Introduction

Functional trait-based approaches are increasingly adopted to explain and understand the distribution and diversity of phytoplankton communities (Litchman and Klausmeier, 2008; Barton et al., 2013; Edwards et al., 2013). Various morphological and physiological traits have been shown to define the ecological niches of phytoplankton species, including size, temperature response and resource acquisition and utilization traits. For example, in planktonic diatoms, which have a key role in marine primary production and biogeochemical cycling (Armbrust, 2009), pronounced species-specific differences in photosynthetic architecture and photophysiological strategies have been documented (for example, Dimier et al., 2007; Key et al., 2010; Schwaderer et al., 2011; Wu et al., 2012) and related to their in situ light environment (Strzepek and Harrison, 2004; Lavaud et al., 2007; Dimier et al., 2009; Petrou et al., 2011). A high capacity for physiological photoprotection is generally observed in highly fluctuating light climates and/or under on average high irradiances. This suggests that photoprotective capacity is an adaptive trait that shapes the distribution of planktonic diatoms in the environment (Lavaud et al., 2007; Dimier et al., 2009; Bailleul et al., 2010; Petrou et al., 2011; Lavaud and Lepetit, 2013).

Benthic marine environments, and especially intertidal environments, are characterized by even more changeable and extreme light climates resulting from the interplay of weather conditions, tides, water column turbidity and sediment composition (and hence light penetration) (Admiraal, 1984; Underwood and Kromkamp, 1999; Paterson and Hagerthey, 2001). Nevertheless, intertidal sediments rank among the most productive ecosystems on Earth, largely owing to the primary production of highly diverse assemblages of benthic diatoms (Underwood and Kromkamp, 1999). To date, however, little is known about the role of functional traits, and especially photophysiological traits, in shaping the structure, dynamics and function of benthic diatom assemblages. In most studies, diatom functional groups are defined on the basis of morphological growth form (for example, Gottschalk and Kahlert, 2012; Larson and Passy, 2012) and not physiological traits. In addition, photoprotective ability (limited to the measurement of the ‘xanthophyll cycle’, XC) and its relationship with ecology has only been studied in natural communities with mixed assemblages of functional groups (for example, Jesus et al., 2009; van Leeuwe et al., 2009; Cartaxana et al., 2011).

In temperate seas, intertidal benthic communities are largely dominated by diatoms (Méléder et al., 2007; Ribeiro et al., 2013), which display a high degree of taxonomic, phylogenetic and functional diversity (Kooistra et al., 2007). Several growth forms can be distinguished, which mainly differ in their attachment mode and degree of motility (for a detailed description, see Ribeiro et al., 2013): (1) the epipelon (EPL) comprises larger (usually>10 μm) motile diatoms, which can move freely in between sediment particles and typically form biofilms (cf. Herlory et al., 2004); (2) the epipsammon (EPM) groups smaller (usually<10 μm) diatoms, which live in close association with individual sand grains; and (3) the tychoplankton (TYCHO), which is an ill-defined and rather enigmatic group of largely non-motile diatoms, which presumably have an amphibious life style (both sediment and water column) (for example, Sabbe et al., 2010). Within the epipsammic group, non-motile (EPM-NM) species are firmly attached (either stalked or adnate) to sand particles, while motile forms (EPM-M) can move within the sphere of individual sand grains. From a phylogenetic perspective, motile forms (that is, all EPL and EPM-M) exclusively belong to the pennate raphid clade (Kooistra et al., 2007), possessing a raphe allowing motility. Most EPM-NM belongs to the pennate araphid lineage, but also includes some raphid pennates, such as Biremis lucens, which firmly attaches to sand grains (Sabbe et al., 1995). TYCHO includes both centric and pennate raphid forms. Intertidal benthic diatom species, but also growth forms, show distinct distribution patterns in time and space, suggesting pronounced (micro)niche differentiation (Sabbe, 1993; Méléder et al., 2007, Ribeiro et al., 2013). For example, EPM dominates non-cohesive sandy sediments (Méléder et al., 2007), while EPL dominates cohesive muddy sediments (Haubois et al., 2005). EPL typically display vertical ‘micromigration’ in the sediment following endogenous tidal/dial rhythms and environmental stimuli (Saburova and Polikarpov, 2003; Consalvey et al., 2004; Coelho et al., 2011): during daylight emersion, they migrate to the sediment surface, while during immersion they migrate to deeper sediment layers.

To prevent photoinhibition (Serôdio et al., 2008), benthic diatoms utilize behavioural and physiological responses (Mouget et al., 2008; van Leeuwe et al., 2009; Perkins et al., 2010b; Cartaxana et al., 2011; Serôdio et al., 2012). Behavioural photoprotection involves motility, allowing cells to position themselves in light gradients and escape from prolonged exposure to excess light (Admiraal, 1984; Kromkamp et al., 1998; Consalvey et al., 2004; Serôdio et al., 2006). In addition, both motile and non-motile species use fast regulatory physiological processes for photoprotection (that is, ‘physiological photoprotection’; Lavaud, 2007; Goss and Jakob, 2010; Depauw et al., 2012; Lepetit et al., 2012). In diatoms, two processes are important in field situations (Lavaud, 2007): photosystem II cyclic electron transfer (PSII CET) and NPQ of chlorophyll (Chl) fluorescence (Depauw et al., 2012; Lepetit et al., 2012; Lavaud and Lepetit, 2013). NPQ is controlled by several regulatory partners including the light-dependent conversion of diadinoxanthin (DD) to diatoxanthin (DT) by the DD de-epoxidase (that is, the XC) (Brunet and Lavaud, 2010; Goss and Jakob, 2010). In benthic diatoms, however, XC-NPQ has only rarely been studied, and mostly in situ: it has been shown to vary with diurnal and tidal cycles, season, latitude (Serôdio et al., 2005; van Leeuwe et al., 2009; Chevalier et al., 2010), the organisms’ position within the sediments and along the intertidal elevation gradient (Jesus et al., 2009; Cartaxana et al., 2011). On the basis of their in situ measurements, the latter authors hypothesized the existence of a trade-off between behavioural and physiological photoprotection mechanisms in benthic diatoms, as a stronger XC was shown to occur in sandy vs muddy sediments. However, at least the sandy sediments contained a mix of both epipsammic and epipelic forms (Jesus et al., 2009; Cartaxana et al., 2011), and even when the latter are not numerically dominant, they can still make a substantial contribution to biomass because of their much larger biovolumes (see for example, Hamels et al. 1998).

Our study represents a comprehensive characterization of fast regulatory physiological photoprotection capacity in typical representatives of the major diatom growth forms occurring in intertidal marine sediments. Given the highly dynamic and often extreme intertidal light climate, we hypothesize that photoprotective features are key traits shaping niche differentiation between benthic growth forms, as has been proposed before for phytoplankton (Huisman et al., 2001; Litchman and Klausmeier, 2008; Dimier et al., 2009; Petrou et al., 2011; Lavaud and Lepetit, 2013). In this respect, we predict that the largely immotile epipsammic life forms are better able to cope with pronounced and rapid changes in light intensity at the physiological level than the motile epipelic forms, which can actively position themselves in the sediment light gradient.

Materials and methods

Diatom culturing and harvesting

Fifteen benthic diatom strains were used (Table 1). All species were assigned to their respective growth form on the basis of microscopical observations on natural assemblages. They were grown in batch cultures at 20 °C in sterile artificial F/2 seawater medium enriched with NaHCO3 (80 mg l−1 final concentration). TYCHO species were also grown in continuously flushed airlift (that is, with air bubbling) to mimic ‘planktonic’ growth conditions. Two light intensities (E, 20 and 75 μmol photons m−2 s−1) were used with a 16-h light:8-h dark photoperiod white fluorescent tubes, L58W/840, OSRAM, Munich, Germany. Cultures were photoacclimated to the above conditions at least 2 weeks before measurements and experiments (see below). Diatom suspensions for the experiments were prepared to a final concentration of 10 μg Chl a ml−1. For this purpose, Chl a concentration was determined according to the Jeffrey and Humphrey (1975) spectrophotometric method. Diatom suspensions were continuously stirred at 20 °C under the growth E (that is, 20 or 75 μmol photons m−2 s−1) at least 1 h before the start of the experiments and all along the course of the experiments (Lavaud et al., 2007). This kept the photosynthetic machinery in an oxidized state and prevented NPQ.

Table 1 List of the 15 diatom species used in this study with their growth form classification, collection number, origin and average biovolume
Full size table

Growth rates and biovolumes

Specific growth rates, μ (day−1), were calculated from regression of the natural logarithm of the number of diatom cells during their exponential growth phase as microscopically determined in a Malassez’s counting chamber. Biovolumes (μm3) were calculated using the formula of Hillebrand et al. (1999) based on measurements performed on 15 specimens per species.

High-performance liquid chromatography pigment analyses

Chl a, Chl c, fucoxanthin, DD, DT and β-carotene content, all normalized to Chl a (that is, expressed as mol. 100 mol Chl a−1), were measured using high-performance liquid chromatography as described in Jakob et al. (1999). One millilitre of diatom suspension was rapidly filtered (Isopore 1.2 μm RTTP filters, Merck Millipore, Darmstadt, Germany) and immediately frozen in liquid nitrogen before extraction in a cold (4 °C) mixture of 90% methanol/0.2 M ammonium acetate (90/10 vol/vol) and 10% ethyl acetate. The pigment extraction was improved by the use of glass beads (diameter 0.25–0.5 mm, Roth, Karlsruhe, Germany) and included several short (20 s) vortexing steps. Supernatants were collected after centrifugation (5 min, 10 000 g, 4 °C) and immediately injected into an high-performance liquid chromatography system (Hitachi Lachrom Elite, Tokyo, Japan) equipped with a cooled auto-sampler and a photodiode array detector (L-2455). Chromatographic separation was carried out using a Nucleosil 120-5 C18 column (125 mm long, 4 mm internal diameter, 5 μm particles, Macherey-Nagel, Düren, Germany) equipped with a pre-column (CC 8/4 Nucleosil, Macherey-Nagel) for reverse phase chromatography during a 25-min elution program. The solvent gradient followed Jakob et al. (1999) with an injection volume of 50 μl and a flow rate of 1.5 ml min−1. Pigments were identified from absorbance spectra (400–800 nm) and retention times (Roy et al., 2011), and their concentrations were obtained from the signals in the photodiode array detector at 440 nm. The de-epoxidation state (DES in %) was calculated as ((DT/DD+DT) × 100), where DD is the epoxidized form and DT is the de-epoxidized form. Chl a concentration per cell was determined during exponential growth based on cell counts (see above) and the Chl a measurements.

Chl fluorescence yield and light curves

For a complete overview of the definition and measurement of the photophysiological parameters, see Table 2. Chl fluorescence yield was monitored with a Diving-PAM fluorometer (Walz, Effeltrich, Germany) on a 2.5 ml stirred and 20 °C controlled diatom suspension (Lavaud et al., 2004). Before measurement, the cells were dark-adapted for 15 min, and a saturating pulse (3600 μmol photons m−2 s−1, duration 0.4 ms) was fired to measure F0, Fm and Fv/Fm. Two types of light curves were performed: non-sequential and rapid light curves (Perkins et al., 2010a). For non-sequential light curves, continuous light (KL-2500 lamp, Schott, Mainz, Germany) was applied for 5 min at different Es (48–1950 μmol photons m−2 s−1); a new diatom suspension was used for each E. At the end of each exposure, Fm’ and NPQ were measured. For rapid light curves, one diatom suspension was exposed to eight successive, incrementally increasing Es (29–1042 μmol photons m−2 s−1) of 30-s each (Perkins et al., 2006; Supplementary Table S1). Rapid light curves allow constructing relative electron transport rate (rETR) vs E and NPQ vs E curves. The NPQ vs E curve is based on a 3-parameter Hill equation model and it is described by the equation NPQ (E)=NPQm × [EnNPQ/(E50NPQ nNPQ+EnNPQ’)] (Serôdio and Lavaud, 2011). From the fitted rETR-E curves (Eilers and Peeters, 1988) and NPQ-E curves (Serôdio and Lavaud, 2011), rETRm, α, Ek, and NPQm, E50NPQ, nNPQ can be derived, respectively. All parameters are described in the Table 2. nNPQ is the Hill coefficient or the sigmoidicity coefficient of the NPQ-E curve (Serôdio and Lavaud, 2011). It informs on the onset of NPQ at moderate Es, that is, when the DT molecules are being ‘activated’ with increasing Es to effectively participate to NPQ: DT ‘activation’ depends on its enzymatic conversion and its binding to the PSII light-harvesting antenna complex in order to promote the antenna switch to a dissipative state of excess energy, which is measurable by NPQ (see Lavaud and Lepetit, 2013). When nNPQ is<1, the NPQ-E curve shows an asymptotic saturation-like increase toward NPQm, while when nNPQ is >1, the NPQ-E curve shows a sigmoidal shape. In the later case, the Hill reaction (that is, NPQ onset) is allosteric (as proposed for the NPQ mechanism, see Lavaud and Lepetit, 2013), nNPQ thus informing on the degree of allostery of the NPQ-E curve. The higher nNPQ, the more positively cooperative the Hill reaction is; nNPQ around 2 being the highest values reported so far (Serôdio and Lavaud, 2011). The same fitting procedure can obviously be used for the DT-E and the DES-E curves, thereby extracting analogous parameters as from the fitted NPQ-E curves.

Table 2 Photophysiological parameters used in this study, their photophysiological meaning and measurement method and conditions
Full size table

O 2 yield and the PSII CET

The relative O2 yield produced during a sequence of single-turnover saturating flashes at a frequency of 2 Hz was measured with a home-made rate electrode (Lavaud et al., 2002). The steady-state O2 yield per flash (YSS) was attained for the last four flashes of a sequence of 20 when the S-state cycle oscillations were fully damped (Lavaud et al., 2002). YSS of 15-min dark-adapted (D) and illuminated (L, samples taken at the end of each non-sequential light curve) cells was used to calculate the PSII CET (Lavaud et al., 2002, 2007) as follows: [{(20 × YSS L)−(Σ (Y1…20)L)}−{(20 × YSS D)−( Σ (Y1…20)D)}]/YSS D.

Statistics

Statistical analyses were conducted using the statistical software package SAS 9.3 (Cary, NC, USA). Species were compared using the general linear model PROC GLM. Growth forms (groups) were compared using the mixed linear model PROC MIXED. Groups were regarded as fixed effects. Data were log transformed or square root transformed when needed to allow the best possible fit. Where necessary, estimated least squares means (lsmeans) and standard errors (s.e.) were back transformed as in Jørgensen and Pedersen (1998).

Results

Growth rate and photosynthetic properties

The Chl a concentration per cell showed an exponential relationship with biovolume with relatively small changes at the smaller cell volumes (Supplementary Figure S1). The average diatom biovolumes were independent of growth form (Table 3, Supplementary Figure S1). Growth rate did not differ significantly between the growth forms at growth E=20 μmol photons m−2 s−1 (Table 3, Supplementary Table S2). Relative concentrations of the light-harvesting pigments Chl c and fucoxanthin were comparable among growth forms (Table 3, Supplementary Table S2). β-Carotene, which is mainly associated with the photosystem cores, was only slightly but significantly higher in EPL than in EPM-NM. DD+DT content was significantly lower in EPL than in the other growth forms. As the cells were grown at low E, DES was generally low, with no significant differences between the growth forms (Table 3, Supplementary Table S2). The highest DD+DT (16.95±2.56 mol 100 mol Chl a−1) and DES (16.4±6.2%) values were observed in Plagiogramma staurophorum (EPM-NM) (Supplementary Table S3). There were no significant differences in Fv/Fm, α, rETRm, Ek and PSII CETmax between the growth forms. Ek was on average three to four times the growth E in all growth forms (Table 3, Supplementary Table S2). PSII CETm was close to 3 (its maximum, Lavaud et al., 2002) for the two EPM growth forms, and about 2 in EPL and TYCHO (Table 3).

Table 3 Growth rate, pigment content and photosynthetic properties of the four growth forms of benthic diatoms
Full size table

NPQ properties

At E values ⩾230 μmol photons m−2 s−1, NPQ was significantly higher in EPM-NM than in both EPL and TYCHO; the same holds true for EPM-M vs EPL and TYCHO at E values ⩾1050 μmol photons m−2 s−1 (Figure 1, Supplementary Table S4). NPQ was also significantly higher in EPM-NM than in EPM-M except at the lowest and highest E values. Similarly, NPQm was significantly higher ( × 3.5 and × 2.4, respectively) in EPM-NM and EPM-M than in EPL and TYCHO (Figure 1, Table 4 and Supplementary Tables S5 and S6). In EPL and TYCHO, the NPQ-E curves showed a lower variability than in the two EPM growth forms (Supplementary Figure S2). EPM-NM had the lowest E50NPQ, significantly lower than all other groups (Table 4, Supplementary Tables S5 and S6). In contrast, TYCHO E50NPQ was significantly higher than in the other groups. EPL and EPM-M E50NPQ did not differ significantly from each other. In contrast, nNPQ was not significantly different and varied around its optimum (that is, 2, Serôdio and Lavaud, 2011) in most species except the TYCHO ones (which is significantly lower than in EPM-NM) (Table 4, Supplementary Tables S5 and S6).

Figure 1
figure 1

NPQ of Chl fluorescence (a), DES of the DD to DT (DES=DT/(DD+DT) × 100) (b) and DT content (c) as a function of light intensity (E from darkness to 1950 μmol photons m−2 s−1, which is equivalent to full sunlight in the field) measured during non-sequential light curves (NSLCs) in the four benthic diatom growth forms (EPL, epipelon; EPM-M, epipsammon motile; EPM-NM, epipsammon non-motile; TYCHO, tychoplankton). Cells were grown at 20 μmol photons m−2 s−1. The NPQ-E curves for the individual species can be found in Supplementary Figure S3. Values are estimated least squares means±estimated s.e. (PROC MIXED procedure).

Full size image
Table 4 NPQ and XC properties of the four growth forms of benthic diatoms
Full size table

XC properties

DES was only significantly different between EPL and both TYCHO and EPM-M at 105 μmol photons m−2 s−1 and between EPL and both EPM forms at 230 μmol photons m−2 s−1 (Figure 1, Supplementary Table S7). DESm varied between 21.2±3.4 for EPL, 22.7±4.4 for TYCHO, 28.7±4.4 for EPM-M and 29.4±3.8 for EPM-NM (lsmeans±s.e.) (Figure 1, Table 4 and Supplementary Tables S5 and S6). The slight difference between EPL and the EPM growth forms, although not significant, in combination with the significantly higher DD+DT in the latter, generated a significantly lower DTm in EPL than in the EPM growth forms (Tables 3 and 4, Supplementary Tables S2 and S6). E50DT was close to the E50NPQ in all growth forms except in TYCHO where it was lower; no significant differences between the EPM and EPL were observed, only EPM-NM and TYCHO E50DT differed significantly (Table 4, Supplementary Tables S5 and S6). nDT was significantly lower in EPM-M and TYCHO than in EPL and EPM-NM. NPQ/DT was about half its optimum (=1 under these experimental conditions) in all groups except EPM-NM (Table 4, Supplementary Tables S5 and S6). It roughly followed the same order as observed for NPQm, that is, EPM-NM >EPM-M>EPL≅TYCHO, with a 2 × higher value in EPM-NM (Table 4). The difference between EPM-NM and the other growth forms, however, was not significant because of the low NPQ/DT value in Plagiogramma staurophorum (Supplementary Table S5). Figure 2 shows that in all growth forms except EPM-M there were species (Seminavis robusta, Fragilaria. cf. subsalina, P. staurophorum, Brockmanniella brockmannii) for which a low NPQ developed without DT synthesis, while two EPM-M species (Amphora sp. and Planothidium delicatulum) showed DT synthesis (0.17±0.03 mol 100 mol Chl a−1) without NPQ (see also Supplementary Table S5). All other species showed a NPQ/DT relationship with an origin close to 0, as expected.

Figure 2
figure 2

NPQ of Chl fluorescence as a function of the amount of DT measured during non-sequential light curves (NSLCs) in the five species of epipelon (EPL) (a), the four species of motile epipsammon (EPM-NM) (b), the three species of tychoplankton (TYCHO) (c) and the three species of non-motile epipsammon (EPM-NM) (d). Cells were grown at 20 μmol photons m−2 s−1. The full names and classification of all species is listed in Table 1.

Full size image

Effect of high light acclimation on the NPQ and XC properties

All species were grown under an E (75 μmol photons m−2 s−1) roughly corresponding to the mean Ek for the low E acclimated cells (20 μmol photons m−2 s−1, Table 3); (Figure 3, Supplementary Table S8). Only EPL had significantly higher growth rates at 75 μmol photons m−2 s−1. DD+DT significantly increased with a factor 1.6–1.7 in EPL and EPM, and 2.3 in TYCHO. There was a significant increase in DES at 75 μmol photons m−2 s−1 in all growth forms except in EPM-M. The increase in DD+DT and DES at the higher light intensity was most pronounced in TYCHO and resulted in a pronounced, significant difference in both parameters between TYCHO and EPL at this light intensity. The comparison of Chl fluorescence yield and light curve parameters could only be performed for a selection of six species (covering all growth forms) and is summarized in Figure 4 (see also Supplementary Table S9). As expected, the Chl a content per cell decreased, roughly with a factor of 2 in all species (except Navicula phyllepta). There was only a slight (up to about 10%) decrease in Fv/Fm in all species, illustrating the unstressed state of the cells (note that in Seminavis robusta and Planothidium delicatulum this decrease was slightly significant). DESm significantly increased in S. robusta only. Together with the overall increase in DD+DT, this resulted in a significant increase in DTm (by a factor of 4) in this species, but also in P. delicatulum and Plagiogrammopsis vanheurckii. The corresponding NPQm did not follow the same trend: it significantly increased in all species (except for P. delicatulum and Opephora sp.) but only by a factor of maximally 2. NPQ/DT remained low (0.2 to 0.5) in all species (and significantly decreased in Opephora sp.). E50NPQ was significantly higher only in the EPM-NM species Plagiogramma staurophorum.

Figure 3
figure 3

Growth rate (μ) (a), DD+DT content (b) and DES of DD to DT (DES=(DT/DD+DT × 100) (c) in the four benthic diatom growth forms (EPL, epipelon, EPM-M; epipsammon motile; EPM-NM, epipsammon non-motile; TYCHO, tychoplankton) for cells grown at light intensities of 20 and 75 μmol photons m−2 s−1, respectively. All parameters were measured on cells in exponential growth and sampled 2 h after the onset of light; growth conditions were 16-h light:8-h dark, 20 °C. The values for all species in 20 and 75 μmol photons m−2 s−1 conditions are found in Supplementary Tables S3 and S8, respectively. Values are estimated least squares means±estimated s.e. (PROC MIXED procedure).

Full size image
Figure 4
figure 4

Comparison of photosynthetic, NPQ of Chl fluorescence and XC parameters measured in diatom species representative of the four benthic diatom growth forms grown at light intensities of 20 and 75 μmol photons m−2 s−1, respectively. For each parameter, the ratio of the values obtained at 75 and 20 μmol photons m−2 s−1−1 was calculated (that is, the 0 line is equal to a 75/20 ratio=1, which is equivalent to no change of values between light intensities). Significant changes between both light intensities are indicated with an asterisk. The values used for the 20 and the 75 μmol photons m−2 s−1 conditions can be found in Supplementary Tables S3/S5 and S8/S9, respectively.

Full size image

Effect of ‘planktonic’ growth on the NPQ and XC properties of tychoplankton

The three TYCHO species were grown under ‘planktonic’ conditions (at 20 μmol photons m−2 s−1) for a comparison with growth under ‘benthic’ conditions (Figure 5, Supplementary Table S10). Brockmaniella brockmannii responded most strongly to a switch from ‘benthic’ to ‘planktonic’ growth: it showed a significantly lower growth rate and a higher DES and DESm but a lower NPQm, suggesting photosynthetic stress and investment of additional DT in other processes than NPQ. Plagiogrammopsis vanheurckii and Cylindrotheca closterium showed very little change, apart from a significantly higher growth rate during planktonic growth in P. vanheurckii, a slight decrease in NPQ/DT in C. closterium, and an increase in DES in both species. The most pronounced and consistent change in TYCHO thus concerned an increase in DES when grown in suspension. Note that there is also an overall decrease in rETRm, but this decrease was just not significant (P=0.08).

Figure 5
figure 5

Comparison of growth, photosynthetic, pigment, NPQ of Chl fluorescence and XC parameters measured in the three tychoplankton (TYCHO) diatom species in ‘benthic’ and ‘planktonic’ conditions. For each parameter, the ratio of the values obtained under benthic and planktonic conditions−1 was calculated (that is, the 0 line is equal to a planktonic/benthic ratio=1, which is equivalent to no change of values between ‘benthic’ and ‘planktonic’ conditions). Chl a per cell (in pg cell−1) and growth rates (in day−1) were measured on cells in exponential growth phase sampled 2 h after the onset of light; growth conditions were 20 μmol photons m−2 s−1, 16-h light:8-h dark, 20 °C. Significant changes between both light intensities are indicated with an asterisk. The values used for the ‘benthic’ and ‘planktonic’ growth conditions can be found in Supplementary Tables S3/S5 and S10, respectively.

Full size image

Discussion

The present work constitutes the first comparative experimental study, using unialgal cultures in standardized conditions, of fast regulatory photoprotective mechanisms in the four main benthic diatom growth forms present in intertidal marine sediments (EPL, EPM-M and EPM-NM and TYCHO). As no sediment was added in our experiments, motile diatoms were not able to position themselves in a light gradient, hence effectively incapacitating their behavioural response. As the growth rate and photosynthetic characteristics (main pigments, Fv/Fm, α, Ek, rETRm) of the studied species were comparable between the growth forms at 20 μmol photons m−2 s−1, we were able to compare their purely physiological light response.

Our study revealed a highly significant and pronounced difference in NPQ between the four growth forms. NPQ was significantly lower in EPL and TYCHO than in EPM species; differences in DES were only observed between EPL and other forms at lower light intensities. Within the EPM, NPQ capacity was significantly higher in the non-motile than in the motile forms. As all growth forms included both small and large species, the functional light response (NPQ capacity) apparently did not depend on biovolume or the Chl a concentration per cell, as has also been observed in situ (Jesus et al., 2009). The absence of significant differences in PSII CET between growth forms underscores the importance of NPQ as the main fast photoprotective process in intertidal benthic diatoms, confirming earlier results for these organisms (Lavaud et al., 2002) but in contrast with planktonic diatoms (Lavaud et al., 2002, 2007). By analogy with previous studies on planktonic diatoms (Strzepek and Harrison, 2004; Lavaud et al., 2007; Dimier et al., 2009; Petrou et al., 2011; Lavaud and Lepetit, 2013), our data suggest that EPL and TYCHO diatoms are adapted to a less fluctuating light climate and/or to a lower average irradiance, and vice versa for EPM diatoms. This result fits well with the ecology of these growth forms. EPL is not only more abundant in muddy cohesive sediments where light penetration is more restricted than in sandy sediments (Paterson and Hagerthey 2001; Cartaxana et al., 2011), but, more importantly, their (micro-) migratory behaviour allows positioning at the optimal irradiance in the vertical light gradient and rapid escape from periodic excess light (Kromkamp et al., 1998; Conn et al., 2004; Consalvey et al., 2004; Serôdio et al., 2006). This alleviates the need to invest in a strong physiological capacity to respond to light stress as previously proposed (Jesus et al., 2009; Cartaxana et al., 2011), although the right balance between motility and physiology still remains essential (van Leeuwe et al., 2009; Perkins et al., 2010b; Cartaxana et al., 2011; Serôdio et al., 2012).

Such balance is more crucial in the EPM-M species, which can move but have only limited control over their immediate light environment as movement is restricted, usually within the sphere of individual sand grains. As expected, they showed a significantly lower NPQ and a higher E50NPQ than EPM-NM, which have no behavioural control over their light environment. An alternative, but not exclusive, explanation could be related to the difference in exopolysaccharide (EPS) secretion between motile and non-motile growth forms. EPS secretion could work as an alternative electron sink under stressful conditions (that is, high light, nutrient limitation etc.) in order to limit the over-reduction of the photosynthetic machinery (‘overflow’ hypothesis; Staats et al., 2000), alleviating the need for a strong NPQ. However, EPS secretion is not as fast as NPQ (minutes/hours vs seconds/minutes) and may not be useful to the cells for responding to rapid light changes but only to cope with prolonged high light exposure. In addition, while the ‘overflow’ hypothesis is often proposed (Underwood and Paterson, 2003; Stal, 2009), it was never clearly proven. A few studies have shown a positive relationship between light intensity and EPS production (Underwood, 2002; Wolfstein and Stal, 2002) but other studies have reported a negative relation with light intensity and no relationship with nutrient limitation (Perkins et al., 2001; Hanlon et al., 2006). To date there is no information on EPS production in different benthic diatom growth forms, and only EPL species have been compared (Underwood and Paterson, 2003), showing no clear relationship between light response and EPS secretion. To our knowledge, there are no reports on a relationship between NPQ-XC capacity and EPS production. Finally, TYCHO typically alternates between resuspension in a highly turbid shallow water column at high tide and deposition and burial in the upper sediment layers of muddy sediments at low tide (deposition in sandy sediments does not occur because of the intense hydrodynamic disturbance in these sediments). As such, the TYCHO resembles planktonic diatoms adapted to subtle light fluctuations and/or on average low irradiance (Bailleul et al., 2010; Lavaud and Lepetit, 2013).

The reason for the NPQ differences between EPL and EPM can be explained by its main control: the XC dynamics. Previous in situ studies reported a consistently stronger DES under light stress in EPM than in EPL diatom communities (that is, in sandy vs muddy sediments) and related growth form with differential (behavioural vs physiological) photoregulatory strategies (Jesus et al., 2009; Cartaxana et al., 2011). As recently shown, a high NPQ is supported by the strong effective involvement of DT, which first depends both on a high DD+DT content and a high DES (Lavaud and Lepetit, 2013). The slope of the NPQ/DT relationship has been proposed as a good indicator of light climate adaptation: the higher the NPQ/DT slope, the better the adaptation to a highly fluctuating and/or on average high light climate (Dimier et al., 2009; Lavaud and Lepetit, 2013). All EPM species, and especially the non-motile ones, showed XC parameter values, which are characteristic for a high NPQ capacity, viz. a higher DD+DT content and DTm, which was 2 × higher than in EPL. EPM-NM also tended to show a higher efficiency in promoting NPQ (NPQ/DT), but this difference was not significant because of high intra-group variability.

Within the EPM, NPQ is clearly more efficient in EPM-NM than EPM-M species. In EPM-M, the discrepancy between E50NPQ and Ek is more pronounced than in non-motile forms: while there is no significant difference in Ek between both growth forms, E50NPQ is significantly higher in the motile growth forms. This suggests a weaker relationship between NPQ development and photochemistry in the latter group, with slower NPQ development with increasing E. Remarkably, E50DT does not significantly differ between both growth forms, and the significantly higher initial induction of DT synthesis (nDT) but not NPQ (nNPQ) in the motile group, together with the fact that some representatives of this group show DT synthesis without NPQ, suggests that either DT is less or not involved in NPQ development, or that the light-dependent built-up of the transthylakoidal proton gradient (which is involved in both the activation of the DD de-epoxidase and the molecular control of NPQ) and the onset of NPQ are uncoupled (Lavaud et al., 2012; Lavaud and Lepetit, 2013). Our observations thus suggest that in contrast to the non-motile group, EPM-M species rely more on a behavioural response (motility) and/or involve DT in other photoprotective processes such as the prevention of lipid peroxidation by reactive oxygen species (Lepetit et al., 2010). The increase in E50NPQ in the EPM-NM species Plagiogramma staurophorum during a shift to higher light illustrates the ability to physiologically modulate the NPQ vs E development kinetics to its light environment in contrast to EPM-M, EPL and TYCHO.

The influence of DT on the inter-group/species NPQ differences was further investigated by the acclimation to higher light (75 μmol photons m−2 s−1, close to the mean Ek for cells acclimated to 20 μmol photons m−2 s−1). High light exposure is known to induce constitutive DT synthesis (Schumann et al., 2007) and in field conditions, DT is usually even present in significant amounts in cells adapted to low/moderate light (Jesus et al., 2009; van Leeuwe et al., 2009; Chevalier et al., 2010; Cartaxana et al., 2011). Acclimation to higher light resulted in a significant increase in XC pigments (DD+DT) and DES in most growth forms, suggesting that although EPL uses behavioural photoprotection, the XC is still important (cf. above). NPQm increased in most of the species examined, mainly because of a higher DTm resulting from a higher DD+DT rather than a higher DESm. The discrepancy between DESm and NPQm as well as the low NPQ/DT may be due to the fact that the additional DT primarily served in the prevention of lipid peroxidation rather than in NPQ as previously reported in high light acclimated diatoms (see also above).

Although under low light conditions, the growth, photosynthetic and steady-state light-response features of TYCHO were similar to those of EPL diatoms (that is, low NPQ, NPQm and DTm), their dynamic light response was significantly different, that is, higher E50NPQ. Surprisingly, E50NPQ was beyond the natural light maximum (2000–2500 μmol photons m−2 s−1) illustrating the inability of TYCHO to strongly and/or continuously develop NPQ in the environmental high light range (a situation also encountered in one epipelic species: Navicula phyllepta). In contrast, its low nNPQ supported a relatively strong onset of NPQ at low Es. Both E50DT and nDT were correspondingly high and low, respectively (and significantly different from EPL for nDT), although E50DT was much lower than E50NPQ suggesting a discrepancy between DT synthesis and NPQ development (cf. above). The response of TYCHO to higher light was much more pronounced, with the strongest increase in XC pigments and DES of all growth forms. However, the NPQm and DTm data (only available, however, for one representative species, Plagiogrammopsis vanheurckii) did not show a similar response, with DTm showing a more pronounced increase than NPQm, suggesting that NPQ development was low and that DT may have mainly been involved in other processes than NPQ. For most parameters, the response of the TYCHO species to growth in suspension (‘planktonic’ growth) was limited and largely species specific, except for a general increase in DES and a decrease (albeit just non-significant) in rETRm. These data suggest that representatives of the TYCHO growth form are well adapted to their amphibious life style, which is characterized by an on average low irradiance (MacIntyre et al., 1996). In contrast, EPL species do not grow well in suspended, turbulent conditions (J Lavaud, personal observation).

Our study for the first time shows that intertidal benthic diatoms display growth form-specific variation in fast regulatory physiological mechanisms for photoprotective capacity (NPQ and the XC), which mirrors their behavioural light response. In EPL motile diatoms, exclusively belonging to the raphid pennate clade, the physiological response is not well developed, as these diatoms appear to largely rely on motility to control their immediate light environment. In the EPM-M, however, the physiological response remains essential because their movement is restricted to the sphere of individual sand grains. The evolution of the raphe system, the hallmark synapomorphy of the raphid pennate diatom clade, which enables locomotion, has therefore been essential for the colonization of intertidal sediments by not only migratory epipelic biofilms but also EPM-M. In contrast, NPQ and XC capacity is high in non-motile araphid pennate diatoms, which passively have to abide often pronounced variations in the intertidal light climate. TYCHO diatoms, which alternate between high tide resuspension in a turbulent and turbid water column, and low tide deposition in muddy sediments, appear to be adapted to an on average low light environment, with low NPQ and XC capacity.

Although we made no formal analysis of the relationship between functional and phylogenetic diversity, it is obvious that despite the fact that a behavioural photoprotective response (motility) is restricted to the raphid pennate diatom clade, differences in the studied physiological traits are more strongly driven by growth form than phylogenetic relatedness. For example, the EPM species Biremis lucens, despite being a raphid pennate species, has a non-motile growth form, and shows a NPQ capacity, which is more similar to EPM-NM than to the (phylogenetically more closely related) EPM-M and EPL. Similarly, photophysiological features of pennate raphid (Cylindrotheca closterium) and centric (Plagiogrammopsis vanheurckii and Brockmanniella brockmannii) TYCHO species were similar as reported before in planktonic centric/pennate species (Lavaud et al., 2004). Raphid pennate diatoms, which have colonized an epipsammic or tychoplanktonic niche, thus display a reverse evolutionary trade-off switch toward a much more performant physiological response. Our observations thus suggest that photoprotective capacity in diatoms is a highly adaptive trait, which is to a certain degree constrained by clade-specific evolutionary innovations (the evolution of the raphe system and hence a behavioural response) but also, and more importantly, by growth form, which ultimately defines the balance between the physiological and behavioural photoprotective response in these organisms. Such differential adaptation is of primary importance for the regulation of the photosynthetic productivity vs light, as has been demonstrated before in planktonic diatoms, where the photochemical vs the photoprotective energy allocation as a function of light is drastically different in species adapted to a fluctuating vs a more stable light environment (Wagner et al., 2006; Lavaud et al., 2007; Petrou et al., 2011; Lavaud and Lepetit, 2013). However, unlike in planktonic environments, the trade-off between a physiological and behavioural response in benthic diatoms allows local co-existence of different growth forms under the same overall light environment.