4

2.2 ± 0.0

6.3 ± 2.0

143 ± 2.3

21.4

CC2

76 ± 172

0.25

245 ± 117

24.1 ± 16.7

1.7 ± 0.0

5.3 ± 3.8

216 ± 96.8

21.7

chemostat

717 ± 324

2.30

−

−

−

−

−

18.8

CC1

825 ± 182

3.17

225 ± 7.3

79.8 ± 1.0

8.7 ± 0.0

18.3 ± 1.9

127 ± 4.5

19.8

CC2

751 ± 239

2.92

632 ± 205

61.9 ± 10.1

5.7 ± 0.0

8.5 ± 5.3

562 ± 221

19.8

CC2

Day 22, µ=0.03 d

-1

CE

CC2

AC

Day 14, µ=0.35 d

-1

9.1 ± 2.6

PT ED

chemostat

Day 25, µ=0.12 d-1

b.d. below detection; ‘−‘ not sampled

35

ACCEPTED MANUSCRIPT

Table 3. Composition of sugars (mol%) for each size fraction (total: tCCHO , <0.45µm-HMW-dCCHO, <1000 kDA-HMW-dCCHO) for each experiment day in the chemostat and replicate Couette flow devices (CC1, CC2).

PT

Also shown is the sum of all sugar concentrations (CCHO) for each sample (µmol L-1), fucose (Fuc), rhamnose (Rha), arabinose (Ara)+galactosamine (GalN), glucosamine (GlcN), galactose (Gal), glucose (Glc), mannose

Ara+ GalN

1.17 2.61 2.25 0.99 1.39 1.10 1.44 3.10 3.51 1.35 1.96

5.83 38.9 39.9 1.75 7.42 0.67 5.27 36.9 40.1 4.56 12.8

6.28 2.37 2.93 5.74 2.12 2.19 7.35 3.34 2.91 7.87 3.09

AC CE P

Glc N

Gal

Glc

Man+ Xyl

URA

0.80 0.35 0.56 0.57 -

9.73 6.37 5.21 4.45 5.40 20.3 6.56 6.40 3.57 5.86 4.82

48.3 11.8 13.1 71.3 81.7 73.2 54.9 23.9 24.4 54.4 72.8

8.96 5.04 5.61 3.17 1.05 2.49 13.4 1.29 1.18 13.9 2.54

19.0 32.9 30.9 12.2 0.97 0.01 10.5 25.1 24.3 11.5 2.03

MA

NU

SC

Rha

2.66 3.05 2.13 2.49 5.79 2.68 2.94 5.06 2.91 2.77 7.35 9.54

5.82 8.62 1.85 6.27 0.88 6.37 20.7 34.4 6.46 7.83 8.79

16.1 2.30 1.09 22.0 9.22 7.63 19.7 8.19 7.14 20.6 15.8 14.4

1.81 1.24 1.10 0.98 -

11.3 4.76 0.70 7.82 17.8 30.4 8.22 3.76 10.2 8.46 15.6 35.5

23.0 10.4 7.77 29.4 41.4 51.2 15.3 5.43 6.53 17.1 23.7 24.5

20.0 2.75 2.94 17.6 12.6 7.18 24.7 7.43 4.92 22.7 19.3 7.24

19.2 76.7 76.7 17.5 6.98 0.08 21.7 49.4 33.9 20.8 10.5 0.02

2.79 2.33 2.82 1.72 1.64 5.22 5.05

2.34 4.40 2.50 1.87 2.98 16.7 4.81 1.97

6.97 1.36 6.76 6.63 8.93 2.43 14.0 11.8

2.93 0.31 2.18 0.27 1.11

7.04 5.20 12.8 13.7 2.07 15.3 17.5

35.1 2.89 66.3 64.6 61.9 16.3 39.6 51.2

33.0 1.79 6.67 5.50 8.83 1.70 21.1 8.19

9.89 89.2 10.3 3.55 1.71 59.2 3.17

TE

Sample ID tCCHO 10 chemostat 37.4 CC1 14.4 CC2 10.3 14 chemostat 44.9 CC1 54.1 CC2 26.9 22 chemostat 94.1 CC1 37.0 CC2 32.3 25 chemostat 92.0 CC1 76.8 <0.45 µm HMW-dCCHO 10 chemostat 10.0 CC1 6.0 CC2 4.7 14 chemostat 7.4 CC1 12.2 CC2 3.6 22 chemostat 26.3 CC1 13.5 CC2 10.7 25 chemostat 27.5 CC1 11.2 CC2 10.1 <1000 kDa-dCCHO 10 CC1 7.3 CC2 16.0 14 CC1 9.2 CC2 5.9 22 CC1 35.1 CC2 21.9 25 CC1 13.2 CC2 10.5

Fuc

D

CCHO (µmol C L-1)

RI

(Man)+xylose (Xyl). URA includes galacturonic acid and glucuronic acid.

36

ACCEPTED MANUSCRIPT

Table 4. Coagulation efficiency (α) determined for each of the four Couette flow device

PT

experiments (Day 10, 14, 22, and 25). α calculated after methods of Kiørboe et al. (1990) modified by equation 3, and Engel (2000). α’ values take ΦTEP into account. Growth rates

SC

α

RI

determined for each experiment are also presented.

α'

Kiørboe et

Growth Rate (d-1)

day

al., 1990†

Engel 2000

0.63 ± 0.03

10

0.42

0.29

0.40

0.28

0.59

0.34

0.55

0.34

0.84

0.38

0.69

0.33

1.19

0.24

0.35

0.09

†

25

AC CE P

0.12 ± 0.08

22

MA D

0.03 ± 0.00

14

TE

0.35 ± 0.00

NU

Experiment

A constant of 8 is used in the coagulation coefficient term of the equation for α

(Kriest and Evans, 1999) instead of 7.824 (Kiørboe et al., 1990).

37

ACCEPTED MANUSCRIPT

Table 5. Cell diameter (Ø) is given for each of the four Couette flow device experiments, and

PT

was estimated as the equivalent spherical diameter (ESD) of the Coulter Counter cell peak at the time coagulation starts (1 min into experiment). Free coccolith per cell estimates were

RI

determined using Coulter Counter size spectra (CCSS), image analysis (IA), and calculations based on Fritz and Balch (1996)† (F&B). All data represent samples at coagulation start time (1

SC

min into experiment). The ratio of detached/attached coccoliths calculated using methods of Fritz

14

5.49

0.1

2.6

NA

NA

7.4

1.5

5.6

0.17

7.0

2.3

7.3

0.23

16.8

6.5

10.2

0.49

5.14

17.1

6.4

9.0

0.28

5.21

5.9

7.4

7.3

0.23

5.28

7.1

6.3

5.6

0.17

5.28 5.21

0.12 ± 0.08

22

5.00

AC CE P

0.03 ± 0.00

25

Coccolith cell-1 IA 1.5

Coccolith cell-1 F&B† NA

detached/ attached

MA

10

Ø (ESD, µm) 5.49

D

0.35 ± 0.00

Day

Coccolith cell-1 CCSS 0.2

TE

Growth rate (d-1) 0.63 ± 0.03

NU

and Balch (1996) is also given.

NA

NA – not applicable; Methods of Fritz and Balch (1996) require knowledge of the change in cell diameter and because day 10 was our first time point of observation, we were not able to determine detached/attached coccolith ratios. †

On day 10 we assumed that cells had 4 layers of coccoliths and a coccosphere

diameter of 5.89 µm. On days 14, 22, and 25, cells were assumed to have 3 layers of coccoliths and a coccosphere diameter of 5.10 µm. In estimating the number of layers of coccoliths on each cell and the diameter of coccospheres, we assumed that a cell with a single layer of coccoliths was 3.5 µm in diameter where coccolith thickness was 0.399 µm (after Balch et al., 1993).

38

ACCEPTED MANUSCRIPT

Table 6. Overview of coagulation efficiencies

PT

(α) measured in this and previous studies. Source

0–0.98

Kiørboe & Hansen (1993)

0.001–0.13

Drapeau et al. (1994)

0.01–0.6

Kiørboe et al. (1994)

8 × 10-5

Passow et al. (1994)

0.03–0.8

Dam & Drapeau (1995)

0–0.7

Hansen & Kiørboe (1997)

~0–0.4

Kiørboe et al. (1998)

0-0.03

Waite et al. (1997)

<0.1–1

Engel (2000)

SC

NU

MA

D

This study

AC CE P

0.35–1.19

Vieira et al. (2008)

TE

0.15–0.37

RI

α

39

ACCEPTED MANUSCRIPT Figure Legends Figure 1. Sampling scheme for this study: Chemostat incubator (temperature 14˚C, 300 µatm

PT

CO2) was first run at a dilution rate (D) of 0.3 d-1 and then changed to D = 0.1 d-1 after 16 days. From this chemostat incubator, material for the two Couette flow device experiments (four days

RI

total, days 10, 14, 22, and 25) was collected from the overflow in a sterile bottle over 1.5 days

SC

(D=0.3 d-1) and 3 days (D=0.1 d-1) and was used to fill the two Couette flow devices (CC1, CC2). On days 10, 14, 22, and 25 the chemostat was sampled for biogeochemical, carbohydrate,

NU

and TEP analyses. Couette flow devices (CC1, CC2) were sampled for particle analysis on the Coulter Counter Multisizer over the duration of the experiment at 9 time points (1, 15, 30, 60, 90,

MA

120, 150, 180, 240 min) to determine changes in particle abundance. Couette flow devices (CC1, CC2) were sampled for carbohydrate and TEP analyses at the end of the experiment (4 h). Environmental conditions (left side) include temperature (T), the light schedule (light:dark cycle,

D

L:D, or dark, D), stirrer speed (rpm) inside the chemostats, and shear rate ( ) inside the Couette

AC CE P

TE

chambers.

Figure 2. Apparent coagulation efficiency (α) vs. growth rate for Emiliania huxleyi during Couette flow device coagulation.

Figure 3. Increase in apparent coagulation efficiency (α) with the abundance of EP, including TEPcolor and AB-stained material on cell surfaces, relative to the total volume of solid particles determined with the Coulter Counter Multisizer (CCP).

Figure 4a–d. Principal components analysis of biogeochemical parameters related to aggregation for each of the four experiment days (D10, D14, D22, and D25): (a) PC1 site scores (53.7% variability explained) and (b) PC1 variable loadings for CAP-covered stained cells; and (c) PC2 site scores (19.5% variability explained) and (d) PC2 variable loadings for AB-stained cells. Parameters include: total coccolith-associated polylsaccharides (tCAP); <0.45 µm high molecular weight dissolved coccolith-associated polysaccharides (<0.45 µm-HMW-dCAP); <1000 kDa high molecular weight dissolved coccolith-associated polysaccharides (<1000 kDa40

ACCEPTED MANUSCRIPT HMW-dCAP); growth rate; non-circular aggregates; circular aggregates; total combined carbohydrates (tCCHO); total uronic acids (tURA); <0.45 µm high molecular weight dissolved uronic acids (<0.45 µm-HMW-dURA); <1000 kDa high molecular weight dissolved uronic acids

PT

(<1000 kDa-HMW-dURA); TEP to chlorophyll-a ratio (TEP:Chl a); detached coccoliths per cell

RI

(Coccoliths:Cell); total alkalinity; and α.

SC

Figure 5a, b: Emiliania huxleyi aggregates formed by cell-to-cell coagulation (a) and aggregates

NU

of E. huxleyi with TEP (b).

Figure 6: Conceptual schematic showing TEP interaction with coccolithoporid cells in

MA

aggregates and resultant effect on settling velocity for (a) aggregates largely consisting of cells and particles embedded within TEP; (b) aggregates of cells and particles bound by TEP; and (c) cells covered with extracellular polysaccharides forming aggregates. The result of excess density

D

(ρ) of aggregates with respect to water density is indicated with upwards (buoyant) or

TE

downwards (sinking) thin arrows. Magnitude of shear indicated by size of the grey arrows. The effect of shear on settling velocity (SV) of an aggregate is indicated with thick arrows. Size of

AC CE P

thick and thin arrows reflects their relative magnitude. Resultant effect of the shear acting on the aggregate is illustrated at the bottom of each figure.

41

0.3 d-1

Couew e flo devices T≈15˚C, D, =0.86 s-1

CC1

SC

Day 14

CC2

0.1 d-1

NU

Day 10

Day 14 Chemical analyses

MA

Experiment

Day 12.5-14

PT ED

Day 8.5-10

Day 10 Chemical analyses

CC1

Day 19-22

Day 17-28 Day 22 Chemical analyses

Day 22 CC2

CE

Overflo w collec on bo le T≈15˚C, L:D

Day 5-16

CC1

Day 22-25

Day 25 Chemical analyses

Day 25 CC2

CC1

CC2

During all 4-h experiments there are 9 me points taken for Par cle analysis on Coulter Counter Mul sizer At end of all 4-h experiments samples are taken for Chemical analyses

AC

Chemostat T=14˚C, L:D, 50 rpm

RI

PT

ACCEPTED MANUSCRIPT

Figure 1

42

AC CE P

TE

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

Figure 2

43

AC CE P

TE

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

Figure 3

44

ACCEPTED MANUSCRIPT

b

PT ED

MA

NU

SC

RI

PT

a

d

AC

CE

c

Fig 4a-d 45

AC

CE

PT ED

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

Fig. 5a, b

46

AC

CE

PT ED

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

Figure 6 a-c

47

ACCEPTED MANUSCRIPT

PT

Highlights

RI

Coagulation efficiency of the coccolithophorid Emiliania huxleyi was determined with Couette flow devices.

SC

Higher coagulation efficiencies of cells were observed at lower growth rates.

AC

CE

PT ED

MA

NU

Coagulation efficiency increases with the extracellular polysaccharides fraction.

48