The ISME Journal

https://doi.org/10.1038/s41396-018-0105-1

ARTICLE

Nutrient limitation suppresses the temperature dependence of

phytoplankton metabolic rates
Emilio Marañón1 María P. Lorenzo1 Pedro Cermeño2 Beatriz Mouriño-Carballido1
● ● ●

Received: 16 November 2017 / Revised: 3 March 2018 / Accepted: 12 March 2018

© The Author(s) 2018. This article is published with open access

Downloaded from https://academic.oup.com/ismej/article/12/7/1836/7475458 by guest on 27 April 2024

Abstract
Climate warming has the potential to alter ecosystem function through temperature-dependent changes in individual
metabolic rates. The temperature sensitivity of phytoplankton metabolism is especially relevant, since these microorganisms
sustain marine food webs and are major drivers of biogeochemical cycling. Phytoplankton metabolic rates increase with
temperature when nutrients are abundant, but it is unknown if the same pattern applies under nutrient-limited growth
conditions, which prevail over most of the ocean. Here we use continuous cultures of three cosmopolitan and
biogeochemically relevant species (Synechococcus sp., Skeletonema costatum and Emiliania huxleyi) to determine the
1234567890();,:
1234567890();,:

temperature dependence (activation energy, Ea) of metabolism under different degrees of nitrogen (N) limitation. We show
that both CO2 fixation and respiration rates increase with N supply but are largely insensitive to temperature. Ea of
photosynthesis (0.11 ± 0.06 eV, mean ± SE) and respiration (0.04 ± 0.17 eV) under N-limited growth is significantly smaller
than Ea of growth rate under nutrient-replete conditions (0.77 ± 0.06 eV). The reduced temperature dependence of metabolic
rates under nutrient limitation can be explained in terms of enzyme kinetics, because both maximum reaction rates and half-
saturation constants increase with temperature. Our results suggest that the direct, stimulating effect of rising temperatures
upon phytoplankton metabolic rates will be circumscribed to ecosystems with high-nutrient availability.

Introduction where R is the mass-specific metabolic rate (in units

of time−1), k is the Boltzmann’s constant (8.62 × 10−5
Temperature is a master variable that controls biological eV K−1), T is temperature in K, a is a normalization con-
activity through its effect on metabolic rates [1–3]. Within stant, and Ea is the activation energy (eV), a measure of how
the temperature range of normal activity, metabolic rates strongly temperature affects the metabolic rate. Increasing
increase with temperature according to the temperature accelerates enzymatic reactions by increasing
Boltzman–Arrhenius function: the proportion of molecules that have sufficient kinetic
energy to react [4, 5]. The fundamental nature of this
R ¼ aeEa=kT ð1Þ thermodynamic mechanism explains that Ea of basal meta-
bolic rate (maintenance respiration) takes relatively similar
values (0.6–0.7 eV) across all organisms from microbes to
plants and animals [2]. The temperature dependence of
metabolic and growth rates can also be expressed by Van’t
Electronic supplementary material The online version of this article Hoff’s Q10 factor:
(https://doi.org/10.1038/s41396-018-0105-1) contains supplementary
material, which is available to authorized users.
Q10 ¼ ðR2 =R1 Þ10=ðT2 T1 Þ ð2Þ
* Emilio Marañón
[email protected] where R2 and R1 are the rates measured at temperatures
1 T2 and T1, respectively. Ea values of 0.6 and 0.7 eV corre-
Departamento de Ecología y Biología Animal, Universidade de
Vigo, 36310 Vigo, Spain spond approximately to Q10 values of 2.2 and 2.6, respec-
2 tively. The temperature dependence of metabolic rates is
Instituto de Ciencias del Mar, Consejo Superior de Investigaciones
Científicas, Passeig Maritim de la Barceloneta 37-49, 08003 one of the foundations of the metabolic theory of ecology
Barcelona, Spain (MTE), which provides a unifying framework for the
 E. Marañón et al.

prediction of ecological processes at multiple levels of temperature dependence of metabolic rates, which means
organization from individuals to ecosystems [3]. that the direct response of phytoplankton primary produc-
Phytoplankton contribute nearly half of the annual global tion to increasing ocean temperatures will differ funda-
primary production and are major drivers of biogeochemical mentally among ecosystems with different nutrient
cycling, sustaining the food webs of most marine ecosys- availability.
tems and contributing to climate regulation through the
uptake and sequestration of atmospheric CO2 [6, 7]. Mean
sea surface temperature is projected to increase between 1 Materials and methods
and 3°C by the end of this century, with the strongest
warming in tropical and subtropical regions [8]. Warmer Chemostats and experimental setup—We maintained
temperatures will likely cause, particularly in low-latitude, monospecific cultures of the diatom Skeletonema costatum
open-ocean regions, a reduction in phytoplankton pro- (strain CCAP 1077/1 C), the coccolithophorid Emiliania
ductivity, as a result of enhanced thermal stratification and huxleyi (strain CCMP 371) and the cyanobacterium Syne-

Downloaded from https://academic.oup.com/ismej/article/12/7/1836/7475458 by guest on 27 April 2024

lower nutrient supply from sub-surface waters [9, 10]. chococcus sp. (strain PCC7002) under nitrogen-limited
However, this indirect effect could be counterbalanced by continuous growth using a Sartorius Biostat Bplus bior-
the direct, stimulating effect of increasing temperature upon eactor. The bioreactor was equipped with two 2-L, double-
phytoplankton growth [11–14]. walled borosilicate culture vessels and an integrated ther-
Ocean ecosystem models typically use Q10 values mostat system with circulation pump that allowed precise
between 1.88 and 2, based on Eppley’s data compilation (0.1 °C) control of growth temperature. Cultures were aer-
[15], to parameterize the relationship between temperature ated through 0.45-µm nylon filters and agitated with a stirrer
and maximum phytoplankton growth rate [13, 16, 17]. shaft rotating at 50 r.p.m. Cells were grown on nitrate-
Importantly, these Q10 values, as well as more recent esti- limited f/4 medium prepared with 0.2-µm filtered and
mates of the temperature dependence of phytoplankton autoclaved seawater (supplemented with Si in the case of S.
growth [18, 19], are all based on measurements from costatum). We modified the nitrate concentration in the
nutrient-saturated, batch cultures, in which nutrient con- medium to obtain a molar N:P ratio of 10 and ensure N-
centrations are typically 2–3 orders of magnitude higher limitation of growth. The concentrations of nitrate, phos-
than those found even in coastal, nutrient-rich waters. Yet, phate and (for S. costatum only) silicate in the final medium
both experimental [20, 21] and observational [22] studies were 181, 18 and 53 µmol L−1, respectively. Fresh medium
with natural communities have found reduced sensitivity of was supplied to the culture vessels with high-precision
phytoplankton metabolic rates to temperature under condi- peristaltic pumps (Watson Marlow 101 U/R). Another set of
tions of low nutrient availability. Interpretation of these peristaltic pumps, integrated in the main bioreactor system
results, however, is not straightforward, due to confounding and activated by a level sensor, controlled outflow rates to
factors such as shifts in species composition across envir- maintain a constant culture volume. Cultures were illumi-
onmental gradients and also because ever-changing growth nated with a LED array delivering white light, under a 12:12
conditions during short-term, batch experiments prevent photoperiod, at a photon flux rate of 200 µmol m−2 s−1,
populations from attaining full physiological acclimation. which has been shown to be saturating for the growth of S.
Thus, the temperature dependence of phytoplankton meta- costatum [28], E. huxleyi [29] and Synechococcus [30].
bolic rates under conditions of steady-state, nutrient-limited We kept our cultures at a range of temperatures and
growth remains unknown. This is a major knowledge gap, dilution rates. The dilution rates used were 0.14, 0.35 and
because phytoplankton experience chronic nutrient limita- 0.60 d−1 for S. costatum; 0.09, 0.34 and 0.60 d−1 for E.
tion and sustain persistently slow growth rates in more than huxleyi; and 0.10 and 0.30 d−1 for Synechococcus. These
80% of the global ocean [23–25]. In addition, major oli- growth rates correspond to the range of phytoplankton
gotrophic regions such as the subtropical gyres are growth rates commonly measured in open-ocean, oligo-
expanding [26] and becoming more nutrient-impoverished trophic regions [24, 31]. For each dilution rate, cultures
[27] as a result of climate warming. were exposed to 4 different temperatures: 8, 12, 16 and 20 °
To determine the temperature dependence of phyto- C for S. costatum; 10, 14, 18 and 22°C for E. huxleyi; and
plankton metabolism under conditions of steady-state 18, 22, 26 and 30 °C for Synechococcus. These temperature
nutrient limitation, we measured Ea of photosynthesis and ranges were selected to avoid supraoptimal temperatures,
respiration in nitrogen-limited continuous cultures of three based on previous studies on the thermal growth response of
widely distributed and biogeochemically significant species S. costatum [32], E. huxleyi [33] and Synechococcus [34].
(the diatom Skeletonema costatum, the coccolithophore All cultures were allowed to reach steady-state (constant
Emiliania huxleyi, and the cyanobacterium Synechococcus biomass over time) and, for each combination of dilution
sp.). Our results show that nutrient limitation suppresses the rate and temperature, sampling for the determination of
Nutrient limitation suppresses the temperature dependence of phytoplankton metabolic rates

elemental composition and metabolic rates took place after were filled with culture. Two bottles were fixed immedi-
an acclimation period of at least 10 days. ately to determine the initial oxygen concentration, whereas
Standing stocks—We obtained cell counts of S. costatum the remaining three bottles were incubated for 5 h. Oxygen
and E. huxleyi under the microscope using Neubauer cham- concentration was measured using the Winkler technique
bers. The abundance of Synechococcus was measured on with a potentiometric endpoint. To obtain respiration rates
fresh samples with a BD Accuri C6 flow cytometer. We in units of carbon, we applied a molar O2 consumption to
determined chlorophyll a concentration fluorometrically on CO2 production ratio of 1.4. Carbon-specific photosynthesis
5-mL samples filtered through GF/F filters and extracted with (PC) and respiration (RC) (units of h−1) were calculated by
90% acetone. The fluorescence signal was measured with a dividing hourly metabolic rates by POC concentration. To
TD-700 Turner fluorometer calibrated with pure chlorophyll allow a direct comparison of the temperature- and nutrient-
a. For the determination of particulate organic carbon (POC) dependence of metabolic rates among the different species,
and nitrogen (PON), duplicate 10-mL samples were filtered we normalized PC and RC data by dividing them by the
through pre-combusted GF/F filters, which were stored at mean rates measured at a similar dilution rate (0.35 d−1 for

Downloaded from https://academic.oup.com/ismej/article/12/7/1836/7475458 by guest on 27 April 2024

−20 °C. For E. huxleyi, filters were exposed to concentrated S. costatum, 0.34 d−1 for E. huxleyi and 0.30 d−1 for
HCl fumes to remove calcium carbonate. Before the analysis, Synechococcus). We calculated the daily respiration to
filters were desiccated at room temperature for 48 h. Samples photosynthesis ratio, R:P, by taking into account that
were analyzed with a Carlo Erba Instruments EA 1108 ele- respiration proceeds during all day (with the assumption
mental analyzer (CE Instruments Ltd, Wigan, UK) using an that respiration is the same in the light and in the dark),
acetanilide standard as a reference. whereas photosynthesis takes place only during the light
Metabolic rates—We measured photosynthetic CO2 phase:
fixation with the 14C-uptake technique, as described before
[35]. Briefly, four 20-mL culture samples (three light and R : P ¼ ðRshort  24Þ=P12h ð4Þ
one dark samples) were amended with 5 µCi of NaH14CO3
and incubated for 2–3h under the same temperature and where Rshort is the hourly rate of respiration measured
irradiance conditions experienced by the chemostat cultures. during a 5-h incubation and P12 h is calculated with Eq. 3.
Experiments started 2 h after the beginning of the light Statistical analyses—We used ordinary least squares
phase of the photoperiod. After incubation, samples were regression to calculate the slope of the linear relationship
filtered under low-vacuum pressure through 0.2-µm poly- between 1/kT and the natural logarithm of mass-specific
carbonate filters, which were then exposed overnight to metabolic rates, which gives the Ea. Throughout the study,
concentrated HCl fumes to remove non-fixed inorganic 14C. when measurement error was present on both independent
After adding 5 mL of scintillation cocktail to each filter, and dependent variables, we used reduced major axis
sample radioactivity (DPM) was determined with a regression to determine the parameters of the linear regres-
1409–012 Wallac liquid scintillation counter. To compute sion. The overall role of temperature and nutrient supply rate
hourly photosynthetic CO2 fixation rates, we subtracted the (dilution rate) as drivers of metabolic rates, as well as the
dark bottle DPM count from the light bottle DPM count and existence of interactive effects, was assessed with multiple
used a constant value of 2142 µmolC L−1 for the dissolved regression analysis on normalized data from all species
inorganic carbon content of seawater. Previous experiments combined, after standardizing the independent variables so
conducted with cultures of 20 phytoplankton species [35] that their effect sizes (coefficients in the linear regression
showed that C fixation rates obtained from short (2–3 h) model) could be comparable. We also applied multiple
incubations with 14C are strongly correlated with daily net regression analysis separately to determine the effect of
POC increase measured in cultures (Fig. S1A in Supple- dilution rate (D) and temperature on carbon-specific photo-
mentary Information). Short-term C fixation rates were also synthesis in each individual species, according to the model:
highly correlated with the C fixation rate measured during
12-h incubations (Fig. S1B). The amount of C fixed during ln PC ¼ c1 þ c2 ln D  Ea ð1=kTÞ ð5Þ
12 h (light phase of the photoperiod), P12 h, is calculated
from the hourly C fixation rate in a short incubation, Pshort, We used the data compiled by [18] to calculate average
using the equation: values of Ea for growth rate in nutrient-replete cultures of S.
  costatum, E. huxleyi and Synechococcus. Ea values in [18]
P12h ¼ 19:5 þ 9:5 Pshort r 2 ¼ 0:94; p<0:001; n ¼ 16 were computed only with data from the growing part of the
ð3Þ temperature response curve, so the relevant temperature
ranges were similar to the ones used in our experiments.
Respiration was measured as the rate of dissolved O2 Also in agreement with our experiments, growth irradiance
consumption in the dark. Five 30-mL borosilicate bottles in the original studies was saturating. Ea values in nutrient-
 E. Marañón et al.

Fig. 1 Relationship between 250 0.3

A S. costatum B
dilution rate and (a) carbon to E. huxleyi
chlorophyll a ratio (C:Chl a) and 200 Synechococcus
(b) carbon-specific

C:Chl a (g:g)
photosynthetic rate (PC) in 0.2
150

PC (h-1)
Skeletonema costatum,
Emiliania huxleyi and
Synechococcus growing under 100
N-limited conditions in 0.1
continuous cultures. The 50
different datapoints at each
dilution rate correspond to
different growth temperatures. 0 0.0
Bars indicate standard deviation 0.0 0.2 0.4 0.6 0.0 0.2 0.4 0.6
-1 -1
Dilution rate (d ) Dilution rate (d )

Downloaded from https://academic.oup.com/ismej/article/12/7/1836/7475458 by guest on 27 April 2024

+ - + -
Degree of N limitation Degree of N limitation

limited versus nutrient-replete cultures were compared nutrient supply had a significant effect on C:N, whereas
using the Mann–Whitney U-test (two-tailed). All statistical temperature had no significant effect (Table S1).
analyses were carried out with SPSS Statistics v. 24. Carbon-specific carbon fixation rate (PC), equivalent to
the biomass turnover rate, increased with increasing dilution
rate (Fig. 1b) (Pearson’s r = 0.61, p = 0.0004, n = 29). Due
Results to the opposite patterns of variability of C:Chl a and PC as a
function of dilution rate, there was no correlation between
Effect of nutrient-limited growth on cellular Chl-specific photosynthesis and dilution rate (Pearson’s r =
composition and carbon fixation −0.24, p > 0.2, n = 28).

The different dilution rates in our chemostats provided a Role of temperature and nutrient limitation in the
range of degrees of nitrogen limitation that reflected upon control of metabolic rates
the biochemical composition and metabolic rate of the
populations. The carbon to chlorophyll a ratio (C:Chl a) Both PC and RC increased markedly with dilution rate and
increased as dilution rates became slower (Fig. 1a). In S. were largely independent of temperature in all species
costatum and E. huxleyi, C:Chl a increased from 10–20 gC (Fig. S7). When normalized PC and RC values are plotted
gChl−1 at the fastest dilution rate (0.6 d−1) to 40–120 gC against temperature and dilution rate for all species together
gChl−1 at the slowest dilution rate (ca. 0.1 d−1). C:Chl a in (Fig. 2), the pattern of nutrient supply-dependent and
Synechococcus took values in the range 50–130 gC gChl−1 temperature-independent metabolic rates is evident. Multi-
at 0.3 d−1 and increased to 130–220 gC gChl−1 at 0.1 d−1. ple regression analysis confirmed a highly significant effect
The log-log relationship between dilution rate and C:Chl a, of nutrient supply rate (dilution rate) upon both PC and RC,
for all species combined, had a slope of −1.5 (Fig. S2). The with R2 values > 0.5, whereas temperature had no significant
effect of nutrient supply (dilution rate) on C:Chl a resulted effect (Table S1). We observed a strong correlation between
from changes in the cellular content of both C and Chl a, PC and RC (Pearson’s r = 0.90, n = 27, p < 0.0001; Fig.
although the latter showed higher variability (Fig. S3). Cells S8A) and consequently the variability in the respiration to
growing at the slowest dilution rate had more C per cell than photosynthesis ratio (R:P), which took a mean value of 0.37
those growing at faster dilution rates, whereas the cellular (95% CI = 0.31, 0.43), was smaller than that of PC and RC
Chl a content tended to increase with dilution rate (Fig. S3). (Fig. 2c). However, photosynthetic carbon fixation
Cell carbon decreased or remained largely unchanged as increased with dilution rate faster than respiration did
temperature increased (Fig. S4). Growth temperature had a (Fig. 2a, b) and as a result R:P showed a moderate but
marked impact on C:Chl a (Fig. S5). In most combinations significant increase with decreasing dilution rate, whereas it
of species and dilution rate, C:Chl a increased by a factor of remained invariant with respect to temperature (Fig. 2c,
1.5 to 2 from the warmest to the coldest temperature. The C: Table S1). The interaction between temperature and dilution
N elemental ratio of particulate organic matter responded to rate was not significant for PC and RC but was marginally
the degree of nutrient limitation, taking lower values at significant for R:P (Table S1). The intercept of the linear
faster dilution rates in both S. costatum and E. huxleyi relationship between growth rate and RC can be used to
(Fig. S6B,D). With all data pooled together, the rate of calculate µ0, the basal metabolic rate, which took a
Nutrient limitation suppresses the temperature dependence of phytoplankton metabolic rates

S. costatum Fig. 2 Normalized PC and RC, and respiration to photosynthesis ratio

E. huxleyi (R:P) in (a) S. costatum, (b) E. huxleyi and (c) Synechococcus under
Synechococcus different temperatures and growth rates. For each species, PC and RC
A data were normalized by dividing them by the mean rate measured
2.0 at a dilution rate of 0.35 d−1 (S. costatum), 0.34 d−1 (E. huxleyi) and
0.30 d−1 (Synechococcus). The smoothed surface was obtained with
local regression (LOESS) using tricube weighting and a polynomial of
1.5
C

degree 2
Normalised P

1.0
The effect of dilution rate on metabolic rates was also
evident when cell-specific rates were examined (Fig. S9).
0.5 Despite the fact that cell carbon tended to decrease or
remain unchanged as dilution rate increased (Fig. S3), cell-
0.0 specific rates of photosynthetic C fixation increased with

Downloaded from https://academic.oup.com/ismej/article/12/7/1836/7475458 by guest on 27 April 2024

12
16 dilution rate (Fig. S9A,C,E). In contrast, cell-specific rates
20 0.6
24 0.4 of respiration remained largely invariant with respect to
28 0.2 nutrient supply (Fig. S9B,D,F).
32
0.0 The multiple regression analysis performed separately
on each species confirmed the strong effect of dilution rate,
B and the non-significant effect of temperature, on
2.0
photosynthetic C fixation (Table S2). The intercept of the
temperature and dilution rate-dependent model was not
1.5
C

significantly different among species, indicating that,

Normalised R

for the same temperature and nutrient supply rate, all spe-
1.0 cies sustained broadly similar rates of biomass-specific
C fixation.
0.5 Given that the cellular content of Chl a was dependent on
both nutrient supply (Fig. S3) and temperature (Fig. S5), we
investigated also the variability in Chl-specific photo-
0.0
12 synthesis. In all species, C fixation per unit Chl a tended to
16
20 0.6 decrease or remain unchanged with increasing temperature
24 0.4 (Fig. 3).
28 0.2
32
0.0
Activation energy of metabolic rates
C
The Arrhenius plots indicated that, with a few exceptions,
both photosynthesis and respiration increase with tempera-
0.6
ture at a much slower pace than predicted by the MTE
(Fig. 4, Table S3). Out of 16 determinations of Ea, only in
one case (respiration in S.costatum at 0.35 d−1) did the
R:P

0.4
obtained estimate differ significantly from 0. For the
0.2
ensemble of all species and dilution rates, Ea took a mean
( ± SE) value of 0.11 ± 0.06 eV for photosynthesis (n = 8)
and 0.04 ± 0.17 eV for respiration (n = 8). Considering all
0.0 metabolic rate measurements pooled together, the mean Ea
Te 1216 was 0.08 ± 0.04 eV (n = 16). The Ea values for photo-
mp 20 0.6
er
atu 24 0.4 -1
synthesis measured under nutrient-limited conditions were
re 28 0.2 (d ) significantly lower (Mann–Whitney U-test) than those
(°C 32 rate
) 0.0 tion
Dilu reported in the literature [18] for nutrient-replete cultures of
the same species (Fig. 5). The mean Ea of growth rate was
0.65 ± 0.07 (n = 11), 0.68 ± 0.09 (n = 7) and 1.09 ± 0.13
value of 0.046 (SE = 0.017), 0.163 (SE = 0.051), and 0.072 (n = 6) eV for nutrient-replete S. costatum, E. huxleyi and
(SE = 0.006) d−1 for S. costatum, E. huxleyi and Synecho- Synechococcus, respectively, while the corresponding Ea
coccus, respectively (Fig. S8B). values for photosynthesis under nutrient limitation in our
 E. Marañón et al.

2.0 ecological niche. Yet, the observed relationships between

A 0.14 d
-1
temperature, nutrient supply and both biochemical compo-
-1
0.35 d sition and metabolic rate depicted a coherent pattern valid
(gC gChl h ) 1.5 0.60 d
-1
-1

for all species. The carbon to chlorophyll a ratio (C:Chl)

-1

tended to increase with decreasing dilution rate, as has often

1.0 been observed in nutrient-limited phytoplankton cultures
[36–38]. This pattern arises because a growing degree of
nitrogen limitation (slower dilution rates) leads to a reduc-
Chl

0.5
P

tion in the size of the nitrogen-rich, light-harvesting appa-

ratus and hence a decreased chlorophyll content [39]. The
0.0 fact that C:Chl increases with decreasing nutrient-limited
5 10 15 20 25
growth rate means that Chl-normalized C fixation rate gives
a biased estimate of biomass turnover rate, and can be

Downloaded from https://academic.oup.com/ismej/article/12/7/1836/7475458 by guest on 27 April 2024

10 misleading when testing hypotheses that concern the rela-
0.09 d
-1 B
-1
tionship between environmental drivers and phytoplankton
0.34 d
8 metabolism and growth [40]. Indeed, we found no corre-
(gC gChl h )

-1
0.60 d
-1

lation between Chl-specific photosynthesis and nutrient

-1

6 supply rate in our study, which illustrates the fact that

phytoplankton respond to nutrient limitation by changing
4 the number and size, but not the efficiency, of photosystems
[41].
Chl
P

2 The C:N elemental ratio of marine phytoplankton has a

major impact on both the transfer of energy through food
0 webs and the efficiency of the biological pump to export
5 10 15 20 25 organic carbon towards deep waters. As reported in pre-
vious studies [42], we found a significant effect of nutrient
5 supply on elemental composition, such that stronger nitro-
C
0.10 d-1 gen limitation led to an increased C:N ratio in both S.
0.30 d-1
4 costatum and E. huxleyi. In contrast, temperature had no
(gC gChl h )
-1

effect upon C:N, which supports the view that nutrient

-1

3 availability alone can explain most of the variability in

phytoplankton elemental stoichiometry [43].
2 Temperature-dependent changes in the allocation of
resources into photosynthetic machinery can affect the
Chl
P

1 relationship between temperature and both individual and

biomass-specific C fixation rates. The MTE predicts that C
0 fixation per unit photosynthetic complex increases with
15 20 25 30 35 temperature [44]. However, in our nutrient-limited cultures,
Temperature (°C)
we found that Chl-specific C fixation tended to decrease or
remain unchanged with increasing temperature. Thus, in
Fig. 3 Temperature dependence of chlorophyll a-specific C fixation spite of having a higher Chl a content, cells under warmer
rate (PChl) in (a) S. costatum, (b) E. huxleyi and (c) Synechococcus
under N-limited continuous growth at different dilution rates. Bars
temperatures sustained similar biomass-specific photo-
indicate standard deviation and dashed lines are the ordinary least synthetic rates than those growing under colder
squares regression fits temperatures.
Temperature-related changes in cell size could also
populations were 0.10 ± 0.17 (n = 3), 0.09 ± 0.06 (n = 3), potentially have an impact on the response of metabolic
and 0.18 ± 0.07 (n = 2) eV. rates to rising temperature. Warmer temperature induces
smaller cell size in protists [45], a pattern we observed in
some of our experiments. Hence, according to the MTE,
Discussion which predicts a faster pace of metabolism with decreasing
body size [3], one would have expected higher metabolic
We conducted our measurements on three species that differ rates at warmer temperatures. Yet, biomass-specific photo-
widely in phylogenetic affiliation, functional traits and synthesis and respiration remained largely unaffected by
Nutrient limitation suppresses the temperature dependence of phytoplankton metabolic rates

Skeletonema costatum Emiliania huxleyi Synechococcus

-1 -1 -1
A 0.14 d-1 B 0.09 d
-1
C -1
-1 0.10 d
0.35 d 0.34 d-1
0.30 d-1
0.60 d-1 0.60 d
-1
-2 -2 -2
Ln PC (h-1)

-3 -3 -3

-4 -4 -4

-5 -5 -5
39 40 41 42 39 40 41 42 38 39 40

Downloaded from https://academic.oup.com/ismej/article/12/7/1836/7475458 by guest on 27 April 2024

-3 -3 -3
D E F

-4 -4 -4
Ln RC (h-1)

-5 -5 -5

-6 -6 -6

-7 -7 -7
39 40 41 42 39 40 41 42 38 39 40
1/kT 1/kT 1/kT
Fig. 4 Arrhenius plots for carbon-specific photosynthesis (P ) and C
rates. Dashed lines represent the linear fit (OLS regression) between
respiration (RC) in S. costatum (left), E. huxleyi (middle) and Syne- inverse temperature (1/kT) and the carbon-specific metabolic rate.
chococcus (right) growing in N-limited chemostats at different dilution Slope values are given in Table S3

temperature in our N-limited populations. Earlier studies

1.5
Nutrient-replete have shown that individual metabolic rates in phyto-
Nutrient-limited plankton do not follow the ¾ power size scaling commonly
Activation energy (eV)

observed in multicellular organisms but instead scale iso-

1.0 metrically with cell size [23, 46]. This means that, as a first-
order approximation, phytoplankton cells of all sizes sustain
broadly comparable maximum rates of biomass-specific
0.5 metabolism. Accordingly, we found that, when compared at
p=0.0055 p=0.0303 p=0.0455 p<0.0001
the same temperature and nutrient supply rate, all three
species had similar rates of photosynthesis per unit biomass.
Carbon-specific carbon fixation rate (PC) increased
0.0
markedly with increasing dilution rate in all species,
reflecting the coupling between population growth rate and
Scos Ehux Syne All spp
biomass turnover rate, which arises from the fact that
Fig. 5 Comparison between the activation energy (Ea) of growth rate reproduction is ultimately fueled by metabolism [3]. A
under nutrient-replete growth conditions in batch cultures and the Ea of
similar covariation between growth rates and PC has been
carbon-specific photosynthesis measured under nutrient-limited
growth in chemostat cultures of S. costatum (n = 3), E. huxleyi (n = observed before in nutrient-limited chemostat cultures
3) and Synechococcus (n = 2). Mean Ea values for S. costatum (n = growing under different dilution rates [36, 47] and in a
11), E. huxleyi (n = 7) and Synechococcus (n = 6) under nutrient- study of the size dependence of phytoplankton metabolism
replete conditions were calculated from the data compilation in [18].
and growth in batch cultures [35].
The significance values (two-tailed) correspond to the Mann–Whitney
U-test carried out to compare the Ea values between nutrient-replete We found that respiration tended to increase with dilu-
and nutrient-limited conditions. Bars indicate the SE tion rate at a slower pace than photosynthesis did, which
 E. Marañón et al.

resulted in enhanced R:P values at the slowest dilution rates. warm temperatures [55]. However, this strategy heavily
These results agree with previous observations showing that increases the cellular demands for nitrogen and thus is
respiratory losses tend to become a larger fraction of pho- unlikely to be used by strongly N-limited cells.
tosynthetic carbon fixation under suboptimal conditions [46, An alternative explanation for the lack of temperature
48, 49]. Such a pattern arises from the existence of basal dependence of metabolic rates in nutrient-limited popula-
metabolic maintenance costs that are largely independent of tions can be found in enzyme kinetics. Under conditions of
biosynthesis rates. The basal metabolic rates we calculated nutrient limitation, intracellular substrate abundance
(0.05, 0.16 and 0.07 d−1 for S. costatum, E. huxleyi and decreases and therefore the temperature dependence of
Synechococcus, respectively) agree well with earlier esti- enzyme half-saturation constant (km) becomes more relevant
mates [48], which gave a mean value of 0.08 d−1 (95% than that of the maximum reaction rate (Vm) [4, 56].
CI = 0.04, 0.13) in nutrient-limited cultures of phyto- Increasing temperature leads to higher kinetic energy of
plankton species from various taxa. reactants and increased rates of collision, as well as higher
R:P is a key variable that determines the efficiency of structural flexibility of enzymes, all of which promote faster

Downloaded from https://academic.oup.com/ismej/article/12/7/1836/7475458 by guest on 27 April 2024

carbon use by primary producers (the fraction of fixed car- catalytic rates [5, 57]. However, higher structural flexibility
bon available for allocation to growth) and their net con- also results in active sites with a reduced ability for ligand
tribution to the carbon cycle [50]. Given that in recognition and binding and lower kinetic efficiency, which
photoautotrophs respiration is ultimately constrained by CO2 results in lower affinity (higher km). The km of most
fixation [44], the temperature dependences of phytoplankton enzymes increases with temperature, with Q10 values
respiration and photosynthesis are similar [51]. However, similar to, or higher than, those of Vm [4, 57]. In phyto-
experimental studies have found that, over temporal scales plankton, the km of Rubisco has a Q10 of approximately 2
of a few generations and under resource-replete conditions, [55], and the km of nitrate uptake and growth in nutrient-
microalgal respiration can be more responsive to tempera- limited cultures has a Q10 higher than 2 [58, 59], while the
ture than photosynthesis [50, 52], such that warmer tem- Q10 of Vm in a wide range of enzymes involved in both
peratures lead to increased R:P. In contrast, our data show anabolic and catabolic pathways takes a mean value of
that temperature has no effect on the R:P of fully-acclimated 2.1 ± 0.4 [54]. If both Vm and km have the same temperature
phytoplankton under nutrient-limited, continuous growth. sensitivity, the realized reaction rates at low substrate con-
Overall, our results suggest that nutrient supply has a larger centration ([S] < km) can be similar at divergent tempera-
role than temperature in controlling the efficiency of pho- tures, and the resulting temperature dependence of
tosynthetic carbon conversion into new biomass. metabolic rate becomes very small (Fig. S10). This
Our determinations of activation energy (Ea) of photo- mechanism, whose importance is well recognized in ter-
synthesis and respiration show that the temperature depen- restrial ecology to explain thermal adaptation of organic
dence of phytoplankton metabolic rates is suppressed by matter decomposition in soils [4, 60, 61], could also explain
nutrient limitation. The mean values of Ea observed in our the lack of temperature sensitivity of phytoplankton meta-
nutrient-limited chemostats were much lower than those bolic rates under conditions of nutrient limitation.
determined for the growth rate of cultures of the same We have provided the first experimental determinations
species growing under nutrient-replete conditions [18]. of the activation energy of metabolic rate in phytoplankton
They were are also lower than the value of 0.32 eV calcu- experiencing chronic nutrient limitation of growth. Our
lated for C3 terrestrial plants [44]. The lack of temperature results stress the need to consider resource limitation when
dependence of metabolic rates under nutrient limitation is using MTE-based approaches to understand the environ-
comparable with the observation that strong light limitation mental control of metabolic activity and, in addition, have
greatly diminishes the temperature sensitivity of phyto- implications for the prediction of climate change impacts on
plankton growth [53]. While our experiments were con- ocean biogeochemistry. Modeling studies have suggested
ducted under saturating light levels, additional studies are that warming can have a direct, stimulating effect on ocean
needed to address the interactive effect of temperature and net primary production, particularly in low-latitude regions
nutrient supply in light-limited conditions. [12, 13]. This increased productivity would be associated
Metabolic acclimation to temperature in photosynthetic with a faster nutrient recycling through the microbial loop
unicells can involve changes in both the abundance and the and an increase in regenerated production. Our results,
specific activity of catalysts [54]. Biomass-specific carbon however, suggest that this effect is unlikely to occur in
fixation rates can be maintained under cold temperatures if ocean regions where phytoplankton growth is severely
the abundance of Rubisco increases sufficiently to com- limited by nutrient availability, such as the nitrogen-limited
pensate for the cold-induced reduction in its substrate subtropical gyres or the iron-limited high-nutrient, low-
turnover rate. For instance, polar diatoms have a relative chlorophyll regions [25]. Hence, it can be expected that
Rubisco content 10 times higher than diatoms growing at direct responses of primary production to warming will vary
Nutrient limitation suppresses the temperature dependence of phytoplankton metabolic rates

widely among regions. The stimulating effect of increasing 6. Falkowski PG, Oliver MJ. Mix and match: how climate selects
temperatures on phytoplankton production and growth may phytoplankton. Nat Rev Microbiol. 2007;5:813–9.
7. Raven JA. The possible roles of algae in restricting the increase in
be significant in coastal and upwelling regions [62], but is
atmospheric CO2 and global temperature. Eur J Phycol.
likely to be minor in oligotrophic waters. Previous studies on 2017;52:506–22.
the interaction between temperature and resources in aquatic 8. Collins M, Knutti R, Arblaster J, Dufresne J-L, Fichefet T, Frie-
ecosystems have emphasized the role of temperature in dlingstein P, et al. Long-term Climate Change: Projections,
Commitments and Irreversibility. In: Stocker TF, Qin D, Plattner
regulating the effect of nutrient supply upon metabolic rates
G-K, Tignor M, Allen SK, Boschung J, et al. (eds). Climate
[1]. In contrast, our results suggest that nutrient availability Change 2013: The Physical Science Basis. Contribution of
controls the temperature dependence of metabolism, such Working Group I to the Fifth Assessment Report of the Inter-
that the direct effect of increasing temperature on metabolic governmental Panel on Climate Change. Cambridge, United
Kingdom and New York, NY, USA: Cambridge University Press;
rates is virtually absent under nutrient-limiting conditions.
2013. p. 1029–136.
Furthermore, we have shown that nutrient supply explains 9. Behrenfeld MJ, O’Malley RT, Siegel DA, McClain CR, Sar-
most of the variability in the photosynthesis and respiration miento JL, Feldman GC, et al. Climate-driven trends in con-

Downloaded from https://academic.oup.com/ismej/article/12/7/1836/7475458 by guest on 27 April 2024

of phytoplankton, whereas temperature plays a much smaller temporary ocean productivity. Nature. 2006;444:752–5.
10. Bopp L, Monfray P, Aumont O, Dufresne J-L, Le Treut H, Madec
role. Indirect effects of temperature upon resource supply are
G, et al. Potential impact of climate change on marine export
therefore likely to dominate the response of phytoplankton production. Glob Biogeochem Cycles. 2001;15:81–99.
growth and productivity to ocean warming. 11. Dutkiewicz S, Scott JR, Follows MJ. Winners and losers: Eco-
logical and biogeochemical changes in a warming ocean. Glob
Acknowledgements We thank P. Chouciño and C. Sobrino for their Biogeochem Cycles. 2013;27:463–77.
help with laboratory work, and R. Geider, M. Huete-Ortega, J. L. 12. Sarmiento JL, Slater R, Barber R, Bopp L, Doney SC, Hirst AC
Otero, and three anonymous reviewers for comments on the manu- et al (2004). Response of ocean ecosystems to climate warming.
script. P.C. was supported by a Ramón y Cajal contract from the Glob Biogeochem Cycles. GB3003. 1–23.
Spanish Ministry of Economy and Competitiveness (MINECO). This 13. Taucher J, Oschlies A (2011). Can we predict the direction of
research was funded by MINECO through grant CTM2014-53582-R marine primary production change under global warming? Geo-
to E.M. phys Res Lett. L02603. 1–6.
14. Thomas MK, Kremer CT, Klausmeier CA, Litchman E. A global
pattern of thermal adaptation in marine phytoplankton. Science.
Compliance with ethical standards 2012;338:1085–8.
15. Eppley RW. Temperature and phytoplankton growth in the sea.
Conflict of interest The authors declare that they have no conflict of Fish Bull. 1972;70:1063–85.
interest. 16. Doney SC, Glover DM, Najjar RG. A new coupled, one-
dimensional biological-physical model for the upper ocean:
Open Access This article is licensed under a Creative Commons applications to the JGOFS Bermuda Atlantic Time-series Study
Attribution 4.0 International License, which permits use, sharing, (BATS) site. Deep Sea Res II. 1996;43:591–624.
adaptation, distribution and reproduction in any medium or format, as 17. Marinov I, Doney SC, Lima ID. Response of ocean phytoplankton
long as you give appropriate credit to the original author(s) and the community structure to climate change over the 21st century:
source, provide a link to the Creative Commons license, and indicate if partitioning the effects of nutrients, temperature and light. Bio-
changes were made. The images or other third party material in this geosciences. 2010;7:3941–59.
article are included in the article’s Creative Commons license, unless 18. Chen B, Liu H, Huang B, Wang J. Temperature effects on the
indicated otherwise in a credit line to the material. If material is not growth rate of marine picoplankton. Mar Ecol Prog Ser.
included in the article’s Creative Commons license and your intended 2014;505:37–47.
use is not permitted by statutory regulation or exceeds the permitted 19. Kremer CT, Thomas MK, Litchman E. Temperature- and size-
use, you will need to obtain permission directly from the copyright scaling of phytoplankton population growth rates: Reconciling the
holder. To view a copy of this license, visit http://creativecommons. Eppley curve and the metabolic theory of ecology. Limnol
org/licenses/by/4.0/. Oceanogr. 2017;62:1658–70.
20. O’Connor MI, Piehler MF, Leech DM, Anton A, Bruno JF.
Warming and Resource Availability Shift Food Web Structure and
References Metabolism. PLoS Biol. 2009;7:e1000178.
21. Staehr PA, Sand-Jensen KAJ. Seasonal changes in temperature
1. Cross WF, Hood JM, Benstead JP, Huryn AD, Nelson D. Inter- and nutrient control of photosynthesis, respiration and growth
actions between temperature and nutrients across levels of eco- of natural phytoplankton communities. Freshw Biol.
logical organization. Glob Change Biol. 2015;21:1025–40. 2006;51:249–62.
2. Gillooly JF, Brown JH, West GB, Savage VM, Charnov EL. 22. Maranon E, Cermeño P, Huete-Ortega M, López-Sandoval DC,
Effects of size and temperature on metabolic rate. Science. Mouriño-Carballido B, Rodríguez-Ramos T. Resource supply
2001;293:2248–51. overrides temperature as a controlling factor of marine phyto-
3. Brown JH, Gillooly JF, Allen AP, Savage VM, West GB. Toward plankton growth. PLoS ONE. 2014;9:e99312.
a metabolic theory of ecology. Ecology. 2004;85:1771–89. 23. Marañón E. Cell size as a key determinant of phytoplankton
4. Davidson EA, Janssens IA. Temperature sensitivity of soil carbon metabolism and community structure. Annu Rev Mar Sci.
decomposition and feedbacks to climate change. Nature. 2015;7:241–64.
2006;440:165–73. 24. Behrenfeld MJ, Boss E, Siegel DA, Shea DM (2005). Carbon-
5. Somero GN. Proteins and temperature. Annu Rev Physiol. based ocean productivity and phytoplankton physiology from
1995;57:43–68. space. Glob Biogeochem Cycles. GB1006, 1–14.
 E. Marañón et al.

25. Moore CM, Mills MM, Arrigo KR, Berman-Frank I, Bopp L, 43. Galbraith ED, Martiny AC. A simple nutrient-dependence
Boyd PW, et al. Processes and patterns of oceanic nutrient lim- mechanism for predicting the stoichiometry of marine ecosys-
itation. Nat Geosci. 2013;6:701–10. tems. Proc Natl Acad Sci. 2015;112:8199–204.
26. Polovina JJ, Howell EA, Abecassis M (2008). Ocean’s least 44. Allen AP, Gillooly JF, Brown JH. Linking the global carbon cycle
productive waters are expanding. Geophys Res Lett. L03618, 1–5. to individual metabolism. Funct Ecol. 2005;19:202–13.
27. Signorini SR, Franz BA, McClain CR (2015). Chlorophyll 45. Atkinson D, Ciotti BJ, Montagnes DJS. Protists decrease in size
variability in the oligotrophic gyres: mechanisms, seasonality and linearly with temperature: ca. 2.5% °C−1. Proc R Soc Lond B Biol
trends. Front Marine Sci Article 1, 1–11. Sci. 2003;270:2605–11.
28. Langdon C. On the causes of interspecific differences in the 46. López-Sandoval DC, Rodríguez-Ramos T, Cermeño P, Sobrino C,
growth-irradiance relationship for phytoplankton. Part 1. A com- Marañón E. Photosynthesis and respiration in marine phyto-
parative study of the growth-irradiance relationship of three marine plankton: Relationship with cell size, taxonomic affiliation, and
phytoplankton species: Skeletonema costatum, Olisthodiscus growth phase. J Exp Mar Bio Ecol. 2014;457:151–9.
luteus and Gonyaulax tamarensis. J Plankton Res. 1987;9: 47. Osborne BA, Geider RJ. Effect of nitrate-nitrogen limitation on
459–82. photosynthesis of the diatom Phaeodactylum tricornutum Bohlin
29. Paasche E. A review of the coccolithophorid Emiliania huxleyi (Bacillariophyceae). Plant Cell Environ. 1986;9:617–25.
(Prymnesiophyceae), with particular reference to growth, cocco- 48. Geider RJ, Osborne BA. Respiration and microalgal growth: A

Downloaded from https://academic.oup.com/ismej/article/12/7/1836/7475458 by guest on 27 April 2024

lith formation, and calcification-photosynthesis interactions. Phy- review of the quantitative relationship between dark respiration
cologia. 2002;40:503–29. and growth. New Phytol. 1989;112:327.
30. Six C, Thomas JC, Brahamsha B, Lemoine Y, Partensky F. 49. Geider RJ. Respiration: Taxation Without Representation? In:
Photophysiology of the marine cyanobacterium Synechococcus Falkowski PG, (ed). Primary Productivity and Biogeochemical
sp. WH8102, a new model organism. Aquat Microb Ecol. Cycles in the Sea. New York and London: Plenum Press; 1992. p.
2004;35:17–29. 333–60.
31. Marañón E. Phytoplankton growth rates in the Atlantic subtropical 50. Padfield D, Yvon-Durocher G, Buckling A, Jennings S, Yvon-
gyres. Limnol Oceanogr. 2005;50:299–310. Durocher G. Rapid evolution of metabolic traits explains thermal
32. Montagnes DJS, Franklin DJ. Effect of temperature on diatom adaptation in phytoplankton. Ecol Lett. 2016;19:133–42.
volume, growth rate, and carbon and nitrogen content: Reconsi- 51. López-Urrutia A, San Martín E, Harris RP, Irigoien X. Scaling the
dering some paradigms. Limnol Oceanogr. 2001;46:2008–18. metabolic balance of the oceans. Proc Natl Acad Sci USA.
33. van Rijssel M, Gieskes WWC. Temperature, light, and the 2006;103:8739–44.
dimethylsulfoniopropionate (DMSP) content of Emiliania huxleyi 52. Staehr PA, Birkeland MJ. Temperature acclimation of growth,
(Prymnesiophyceae). J Sea Res. 2002;48:17–27. photosynthesis and respiration in two mesophilic phytoplankton
34. Mackey KRM, Paytan A, Caldeira K, Grossman AR, Moran D, species. Phycologia. 2006;45:648–56.
McIlvin M, et al. Effect of Temperature on Photosynthesis and 53. Edwards KF, Thomas MK, Klausmeier CA, Litchman E. Phyto-
Growth in Marine Synechococcus spp. Plant Physiol. plankton growth and the interaction of light and temperature: A
2013;163:815–29. synthesis at the species and community level. Limnol Oceanogr.
35. Marañón E, Cermeño P, López‐Sandoval DC, Rodríguez‐Ramos 2016;61:1232–44.
T, Sobrino C, Huete‐Ortega M, et al. Unimodal size scaling of 54. Raven JA, Geider RJ. Temperature and algal growth. New Phytol.
phytoplankton growth and the size dependence of nutrient uptake 1988;110:441–61.
and use. Ecol Lett. 2013;16:371–9. 55. Young JN, Goldman JAL, Kranz SA, Tortell PD, Morel FMM.
36. Geider RJ. Quantitative phytoplankton physiology: implications Slow carboxylation of Rubisco constrains the rate of carbon
for primary production and phytoplankton growth. ICES Mar Sci fixation during Antarctic phytoplankton blooms. New Phytol.
Symp. 1993;197:52–62. 2015;205:172–81.
37. Geider RJ, MacIntyre HL, Kana TM. Dynamic model of phyto- 56. Davidson EA, Janssens IA, Luo Y. On the variability of respira-
plankton growth and acclimation: responses of the balanced tion in terrestrial ecosystems: moving beyond Q10. Glob Change
growth rate and the chlorophyll a:carbon ratio to light, nutrient- Biol. 2006;12:154–64.
limitation and temperature. Mar Ecol Prog Ser. 57. Somero GN. Adaptation of enzymes to temperature: searching for
1997;148:187–200. basic “strategies”. Comp Biochem Physiol B. 2004;139:321–33.
38. Sakshaug E, Andresen K. A steady state description of growth and 58. Eppley RW, Rogers JN, McCarthy JJ. Half-saturation constants
ligth absorption in the marine planktonic diatom Skeletonema for uptake of nitrate and ammonium by marine phytoplankton.
costatum. Limnol Oceanogr. 1989;34:198–205. Limnol Oceanogr. 1969;14:912–20.
39. Halsey KH, Jones BM. Phytoplankton strategies for photo- 59. Thomas WH, Dodson AN. Effect of interactions between tem-
synthetic energy allocation. Annu Rev Mar Sci. 2015;7:265–97. perature and nitrate supply on the cell-division rates of two marine
40. Kruskopf M, Flynn KJ. Chlorophyll content and fluorescence phytoflagellates. Mar Biol. 1974;24:213–7.
responses cannot be used to gauge reliably phytoplankton bio- 60. German DP, Marcelo KRB, Stone MM, Allison SD. The
mass, nutrient status or growth rate. New Phytol. Michaelis–Menten kinetics of soil extracellular enzymes in
2006;169:525–36. response to temperature: a cross-latitudinal study. Glob Change
41. Milligan AJ, Halsey KH, Behrenfeld MJ. Advancing interpreta- Biol. 2012;18:1468–79.
tions of 14C-uptake measurements in the context of phyto- 61. Blagodatskaya Е, Blagodatsky S, Khomyakov N, Myachina O,
plankton physiology and ecology. J Plankton Res. 2015;37: Kuzyakov Y. Temperature sensitivity and enzymatic mechanisms
692–8. of soil organic matter decomposition along an altitudinal gradient
42. Geider R, La Roche J. Redfield revisited: variability of C:N:P in on Mount Kilimanjaro. Sci Rep. 2016;6:22240.
marine microalgae and its biochemical basis. Eur J Phycol. 62. Banse K. Rates of phytoplankton cell division in the field and in
2002;37:1–17. iron enrichment experiments. Limnol Oceanogr. 1991;36:1886–98.