The trade-off between the light-harvesting and photoprotective functions of fucoxanthin-chlorophyll proteins dominates light acclimation in Emiliania huxleyi (clone CCMP 1516)



  • Mechanistic understanding of the costs and benefits of photoacclimation requires knowledge of how photophysiology is affected by changes in the molecular structure of the chloroplast.
  • We tested the hypothesis that changes in the light dependencies of photosynthesis, nonphotochemical quenching and PSII photoinactivation arises from changes in the abundances of chloroplast proteins in Emiliania huxleyi strain CCMP 1516 grown at 30 (Low Light; LL) and 1000 (High Light; HL) μmol photons m−2 s−1 photon flux densities.
  • Carbon-specific light-saturated gross photosynthesis rates were not significantly different between cells acclimated to LL and HL. Acclimation to LL benefited cells by increasing biomass-specific light absorption and gross photosynthesis rates under low light, whereas acclimation to HL benefited cells by reducing the rate of photoinactivation of PSII under high light. Differences in the relative abundances of proteins assigned to light-harvesting (Lhcf), photoprotection (LI818-like), and the photosystem II (PSII) core complex accompanied differences in photophysiology: specifically, Lhcf:PSII was greater under LL, whereas LI818:PSII was greater in HL.
  • Thus, photoacclimation in E. huxleyi involved a trade-off amongst the characteristics of light absorption and photoprotection, which could be attributed to changes in the abundance and composition of proteins in the light-harvesting antenna of PSII.


Photosynthetic organisms must maintain a balance between the interception of light energy, the supply of this excitation energy to the photosynthetic reaction centres, the production of NADPH and ATP within the thylakoid membranes, and the utilization of NADPH and ATP for CO2 fixation and for biosynthesis. Changes in photosynthetic photon flux density (PFD) on timescales of seconds to days affect this balance, and as a consequence plants and algae must employ a range of mechanisms to adjust their physiologies to maintain this balance (Raven & Geider, 2003; Falkowski & Raven, 2007).

Photoacclimation is the term used to designate the process of adjustment of the phenotype of plants, algae and cyanobacteria to changes in the light environment. It operates on timescales of hours to days, which are intermediate between rapid changes involved in the regulation of enzyme activities and long timescales associated with changes in adaptation of the genotype (Raven & Geider, 2003). Photoacclimation involves changes in ultrastructure, morphology, biochemical and elemental composition, pigment content (Falkowski & La Roche, 1991), and the composition of the proteome (Pandhal et al., 2007; McKew et al., 2013).

The contrasting demands of aquatic and terrestrial environments, and the contrasting lifestyles of multicellular vascular plants and unicellular microalgae have led to differences in evolutionary adaptations and strategies of physiological acclimation (Geider et al., 2001). For example, many vascular plant leaves photoacclimate to high light by increasing the abundance of Calvin cycle enzymes per unit area whilst maintaining light absorption (Seemann et al., 1987; Givnish, 1988), whereas microalgae reduce light harvesting without increasing Calvin cycle enzymes (Sukenik et al., 1987; McKew et al., 2013). These differences in acclimatory strategy are likely to be a consequence of the fact that vascular plant leaves are optically thick and absorb > 90% of incident solar radiation (Knapp & Carter, 1998), whereas microalgae are typically optically thin and absorb < 50% of the light that falls on their surfaces (Morel & Bricaud, 1981). Additionally, vascular plants can often be characterized into sun and shade ecotypes (Givnish, 1988), whereas most microalgae and cyanobacteria are evolutionarily adapted to low light (Richardson et al., 1983), although a few notable exceptions of high and low light ecotypes have been identified (Moore & Chisholm, 1999). Within the microalgae and cyanobacteria, the rate of photosynthesis is typically saturated at c. 50–400 μmol photons m−2 s−1 (Richardson et al., 1983), which is significantly less than full noon sunlight. In most ocean environments photosynthesis shows a subsurface maximum at a depth corresponding to c. 10–30% of surface PFD (Behrenfeld & Falkowski, 1997), whereas in many vascular plants photosynthesis of ‘sun’ leaves does not saturate even in full sunlight (Bjorkman, 1981).

Photoacclimation on timescales of hours to days affects important physiological processes including the light dependence of photosynthesis and the susceptibility of photosystem II (PSII) reaction centres (RCII) to photoinactivation (Richardson et al., 1983; Falkowski & La Roche, 1991; Raven & Geider, 2003). During acclimation to incident light, microalgae modify components associated with PSII and photosystem I (PSI), (Falkowski & Raven, 2007). These adjustments appear to be driven, in part, through redox and reactive oxygen species (ROS) signalling that arise from an imbalance between the rate of light absorption and the rate of energy consumption in photosynthesis and biosynthesis (Escoubas et al., 1995; Pfannschmidt et al., 2009), although blue-light receptors have also been implicated in the photoacclimation of diatoms (Schellenberger Costa et al., 2013).

Knowledge of the cost–benefit trade-offs involved in photoacclimation is essential for developing a mechanistic understanding of how photosynthetic organisms acclimate to the light environment in order to improve fitness (Geider et al., 2009). These costs and benefits can be evaluated within the context of three design considerations for structure–function relationships in chloroplasts (Raven, 1980). These are the energetic efficiency of photosynthesis (work output divided by energy input), the catalytic efficiency of photosynthesis (work output per unit catalytic and structural material contained in the chloroplast) and the provision of mechanisms to ensure safe operation of the photosynthetic apparatus. Such an analysis is facilitated by the fact that all oxygenic photoautotrophs possess a highly conserved light-driven energy transduction pathway linking O2 evolution by PSII to the production of triose phosphates in the Calvin cycle (Blankenship, 2002). By contrast, considerable phylogenetic diversity is evident upstream of PSII in the light-harvesting pigments and proteins amongst the cyanobacteria and the green and red algal lineages (Blankenship, 2002), as well as in the downstream metabolism of the products of CO2 fixation (Wilhlem et al., 2006). This includes differences in organic C storage products, with carbohydrates or neutral lipids the major storage products depending on species and growth condition (Lacour et al., 2012).

In this paper, we examine photoacclimation of the marine haptophyte Emiliania huxleyi, calcifying strain CCMP 1516, which was the first coccolithophorid to have its genome sequenced by JGI (Read et al., 2013). The haptophytes significantly affect ocean biogeochemistry through their roles in the carbon and sulphur cycles. Emiliania huxleyi is a widely distributed species that typically blooms in high-latitude seas when and where the mixed layer is relatively shallow and thus characterized by high light (Iglesias-Rodriguez et al., 2002). Although its tolerance of high light is one of the attributes considered to contribute to blooms and the success of this species (Nanninga & Tyrrell, 1996), E. huxleyi can grow over a wide range of PFDs (Suggett et al., 2007). As in other photoautotrophs, growth of E. huxleyi in low light environments requires maximizing both the rate of light absorption per unit biomass and the quantum efficiency of photosynthesis, whilst tolerance of high light requires that oxidative stress and consequent cellular damage be minimized by reducing the production of ROS as well as increasing the scavenging of unavoidably produced ROS.

We investigated the functional consequences for light absorption, photosynthesis, nonphotochemical excitation energy quenching and photoinactivation that accompany the structural remodelling of the chloroplast proteome of E. huxleyi in response to growth at suboptimal and supraoptimal PFDs. We tested the hypotheses that changes in the efficiency of PSII photochemistry, the capacity for nonphotochemical quenching, and the susceptibility of RCII to photoinactivation could be explained as consequences of changes in the abundances of key chloroplast proteins. To this end, we used selected results from our analysis of the E. huxleyi proteomes under suboptimal and supraoptimal PFDs (McKew et al., 2013) together with new additional measurements of a suite of photophysiological characteristics. Unlike many previously published investigations, which focused on individual photophysiological processes in isolation, we have brought together observations of photoacclimation of light harvesting, nonphotochemical quenching and photoinactivation in conjunction with changes in the chloroplast proteome to assess the mechanistic basis of the cost–benefit trade-offs involved in photoacclimation.

Materials and Methods

Culture conditions and proteomics

Full details of culture conditions and the proteomic analysis are described in McKew et al. (2013). Briefly, cultures of calcifying Emiliania huxleyi CCMP 1516 were grown to steady-state in nutrient-replete turbidostat cultures on a 16 h : 8 h, light : dark cycle at PFDs (λ = 400–700 nm) of 30 μmol photons m−2 s−1 (Low Light, LL) or 1000 μmol photons m−2 s−1 (High Light, HL), and nontargeted proteomics of trypsin digested unfractionated cell extracts was conducted using LC-electrospray-ionization tandem mass spectrometry on a hybrid high-resolution LTQ/Orbitrap Velos instrument.

Pigment content

Samples for pigment analysis were collected on MF300 glass microfibre filters (Fisher Scientific, Loughborough, UK). In addition to measuring Chla on methanol extracts (McKew et al., 2013), High Performance Liquid Chromatography (HPLC) analysis was conducted on samples stored under liquid nitrogen. For HPLC analysis, pigments were extracted in 3.0 ml 90% acetone using sonication. Extracts were centrifuged to remove debris, and 100-μl aliquots analysed on a 3 μm Hypersil MOS2 C8 column using a ThermoFinnigan Spectra HPLC system with Thermo Separations AS3000 autosampler, Thermo Separations UV6000 diode array absorbance detector, and PC1000 and ChromQuest chromotography software (Barlow et al., 1997a,b). Chla (Sigma) and other pigment standards (DHI Institute for Water and Environment, Hørsholm, Denmark), along with retention time and light absorption spectra were used to identify pigments, whilst concentrations were calculated from peak area.

In vivo light absorption

Cells were concentrated by vacuum filtration of 300 ml of culture onto Whatman 3 μm pore polycarbonate filters (GE Healthcare, Little Chalfont, UK) and resuspended in 5 ml of the filtrate for measurement of the spectral dependence (380–800 nm) of light absorption using a Hitachi U-3000 spectrophotometer fitted with a 60 mm diameter integrating sphere (Hitachi High-Technologies, Berkshire, UK) correcting for residual scattering as previously described (Suggett et al., 2007). Chla was also measured on the concentrates to allow Chla-specific light absorption coefficients to be calculated.

Chla-specific light absorption spectra were also reconstructed from the measurements of the ratios of accessory pigments to Chla and the mass-specific absorption spectra reported for chlorophylls, photosynthetic carotenoids and nonphotosynthetic carotenoids (Bidigare et al., 1990) as described in Suggett et al. (2004) with corrections made for the package effect from measurements of equivalent spherical cell diameter and cellular Chla content (Morel & Bricaud, 1981).

Fast repetition rate (FRR) fluorometry

Fast repetition rate (FRR) fluorometry (Kolber et al., 1998) was used to obtain fluorescence light curves (FLCs) using the FastTracka II fluorometer and Fast Act laboratory system (both CTG Ltd, West Molesey, UK). Samples were dark adapted for 20 min before each FLC was started. The PFDs used for all FLCs were 5, 19, 34, 48, 121, 237, 444, 853 and 1413 μmol m−2 s−1. Each PFD step was run for 5 min. A 20-s dark interval was placed between all adjacent PFD steps, during which a single dark acquisition was made. FastPro software (CTG Ltd) was used to calculate the dark-adapted values of minimum and maximum fluorescence (Fo, Fm), the maximal photochemical efficiency in the dark-adapted state, Fv/Fm and the PSII effective cross-section (σPSII). At each PFD step the fluorescence under actinic illumination (F ′), the maximum fluorescence under actinic illumination (math formula) and the parameters math formula, math formula and math formula were obtained: By convention math formula is the overall actual operating efficiency of PSII under actinic light, whereas math formula provides maximum efficiency of PSII photochemistry in the light, and math formula is the photochemical factor, which is nonlinearly related to the proportion of RCII that are in the open state and thus capable of performing photochemistry.

Rates of photoinactivation and repair

Rates of PSII photoinactivation were determined at four PFDs from 150 to 1200 μmol photons m−2 s−1 by measuring the time dependence of changes in Fv/Fm. Values of Fv/Fm were measured after 30 min dark adaptation, following 5, 15, 30 or 60 min light treatment at each of the four PFDs. Gross rates of PSII photoinactivation were obtained from measurements on samples incubated in the presence of lincomycin (a specific inhibitor of chloroplast and mitochondrial protein synthesis, Ragni et al., 2008). Rate constants for photoinactivation were obtained by nonlinear curve fitting (SIGMAPLOT®, Systat Software Inc., London, UK) to the equation:

display math(Eqn 1)

(Fv(t)/Fm(t), values measured after 5, 15, 30 and 60 min light exposure in the presence of lincomycin; ki, inactivation rate constant; t, time). The rate constants for net photoinactivation were obtained from similar measurements made in the absence of lincomycin, and recovery was obtained by difference.

Gross and net oxygen exchange rate

The light response curves for gross O2 evolution and O2 uptake were determined using membrane inlet mass spectrometry (MIMS; model QMG 422; Pfeiffer Vacuum, Asslar, Germany) modified from Kana (1990) as described in Brading et al. (2011). Briefly, samples to which 18O2 was added were illuminated for 60–90 min at PFDs between 0 and 1200 μmol photons m−2 s−1. The gross O2 evolution rates were calculated from measurements of changes in 36O2 and 32O2 using isotope dilution equations (Kana, 1990). These rates were then normalized to cell concentration and parameters of the photosynthesis–light (PE) response curve (Jassby & Platt, 1976) were determined by nonlinear least squares (SIGMAPLOT®):

display math(Eqn 2)

(PCell, cell-specific photosynthesis rate; math formula, light-saturated rate; αCell, initial slope of the PE curve; E, photon flux density). As expected, and consistent with Eqn 2, PCell was not significantly different from zero at E = 0.


Growth rate and pigment composition

Emiliania huxleyi 1516 grew over three times faster under HL (1000 μmol photons m−2 s−1) than LL (30 μmol photons m−2 s−1), and, although HL cells were significantly larger, they contained significantly less Chla (Table 1). The β-carotene-to-Chla ratio, although low in absolute values, was c. 30% higher in HL than in LL (Table 1) and the total xanthophyll-to-Chla ratio was c. 20% greater in HL than in LL (Table 1). The major light-harvesting xanthophyll 19′-hexanoyloxyfucoanthin accounted for 55% of the total xanthophylls in LL, but declined to 48% in HL (Table 1). The ratio of fucoxanthin-to-Chla was c. 10-fold lower, and the ratio of diadinoxanthin-to-Chla c. 7-fold higher in HL compared to LL (Table 1).

Table 1. Response of growth rate, biomass variables and photophysiological variables to low light (30 μmol photons m−2 s−1) or high light (1000 μmol photons m−2 s−1) in Emiliania huxleyi (1516)
 Low LightHigh Light P
  1. Shown are the mean ± SE for = 3 and P values from two-tailed T-test; ns, nonsignificant difference (> 0.05). Total carotenoids are the sum of trace amounts of 19’-butanoyloxyfucoxanthin and lutein (not shown) and the large amounts of fucoxanthin, 19’-hexanoyloxyfucoxanthin and β-carotene, which are reported separately in the Table. Diatoxanthin was not detected. Chla, chlorophyll a; POC, particulate organic carbon; PN, particulate nitrogen; math formula, light saturated photosynthesis rate; αCell, initial slope of the photosynthesis–light curve; Ek, math formulaCell, light saturation parameter; math formula, dark respiration rate; math formula, O2 uptake rate at light saturation; Fv/Fm, ratio of variable to maximum fluorescence of dark-adapted cells; σPSII,effective cross-section of photosystem II in the dark adapted state; 1/τ, inverse of the turnover time for photosystem II in the dark-adapted state; aPSII, optical absorption cross-section of PSII, calculated from σ/( Fv/Fm).

Growth rate (d−1)0.19 ± 0.030.67 ± 0.03< 0.001
Cell diameter (μm)4.17 ± 0.015.13 ± 0.140.001
Chla (fg per cell)215 ± 5.2120 ± 15.90.002
Chla:POC (mg g−1)23.3 ± 2.48.8 ± 0.80.002
Chla:PN (μg g−1)0.131 ± 0.0080.062 ± 0.0080.002
math formula (fmol O2 cell−1 h−1)84 ± 9.1127 ± 18ns
αCell ((fmol O2 cell−1 h−1)(μmol photons m−2 s−1)−1)1.33 ± 0.220.68 ± 0.13ns
Ek (μmol photons m−2 s−1)65 ± 6191 ± 12< 0.001
math formula (fmol O2 cell−1 h−1)12 ± 415 ± 2ns
math formula (fmol O2 cell−1 h−1)43 ± 733 ± 5ns
Fv/Fm (dimensionless)0.506 ± 0.0060.390 ± 0.0140.013
σPSII (nm2)2.48 ± 0.061.93 ± 0.090.003
1/τ (s−1)1506 ± 751784 ± 320.013
aPSII (nm2)4.90 ± 0.134.95 ± 0.25ns
Total carotenoids:Chla (g g−1)0.896 ± 0.0311.05 ± 0.0120.01
Fucoxanthin:Chla (g g−1)0.314 ± 0.0080.038 ± 0.002< 0.001
19’ hexanoyloxyfucoxanthin:Chla (g g−1)0.470 ± 0.0220.475 ± 0.007ns
Diadinoxanthin:Chla (g g−1)0.073 ± 0.0070.489 ± 0.01< 0.001
β-carotene:Chla (g g−1)0.036 ± 0.0010.043 ± 0.0010.004

Light absorption

The Chla-specific light absorption coefficient, aChl(λ), was c. 35% greater in HL than LL at the red (λ = 675 nm) peak, and c. 60% greater at the blue peak (λ = 440 nm) (Fig. 1). By contrast, values of aChl(λ) were very similar in the region where absorption was minimal (λ ≈ 575–650 nm). These observations are entirely consistent with the package effect (Kirk, 1976; Morel & Bricaud, 1981). Specifically, the difference in aChl(675) can be attributed to the greater optical thickness of the LL cells, which contained a higher concentration of Chla, while differences in the spectra at wavelengths < 550 nm can be attributed to both the package effect and to the higher ratios of xanthophylls-to-Chla in HL (Table 1). The differences in pigment composition are also evident in the spectra from the methanol extracts (Fig. 1b): The HL spectrum has a sharper peak at 440 nm and lower absorption in the 490–600 nm range. Across the photosynthetically active range (λ = 400–700 nm), aChl was c. 40% greater in HL than LL.

Figure 1.

Light absorption spectra for Emiliania huxleyi (1516) grown under low light (LL, closed circles) and high light (HL, open circles). (a) Intact cell, Chla-specific light absorption. (b) Spectra of methanol extracts, normalized to unity at the red (664 nm) absorption peak). Shown are the mean ± SE for = 3. (c) Reconstructed spectra from cellular pigment content, pigment composition cell diameter allowing for the package effect. The reconstruction uses the pigment-specific absorption coefficient given in Bidigare et al. (1990) and accounts for the package effect following Morel & Bricaud (1981).

We calculated Chla-specific light absorption spectra from observations of cell size and cell pigment content assuming spherical homogenous cells. These calculated spectra were similar in magnitude to the observed absorption spectra of intact cells, although they differed slightly in shape (compare Fig. 1c with 1a). Across the wavelength range from 400 to 700 nm, Chla and the photosynthetic carotenoids accounted for c. 90% of aChl in LL but only 50% of aChl in HL E. huxleyi (Supporting Information Fig. S1).

Biophysical characteristics of photosystem II

The maximum PSII photochemical efficiency in the dark-adapted state (Fv/Fm) was c. 20% lower in the HL than in LL (Table 1), as was the effective cross-section of PSII photochemistry (σPSII) (Table 1). As a consequence, the calculated value for the light absorption cross-section of PSII, aPSII = σPSII/(Fv/Fm), was unchanged between LL and HL (Table 1). The rate constant for reopening of closed RCII (designated 1/τ) in the dark-adapted state was higher for HL than LL cells (Table 1). Although this could indicate that there are a larger number of plastoquinone molecules per PSII within HL cells, it may simply indicate a slightly more reduced plastoquinone pool within dark-adapted LL cells.

The PSII photochemical efficiency under ambient light is often estimated as math formula (Genty et al., 1989). This parameter is the product of two other parameters, math formula and math formula, which (respectively) track the nonphotochemical and photochemical constraints on the overall efficiency of PSII photochemistry (Baker & Oxborough, 2005). The lower values of math formula for HL cells (Fig. 2a) indicate higher effective rate constants for nonphotochemical energy dissipation throughout the entire range of PFDs. By contrast, math formula decreased much more slowly with increasing PFD in the HL than LL cells (Fig. 2b). As a consequence LL cells showed a higher math formula at low PFDs, whereas HL cells achieved a higher math formula at high PFDs (Fig. 2c).

Figure 2.

Light dependence of indices of: (a) the efficiency of excitation energy transfer from the light-harvesting pigment bed to the reaction centres, designated math formula; (b) the efficiency of charge separation within PSII reaction centres (photochemical efficiency), designated math formula; (c) the overall operating efficiency of PSII, designated math formula, which is the product of math formula and math formula; (d) the Stern-Volmer coefficient for quenching of maximum fluorescence, SVm = math formula; and (e) the normalized Stern-Volmer quenching coefficient, NVS = math formula (K. Oxborough, unpublished) for Emiliania huxleyi (1516) grown under low (closed circles) and high (open circles) light conditions. Shown are the mean ± SE for = 3. The precise relationship between these fluorescence parameters and the biophysical properties of PSII depends on the molecular architecture of PSII, in particular on the extent to which reaction centres do or do not share a common antenna (see Oxborough et al., 2012).

The phenomenon of nonphotochemical excitation energy quenching is commonly quantified by the Stern-Volmer coefficient for quenching of maximum fluorescence, which we designate as SVm but others often refer to as NPQ. A major problem with use of NPQ = SVm = (Fmmath formula) to quantify nonphotochemical quenching is that a comparison of values of SVm in samples with different Fv/Fm (Table 1) can be misleading. In fact, using SVm (Fig. 2d) hides an important difference between the LL and HL E. huxleyi cells; namely that the dark level of nonphotochemical excitation energy dissipation in the HL cells was significantly greater than in LL cells, thus accounting for the lower math formula values at low PFD (Fig. 2c). As a consequence, the differences in downregulation of excitation energy transfer between HL and LL are not elucidated by SVm simply because the degree of nonradiative decay for the HL sample in the dark is much higher than for the LL sample. A new parameter, the normalized Stern-Volmer coefficient, defined as NSV = math formula, corrects this deficiency (K. Oxborough, unpublished): NSV is the ratio of the total nonphotochemical dissipation in the light-adapted state to the rate constant for photochemistry. Plots of NSV vs PFD clearly show the greater nonphotochemical quenching of HL relative to LL cells at all PFDs from those that are light-limiting to those that are light-saturating for photosynthesis (Fig. 2e).

Photoinactivation kinetics

The first-order rate constant for gross photoinactivation, ki, appeared to be insignificant at PFDs < 140 μmol photons m−2 s−1, but increased linearly with the PFDs above this threshold (Fig. 3a). At PFDs above 140 μmol photons m−2 s−1, ki was approximately two-fold greater in LL cells than in HL cells (Fig. 3a). Significant net photoinactivation was not observed in HL cells at any PFD, and was only observed in LL cells when they were exposed to a PFD of 1200 μmol photons m−2 s−1 (Table S1). Even in this extreme case, the net photoinactivation rate was < 20% of the gross photoinactivation rate (Table S1).

Figure 3.

(a) Dependence of the rate constant for gross photoinhibition of Fv/Fm (ki) on the PFD for Emiliania huxleyi (1516) grown under low (closed circles) and high (open circles) light. Linear regression of ki on PFD yielded x-intercepts of 136 ± 18 and 146 ± 31 μmol photons m−2 s−1 (intercept ± SD) for HL and LL, respectively. (b) Dependence of ki on the product of PFD and the proportion of RCII that are reduced. The linear relationship which extrapolates to the origin provides a test of Eqn 5.

A strict photon-dose response would result in ki being directly proportional to the product of the absorption cross-section for target, the probability that an absorbed photon will inactivate the reaction centre, and the photon flux density:

display math(Eqn 3)

(ki, rate constant for photoinactivation (s−1); atarget, absorption cross-section of the putative target; Ψ, probability of a PSII being photoinactivated by an incident photon (photon−1); E, photon flux density (mol photons m−2 s−1)). Although we observed a linear relationship between ki and PFD, this applied only above a threshold of c. 140 μmol photons m−2 s−1 (Fig. 3). Consequently, our data are not consistent with this model if both atarget and Ψ are assumed to be constant (see Fig. S2a). This conclusion prompted us to examine an alternative model in which the trigger for inactivation is double reduction of QA, as proposed by Vass et al. (1992). This would require that the observed rate of photoinactivation is proportional to the absorption cross-section of PSII light harvesting at closed RCIIs, corrected for downregulation, as is given by Eqn 4:

display math(Eqn 4)

(aPSII, absorption cross-section of PSII light harvesting; math formula/Fm, correction for downregulation; (1 − (math formula)), ratio of the sum of aPSII for light harvesting systems serving only closed RCII to the sum of aPSII for light harvesting systems serving all RCII (Oxborough et al., 2012; Notes S1)). When we calculated Ψ from rearrangement of Eqn 4 we found a systematic decline with increasing PFD (Fig. S2b), indicating that our observations are inconsistent with this model as well. Finally, we considered a scenario where we assumed that the photoinactivation target is independent of PSII light harvesting, but is only active within a PSII complex with a closed RCII:

display math(Eqn 5)

In this scenario, (1 −math formula) provides an estimate of the fraction of RCII in the closed state, and although this estimation is connectivity-dependent, our fluorescence data are entirely consistent with the use of math formula in this instance. Our observations are in very good agreement with this model (Figs 3b, S2c), which is, to some extent, a hybrid of Eqns 3 and 4.

Photosynthesis–light response curve

Cellular gross light-saturated photosynthesis (math formula) and dark respiration rates (math formula) were not significantly different in LL and HL cells (Table 1), nor was the rate of oxygen uptake (math formula) at a saturating PFD of 1000 μmol photons m−2 s−1 (Table 1). By contrast, the Chla-specific light-saturated rate of gross photosynthesis (math formula) was three-fold higher in HL, relative to LL (Fig. 4; Table 2), as was the PFD required to saturate gross photosynthesis, EK (Table 2). The Chla-specific initial slopes (αChl) of the photosynthesis–light (PE) response curves were not significantly different between LL and HL (Fig. 4; Table 2), but the maximum quantum efficiency (ϕm = αChl/aChl) of photosynthesis was lower in HL than LL (Table 2).

Table 2. Response of photophysiological variables to growth under low light (30 μmol photons m−2 s−1) or high light (1000 μmol photons m−2 s−1) in Emiliania huxleyi (1516)
VariableDimensionsLow LightHigh Light
  1. Chla specific, carbon-specific photosynthesis rates were calculated from cell-specific data (Table 1) with error propagated assuming uncorrelated errors.

  2. math formula, the Chla-specific light saturated gross oxygen evolution rate; αChl, the initial slope of the PChl vs light curve; aChl, the Chla-specific light absorption coefficient; ϕm, the maximum quantum efficiency of gross oxygen evolution; math formula, light saturated gross oxygen evolution rate; αC, the initial slope of the PC vs light curve; aC, the organic C-specific light absorption coefficient.

math formula mol O2 (g Chla)−1 h−10.39 ± 0.041.18 ± 0.24
αChlmol O2 (mol photons)−1 m2 (g Chla)−11.7 ± 0.31.7 ± 0.4
a Chl m2 (g Chla)−114.6 ± 1.220.3 ± 0.5
ϕmmol O2 (mol photons)−10.12 ± 0.020.08 ± 0.02
math formula mol O2 (mol C)−1 h−10.11 ± 0.020.11 ± 0.02
αCmol O2 (mol photons)−1 m2 (g C)−10.14 ± 0.030.050 ± 0.011
a C m2 (mg C)−10.34 ± 0.050.18 ± 0.02
Figure 4.

Typical light responses of (a) gross photosynthetic O2 evolution and (b) O2 consumption for Emiliania huxleyi (1516) grown under low (closed circles) and high (open circles) light. Summary statistics for cell-specific photosynthetic parameters and dark respiration rates are given in Table 1, and summary statistics for Chla-specific and organic carbon-specific parameters are given in Table 2.

Impact of light on photosynthetic proteins

Our recent findings (McKew et al., 2013) indicated that the photosynthetic proteins accounted for a greater proportion of the total proteome in LL, and in addition that the relative abundances of specific photosynthetic proteins also changed significantly between HL and LL. We reanalysed these data by first normalizing to the abundance of PSII core complex proteins (sum of spectral counts for detected proteins PsbA, B, C, D, E, H and V) as an internal photosynthetic unit reference before taking the ratio of these PSII-normalized relative abundances between LL and HL (Fig. 5) to facilitate interpretation of the photophysiological data. The ratios PSI:PSII and cyt b6/f:PSII were identical in LL and HL treatments, and Lhcf(Red):PSII did not change significantly between LL and HL treatments (Fig 5). Relative to PSII, there were significant decreases in diatom/haptophyte groups I and II Lhcfs in HL, countered by significant increases in the LI818-like group of fucoxanthin chlorophyll binding proteins. Similarly Lhcz:PSII was greater in HL although the difference was not significant. The ratios Calvin cycle:PSII and ATPsynthase:PSII were also significantly lower in HL.

Figure 5.

(a) Changes in the ratios of Lhcfs and LI818-like proteins to PSII between low light (LL) and high light (HL) Emiliania huxleyi (1516) expressed as the LL-to-HL ratio. (b) Changes in the ratios of thylakoid membrane complexes and Calvin cycle enzymes to PSII between LL and HL. (Shown are the mean ± SE of three technical replicates across three replicate cultures of each light treatment; = 3). *, Significant difference at < 0.05. Calculated from data reported for spectral counts in McKew et al. (2013).


We evaluated changes in the photophysiological characteristics of Emiliana huxleyi within the context of three design considerations for the structure–function relationship in chloroplasts identified by Raven (1980). These considerations are: the energetic efficiency of photosynthesis; the catalytic efficiency of photosynthesis; and the provision of mechanisms to ensure safe operation of the photosynthetic apparatus. Protein typically accounts for c. 50% of the mass of a microalga under nutrient-replete conditions, and as such the proteome is expected to adjust and respond to selective pressures associated with changes in the environment. The greatest response of the E. huxleyi proteome to HL and LL conditions was observed in the antennae proteins of PSII: these proteins are involved in both light harvesting and excitation energy dissipation (McKew et al., 2013). Here, we discuss the implications of these changes in the proteome for the photophysiology of E. huxleyi, focusing on characteristics that map onto Raven's (1980) three design considerations (Table 3).

Table 3. Summary of responses of differences in photophysiological characteristics of Emiliania huxleyi (1516) between low light (LL) and high light (HL) treatments categorized according to Raven's (1980) design considerations for structure–function relationships in eukaryotic chloroplasts
Design considerationPhotophysiological characteristic assessedSymbolTreatment effectChange in chloroplast proteins
Energetic efficiency of the process defined as the ratio of useful work output to energy inputMaximum quantum efficiency of photosynthesisϕmLL > HLIncrease of LI818:Lhcf in HL reduces energetic efficiency
Maximum efficiency of PSII in the dark-adapted state F v /F m LL > HL
Catalytic efficiency of the process defined as work output per unit catalytic and structural material usedChla-specific light absorption coefficient a Chl LL < HLIncrease of Lhcf:RCII in LL increases rate of light absorption
Organic carbon-specific light absorption coefficient a C LL > HL
Light harvesting (light absorption) cross section of PSII a PSII LL = HL
Organic carbon-specific light-saturated photosynthesis rate math formula LL = HL
Provision of mechanisms to ensure safe operation of the photosynthetic apparatusRate constant for photoinactivation of RCII at a given PFD k i LL > HLIncrease of LI818:RCII in HL maintains more RCII in the open state
Susceptibility of reduced RCII to photoinactivationatarget ΨLL = HL
Proportion of PSII reaction centres in the ‘open’ state at high PFD math formula LL < HL

The Chla-specific initial slope, αChl, of the photosynthesis–light response curve (PE curve) was unaffected by acclimation to low or high light

The PE curve (Fig. 4a) has a central role in elucidating and understanding different strategies of photoacclimation in microalgae (Richardson et al., 1983). The initial slope (αChl) of this curve quantifies the linear increase of the rate of photosynthesis with PFD under light-limited conditions (Jassby & Platt, 1976; MacIntyre et al., 2002), and is the product of the light absorption coefficient (aChl) and the maximum quantum efficiency of photosynthesis (ϕm):

display math(Eqn 6)

In the context of Raven's (1980) design considerations, aChl is a catalytic efficiency, which when multiplied by the PFD gives the rate of photon absorption, whereas ϕm is an energetic efficiency that relates the output of O2 to the input of photons. We calculated that ϕm was 25% lower in HL than LL from our measurements of αChl and aChl. This decrease is consistent with the 22% lower value of Fv/Fm for the HL cells (Table 1), which along with the lower values of math formula in the light (Fig. 2) can be ascribed to the increase in the contribution of the photoprotective pigments to light absorption in the blue region of the spectrum (Fig. S1).

The independence of αChl from growth PFD in E. huxleyi results from compensatory changes between the catalytic efficiency of light absorption and the energetic efficiency of excitation energy transfer from the light harvesting pigments to the PSII reaction centres. Similar reciprocal changes of aChl and ϕm in response to growth at different PFDs have been observed in other microalgae (Geider et al., 1998; Anning et al., 2000). This invariance of αChl within a taxon facilitates the formulation of simple models of algal growth and acclimation (e.g. Geider et al., 1998); however, it is not immediately obvious why this compensation occurs. To begin to address this issue, we now turn to the bio-optical and biophysical mechanisms that underpin changes of cellular pigment content, aChl and ϕm.

Acclimation of E. huxleyi to low light increased both the size and number of photosynthetic units

Changes in the protein composition of the thylakoid membranes that accompany increased cellular pigment content in low light can involve: an increase in the number of photosynthetic units of constant ‘size’; or an increase in the sizes of a constant number photosynthetic units; or some combination of the two (Falkowski & La Roche, 1991). The number of RCII per cell is typically quantified by measuring the abundances of PSII and PSII reaction centre proteins using either biophysical techniques (Falkowski et al., 1981) or Western blotting (Six et al., 2008), whereas photosynthetic unit size is typically quantified by the ratio Chla:RCII. In E. huxleyi the proportion of the proteome attributable to PSII was greater in LL than HL (McKew et al., 2013), and the ratio Lhcf:PSII was also significantly greater in LL compared to HL (Fig. 5), consistent with previous observations that photoacclimation in this species involves changes of both photosynthetic unit size and number (Suggett et al., 2007).

The higher photosynthetic unit number and photosynthetic unit size in LL E. huxleyi accounted for c. 55% of the 2.8-fold greater C-specific initial slope of the PE curves in LL treatments, with the remaining 45% of the increase attributable to the greater ϕm in LL than HL. Thus, by increasing pigment content under LL E. huxleyi achieved a higher growth rate under LL than would have been possible in the absence of photoacclimation, thereby contributing to increased fitness.

Acclimation of E. huxleyi to low light increased the effective cross-section of PSII photochemistry (σPSII) but not the absorption cross-section of PSII light harvesting (aPSII)

Our proteomic and biophysical data yielded markedly different conclusions regarding the response of PSU ‘size’ to the light treatments. The 2.3-fold greater Lhcf:PSII in LL than HL (Fig. 5) indicates that the number of pigment molecules, and mass of protein, associated with each PSII was markedly greater in LL than HL. By contrast, there was no significant change in the absorption cross-section of PSII light harvesting between LL and HL (aPSII = σPSII/[Fv/Fm] = 4.9 in both LL and HL) because the 1.3-fold higher σPSII under LL was offset by an equivalent increase in Fv/Fm (Table 1).

To summarize, despite the markedly higher resource allocation to Lhcfs in LL cells, the average light absorption cross-section of each individual PSII light-harvesting system was the same for both LL and HL cells. The likely explanation for this result is that the extent of intracellular self-shading due to the pigment packaging effect, which quantifies the decrease in the efficiency of light absorption by pigments when present at extremely high concentrations in cells compared with the same mass of pigment when more diffusely distributed in solution (Morel & Bricaud, 1981), was much greater in highly pigmented and slightly smaller LL cells than in less pigmented, larger HL cells. Thus, aPSII was unchanged between LL and HL because the decline in aChl was offset by the increase in Lhcf:PSII. Our data indicate that to maintain the same PSII light absorption cross-section (when PSII abundance was 1.7-fold greater) in LL relative to HL required, on average, that each PSII of a LL E. huxleyi cell contained about a 2.3-fold greater amount of Lhcfs. In conclusion, there was a marked decline in the marginal rate of increase in light interception per unit investment of the proteome into Lhcfs and PSII of LL E. huxleyi due to the package effect, with each pigment molecule only c. 50% as effective in LL as in HL cells.

Our observations differ from those reported for Ostreococcus where changes in σPSII were correlated with changes in Lhcp:PSII (Six et al., 2008). Cells of Ostreococcus are much smaller than those of E. huxleyi, and the package effect plays a less significant role in such small cells (Morel & Bricaud, 1981). Thus, in larger organisms like E. huxleyi, it is likely that the diminishing return on investment in Lhcfs is greater than in picoplankton like Ostreococcus species. This has implications for integrating optimality considerations (Smith et al., 2011) into trait-based models of phytoplankton ecophysiology (Litchman & Klausmeier, 2008).

Different costs and benefits are associated with acclimation of PSU number and size

In their study of photoacclimation of two Ostreococcus ecotypes, Six et al. (2008) contrasted n-type photoacclimation, which involved changes in PSU number per cell, with σ-type photoacclimation, which involved changes in the effective cross-section of PSII photochemistry (σPSII). Six et al. (2008) suggested that the photoacclimation strategy employed by a microalga can have significant implications for the capital costs of investment in thylakoid proteins, as well as the operating and/or opportunity costs associated with the D1 damage and repair cycle. Increasing the number of photosynthetic units in response to low light is more nitrogen-expensive than adding peripheral light harvesting complexes to the PSII and PSI cores (Dubinsky et al., 1986; Moore et al., 2006; Six et al., 2008). Similarly, in E. huxleyi investment in Lhcfs with a ratio of c. 0.37 mol Chla (kg protein)−1 will be more efficient in light absorption per unit nitrogen than investment in core PSII complex which have a ratio of c. 0.10 mol Chla (g protein)−1 (see Table S2). Exacerbating the difference in Chla:protein between antennae Lhcfs and the PSII core is the fact that much of light absorption by E. huxleyi is due to xanthophylls (Fig. S1), which are bound to Lhcfs.

There are, however, potential costs of increasing the ratio of Lhcfs to PSII in a σ-type strategy. For example, increasing aPSII may increase the susceptibility to photoinhibition by increasing the rate of excitation delivery to the reaction centres (Moore et al., 2006; Six et al., 2008). Although this may be the case in Ostreococcus (Six et al., 2008), aPSII was not significantly different between LL and HL E. huxleyi. Six et al. (2008) also suggested that a large antenna may impede the rate of RCII repair following photoinactivation, which would have consequences for intermittent exposure to supraoptimal light often experienced in nature. By contrast, a large cellular pool of RCIIs with smaller light harvesting absorption cross-sections may provide a buffer against the accumulation of inactive PSII units (Behrenfeld et al., 1998). Taking all of these considerations into account, Six et al. (2008) suggested that σ-type photoacclimation is expected in organisms evolutionally adapted to stable low-light environments with n-type strategy expected in organisms adapted to environment characterized by high variability in light.

The carbon-specific light-saturated photosynthesis rate, math formula, was unaffected by acclimation to low or high light

Although the Chla-specific light-saturated photosynthesis rate (math formula) was about three-fold greater in HL than LL E. huxleyi (Fig. 4a), the organic C-specific light-saturated photosynthesis rate (math formula) did not differ between LL and HL cells (Table 2). The limited variability of math formula in response to growth at different PFDs is also observed in other microalgae (MacIntyre et al., 2002), and is consistent with observations that the abundance of the Calvin cycle enzyme Rubisco is largely unaffected by growth at different PFDs in a number of species (Sukenik et al., 1987; Fisher et al., 1989; Six et al., 2008; Harris et al., 2009; Lefebvre et al., 2010). In fact, our own data suggest that Rubisco is slightly more abundant in LL than in HL acclimated cells of E. huxleyi 1516 (McKew et al., 2013).

The invariance of math formula in E. huxleyi was accompanied by a two-fold reduction in the cell-specific initial slope of the PE curve (Table 1), and a three-fold reduction in the carbon-specific initial slope, in HL relative to LL (Table 2). Thus, photoacclimation to HL does not appear to benefit E. huxleyi, or most other microalgae that have been examined (Anning et al., 2000; MacIntyre et al., 2002), by increasing the biomass-specific photosynthesis rate at the growth PFD. Thus, we must seek an alternative explanation to account for the benefit of acclimation to HL in E. huxleyi. This benefit is likely to be a reduction in the susceptibility of RCII to photoinactivation, although it may also include mechanisms that reduce the production of ROS and associated oxidative stress.

Blooms of E. huxleyi tend to occur in water columns where incident solar radiation exceeds a mean of c. 500 μmol photons m−2 s−1 over the day (Nanninga & Tyrrell, 1996). Thus, conditions under which E. huxleyi or other microalgae would experience prolonged exposure to PFDs as high as 1000 μmol photons m−2 s−1 will be limited to water columns where the surface layer becomes diurnally stratified under clear skies. Under these conditions, PFD can exceed 1000 μmol photons m−2 s−1 for several hours at a time. However, nocturnal destratification will redistribute cells vertically, making it unlikely that the same cells (or their daughter cells) will experience such high PFDs for several days running. Thus, there would appear to be little selection pressure in nature to develop an acclimation strategy to prolonged exposure to such high PFDs. Instead, the acclimation strategy is likely to involve development of tolerance to episodic exposure to high PFDs whilst maintaining the capacity to photosynthesize efficiently in low light.

Acclimation of E. huxleyi to high light reduced the susceptibility of RCII to photoinactivation by increasing energy dissipation in the antenna complexes

It is inevitable that some excitation energy is dissipated by uncontrolled reactions which generate free radicals, which in turn can cause photo-oxidative damage (Salin, 1987). The D1 reaction centre protein of PSII is one of most rapidly turned over proteins in the photosynthetic electron transfer chain because it is particularly susceptible to photo-oxidative damage (Raven, 2011). When the rate of D1 photoinactivation exceeds the rate of repair, photoinhibition occurs. We observed that the rate of D1 damage was about two-fold greater in E. huxleyi growing under LL than in E. huxleyi grown in HL (Fig. 3).

As discussed by Anderson et al. (1998) and Oxborough & Baker (2000), one mechanism to reduce the susceptibility to photoinactivation during prolonged exposure to supra-optimal light is to increase the capacity for nonphotochemical quenching, a process that is referred to as downregulation (Krause & Behrend, 1986; Oxborough & Horton, 1988; Li et al., 2002). Downregulation operates on timescales of seconds to minutes, to adjust the level of nonphotochemical de-excitation in response to changes in both PFD and the requirement for NADPH and ATP. This process increases the effective rate constant for nonradiative decay, which competes directly with both PSII photochemistry and Chla fluorescence for excitons within the PSII pigment matrix. The result is an increase in the level of nonphotochemical de-excitation, which shortens exciton lifetime within the PSII pigment matrix. Invariably, down regulation results in a nonphotochemical quenching of the fluorescence signal (Schreiber et al., 1986), which decreases maximum fluorescence (Fm in the dark-adapted state, math formula in the light-adapted state). Despite the fact that downregulation competes directly with PSII photochemistry, the process may have minimal impact on the rate of PSII photochemistry, particularly under high light conditions where the rate becomes limited on the reducing side of PSII.

In E. huxleyi, as in other microalgae, different Chla binding proteins contribute to light-harvesting and nonphotochemical quenching. Fucoxanthin-Chl binding proteins in the groups I and II Lhcf clades and the Lhcf(red) clade contribute to light-harvesting and efficient excitation energy transfer to the reaction centres, and the abundances of these complexes were up-regulated in LL (Fig. 5a). By contrast, LI818 and LI818-like genes of various microalgal species are upregulated under high light (Lefebvre et al., 2010; Park et al., 2010; Zhu & Green, 2010). Consistent with the photoprotective role of downregulation, the contribution of LI818 proteins to spectral counts of all of the fucoxanthin binding proteins increased from c. 3% in LL to 25% in HL (see Table S2 in McKew et al., 2013). This difference was associated with an increase in diadinoxanthin:Chla (Table 1), decrease in Fv/Fm (Table 1), decrease of math formula (Fig. 2a), and a decrease in the rate constant for gross photoinactivation (Fig. 3a). Thus, in E. huxleyi, as previously shown for Chlamydomonas (Peers et al., 2009), there is a direct functional correlation between accumulation of LI818 proteins and photoprotection. Our data also suggest a photoprotective role for the Lhcz-like proteins, which were more abundant (relative to PSII) in HL than LL (Fig. 5), although the difference was not as pronounced as for the LI818-like proteins.

Significantly, our data indicate that photoinactivation of Fv/Fm in E. huxleyi can be explained by a process in which some component of a closed RCII is the target (Fig. 3c). Thus, increased protection of RCII from photoinactivation in HL cells appears to be conferred by higher capacity of these cells to downregulate excitation energy transfer from the light-harvesting antenna to RCII, thus maintaining a greater proportion of RCII in the open state.


Differences in photophysiology of E. huxleyi between low-light and high-light treatments were associated primarily with changes in the abundances of different groups of fucoxanthin-Chl binding proteins. The acclimation of photophysiology reported in this paper supports our conclusion based on whole-cell proteomic data (McKew et al., 2013) that the benefit to E. huxleyi of acclimation to light energy-limited conditions is an increase in the rate of gross photosynthesis at the growth irradiance, whereas the benefit of acclimation to supra-optimal, light-saturating conditions is a reduction in the rate of photoinactivation, and possibly other sources of photodamage, at the growth PFD. Significantly, the decreased carbon-specific initial slope of the PE curve that accompanied reduced pigment:biomass in E. huxleyi acclimated to high light was not accompanied by an increase in the carbon-specific light-saturated photosynthesis rate.

Photoacclimation was most evident in the protein and pigment composition of the antenna of PSII and in the ratio of peripheral antenna to the PSII core. These changes reflect differences in resource allocation between the light-harvesting and photoprotective functions within the PSII antenna. The functional consequences of these changes in the molecular structure of PSII included reducing the susceptibility of RCII to photoinactivation in cells acclimated to high-light and increasing the delivery of excitation energy to RCII in cells acclimated to low light. The specific changes are as follows: an increase of LI818-like proteins and xanthophyll cycle pigments in HL reduced the energetic efficiency of photosynthesis; an increase of Lhcf proteins and photosynthetic pigments in LL increased light harvesting and the light-limited rate of photosynthesis; an increase in the ratio LI818:Lhcf in HL reduced susceptibility of RCII to photoinactivation by maintaining a greater proportion of RCII in the open state (Table 3).

Our results support Raven's (1980) analysis that structure–function relationships in chloroplasts of eukaryotic cells are constrained by trade-offs to maximize the energetic and catalytic efficiencies of photosynthesis, subject to the provision of mechanisms to ensure safety from photooxidative stress. Quantifying the cost–benefit trade-offs associated with different acclimation and adaptation strategies is still in its infancy. This is despite the need for better mechanistic descriptions of algal growth based on optimal allocation principles (Smith et al., 2011) for inclusion in larger scale models of open ocean ecology and biogeochemistry (Follows & Dutkiewicz, 2011). Our results confirm that Raven's (1980) design considerations of energetic efficiency, catalytic efficiency and safety provide a framework that can be used to inform optimality-based models of chloroplast function and algal growth.


This research was supported by NERC grant NE/G003688/1 awarded to R.J.G., C.A.R. and M.V.M. J.H. and S.J.F. were supported by NERC studentships.