Limnol. Oceanogr. 9999, 2022, 1–16 © 2022 The Authors. Limnology and Oceanography published by Wiley Periodicals LLC on behalf of Association for the Sciences of Limnology and Oceanography. doi: 10.1002/lno.12204 Nutritional response of a coccolithophore to changing pH and temperature Roberta Johnson ,1* Gerald Langer,1 Sergio Rossi,1,2 Ian Probert,3 Marta Mammone,2 Patrizia Ziveri1,4 1 Institut de Ciència i Tecnologia Ambientals, Universitat Autònoma de Barcelona, Barcelona, Spain 2 Department of Biological and Environmental Sciences and Technologies, University of Salento, Lecce, Italy 3 Sorbonne Université, CNRS, Roscoff Culture Collection, Station Biologique de Roscoff, Roscoff, France 4  Catalana de Recerca i Estudis Avancats (ICREA), Barcelona, Spain Institucio Abstract Coccolithophores are a calcifying unicellular phytoplankton group that are at the base of the marine food web, and their lipid content provides a source of energy to consumers. Coccolithophores are vulnerable to ocean acidification and warming, therefore it is critical to establish the effects of climate change on these signifi- cant marine primary producers, and determine potential consequences that these changes can have on their consumers. Here, we quantified the impact of changes in pH and temperature on the nutritional condition (lipid content, particulate organic carbon/nitrogen), growth rate, and morphology of the most abundant living coccolithophore species, Emiliania huxleyi. We used a regression type approach with nine pH levels (ranging from 7.66 to 8.44) and two temperatures (15 C and 20 C). Lipid production was greater under reduced pH, and growth rates were distinctly lower at 15 C than at 20 C. The production potential of lipids, which estimates the availability of lipids to consumers, increased under 20 C, but decreased under low pH. The results indicate that, while consumers will benefit energetically under ocean warming, this benefit will be mitigated by ocean acidifi- cation. The carbon to nitrogen ratio was higher at 20 C and low pH, indicating that the nutritional quality of coccolithophores for consumers will decline under climate change. The impact of low pH on the structural integrity of the coccosphere may also mean that coccolithophores are easier to digest for consumers. Many responses suggest cellular stress, indicating that increases in temperature and reductions in pH may have a nega- tive impact on the ecophysiology of coccolithophores. Coccolithophores are unicellular calcifying phytoplankton formation and dissolution of their calcium carbonate skeleton. found in all of the world’s oceans. They influence seawater An estimated 83% of the organic carbon flux to the seafloor is chemistry and the exchange of carbon dioxide between the associated with their calcium carbonate ballast (Klaas and atmosphere and the ocean via photosynthesis and the Archer 2002), establishing coccolithophores as major contrib- utors to the carbon cycle in the oceans (Ziveri et al. 2007; Falkowski et al. 2008; Lefebvre et al. 2011). *Correspondence: roberta.johnson@uab.cat Coccolithophores form part of the base of the food web, and their nutritional quality, particularly in terms of their This is an open access article under the terms of the Creative Commons Attribution-NonCommercial License, which permits use, distribution and organic carbon to nitrogen ratio and lipid content, are impor- reproduction in any medium, provided the original work is properly cited tant predictors of the nutritional condition for consumers at and is not used for commercial purposes. higher trophic levels (Pond and Harris 1996; Mitra and Additional Supporting Information may be found in the online version of Flynn 2005; Schlüter et al. 2014). In phytoplankton, lipids this article. (total and fatty acids) serve as structural molecules within cells and as energy storage units (Fuentes-Grünewald et al. 2012). Author Contribution Statement: R.J. conducted the experiment, ana- lyzed the carbonate chemistry data, the morphological data, performed These are key determinants of food quality, and, in turn, the the statistical analyses, and drafted the manuscript. G.L. assisted with the health and functioning of marine ecosystems (Jin et al. 2020). experiment and experimental design and assisted with writing the manu- script. S.R. analyzed the lipid samples, assisted with writing the manu- Chlorophyll a (Chl a), an indicator of photosynthetic capac- script, and created the original concept. I.P. assisted with the experiment, ity, contains a large amount of lipids, and, as such, increases and with writing the manuscript. M.M. analyzed the chlorophyll in Chl a are usually associated with an increase in lipids a samples, the carbon and nitrogen samples, and assisted with writing the manuscript. P.Z. assisted with the experimental design, writing the (Woodworth et al. 2015) as well as photosynthetic perfor- manuscript, conceptualization, resources, and supervision. mance. These lipid macromolecules provide a source of energy 1 Johnson et al. Nutr response of e. hux to pH and temp to phytoplankton consumers higher on the trophic ladder photosynthetic ability (Bach et al. 2013; Fukuda et al. 2014), and (Broglio et al. 2003; Litzow 2006), and changes in the avail- growth rate (Lefebvre et al. 2011; Bach et al. 2013; Schlüter ability of essential fatty acids can have a significant impact on et al. 2014). consumer productivity (Fraser et al. 1989; Breteler et al. 2005). Global sea surface temperature is expected to increase by Suspension feeders and zooplankton depend on primary pro- approximately 3 C on average by the end of this century, ducers as a food source (Sailley et al. 2013; El-Hady though this will vary by region (RCP 8.5; IPCC 2019). Ocean et al. 2016), and food quality and availability is a major factor warming has been shown to impact coccolithophore growth affecting reproduction and survival in these organisms (Gili (Schlüter et al. 2014; Rosas-Navarro et al. 2016; Feng and Coma 1998; Broglio et al. 2003; Gori et al. 2013). et al. 2017; Krumhardt et al. 2017), calcification (Rosas- Any detrimental effects of climate change on phytoplank- Navarro et al. 2016; Feng et al. 2017; Krumhardt et al. 2017), ton physiology may have cascading effects on other compo- and morphology (Rosas-Navarro et al. 2016). The combined nents of the ecosystem (Chavez et al. 2010; Guinder and impacts of ocean acidification and warming have also been Molinero 2013). For instance, increased pCO2 has been shown shown to have an interactive effect on coccolithophore to effect trophic transfer efficiency between phytoplankton growth (Arnold et al. 2013; Sett et al. 2014; D’Amario and their consumer, copepod Acartia tonsa (Cripps et al. 2017), morphology (Milner et al. 2016; D’Amario et al. 2016). In addition, four species of phytoplankton grown et al. 2017), calcification (Schlüter et al. 2014; Sett et al. 2014), under high CO2 conditions (1000 ppm) showed reduced and photosynthesis (Sett et al. 2014). essential fatty acid content, and the copepods consuming this Although there has been some focus on the effects of cli- plankton experienced reduced egg production, hatching suc- mate change stressors on the quality of some primary pro- cess, and egg viability (Meyers et al. 2019). The diatom ducers as a food source (Klauschies et al. 2012; Guinder and Thalassiosira pseudonana experienced reduced fatty acid con- Molinero 2013), there has been no published work that tent under elevated pCO2, which translated to a 10-fold investigates how ocean acidification combined with ocean decrease in fatty acids for the consumer, copepod A. tonsa, warming conditions will alter the quality and quantity including a decrease in somatic growth and egg production (in terms of food availability) of coccolithophores as a source from 34 to 5 eggs female1 d1 (Rossoll et al. 2012). When of nutrition for higher trophic levels. This study aims to fill considering such drastic impacts that changes in phytoplank- this gap. We focus on E. huxleyi here due to its extensive use ton nutritional status can have on consumers, it is critical to in previous studies, making this an ideal species for compari- establish the effects of climate change stressors on this signifi- son. The hypothesis tested here is that reduced pH (in line cant marine phytoplankton, and the consequences that they with ocean acidification) and increased temperature (ocean may have on the food web (Rossi et al. 2019). warming) will likely affect the quality of coccolithophores as Emiliania huxleyi is one of the most abundant a food source (Guinder and Molinero 2013) by impacting coccolithophores in the oceans and is responsible for giant sea- their investment in stored energy (in the form of lipids). This sonal algal blooms (Brown and Yoder 1994) that are visible from is investigated here by measuring cellular lipid content space (Holligan et al. 1993). Such blooms can provide a substan- (Pond and Harris 1996), as well as carbon and nitrogen tial food source to zooplankton grazers (Pond and Harris 1996). ratios. We also explore the impact of pH and temperature on E. huxleyi is known to produce stable lipid compounds, including food quantity through a parameter known as production alkenones, alkyl alkeonates, and alkene, which can be used as cli- potential, which combines cell growth with cellular lipid matic proxies to evaluate the impacts of a changing climate content. Production potential translates cellular lipid pro- (Bendle et al. 2005; Malinverno et al. 2008). E. huxleyi has been duction to community production and estimates the avail- studied extensively and found to be sensitive to ocean acidifica- ability of lipids in a coccolithophore community (Gafar tion conditions (Riebesell et al. 2000; Beaufort et al. 2011; Lefeb- et al. 2018; Klintzsch et al. 2019). These points have rarely vre et al. 2011; Schlüter et al. 2014). The most relevant climate been addressed, particularly in a combined ocean acidifica- change impacts for coccolithophores are ocean acidification and tion and warming scenario. In addition, we investigate the ocean warming. impacts of pH and temperature on important parameters Ocean acidification is the result of increasing atmospheric related to coccolithophore nutritional quality, including CO2, which is absorbed by the ocean, resulting in an increase in growth rate, particulate organic and inorganic content, and [H+] and a decrease in ocean pH, as well as a decrease in the con- coccosphere morphology (Table 1). centration of carbonate ions (Fabry et al. 2008; Doney et al. 2009). Global open ocean surface seawater pH is projected to decrease by approximately 0.3 units by 2100 under RCP 8.5 Materials and methods (business as usual scenario; IPCC 2019). The reduction Culture medium preparation in seawater pH has been shown to impact E. huxleyi calcification Seawater collected from 3 km off the coast of Roscoff (Riebesell et al. 2000; Lefebvre et al. 2011; Schlüter et al. 2014), (Brittany, France) on the 3rd and 10th of June 2019 was mixed morphology (Langer et al. 2009; Lefebvre et al. 2011), homogenously, pre-filtered using 0.7-μm nominal pore size 2 Johnson et al. Nutr response of e. hux to pH and temp Table 1. Information on how each response variable relates to nutritional quantity or quality, and other parameters indicating fitness and cellular function for Emiliania huxleyi. The response variables include: lipids (pg cell1), POC (pg cell1), PIC (pg cell1), Chl a (pg cell1), growth rate (day1), lipid production (pg cell1 d1), POC production (pg cell1 d1), PIC production (pg cell1 d1), Chl a production (pg cell1 d1), production potential—lipids (ng), POC : N, PIC : N, PIC : POC, lipid : POC (cellular lipid content : cellular POC content), coccosphere diameter (μm), coccolith distal shield length (μm), coccolith distal shield width (μm), inner circle diameter (μm), tube width (μm), collapsed coccosphere (%). Response variable Parameter relevance Cellular quotas Lipid quota High-energy carbon source Increase indicates higher associated energy to catabolize (food quantity) POC quota Organically digestible carbon—comprised of both low and high energy (i.e., lipids) compounds PIC quota Increase might indicate lower food quality Chl a quota Increase can indicate higher capability of photosynthesis—effect on food quality uncertain Production rates Growth rate Main driver of food production (quantity) Lipid production Food quantity (energy) production Note that this is cellular level and less important for consumers (see also production potential) POC production Food quantity (see also production potential) Alone insufficient to indicate food quality PIC production Increase tends to lower food quality but ratios more informative Chl a production Effect on food quality/quantity uncertain Production potential— Best indication of food quantity for coccolithophore consumers lipids Amount of lipids available to consumers after a given period of growth Carbon ratios POC : N High-quality food (low ratio), that is, less nitrogen Nitrogen is involved in the production of amino acids and proteins, including chlorophyll, and is essential for cellular function PIC : N High food quality (low ratio) PIC has no nutritional value but might hamper digestion PIC : POC High food quality (low ratio), that is, less PIC Lipid : POC High food quality (high ratio), that is, more lipids Morphological characteristics Coccosphere diameter Reduced size may mean the coccosphere is more easily digestible for nonselective grazers Distal shield length General coccolith morphological parameter—differences indicate change in growth and/or calcification Distal shield width General coccolith morphological parameter—differences indicate change in growth and/or calcification Inner circle diameter General coccolith morphological parameter—differences indicate change in growth and/or calcification Tube width General coccolith morphological parameter—differences indicate change in growth and/or calcification Collapsed coccospheres Weak coccospheres may mean cell contents are more readily digestible for selective grazers glass fiber filters (Whatman GF/F) and 0.2-μm mixed cellulose light/dark cycle in culture cabinets set to either 15 C or 20 C. ester membrane filters (Millipore), heated to 80 C for 10 min, Using these initial temperature acclimated pre-cultures, and then cooled overnight. The seawater was enriched acclimation cultures were inoculated with 500 cells mL1 in with 100 μmol L1 nitrate, 6.25 μmol L1 phosphate, trace culture flasks (300 mL) and held in experimental conditions metals, and vitamins as in K/2 medium (Keller et al. 1987). (two temperatures: 15 C and 20 C, 9 pH levels targeted to The culture medium was then filter sterilized using 0.2 μm Fast approximately: 8.4, 8.3, 8.2, 8.1, 8.0, 7.9, 7.8, 7.7, 7.6, light Flow Polyethersulfone (PES) Express PLUS filter modules regime as above) until cell concentrations reached between (Millipore). 50,000 and 100,000 cells mL1. For the experimental cultures, Cultures of E. huxleyi RCC 1832 (Western Mediterranean polycarbonate Nalgene bottles (2.95 liter) completely filled strain collected from latitude 39 1000 N and longitude 5 3500 E; with sterile culture medium with target pH levels (refer to Type A morphotype) from the Roscoff Culture Collection “Carbonate system” section) were then inoculated with either (www.roscoff-culture-collection.org) were initially grown for 500, 1000, or 2000 cells mL1 from the corresponding 1 week at a light intensity of 150 μmol m2 s1 in a 16/8 h acclimation culture, and incubated in the same temperature 3 Johnson et al. Nutr response of e. hux to pH and temp and light conditions as the acclimation cultures (without rep- minimize gas exchange. Cell growth rates were determined lication). Experimental cultures were sampled for the response using the equation growth rate = [Ln(Df)  Ln(D0)]/t, where Df variables once cell counts reached approximately 50,000– is the final concentration, D0 is the initial concentration, and 100,000 cells mL1. This so-called dilute batch approach t is the time in days. At the end of the experiment, all samples ensures a quasi-constant seawater carbonate chemistry (Langer were taken within 3 h of conducting cell counts. et al. 2009). As a western Mediterranean Sea strain of E. huxleyi was used Lipids here, the temperature levels were chosen to reflect current and For total lipid extraction, triplicate 0.15 liter samples were fil- future projected temperatures for this region. The optimum tered onto precombusted (8 h, 450 C) GF/F Whatman glass-fiber growth temperature for this strain is likely within the range of filters, immediately frozen in liquid N2 and then stored at 20 C 22–25 C based on other strains of E. huxleyi collected from the until analysis. Lipids were extracted in chloroform : methanol western Mediterranean Sea. The higher temperature of 20 C (2 : 1) following Barnes and Blackstock (1973) (colorimetry). Sam- was chosen to ensure that thermal damage to lipids and pro- ples were evaporated and sulfuric acid was added; the final reac- teins did not occur, and that the cultures did not crash, partic- tion was performed using vanillin. Cholesterol was used as a ularly as the combined, and possible synergistic, effect of standard, and measurements were read using a spectrophotome- temperature and pH was not known. This is particularly ter (UV mini1240, Shimadzu). Lipid cellular quotas are reported important when investigating the impact of multiple stressors, as pg cell1. The production rate for lipids (pg cell1 d1) was cal- as reduced pH and increased temperature are known to have culated as: negative interactive effects on multiple physiological parame- ters for coccolithophores (Arnold et al. 2013; Sett et al. 2014; Lipid production ¼ growth rate  cellular lipid content pg cell1 : D’Amario et al. 2017). In the Algerian Basin, where this cul- ture originated from, recorded sea surface temperatures range Production potential is a parameter that extrapolates the cellu- from 14 C to 16 C in winter, 18 C to 18.4 C in spring (for lar lipid production of a community of coccolithophores. This 80% of the basin), 19.5 C to 19.9 C in autumn (for 50% of parameter assumes exponential growth and starts from a the basin), and 24.5 C to 24.9 C in summer (for 70% of the known cell density (1 cell). The corresponding lipid produc- basin; Shaltout and Omstedt 2014). Anticipating a 3 C tion of the community can be calculated using the growth warming of sea surface temperatures by 2100, 20 C is a useful rate (μ) and cellular lipid content (pg cell1) to determine the temperature to investigate the potential impacts of ocean amount of lipids available to consumers after 1 week of warming, as by 2100, average temperatures in the region that growth (Gafar et al. 2018; Klintzsch et al. 2019). The following this strain of E. huxleyi was collected will be closer to 20 C for equation was used to calculate the production potential of most of the year. A 5 C difference between temperatures was lipids: chosen to ensure a clear, observable effect (if there was one) of temperature on the response variables. mðlipidsÞ PPlipids ¼ N 0  eμt  , We utilized a large range of pH levels (ranging from 7.66 to cell 8.44) and two temperatures (15 C and 20 C) to create a response curve for the dependent variables over the two tem- where PPlipids is the production potential of lipids after 7 d, N0 perature levels. The lower end of the pH range is in line with, is the assumed cell density of the community (using a starting and exceeds, future predictions of ocean acidification over the cell density of 1), e is the exponential growth factor, μ is the next century, the middle range is in line with current ocean growth rate, t is the time in days (here we use 7 d), m (lipids) pH levels, and the upper range includes preindustrial pH levels is the cellular lipid content (pg cell1) of each experimental and above. A higher number of treatment levels for pH were culture, and cell is the final cell concentration. used at the expense of replication and a generalized linear model (GLM) including linear and nonlinear terms was used Particulate inorganic/organic carbon and nitrogen ratios to analyze the data (Cottingham et al. 2005). For the determination of total particulate carbon (TPC includes both inorganic and organic carbon; duplicates) and Growth rates particulate organic carbon (POC; duplicates), 0.25 liter of Cell counts were conducted using a Guava easyCyte HT water was filtered onto precombusted (8 h, 450 C) 0.7-μm flow cytometer (software guavaSoft 3.1.1—detected via chloro- nominal pore size glass fiber filters (Whatman GF/F) and phyll fluorescence) at the same time each day (after cell counts stored at 20 C. The filters for POC analysis were dried at were estimated to be above 4000 cells mL1). Cells were 60 C for 24 h, then exposed to HCl vapors for 48 h to dissolve suspended in the Nalgene bottles to ensure cell concentrations the calcite and convert all particulate inorganic carbon (PIC) were homogenous by inverting the bottle 5–10 times. Cells to CO2, leaving only POC to be measured (Rossi and counts were conducted in duplicate (200 μL). The Nalgene Gili 2007). These filters were then redried at 60 C for 24 h. bottle lids were covered in parafilm between each sampling to The analyses were performed with an elemental analyzer 4 Johnson et al. Nutr response of e. hux to pH and temp (Elementar Vario Pyro Cube EA CNS; Elementar Anal- ysensysteme GmbH, Hanau, Germany). To calculate PIC, which represents the calcite shell of the coccolithophore, POC was subtracted from TPC. PIC and POC (pg cell1) were calcu- lated as well as C : N ratios (particulate inorganic or organic carbon : nitrogen). Rates of production for particulate inorganic carbon (pg cell1 d1) and particulate organic carbon (pg cell1 d1) were calculated as: PIC production ¼ growth rate  PIC content pg cell1 , and POC production ¼ growth rate  POC content pg cell1 : The ratio between lipids (pg cell1) and POC (pg cell1) was used to explore the relationship between cellular lipids and POC content (lipid : POC * 100). Chlorophyll a To determine the Chl a concentration, 3  100 mL repli- cates were filtered through GF/F precombusted glass fiber filters and stored at 20 C. Chlorophyll was extracted in 8 mL 90% acetone and left for 24 h covered with aluminum foil at + 4 C. Samples were then centrifuged for 15 min at 10,000g and 4 C, and then analyzed using a spectrophotometer (UV mini1240, Shimadzu). The absorbance was read at 630, 663, and 750 nm. Chl a concentration was calculated according to the spectrometric equations reported in Jeffrey and Humphrey (1975). Chl a cellular quotas are reported as pg cell1. The production rate for Chl a (pg cell1 d1) was calcu- lated as: Chl a production ¼ growth rate Fig. 1. (A) An intact coccosphere and (B) a collapsed coccosphere of Emiliania huxleyi. Images taken using a desktop Phenom G2 pro scanning  cellular Chl a content pg cell1 : electron microscope. Morphology interlocked coccoliths; Fig. 1) and intact coccospheres (Fig. 1) Malformations in coccolith formation might lead to were conducted using a scanning electron microscope (Merlin, unstable coccospheres, which in turn may lead to impaired Zeiss). The results are reported as a percentage of the total cell protection (Monteiro et al. 2016; Kottmeier et al. 2022). To count. investigate potential malformation, 1 mL samples for scan- ning electron microscopy were filtered onto 0.8-μm pore Carbonate system size polycarbonate filters (Millipore) and dried in a drying The pH of experimental cultures was adjusted through cal- cabinet at 55 C for 24–48 h. A desktop Phenom G2 pro culated additions of either HCl or NaOH. Samples for total scanning electron microscope was used to take images for alkalinity (TA) were collected from the filtrate of the TPC/POC subsequent analysis (using ImageJ software) of coccosphere filters and stored at 4 C for a maximum of 7 d before diameter (along the longest axis) and coccolith morphologi- processing. The TA was measured in duplicates via titration cal features (distal shield length, distal shield width, inner with Metrohm 877 Titrino plus (software tiBase 1.1) at 25 C circle diameter, tube width; Supporting Information using Dickson Certified Standards (Batch #100). Fig. S1). Samples for pH were collected using unfiltered experimen- To determine whether the integrity of the coccolithophore tal medium and stored in gas tight bottles, without air cell was maintained, cell counts of collapsed (four or more bubbles. All samples were measured spectrophotometrically in 5 Johnson et al. Nutr response of e. hux to pH and temp triplicate within 3 h of collection in the method of Liu et al. Response variable  temperature þ pH þ pH2 þ pH3 (2011) at 25 C (PerkinElmner UV/VIS Lambda 365). þ temperature pH: The carbonate system was calculated with TA and pH using CO2SYS (Pierrot et al. 2011). The TA and pH were measured at The model was run with 100 iterations and Wald chi- 25 C and used to calculate DIC. Following this, TA and DIC squared statistics and confidence intervals were used. For the were used to calculate the system at the target temperature covariance matrix, the robust estimate was used which (either 15 C or 20 C; Supporting Information Table S1). provides a more conservative model. A Type III analysis was chosen as it holds all the variables constant relative to each Statistics other. Data from each sample were averaged to create a Data were analyzed using a GLM (Nelder and single data point for each treatment. Assumptions of normal- Wedderburn 1972; Zuur et al. 2009a). Each model was run ity were satisfied for each dependent variable (Shapiro–Wilk using a gamma distribution with log link (continuous test). Outliers were defined as any data point outside of the response variable Y that has positive values > 0; Zuur following ranges—3rd quartile + 1.5 * interquartile range or 1st et al. 2009b), and chosen following a backward selection quartile—1.5 * interquartile range, and were removed from criteria and based on the lowest Akaike information criterion the dataset for the statistical analysis. The reference category is (AIC) score (Aho et al. 2014). In the model, temperature was included for comparison against the other predictors. The selected as the fixed factor and pH as the co-variate, including models were run using IBM SPSS v22. The GLM results are their interaction and possible nonlinear responses (with a reported in detail in the Supporting Information (Table S2). maximum nonlinearity of a cubic term—pH3). Significant var- iables (p < 0.05) for the chosen models were included. Only significant terms were kept in the model except in instances Results where models including nonsignificant variables had lower Cellular quotas AIC scores. The following linear and nonlinear predictors were The average cellular lipid content for both temperatures tested in the GLM: across all pH levels was 44.75 pg cell 1 (SD = 9.25; Fig. 2. Cellular quotas in response to pH levels and temperatures. (A) Lipids (pg cell1), (B) POC (pg cell1), (C) PIC (pg cell1), and (D) Chl a (pg cell1). Error bars indicate data standard deviation within each treatment, not between replicates. 6 Johnson et al. Nutr response of e. hux to pH and temp Fig. 3. Production rates in response to pH levels and temperatures. (A) Growth rate (day1), (B) lipid production (pg cell1 d1), (C) POC production (pg cell1 d1), (D) PIC production (pg cell1 d1), (E) Chl a production (pg cell1 d1), and (F) production potential—lipids (ng). Error bars indicate data standard deviation within each treatment, not between replicates. range = 29.42–61.34 pg cell1). The average cellular lipid con- Cellular POC content (pg cell1) followed a parabolic tent at 15 C (47.66 pg cell1; SD = 6.61) was slightly higher than trend at 20 C (Fig. 2B), with a higher average of 27.50 pg at 20 C (41.85 pg cell1; SD = 10.92). Cellular lipid content cell1 (SD = 7.25) than at 15 C (25.64 pg cell1; SD = 3.10). increased as pH decreased and this trend appeared to be similar At 15 C, cellular POC content had a linear trend, which across both temperatures (Fig. 2A). The quadratic term for pH was slightly decreased toward higher pH (Fig. 2B). The selected significant in the selected model (Supporting Information model includes a significant quadratic term for pH Table S2). Temperature, pH, and their interaction were significant (Supporting Information Table S2). Temperature was not a predictors for cellular lipid content (Supporting Information significant predictor yet is still included in the model (lowest Table S2). AIC value). 7 Johnson et al. Nutr response of e. hux to pH and temp Fig. 4. Carbon ratios in response to pH levels and temperatures. (A) POC : nitrogen ratio (POC : N), (B) PIC : nitrogen ratio (PIC : N), (C) PIC : POC, and (D) cellular lipid quota : cellular particulate organic carbon quota (lipid : POC). Error bars indicate data standard deviation within each treatment, not between replicates. Cellular PIC content (pg cell1) followed a similar parabolic The growth rate at 20 C was distinctly higher with an average trend at both temperatures (Fig. 2C). Average cellular PIC content rate of 0.86 (day1; SD = 0.07; range = 0.75–0.98) compared at 20 C (13.48 pg cell1; SD = 4.95) was slightly higher than at with 0.62 (day1; SD = 0.04; range = 0.55–0.67) at 15 C. 20 C (12.22 pg cell1; SD = 4.38), both similar to the overall aver- Growth rate showed no clear trend from low to high pH age of 12.85 pg cell1 (SD = 4.58). The highest cellular PIC con- (Fig. 3A). The growth rate at 20 C was overall higher than at tent was 21.34 pg cell1 at 20 C and pH 7.97 (Fig. 2C). PIC 15 C, with a visible trend of decreasing growth rate with content responded in a cubic relationship with pH, which was decreasing pH. Temperature, pH (quadratic), and the interac- the only significant predictor (Supporting Information Table S2). tion between temperature and pH were significant predictors Temperature and the interaction between temperature and pH of the growth rate response for E. huxleyi (Supporting Informa- are still included in the model (lowest AIC score). tion Table S2). Cellular Chl a content followed distinctly different trends The average lipid production for both temperatures was at 15 C and 20 C and was stable across the pH range at 15 C, 32.68 pg cell1 d1 (SD = 6.95; range = 20.19–46.31 pg while 20 C exhibited a clear negative trend from low to high cell1 d1), with each temperature group having similar aver- pH (Fig. 2D). Average cellular Chl a content at 15 C was age values (15 C = 29.86 and 20 C = 35.49 pg cell1 d1; 0.27 pg cell1 (SD = 0.03) and 0.44 pg cell1 (SD = 0.18) at Supporting Information Table S2). Lipid production 20 C. A linear term for pH was significant in the selected increased as pH decreased; however, this response varied model. Temperature, pH, and their interaction were significant between temperatures, particularly at the lower range of pH predictors for cellular Chl a content (Supporting Information (Fig. 3B). Temperature and pH (quadratic) were significant Table S2). predictors for lipid production in E. huxleyi (Supporting Information Table S2). Production rates The average POC production at 20 C (23.62 pg cell1 d1; The average growth rate (day1) over both temperatures SD = 6.34) was higher than at 15 C (16.01 pg cell1 d1; was 0.74 (day1; SD = 0.13) with a range of 0.55–0.98 (day1). SD = 2.44). There was a bell curve trend at 20 C, with a higher 8 Johnson et al. Nutr response of e. hux to pH and temp Fig. 5. Coccosphere morphological characteristics in response to pH levels and temperatures. (A) Coccosphere diameter (μm), (B) distal shield length (μm), (C) distal shield width (μm), (D) inner circle diameter (μm), (E) tube width (μm), and (F) collapsed coccospheres (%). Error bars indicate data stan- dard deviation within each treatment, not between replicates (not applicable for collapsed coccospheres as percentage was calculated from a single filter). POC production within the middle of the pH range, while (Fig. 3D). The selected model included a significant cubic term there was no clear trend at 15 C (Fig. 3C). The selected model for pH, and temperature and pH were both significant predic- included a significant quadratic term for pH, which was the tors for POC production (Supporting Information Table S2). only significant predictor of POC production for E. huxleyi Chl a production (pg cell1 d1) remained stable across all (Supporting Information Table S2). pH levels at 15 C (average = 0.15 pg cell1; SD = 0.04), while The average PIC production (pg cell1 d1) was higher at it gradually declined from low to high pH at 20 C 20 C (11.57 pg cell1 d1; SD = 4.29) than at 15 C (7.71 pg  (average = 0.32 pg cell1; SD = 0.15; Fig. 3E). Temperature cell1 d1; SD = 2.94), with a bell curve trend across the pH and pH (quadratic) were significant predictors for Chl range, which was more pronounced at the higher temperature a production (Supporting Information Table S2). 9 Johnson et al. Nutr response of e. hux to pH and temp The production potential of lipids showed a clear difference predictors of coccosphere diameter for E. huxleyi (Supporting between 15 C and 20 C, and there was a trend of increasing Information Table S2). production potential as pH increased, but only at 20 C Distal shield length (μm) was greater at 15 C than 20 C. It (Fig. 3F). This largely followed the pattern of growth rate also exhibited slight bell curve response at 15 C (Fig. 5B). A (Fig. 3A), showing that growth rate was the major factor driv- quadratic term for pH was significant in the selected model ing the production potential for E. huxleyi here. Similar to and temperature and pH were significant predictors for distal growth rate, the quadratic term for pH was significant in the shield length (Supporting Information Table S2). Distal shield selected model. Temperature, pH, and their interaction were width (μm) followed a similar pattern to distal shield length; significant predictors of the production potential of lipids however, here only temperature and the interaction between (Supporting Information Table S2). temperature and pH were significant predictors (Fig. 5C; Supporting Information Table S2). Although pH was a nonsig- Carbon ratios nificant term for the model, it was retained as it is included in Mean POC : N ratio for all treatments was 6.28 (range 4.9– the interaction. It may be the effect of high variability associ- 9.2). POC : N ratios were higher at 20 C (6.74; SD = 1.35) ated with distal shield length that may explain why pH was than at 15 C (6.74; SD = 1.35). Both temperatures followed a not a significant predictor here. similar trend of decreasing POC : N with increasing pH Average inner circle diameter (μm) remained largely stable (Fig. 4A). This trend was stronger at 20 C. The cubic term for across pH at both temperatures, however, was highly variable. pH was significant in the selected model, and pH, tempera- Inner circle diameter had no significant predictors (Fig. 5D; ture, and their interaction were significant predictors of the Supporting Information Table S2). Coccolith tube width POC : N response in E. huxleyi (Supporting Information increased with increasing pH for both temperatures, and pH Table S2). was a significant predictor (Fig. 5E; Supporting Information Similar to POC : N, PIC : N ratios were higher at 20 C Table S2). (x̅> = 3.77; SD = 0.52) than at 15 C (x̅ = 3.15; SD = 0.55); Overall, collapsed coccospheres (%) made up approxi- however, this difference was not as distinct. At low pH, mately 37% of the cells. This differed with temperature, with PIC : N showed similar values for both temperatures; however, an average of 35% at 15 C, and 39% at 20 C. A cubic term for at approximately pH 8.0 and above, the response curves pH was significant in the selected model and is the only signif- started to diverge, with PIC : N values higher at 20 C than at icant predictor for collapsed coccospheres (Supporting Infor- 15 C (Fig. 4B). Temperature and the interaction between tem- mation Table S2), with the percentage of collapsed perature and pH were significant predictors of the PIC : N coccospheres increasing toward the lower pH range (Fig. 5F). response of E. huxleyi (Supporting Information Table S2). For Temperature and the interaction between temperature and pH PIC : N, pH in still included in the selected model and was were not significant; however, they were included in the minorly significant (p < 0.1). selected model (lowest AIC score). The average PIC : POC ratio was similar across both temper- atures (15 C = 0.49, SD = 0.14; 20 C = 0.48, SD = 0.09); how- ever, PIC : POC had a parabolic relationship with pH, with the Discussion highest ratios in the middle range of the pH range, where it Consumer impacts reached a maximum of 0.69 at pH 8.23 before declining Food quantity toward low and high pH (Fig. 4C). Both temperatures followed In this study, we show that temperature is an important this relationship curve, and the results of the GLM indicate predictor for the production potential of lipids, indicating that pH (cubic term) was a significant predictor of the increases in lipid availability under ocean warming. This PIC : POC response of E. huxleyi (Supporting Information increase will likely be energetically favorable for consumers. Table S2). pH is also a significant predictor of the production potential The lipid : POC ratio overall was higher at 15 C (aver- of lipids, and the results suggest that this increase in availabil- age = 1.87, SD = 0.21) than at 20 C (average = 1.6, ity will be mitigated by ocean acidification. The availability of SD = 0.63), except at pH 7.74, where the ratio distinctly lipids in coccolithophores for their consumers is important increases at 20 C (Fig. 4D). Temperature was a significant pre- when investigating potential energy transfer between trophic dictor for lipid : POC (Supporting Information Table S2). levels and to understand how this may change under ocean acidification and warming. The production potential of lipids Morphology is strongly influenced by the growth rate, and it is distinctly Coccosphere diameter (μm) followed similar trends for both higher under the warmer temperature (Fig. 3F). The decrease temperatures in the higher pH range; however, at low pH, in production potential under low pH was mostly related to coccosphere diameter was smaller at 20 C (Fig. 5A). The the impact of pH on growth rate, rather than the impact of selected model includes a significant quadratic term for pH pH on cellular lipid quota, which increased under lower pH and temperature, and their interaction was significant (Fig. 2A). At any rate, this suggests that consumers may be 10 Johnson et al. Nutr response of e. hux to pH and temp positively affected by the increased quantity of available cells, microalgae and macroalgae (Kuwata et al. 1993; Fuentes-Grü- and in warmer conditions, there will be greater lipid availabil- newald et al. 2012; Prabhu et al. 2019). ity, which is likely to be energetically beneficial for coccolithophore consumers. Coccosphere integrity and PIC : POC Ratio Coccoliths can diminish the nutritional value of coccolithophores in three ways. First, through mechanical Food quality protection, making it more difficult for grazers to access the The interaction between pH and temperature was an protoplast (Haunost et al. 2021). Second, through the need to important driver for POC : N which was higher at low pH counteract the increase in the pH of guts or food vacuoles due ( 7.9–7.7) and 20 C (Fig. 4A), indicating that the availability to carbonate dissolution (Harvey et al. 2015). Third, through of nitrogen for coccolithophore consumers will be reduced reduction in the percentage of digestible material (Haunost under ocean acidification and warming. Nitrogen is essential et al. 2021). While the first point hinges on morphology, the for cellular function and is involved in the production of mac- last two points center on the amount of calcite relative to romolecules, such as proteins or amino acids, including chlo- organic material. To assess the first point, we focus on the per- rophyll (Riegman et al. 2000). A higher POC : N suggests a centage of collapsed coccospheres, which indicates reduced lower quality food as a higher ratio of carbon can indicate mechanical stability (Jaya et al. 2016), and to assess the last nitrogen limitation. The mean POC : N ratio across all treat- two points, we analyzed the PIC : POC ratio. ments is similar to the Redfield ratio of 6.3 (Martiny The percentage of collapsed coccospheres increased with et al. 2014); however, this average differs with temperature. decreasing pH at both temperatures, indicating reduced struc- The range of POC : N is similar to what is expected in natural tural integrity of the coccosphere (Fig. 5F). Copepods do not populations of phytoplankton (Geider and La Roche 2002) need to destroy the coccosphere as they have to in the case of and experimental work with coccolithophores (Fiorini the diatom frustule (Langer et al. 2007; Jansen 2008); how- et al. 2010). Contrary to our results, there was no significant ever, dinoflagellates need to penetrate the coccosphere to effect of elevated pCO2 (760 μatm corresponding to pH 7.81) digest the cell contents, and then egest the remaining indi- on POC : N for E. huxleyi strain (AC472) from the South Pacific gestible calcium carbonate (Haunost et al. 2021). This means (Fiorini et al. 2010). However, the difference in pH levels was that if the interlocking between coccoliths is reduced, dinofla- not as pronounced as they are in this experiment. For gellates may be able to more readily access coccolithophore instance, an increase in POC : N was also reported under ele- cell contents, potentially aiding in digestion. vated pCO2 for E. huxleyi using a wider range of pH levels The PIC : POC ratio was unaffected by temperature; how- (6.8–8.3; Iglesias-Rodriguez et al. 2008—morphotype R). ever, we observed a distinct bell-curve in response to pH under Under ocean acidification conditions, the reduced nitrogen both temperatures (Fig. 2H). Although PIC : POC response pat- availability might have negative impacts on micrograzers that terns are strain specific (Langer et al. 2009; Rosas-Navarro have issues with nitrogen limitation. et al. 2016), a meta-analysis suggests a negative correlation Lipid content is also an important indicator of the nutri- between PIC : POC and pH (Meyer and Riebesell 2015). Taken tional quality of coccolithophores for secondary consumers together, the PIC : POC ratio and the coccosphere integrity (Pond and Harris 1996), as also seen for diatoms in particular, response suggest that ocean acidification will make it easier and plankton in general (Rossoll et al. 2012; Meyers for grazers to digest coccolithophores and access cell contents. et al. 2019). Although we see greater availability of E. huxleyi Therefore, macrograzers such as copepods will likely benefit in terms of the production potential of lipids (Fig. 3F), given from the higher percentage of digestible material, and less so that the availability of nitrogen under both ocean acidification from both the reduced structural integrity of the coccosphere and warming will be reduced (Fig. 4A), the quality of and the potentially negligible gut pH change (Harris 1994; coccolithophores as a food source will likely be negatively Nejstgaard et al. 1997; Langer et al. 2007b; White et al. 2018; impacted (Conde-Porcuna et al. 2002; Mitra and Flynn 2005; Mayers et al. 2020). Micrograzers such as dinoflagellates will Iglesias-Rodriguez et al. 2008). However, as cellular lipid con- likely benefit from all three aspects of coccolith-related nutri- tent is predicted to either be unaffected (Fiorini et al. 2010) or tional value change (Harvey et al. 2015; Jaya et al. 2016; increase under ocean acidification conditions, as is the case Haunost et al. 2021). here (Fig. 2A), the nutritional impacts on coccolithophore consumers will likely be varied. From the perspective of Individual response lipid : POC, however, food quality appears unaltered Physiological rates (Kharbush et al. 2020) except under the lowest pH at 20 C, The response to ocean acidification conditions in terms of which is exceptionally high (Fig. 4D). This increase might physiological rates generally shows high inter- and intra- indicate a crossed threshold under ocean acidification condi- species variability (Langer et al. 2009; Hoppe et al. 2011). After tions at 20 C involving a carbon storage strategy related to initial uncertainty as to the interpretation of this variability, extreme stress which has been previously reported for both the work of Bach and co-authors has produced a plausible and 11 Johnson et al. Nutr response of e. hux to pH and temp now widely accepted interpretation using substrate–inhibitor difficulty digesting calcite may benefit from decreases in concept (Bach et al. 2011, 2013, 2015; Paul and Bach 2020), PIC : POC under ocean acidification. which states that with increasing substrate (CO2) concentra- Phytoplankton biomass is predicted to have a varied tion, POC production increases up to the point when the response to ocean acidification and warming, and this inhibitory effect of increasing H+ concentration then causes response will largely vary depending on latitude and taxo- a decrease in POC production (ca. pH 8 in our data, Fig. 3C). nomic group (Seifert et al. 2020). Due to the calcifying nature In these studies, it was shown that the response of of coccolithophores, they will have the additional stress of coccolithophores to seawater carbonate chemistry changes is reduced seawater pH (Guinder and Molinero 2013), affecting best thought of as a bell curve, with different strains featuring their ability to calcify (Riebesell et al. 2000; Engel et al. 2005; different optima. In our case, we can see the bell curve in the Hoppe et al. 2011; Lefebvre et al. 2011; Schlüter et al. 2014) PIC and POC production, while other rates display only part and potentially pushing them closer to a stress response than of the full bell curve (Fig. 4A,B,F,E). Growth rate, for instance, noncalcifying plankton. Although adaptation to ocean acidifi- shows a somewhat atypical increase at 20 C (Iglesias- cation and warming has been shown to occur relatively rap- Rodriguez et al. 2008; Langer et al. 2009—morphotypes A, B, idly in E. huxleyi (1 yr for PIC and POC to return to present- R; Mackey et al. 2015); however, the dependence of growth day levels; growth 16% higher than nonadapted controls after rate on temperature has been seen in several coccolithophorid 1 yr; Schlüter et al. 2014), for the coccolithophore strains (Buitenhuis et al. 2008; Fielding 2013). The response Gephyrocapsa oceanica, growth rate (after an initial increase), patterns over the pH range tested here do not appear to be POC production, and nitrogen production all decreased over related to morphotype or other easily identifiable strain fea- 2000 generations (approximately 1400 d) under high CO2 tures (see also Krumhardt et al. 2017). conditions, indicating that resilience to ocean acidification conditions can reduce over time (Jin and Gao 2016). In Morphological parameters as PIC production proxy another experiment, E. huxleyi cells adjusted their chlorophyll The decrease in PIC production does not correlate with content more rapidly than other coccolithophore species changes in any morphological characteristics at both tempera- (Lewis et al. 1984), potentially as an adaptive photosynthetic tures or across the pH range. Coccosphere size and coccolith trait to changing ocean conditions (Feng et al. 2008). This sug- weight (usually correlated with coccolith size) have been pro- gests that the ability to adapt to ocean acidification conditions posed as proxies for PIC production (Beaufort et al. 2011; Bach will likely vary depending on the species and physiological et al. 2012; Gibbs et al. 2013). In contrast to our study, Rosas- parameter in question. Navarro et al. (2016) described a positive correlation of We may see future shifts in E. huxleyi ranges as they shift to coccolith size and PIC production in three E. huxleyi strains. warmer areas where they achieve a more favorable growth rate This discrepancy could stem from strain specificity but seems (Neukermans et al. 2018), and this may further affect other unrelated to morphotype because all strains considered here organisms that rely on coccolithophores as a food source. are Morphotype A. A widespread strain specificity in the rela- Despite inter- and intra-species variability in response patterns tionship between coccolith size and PIC production would (Langer et al. 2006, 2009), our results suggest that combined make the application of this proxy difficult. low pH and an increased temperature has the potential to increase coccolithophore lipid standing stock, but decrease food quality, as inferred from an increased POC : N ratio. Future ramifications The short-term experiment presented here provides us with Future coccolithophore consumers will likely benefit from important information about the immediate response of an energetic standpoint, due to the increase in lipid availabil- coccolithophores to changes in temperature and pH, and this ity under ocean warming; however, this increase will be miti- can be decisive in field scenarios, particularly in cases where gated by ocean acidification. As this experiment was done populations are not given the opportunity to adapt. Further under nutrient replete exponential growth, concurrent work investigating the long-term impact of ocean acidification changes in nutrient availability may alter these impacts on and warming on the nutritional content (in terms of both the nutritional condition of E. huxleyi (Müller et al. 2017). quality and quantity) of coccolithophores, a globally impor- From a functional point of view, a higher POC : N will not tant phytoplankton at the base of the marine food web, will contribute as much to cell function for consumers, and as be a key part of understanding the impacts of climate change such, the capacity of coccolithophores to provide a healthy on ocean trophic dynamics. food source may be reduced. The nutritional quality of coccolithophores as indicated by POC : N, lipid : POC, and PIC : POC will likely have varied impacts, depending on an Data availability statement organism’s specific requirements. For example, organisms with The data that support the findings of this study will be limited nutrient supplies could be affected by increases in available from PANGAEA.de and are also available from the POC : N under future conditions, while organisms that have corresponding author, R.J., upon request. 12 Johnson et al. Nutr response of e. hux to pH and temp References of temperature. Limnol. Oceanogr. 53: 1181–1185. doi: Aho, K., D. Derryberry, and T. Peterson. 2014. Model selection 10.4319/lo.2008.53.3.1181 for ecologists: The worldviews of AIC and BIC. Ecology 95: Chavez, F. P., M. Messié, and J. T. Pennington. 2010. Marine 631–636. doi:10.1890/13-1452.1 primary production in relation to climate variability and Arnold, H. E., P. Kerrison, and M. Steinke. 2013. Interacting change. Ann. Rev. Mar. Sci. 3: 227–260. doi:10.1146/ effects of ocean acidification and warming on growth and annurev.marine.010908.163917 DMS-production in the haptophyte coccolithophore Conde-Porcuna, J. M., E. Ramos-Rodriguez, and C. Perez- Emiliania huxleyi. Global Chang. Biol. 19: 1007–1016. doi: Martinez. 2002. Correlations between nutrient concentra- 10.1111/gcb.12105 tions and zooplankton populations in a mesotrophic reser- Bach, L. T., U. Riebesell, and K. G. Schulz. 2011. Dis- voir. Freshw. Biol. 47: 1463–1473. doi:10.1046/j.1365-2427. tinguishing between the effects of ocean acidification and 2002.00882.x ocean carbonation in the coccolithophore Emiliania Cottingham, K. L., J. T. Lennon, and B. L. Brown. 2005. huxleyi. Limnol. Oceanogr. 56: 2040–2050. doi:10.4319/lo. Knowing when to draw the line: Designing more infor- 2011.56.6.2040 mative ecological experiments. Front. Ecol. Environ. 3: Bach, L. T., C. Bauke, K. J. S. Meier, U. Riebesell, and K. G. Schulz. 145–152 doi:10.1890/1540-9295(2005)003[0145: 2012. Influence of changing carbonate chemistry on morphol- KWTDTL]2.0.CO;2. ogy and weight of coccoliths formed by Emiliania huxleyi. Cripps, G., K. J. J. Flynn, and P. K. K. Lindeque. 2016. Ocean Biogeosciences 9: 3449–3463. doi:10.5194/bg-9-3449-2012 acidification affects the phyto-zoo plankton trophic trans- Bach, L. T., L. C. M. Mackinder, K. G. Schulz, G. Wheeler, fer efficiency. PLoS One 11: e0151739. D. C. Schroeder, C. Brownlee, and U. Riebesell. 2013. Dis- D’Amario, B., P. Ziveri, M. Grelaud, A. Oviedo, and M. Kralj. secting the impact of CO2 and pH on the mechanisms of 2017. Coccolithophore haploid and diploid distribution photosynthesis and calcification in the coccolithophore patterns in the Mediterranean Sea: Can a haplo-diploid life Emiliania huxleyi. New Phytol. 199: 121–134. doi:10.1111/ cycle be advantageous under climate change? J. Plankton nph.12225 Res. 39: 781–794. doi:10.1093/plankt/fbx044 Bach, L. T., U. Riebesell, M. A. Gutowska, L. Federwisch, and Doney, S. C., V. J. Fabry, R. A. Feely, and J. A. Kleypas. 2009. K. G. Schulz. 2015. A unifying concept of coccolithophore Ocean acidification: The other CO2 problem. Ann. Rev. sensitivity to changing carbonate chemistry embedded in Mar. Sci. 1: 169–192. doi:10.1146/annurev.marine.010908. an ecological framework. Prog. Oceanogr. 135: 125–138. 163834 doi:10.1016/j.pocean.2015.04.012 El-Hady, H., S. Fathey, G. Ali, and Y. Gabr. 2016. Biochemical Barnes, H., and J. Blackstock. 1973. Estimation of lipids in profile of phytoplankton and its nutritional aspects in some marine animals and tissues: Detailed investigation of the khors of Lake Nasser, Egypt. Egypt. J. Basic Appl. Sci. 3: sulphophosphovanilun method for ‘total’ lipids. J. Exp. Mar. 187-193. doi:10.1016/j.ejbas.2016.03.002 Bio. Ecol. 12: 103–118. doi:10.1016/0022-0981(73)90040-3 Engel, A., and others. 2005. Testing the direct effect of CO2 Beaufort, L., and others. 2011. Sensitivity of coccolithophores concentration on a bloom of the coccolithophorid to carbonate chemistry and ocean acidification. Nature Emiliania huxleyi in mesocosm experiments. Limnol. 476: 80–83. doi:10.1038/nature10295, 7358 Oceanogr. 50: 493–507. doi:10.4319/lo.2005.50.2.0493 Bendle, J., A. Rosell-Melé, and P. Ziveri. 2005. Variability of unusual Fabry, V. J., B. A. Seibel, R. A. Feely, and J. C. Orr. 2008. distributions of alkenones in the surface waters of the Nordic Impacts of ocean acidification on marine fauna and ecosys- seas. Paleoceanography 20. doi:10.1029/2004PA001025 tem processes. ICES J. Mar. Sci. 65: 414–432. doi:10.1093/ Breteler, W. C. M., N. Schogt, and S. Rampen. 2005. Effect of icesjms/fsn048 diatom nutrient limitation on copepod development: Role Falkowski, P. G., T. Fenchel, and E. F. Delong. 2008. The of essential lipids. Mar. Ecol. Prog. Ser. 291: 125–133. doi: microbial engines that drive earth’s biogeochemical cycles. 10.3354/meps291125 Science 320: 1034–1039. doi:10.1126/science.1153213 Broglio, E., S. Jonasdottir, A. Calbet, H. Jakobsen, and E. Saiz. Feng, Y., M. E. Warner, Y. Zhan, J. Sun, F. Fu, J. M. Rose, and 2003. Effect of heterotrophic versus autotrophic food on D. A. Hutchins. 2008. Interactive effects of increased pCO2, feeding and reproduction of the calanoid copepod Acartia temperature and irradiance on the marine coccolithophore tonsa: Relationship with prey fatty acid composition. Emiliania huxleyi (Prymnesiophyceae). Eur. J. Phycol. 43: Aquat. Microb. Ecol. 31: 267–278. 87–98. doi:10.1080/09670260701664674 Brown, C. W., and J. A. Yoder. 1994. Coccolithophorid Feng, Y., M. Y. Roleda, E. Armstrong, P. W. Boyd, and C. L. blooms in the global ocean. J. Geophys. Res. Ocean 99: Hurd. 2017. Environmental controls on the growth, photo- 7467–7482. doi:10.1029/93JC02156 synthetic and calcification rates of a Southern Hemisphere Buitenhuis, E. T., P. T, D. J. Franklin, and others. 2008. strain of the coccolithophore Emiliania huxleyi. Limnol. Growth rates of six coccolithophorid strains as a function Oceanogr. 62: 519–540. doi:10.1002/lno.10442 13 Johnson et al. Nutr response of e. hux to pH and temp Fielding, S. R. R. 2013. Emiliania huxleyi specific growth rate coccolithophores on ingestion and growth of a dinoflagel- dependence on temperature. Limnol. Oceanogr. 58: 663– late predator. Front. Mar. Sci: 8. 666. doi:10.4319/lo.2013.58.2.0663 Holligan, P. M., and others. 1993. A biogeochemical study of Fiorini, S., J.-P. Gattuso, P. van Rijswijk, and J. J. Middelburg. the coccolithophore, Emiliania huxleyi, in the North Atlan- 2010. Coccolithophores lipid and carbon isotope composi- tic. Global Biogeochem. Cycl. 7: 879–900. doi:10.1029/ tion and their variability related to changes in seawater car- 93GB01731 bonate chemistry. J. Exp. Mar. Biol. Ecol. 394: 74–85. doi: Hoppe, C. J. M., G. Langer, and B. Rost. 2011. Emiliania huxleyi 10.1016/j.jembe.2010.07.020 shows identical responses to elevated pCO2 in TA and DIC Fraser, A. J., J. R. Sargent, J. C. Gamble, and D. D. Seaton. manipulations. J. Exp. Mar. Biol. Ecol. 406: 54–62. doi:10. 1989. Formation and transfer of fatty acids in an enclosed 1016/j.jembe.2011.06.008 marine food chain comprising phytoplankton, zooplank- Iglesias-Rodriguez, M. D., and others. 2008. Phytoplankton ton and herring (Clupea harengus L.) larvae. Mar. Chem. 27: calcification in a high-CO2 world. Science 320: 336–340. 1–18. doi:10.1016/0304-4203(89)90024-8 doi:10.1126/science.1154122 Fuentes-Grünewald, C., E. Garcés, E. Alacid, N. Sampedro, S. IPCC. 2019. IPCC special report on the ocean and cryosphere Rossi, and J. Camp. 2012. Improvement of lipid production in a changing climate. IPCC. in the marine strains Alexandrium minutum and Heterosigma Jansen, S. 2008. Copepods grazing on Coscinodiscus wailesii: A akashiwo by utilizing abiotic parameters. J. Ind. Microbiol. question of size? Helgoland Mar. Res. 62: 251–255. doi:10. Biotechnol. 39: 207–216. doi:10.1007/s10295-011-1016-6 1007/s10152-008-0113-z Fukuda, S. Y., Y. Suzuki, and Y. Shiraiwa. 2014. Difference in Jaya, B. N., R. Hoffmann, C. Kirchlechner, G. Dehm, C. Scheu, physiological responses of growth, photosynthesis and and G. Langer. 2016. Coccospheres confer mechanical pro- calcification of the coccolithophore Emiliania huxleyi to tection: New evidence for an old hypothesis. Acta acidification by acid and CO2 enrichment. Photosynth. Biomater. 42: 258–264. doi:10.1016/j.actbio.2016.07.036 Res. 121: 299–309. doi:10.1007/s11120-014-9976-9 Jeffrey, S. W., and G. F. Humphrey. 1975. New spectrophoto- Gafar, N. A., B. D. Eyre, and K. G. Schulz. 2018. A conceptual metric equations for determining chlorophylls a, b, c1 and model for projecting coccolithophorid growth, calcification c2 in higher plants, algae and natural phytoplankton. and photosynthetic carbon fixation rates in response to Biochem. Physiol. Pflanz. 167: 191–194. doi:10.1016/ global ocean change. Front. Mar. Sci. 4: 433. S0015-3796(17)30778-3 Geider, R., and J. La Roche. 2002. Redfield revisited: Variability Jin, P., and K. Gao. 2016. Reduced resilience of a globally dis- of C:N:P in marine microalgae and its biochemical basis. tributed coccolithophore to ocean acidification: Confirmed Eur. J. Phycol. 37: 1–17. doi:10.1017/S0967026201003456 up to 2000 generations. Mar. Pollut. Bull. 103: 101–108. Gibbs, S. J., and others. 2013. Species-specific growth response doi:10.1016/j.marpolbul.2015.12.039 of coccolithophores to Palaeocene–Eocene environmental Jin, P., D. A. A. Hutchins, and K. Gao. 2020. The impacts of change. Nat. Geosci. 6: 218–222. doi:10.1038/ngeo1719 ocean acidification on marine food quality and its potential Gili, J.-M., and R. Coma. 1998. Benthic suspension feeders: food chain consequences. Front. Mar. Sci: 7. Their paramount role in littoral marine food webs. Trends Keller, M. D., R. C. Selvin, W. Claus, and R. R. L. Guillard. Ecol. Evol. 13: 316–321. doi:10.1016/S0169-5347(98) 1987. Media for the culture of oceanic ultraphytoplankton. 01365-2 J. Phycol. 23: 633–638. doi:10.1111/j.1529-8817.1987. Gori, A., and others. 2013. Effects of food availability on the tb04217.x sexual reproduction and biochemical composition of the Kharbush, J. J., and others. 2020. Particulate organic carbon Mediterranean gorgonian Paramuricea clavata. J. Exp. Mar. deconstructed: Molecular and chemical composition of par- Biol. Ecol. 444: 38–45. doi:10.1016/j.jembe.2013.03.009 ticulate organic carbon in the ocean. Front. Mar. Sci. Guinder, V., and J. Molinero. 2013. Climate change effects on 7: 518. marine phytoplankton, p. 68–90. In Marine ecology in a Klaas, C., and D. E. Archer. 2002. Association of sinking changing world. CRC Press. organic matter with various types of mineral ballast in Harris, R. P. 1994. Zooplankton grazing on the the deep sea: Implications for the rain ratio. Global coccolithophore Emiliania huxleyi and its role in inorganic Biogeochem. Cycl. 16: 63-1–63-14. doi:10.1029/ carbon flux. Mar. Biol. 119: 431–439. doi:10.1007/ 2001GB001765 BF00347540 Klauschies, T., B. Bauer, N. Aberle-Malzahn, U. Sommer, and Harvey, E. L., K. D. Bidle, and M. D. Johnson. 2015. Conse- U. Gaedke. 2012. Climate change effects on phytoplankton quences of strain variability and calcification in Emiliania depend on cell size and food web structure. Mar. Biol. 159: huxleyi on microzooplankton grazing. J. Plankton Res. 37: 2455–2478. doi:10.1007/s00227-012-1904-y 1137–1148. doi:10.1093/plankt/fbv081 Klintzsch, T., G. Langer, G. Nehrke, A. Wieland, K. Lenhart, Haunost, M., U. Riebesell, F. D’Amore, O. Kelting, and L. T. T. and F. Keppler. 2019. Methane production by three wide- Bach. 2021. Influence of the calcium carbonate shell of spread marine phytoplankton species: Release rates, 14 Johnson et al. Nutr response of e. hux to pH and temp precursor compounds, and potential relevance for the envi- Martiny, A. C., J. A. Vrugt, and M. W. Lomas. 2014. Concen- ronment. Biogeosciences 16: 4129–4144. doi:10.5194/bg- trations and ratios of particulate organic carbon, nitrogen, 16-4129-2019 and phosphorus in the global ocean. Sci. Data 1: 140048. Kottmeier, D. M., A. Chrachri, G. Langer, K. E. Helliwell, G. L. doi:10.1038/sdata.2014.48 Wheeler, and C. Brownlee. 2022. Reduced H+ channel Mayers, K. M. J., and others. 2020. The possession of activity disrupts pH homeostasis and calcification in coccoliths fails to deter microzooplankton grazers. Front. coccolithophores at low ocean pH. Proc. Natl. Acad. Sci. Mar. Sci. 7. 119: e2118009119. doi:10.1073/pnas.2118009119 Meyer, J., and U. Riebesell. 2015. Reviews and syntheses: Krumhardt, K. M., N. S. Lovenduski, M. D. Iglesias-Rodriguez, Responses of coccolithophores to ocean acidification: A and J. A. Kleypas. 2017. Coccolithophore growth and calci- meta-analysis. Biogeosciences 12: 1671–1682. doi:10.5194/ fication in a changing ocean. Prog. Oceanogr. 159: 276– bg-12-1671-2015 295. doi:10.1016/j.pocean.2017.10.007 Meyers, M. T., W. P. Cochlan, E. J. Carpenter, and W. J. Kuwata, A., T. Hama, and M. Takahashi. 1993. Ecophysiologi- Kimmerer. 2019. Effect of ocean acidification on the nutri- cal characterization of two life forms, resting spores and tional quality of marine phytoplankton for copepod repro- resting cells, of a marine planktonic diatom, Chaetoceros duction. PLoS One 14: e0217047. doi:10.1371/journal. pseudocunisetus, formed under nutrient depletion. Mar. pone.0217047 Ecol. Prog. Ser. 102: 245–255. doi:10.3354/meps102245 Milner, S., G. Langer, M. Grelaud, and P. Ziveri. 2016. Ocean Langer, G., M. Geisen, K. H. Baumann, J. Klas, U. Riebesell, S. warming modulates the effects of acidification on calcifica- Thoms, and J. R. Young. 2006. Species-specific responses of calci- tion and sinking. Limnol. Oceanogr. 61: 1322–1336. doi: fying algae to changing seawater carbonate chemistry. Geochem. 10.1002/lno.v61.4 Geophys. Geosyst. 7: Q09006. doi:10.1029/2005GC001227 Mitra, A., and K. J. Flynn. 2005. Predator–prey interactions: Is Langer, G., G. Nehrke, and S. Jansen. 2007. Dissolution of Cal- “ecological stoichiometry” sufficient when good food goes cidiscus leptoporus coccoliths in copepod guts? A morpho- bad? J. Plankton Res. 27: 393–399. doi:10.1093/plankt/ logical study. Mar. Ecol. Ser. 331: 139–146. doi:10.3354/ fbi022 meps331139 Monteiro, F. M., and others. 2016. Why marine phytoplank- Langer, G., G. Nehrke, I. Probert, J. Ly, and P. Ziveri. 2009. ton calcify. Sci. Adv. 2: e1501822. doi:10.1126/sciadv. Strain-specific responses of Emiliania huxleyi to changing 1501822 seawater carbonate chemistry. Biogeosciences 6: 2637– Müller, M. N., T. W. Trull, and G. M. Hallegraeff. 2017. Inde- 2646. doi:10.5194/bg-6-2637-2009 pendence of nutrient limitation and carbon dioxide Lefebvre, S. C., and others. 2011. Nitrogen source and pCO2 syn- impacts on the Southern Ocean coccolithophore Emiliania ergistically affect carbon allocation, growth and morphology huxleyi. ISME J. 11: 1777–1787. doi:10.1038/ismej.2017.53 of the coccolithophore Emiliania huxleyi: Potential implica- Nejstgaard, J. C., I. Gismervik, and P. T. Solberg. 1997. tions of ocean acidification for the carbon cycle. Glob. Chang. Feeding and reproduction by Calanus finmarchicus, and Biol. 18: 493–503. doi:10.1111/j.1365-2486.2011.02575.x microzooplankton grazing during mesocosm blooms of dia- Lewis, M. R., J. J. Cullen, and T. Platt. 1984. Relationships toms and the coccolithophore Emiliania huxleyi. Mar. Ecol. between vertical mixing and photoadaptation of phyto- Prog. Ser. 147: 197–217. plankton: Similarity criteria. Mar. Ecol. Prog. Ser. 15: 141– Nelder, J. A., and R. W. M. Wedderburn. 1972. Generalized 149. doi:10.3354/meps015141 linear models. J. Roy. Stat. Soc. Ser. A 135: 370–384. doi:10. Litzow, M. 2006. Climate regime shifts and community reor- 2307/2344614 ganization in the Gulf of Alaska: How do recent shifts com- Neukermans, G., L. Oziel, and M. Babin. 2018. Increased pare with 1976/1977? ICES J. Mar. Sci. 63: 1386–1396. doi: intrusion of warming Atlantic water leads to rapid expan- 10.1016/j.icesjms.2006.06.003 sion of temperate phytoplankton in the Arctic. Glob. Liu, X., M. C. Patsavas, and R. H. Byrne. 2011. Purification Chang. Biol. 24: 2545–2553. doi:10.1111/gcb.14075 and characterization of meta-cresol purple for spectropho- Paul, A. J., and L. T. Bach. 2020. Universal response pattern of tometric seawater pH measurements. Environ. Sci. Technol. phytoplankton growth rates to increasing CO2. New 45: 4862–4868. doi:10.1021/es200665d Phytol. 228: 1710–1716. doi:10.1111/nph.16806 Mackey, K. R. M., J. J. Morris, F. M. M. Morel, and S. A. Kranz. Pierrot, D., D. Wallace, E. Lewis, R. Wallace, and W. Wallace. 2015. Response of photosynthesis to ocean acidification. 2011. MS Excel Program Developed for CO2 System Calcu- Oceanography 28: 74–91. doi:10.5670/oceanog.2015.33 lations. ORNL Environmental Sciences Division. doi: Malinverno, E., F. G. Prahl, B. N. Popp, and P. Ziveri. 2008. 10.3334/CDIAC/otg Alkenone abundance and its relationship to the coccolithophore Pond, D. W., and R. P. Harris. 1996. The lipid composition of assemblage in Gulf of California surface waters. Deep Sea Res. the Coccolithophore Emiliania huxleyi and its possible Part I Oceanogr. Res. Pap. 55: 1118–1130. doi:10.1016/j.dsr. ecophysiological significance. J. Mar. Biol. Assoc. UK 76: 2008.04.007 579–594. doi:10.1017/S0025315400031295 15 Johnson et al. Nutr response of e. hux to pH and temp Prabhu, M., and others. 2019. Starch from the sea: The green Shaltout, M., and A. Omstedt. 2014. Recent sea surface tem- macroalga Ulva ohnoi as a potential source for sustainable perature trends and future scenarios for the Mediterranean starch production in the marine biorefinery. Algal Res. 37: Sea. Oceanologia 56: 411–443. doi:10.5697/oc.56-3.411 215–227. doi:10.1016/j.algal.2018.11.007 White, M. M., J. D. Waller, L. C. Lubelczyk, D. T. Drapeau, Riebesell, U., I. Zondervan, B. Rost, P. D. Tortell, and F. M. M. B. C. Bowler, W. M. Balch, and D. M. Fields. 2018. Morel. 2000. Reduced calcification of marine plankton in Coccolith dissolution within copepod guts affects fecal pel- response to increased atmospheric CO2. Nature 407: 364– let density and sinking rate. Sci. Rep. 8: 9758. doi:10.1038/ 367. doi:10.1038/35030078 s41598-018-28073-x Riegman, R., W. Stolte, A. A. M. Noordeloos, and D. Slezak. 2000. Woodworth, B. D., R. L. Mead, C. N. Nichols, and D. R. J. Nutrient uptake and alkaline phosphatase (ec 3:1:3:1) activity Kolling. 2015. Photosynthetic light reactions increase total of Emiliania huxleyi (prymnesiophyceae) during growth under lipid accumulation in carbon-supplemented batch cultures n and p limitation in continuous cultures. J. Phycol. 36: 87– of Chlorella vulgaris. Bioresour. Technol. 179: 159–164. doi: 96. doi:10.1046/j.1529-8817.2000.99023.x 10.1016/j.biortech.2014.11.098 Rosas-Navarro, A., G. Langer, and P. Ziveri. 2016. Temperature Ziveri, P., B. de Bernardi, K.-H. Baumann, H. M. Stoll, and affects the morphology and calcification of Emiliania P. G. Mortyn. 2007. Sinking of coccolith carbonate and huxleyi strains. Biogeosciences 13: 2913–2926. doi:10.5194/ potential contribution to organic carbon ballasting in the bg-13-2913-2016 deep ocean. Deep Sea Res. Part II Top. Stud. Oceanogr. 54: Rossi, S., and J.-M. Gili. 2007. Short-time-scale variability of 659–675. doi:10.1016/j.dsr2.2007.01.006 near-bottom seston composition during spring in a warm Zuur, A. F., E. N. Ieno, N. J. Walker, A. A. Saveliev, and G. M. temperate sea. Hydrobiologia 575: 373–388. doi:10.1007/ Smith. 2009a. GLM and GAM for count data. Springer, s10750-006-0390-y p. 209–243. Rossi, S., and others. 2019. Changes of energy fluxes in marine Zuur, A. F., E. N. Ieno, N. J. Walker, A. A. Saveliev, and G. M. animal forests of the Anthropocene: Factors shaping the Smith. 2009b. Meet the exponential family, p. 193–208. In future seascape. ICES J. Mar. Sci. 76: 2008–2019. doi:10. A. F. Zuur, E. N. Ieno, N. Walker, A. A. Saveliev, and G. M. 1093/icesjms/fsz147 Smith [eds.], Mixed effects models and extensions in ecol- Rossoll, D., R. Bermúdez, H. Hauss, K. G. Schulz, U. Riebesell, U. ogy with R. Springer. Sommer, and M. Winder. 2012. Ocean acidification-induced food quality deterioration constrains trophic transfer. PLoS Acknowledgments This work was funded by the ASSEMBLE Plus grant program (Project One 7: e34737. doi:10.1371/journal.pone.0034737 730984_A+), the Spanish Ministry of Science and Innovation, BIOCAL Project Sailley, S., M. Vogt, S. Doney, and others. 2013. Comparing (PID2020-113526RB-I00) and CALMED project (CTM2016-79547-R). Roberta food web structures and dynamics across a suite of global Johnson received a fellowship from MINECO (FPI/BES-2017-080469). marine ecosystem models. Ecol. Model. 261–262: 43–57. This work is contributing to the ICTA “Unit of Excellence” (MINECO, doi:10.1016/j.ecolmodel.2013.04.006 MDM2015-0552; CEX2019-000940-M), and The Marine and Environ- Schlüter, L., K. T. Lohbeck, M. A. Gutowska, J. P. Gröger, U. mental Biogeosciences research group of the Generalitat de Catalunya (2017 SGR-1588). The authors would like to thank the specialized team Riebesell, and T. B. H. Reusch. 2014. Adaptation of a globally at Station Biologique De Roscoff who helped us conduct the experiment. important coccolithophore to ocean warming and acidifica- The authors would also like to thank the anonymous reviewers for their tion. Nat. Clim. Chang. 4: 1024–1030. doi:10.1038/ comments that helped to greatly enhance and refine the manuscript. nclimate2379 Seifert, M., B. Rost, S. Trimborn, and J. Hauck. 2020. Meta- analysis of multiple driver effects on marine phytoplankton Conflict of interest highlights modulating role of pCO2. Glob. Chang. Biol. 26: The authors declare that there is no conflict of interest. 6787–6804. doi:10.1111/gcb.15341 Sett, S., L. T. Bach, K. G. Schulz, S. Koch-Klavsen, M. Lebrato, and Submitted 28 November 2021 U. Riebesell. 2014. Temperature modulates coccolithophorid Revised 25 March 2022 sensitivity of growth, photosynthesis and calcification to Accepted 03 August 2022 increasing seawater pCO2. PLoS One 9: e88308. doi:10.1371/ Associate editor: Susan Menden-Deuer journal.pone.0088308 16