Introduction

Coral bleaching is a globally important phenomenon, which has a devastating impact upon (coral) reef community structure and function. While coral bleaching, which is broadly defined as a loss of coral colour through the loss of symbionts and/or pigments (i.e. Fitt et al. 2001; Smith et al. 2005), has been the subject of intense research in recent years, the mechanisms that underlie bleaching susceptibility remain largely unknown. Consequently, it is currently unknown which corals within a community are the most susceptible to future environmental stressors. Some work suggests that massive scleractinian corals are relatively tolerant of environmental disturbance compared with many non-massive species (Marshall and Baird 2000; Loya et al. 2001; West and Salm 2003; Kenyon et al. 2006). However, this idea is largely developed from Pacific coral species and this pattern is not ubiquitous across all species (see Obura 2001), as considerable variability in bleaching tolerance exists within morphological types. Given the importance of massive and submassive coral species in the long-term stability of coral reef systems, particularly in terms of reef resilience (Hoegh-Guldberg and Salvat 1995; Loya et al. 2001; Kenyon et al. 2006), they represent an important group to study. Additionally, emerging evidence highlights the importance of understanding determinants of bleaching susceptibility in a range of coral species of all morphologies, as different bleaching characteristics can actually drive long-term coral community structure and function.

Many different mechanisms exist for coral bleaching, and both the host and the symbiotic microalgae that live within the host tissue (Symbiodinium spp.) can play a significant role (Fitt and Warner 1995; Warner et al. 1999; Fitt et al. 2001; Smith et al. 2005; Baird et al. 2009; Anthony et al. 2009). However, there is ample evidence that, in many cases, the responses of Symbiodinium spp. to bleaching stressors may act as a precursor to further effects exhibited throughout the host (Smith et al. 2005; Suggett et al. 2008; Weis 2008). One such precursor could be the increase in the production of reactive oxygen species (ROS) by the Symbiodinium, which can lead to a cascade of detrimental effects both within the host and the Symbiodinium (Lesser 1996, 1997; Smith et al. 2005; Suggett et al. 2008). Thermal- and light-induced stress can cause the production of these ROS by Symbiodinium as they become photoinhibited, which can damage several cellular targets, including photosystem II (PSII) (Warner et al. 1996, 1999; Ragni et al. 2010). The reaction centres within photosystem II (PSII), and in particular the D1 protein in PSII, are particularly susceptible to damage (Warner et al. 1999; Takahashi et al. 2004, 2009). Importantly, maintaining photosynthetic activity (and preventing or minimising net photoinhibition) is ultimately a careful balance between net rates of photoprotection, photodamage and photorepair (Ragni et al. 2010), which correlates with D1 content in cultured Symbiodinium and in intact corals (Warner et al. 1999; Hill et al. 2011).

There is substantial genetic diversity within the genus Symbiodinium (LaJeunesse 2001), as well as a broad array of strategies to adjusting to changing light intensity via photoacclimation. For in vitro Symbiodinium studied to date, the dominant pathway for photoacclimating to high light is a reduction in the number of photosynthetic units (Iglesias-Prieto and Trench 1997; Hennige et al. 2009). Many types of cultured Symbiodinium display different rates of photodamage and repair via chloroplast protein turnover (Warner et al. 1999; Takahashi et al. 2009; Ragni et al. 2010). This is important, since if the rate of photodamage occurs faster than photorepair, long-term photoinactivation and functional PSII reaction centre loss will result (Adir et al. 2003; Ragni et al. 2010). Photorepair is consequently extremely important for Symbiodinium and has been speculated to underlie differences in coral bleaching susceptibility (Warner et al. 1999; Takahashi et al. 2004, 2009; Ragni et al. 2010), but remains to be thoroughly tested in Symbiodinium while in hospite. Likewise, given recent evidence for the possibility of negligible D1 protein turnover in two isolates of cultured Symbiodinium (Takahashi et al. 2009) and the unique nature of the structure and transcription of the chloroplast genome in dinoflagellates (Howe et al. 2008; Dang and Green 2010), it would not be surprising if chloroplast protein repair pathways in Symbiodinium in general may deviate from model organisms such as green algae and cyanobacteria.

It is recognised that both high light and high temperature conditions can cause net photoinhibition (Jones and Hoegh-Guldberg 2001; Gorbunov et al. 2001; Ragni et al. 2010) and that both stressors can act cumulatively as well as independently. For shallow water corals, high light is a daily stressor at midday, whereas higher temperatures are usually seasonal, such that a certain percentage of PSII reaction centres remain inactive over a daily cycle, while this number may rise on a seasonal basis (Gorbunov et al. 2001; Warner et al. 2002). Consequently, in hospite Symbiodinium that show increased net photoinhibition under ‘normal’ conditions may be more susceptible to thermally induced coral bleaching. Conversely, down-regulation of PSII reaction centres may provide a means to quench excess light energy during enhanced stress. This study investigated the rate of photoinhibition and photorepair, and the efficiency of preventative mechanisms Symbiodinium can use to prevent net photoinhibition (such as non-photochemical quenching) in a coral that is historically sensitive to thermally induced bleaching (M. faveolata) and a coral that is known to be more tolerant to thermal stress (P. astreoides).

Materials and methods

Sample collection

Samples were collected near the Institute for Marine and Freshwater Science (UNAM) in Puerto Morelos, Mexico. Two coral species were used in this investigation: M. faveolata and P. astreoides. M. faveolata is a primary frame-work building massive coral in the Caribbean, while P. astreoides is represented by smaller encrusting to mounding colonies that do not contribute as largely to reef topography overall, but are a common species in the back-reef lagoon in this area. Although not considered a reef framework builder, P. astreoides is increasing in ecological dominance in some parts of the Caribbean (Green et al. 2008; Edmunds 2010). While both species are noted to bleach (Brandt 2009; DeSalvo et al. 2008), M. faveolata typically has a lower ‘bleaching threshold’ in terms of temperature and exposure time and displays a greater bleaching prevalence as compared to P. astreoides on the reef (Brandt 2009). Similarly, the endosymbionts within M. faveolata tend to show higher levels of PSII photoinactivation during initial experimental thermal treatment (Warner and Berry-Lowe 2006; Warner and Grottoli, personal observation).

In August 2009, parent colonies of both species were identified at 3–4 m depth (approximately 800 μmols photons m−2 s−1 recorded at peak midday irradiance with underwater light loggers, Odyssey, NZ) off the coast of Puerto Morelos, approximately 2 km North of the Institute of Marine Sciences and Limnology of the University of Mexico. A fragment ca. 10 cm2 was collected from 5 parent colonies and transported immediately back to the laboratory facility. Each colony sample was then split into seven smaller fragments. One fragment was immediately processed for Symbiodinium genetic analysis. The remaining six fragments from each colony (30 total of each species) were allowed to recover from sampling stress in a shaded holding tank for 5 days with flow through reef seawater at ambient temperature to aid any recovery from fragmentation stress. Water temperature matched that of the back-reef lagoon where the corals were collected (maximum 30.5 ± 0.01°C, with submersible temperature loggers, Onset USA), and the peak average light level was slightly lower relative to the collection site (589 ± 0.34 μmols photons m−2 s−1, recorded with a 2π sensor LI-COR, USA).

Chlorophyll a fluorometry

PSII activity was assessed via single-turnover chlorophyll fluorescence as outlined by Hennige et al. (2009) recorded from a fluorescence induction and relaxation (FIRe) fluorometer (Satlantic) fitted with a Satlantic fibre optic probe attachment. The probe was placed against the coral in the water to avoid air–water signal dispersion on a surface parallel to the water surface. Fluorescence transients comprised of a 2-step sequence that included a 100 μs saturating single-turnover (ST) excitation light pulse, followed by a 60 μs relaxation phase from a series of weaker light pulses used to follow the reoxidation of PSII. Fifteen iterations of this sequence were averaged into a single fluorescence transient to increase the signal-to-noise ratio for each sample. Fluorescence parameters were calculated in FIREPRO software (Satlantic Inc.), based on the biophysical (KPF) model of Kolber et al. (1998) with specific gain excitation profiles for the fibre optic probe to curve fit several physiological components of PSII activity under dark-acclimated conditions, including minimum fluorescence, maximum fluorescence and absorption cross section of PSII (\( F_{o} ,F_{m} \,{\text{and}}\,\sigma_{\text{PSII}} \), respectively, Table 1). Under actinic light, these terms are \( F_{o}{^\prime } ,F_{m}{^\prime } \,{\text{and}}\,\sigma_{\text{PSII}}{^\prime} \), respectively (Table 1; Kromkamp and Forster 2003; Baker and Oxborough 2005; Cosgrove and Borowitzka 2010). \( F_{o}{^\prime} \) is problematic to measure directly (Kromkamp and Forster 2003) and was calculated according to Suggett et al. (2003) as \( F_{o} /\left[ {\left( {F_{v} /F_{m} } \right) + (F_{o} /F_{m}{^\prime } )} \right] \).

Table 1 Table of terms, their definitions and their units
Full size table

Dark-acclimated fluorescence indicates the physiological ‘state’ when it is assumed the majority of photoprotective mechanisms are relaxed, PSII reaction centres are ‘open’ and the electron transport chain is oxidised (Kromkamp and Forster 2003; Ralph and Gademann 2005; Hennige et al. 2009). Fluorescence transients collected under actinic light exposure indicate the capacity of cells to utilise absorbed excitation energy (Ralph and Gademann 2005; Falkowski and Raven 1997). The maximum PSII photochemical efficiency (also referred to as the maximal quantum yield of PSII) was determined as \( (F_{m} - F_{o} )/F_{m} = F_{v} /F_{m} \) (dimensionless) following the terminology of Kromkamp and Forster (2003). Similarly, the effective photochemical efficiency of PSII photochemistry (or effective quantum yield) at any given actinic irradiance is determined as \( (F_{m}{^\prime } - F^{\prime } )/F_{m}{^\prime } ({\text{termed}}\,F_{q}{^\prime } /F_{m}{^\prime } ) \). The photochemical quenching parameter (qP), calculated as \( (F_{m}{^\prime } - F{^\prime } )/(F_{m}{^\prime } - F_{o}{^\prime } )\,{\text{or}}\,F_{q}{^\prime } /F_{v}{^\prime } \), can be used to describe the decrease in effective photochemical efficiency of PSII due to changes in \( F_{q}{^\prime} \) which decreases with the increased closure of PSII reaction centres (Hennige et al. 2008a). The quantum yield of non-photochemical quenching at any actinic irradiance is \( (F_{m} - F_{m}{^\prime} )/F_{m}{^\prime } \), termed as NPQ hereafter (Table 1). NPQ describes the quenching of maximum fluorescence, Fm, through many possible pathways such as xanthophyll cycling or inactive reaction centres (see Kromkamp and Forster 2003). The proportion of non-photochemical quenching specifically through the antennae was estimated by normalising the light-acclimated functional absorption cross section of PSII, \( \sigma_{\text{PSII}}{^\prime} \), to the dark-acclimated functional absorption cross section of PSII, as \( 1- (\sigma_{\text{PSII}}{^\prime} /\sigma_{\text{PSII}} ) \) (Gorbunov et al. 2001; Hennige et al. 2009).

Experimental treatment and sampling regime

At the start of the experiment, 2 fragments from 5 colonies of each species were shifted from the holding tanks to each of three light levels: low light (LL), medium light (ML) and high light (HL), which were 100, 200 and 589 μmols photons m−2 s−1, respectively, and placed in individual plastic beakers (400 ml). One fragment from each colony, species, and light level was assigned to the experimental treatment of chloramphenicol addition (+CAP), and one was assigned to the control treatment (no CAP). All light levels were controlled with neutral density screening to modify ambient irradiance. Within each light level, chlorophyll fluorescence measurements of control and experimental fragments were taken immediately preceding the addition of CAP, and then 1 and 2 h following the addition. As the efficiency of CAP decreases over time, experiments were terminated after 2 h when changes in efficiency were no longer detectable. All fluorescence measurements first were taken on light acclimated samples, i.e., samples that were in the light (LL, ML or HL). At the end of the 2-h experiment, samples were dark acclimated for 30 min (using a black shade cloth over the tank), before an additional fluorescence measurement was taken from all coral fragments. During the course of the 2-h experiment, natural variations in sun light intensity resulted in HL tanks having an approximate variation of ±50 μmols photons m−2 s−1, ML tanks ~20 μmols photons m−2 s−1 and LL tanks ~10 μmols photons m−2 s−1. The exposure time was chosen for since it represents the point of peak irradiance with relatively little variation. Potential photodamage is therefore high over this time period.

At time 0 (11:30 h), chloramphenicol (CAP) dissolved in 99% ethanol, which blocks de novo translation of chloroplast encoded proteins (including the psbA or D1 protein) (Warner et al. 1999), was added to the 5 treatment (+CAP) beakers at each light level to a final concentration of 386 μM (1.17 ml from a stock of 0.13 M). Previous work has confirmed that CAP additions result in the loss of D1 protein content in Symbiodinium in vitro (Warner et al. 1999) and in hospite (personal observation). The CAP concentration required to illicit a significant decline in photochemical activity was determined by preliminary experiments, and inclusion of 1.17 ml of 99% ethanol alone did not significantly impact the photosynthetic activity of either coral species (data not shown).

Calculation of gross and net photoinhibition

Values of \( F_{v} /F_{m} \) from CAP-treated and CAP-untreated samples at the end of 2 h of incubation were used to calculate net photoinhibition (NPiR, Eq. 1), gross photoinhibition (GPiR, Eq. 2) and repair rate (RR, Eq. 3) following equations from Ragni et al. (2008).

$$ NPiR(E) = - \frac{{Ln\frac{{F_{v} /F_{m} (E)}}{{F_{v} /F_{m} (DA)}}}}{\Updelta T} $$
(1)
$$ GPiR(E) = - \frac{{Ln\frac{{F_{v} /F_{mCAP} (E)}}{{F_{v} /F_{m} (DA)}}}}{\Updelta T} $$
(2)
$$ {\text{RR }} = GPiR \, - \, NPiR $$
(3)

where E is the light level that samples were subject to, ΔT is the duration in minutes of the light incubation period (120 min), Fv /Fm(DA) is the Fv /Fm measured on dark-acclimated samples at night, a few hours after sunset (to ensure full relaxation of NPQ), Fv /Fm(E) are the values of Fv /Fm measured at the end of light exposure at different Photon Flux Densities (PFD) levels (either LL, ML or HL) and Fv /FmCAP (E) is the same parameter measured on chloramphenicol-treated samples (Ragni et al. 2010). These equations assume that the change in Fv /Fm occurs at a constant specific rate (e.g., \( {\text{d}}\left( {F_{v} /F_{m} } \right)/{\text{d}}t = {\text{k}}F_{v} /F_{m} \)) (Six et al. 2008; Ragni et al. 2010). Two-sample T tests were used to compare fluorescence and photoinhibition (net and gross) results between both coral species in CAP-treated and CAP-untreated samples. Fluorescence parameters were compared with each other (to test for potential relationships) by using a Pearson’s Correlation test.

Light curve experiment

In the days following outdoor incubation experiments, control fragments from HL (fragments not exposed to CAP) of each coral species were exposed to a series of light levels using an independent light source to calculate light response curves. A halogen light source was (Schott KL1500 LCD) fitted with a DT-cyan filter (to provide irradiance from 400 to 600 nm) in conjunction with a Schott annular ring light to expose the coral fragments to actinic light. The FIRe fibre optic probe was fitted into the centre of the ring light and enabled fluorescence readings to be taken on an illumination ‘spot’ approximately 1 cm in diameter on the coral. Eleven light levels were used: 0, 40, 60, 100, 160, 240, 400, 660, 1,100, 1,600 and 2,200 μmols photons m−2 s−1, with each light step lasting 4 min. All light curves were carried out on dark-acclimated (30 min) samples at 08:00 h on sequential days to ensure that light curves accurately reported the kinetics of electron transport from samples wherein the majority of PSII was in the oxidised state, and was not influenced by ‘midday depression’ (Lesser and Gorbunov 2001; Warner et al. 2002), as dark acclimation at midday may not allow full reversal of NPQ (per obs).

The electron transport rate (ETR) specific to the functional absorption cross section and PSII reaction centres (RC) (ETRRCII, mol e− mol RCII−1 h−1, Table 1) was calculated as in Ragni et al. 2010 (Eq. 4):

$$ {\text{ETR}}^{\text{RCII}} = {\text{ PFD}} \times \sigma_{\text{PSII}} \times F_{q}{^\prime } /F_{m}{^\prime } \times 2 1. 6 8 3 $$
(4)

where 21.683 converts seconds to hours, μmol e− mol e− and Å2 quanta−1 to m2 mol RCII−1 (Suggett et al. 2003, 2006). \( F_{q}{^\prime } /F_{m}{^\prime } \) was also used to calculate the light saturation coefficient of electron transport, Ek (μmols photons m−2 s−1) for both coral species following methodology from Hennige et al. (2008a) using least-squares non-linear regression. Ek, which is also known as the minimum saturating irradiance (Hill et al. 2004), describes the transition between light-limited and light-saturated photochemical efficiency (Hennige et al. 2008a).

To account for changes in light quality and hence the amount of photosynthetically useable radiation (PUR) (Kirk 1994; Hennige et al. 2009; Ragni et al. 2010) between the halogen light source and sunlight, the spectrum and intensity of the halogen light source were measured using a fibre optic spectrometer (Ocean Optics) and weighted to sunlight using a conversion factor calculated according to Ragni et al. (2008). The PUR of the filtered halogen light source was 99% that of the sun at any given PAR.

Symbiodinium identification

Algal DNA extractions were performed on coral fragments using the Wizard® isolation (Promega) protocol adapted by LaJeunesse et al. (2003). For each sample, the ribosomal internal transcribed spacer 2 (ITS2) region was amplified using the ‘ITSintfor2’ and ‘ITS2 clamp’ primers under a touch-down PCR protocol (LaJeunesse et al. 2002). The resulting products were subjected to denaturing gradient gel electrophoresis (DGGE; 45–80% denaturing gradient) for 15–16 h using a CBS Scientific system (Del Mar, California). Fingerprints for each sample were compared with an existing database for corals sampled in this region (LaJeunesse 2001; Warner and Berry-Lowe 2006).

Results

Photoinhibition and photorepair

Both coral species exhibited at least a twofold decrease in the operating efficiency of PSII, \( F_{q}{^\prime } /F_{m}{^\prime } \), between control fragments and CAP-treated fragments after 2 h at all three light intensities (LL, ML and HL) (Fig. 1a, b). As expected, there was a progressive decline in \( F_{q}{^\prime } /F_{m}{^\prime } \) in CAP-treated fragments across LL, ML and HL. No significant change in \( F_{q}{^\prime } /F_{m}{^\prime } \) was observed at any time in HL or ML control (no CAP) fragments. However, LL control fragments, which were subjected to a fivefold decrease in light intensity relative to pre-experiment conditions, had significantly higher \( F_{q}{^\prime } /F_{m}{^\prime } \) after 2 h than at T0 (t8 = 4.42, P = 0.01; t8 = 16.88, P = 0.00 for M. faveolata and P. astreoides, respectively).

Fig. 1
figure 1

(a–c) Operating efficiency of PSII, (\( F_{q}{^\prime } /F_{m}{^\prime } \)) and (d–f) effective absorption cross section of PSII, (σPSII′) in M. faveolata and P. astreoides coral fragments at LL, ML and HL with (+C) and without (−C) added chloramphenicol at Time 0, T + 1 and T + 2 h (n = 5 ± SE)

Full size image

In the presence of CAP and HL, P. astreoides displayed a more rapid decline in the light-acclimated PSII operating efficiency relative to M. faveolata. However, by the end of the HL incubation, Fv/Fm and \( F_{q}{^\prime } /F_{m}{^\prime } \) in fragments from both species did not significantly differ (Fig. 1a–c, Table 2), despite the initially higher \( F_{q}{^\prime } /F_{m}{^\prime } \) in P. astreoides (Fig. 1a, b) (t8 = 3.07, P = 0.02) and higher Fv/Fm in dark-acclimated control fragments (Table 2).

Table 2 Mean values (n = 5 ± SE) of fluorescence-derived parameters of high light (589 μmols photons m−2 s−1) coral fragments with and without chloramphenicol (CAP)
Full size table

In both species, addition of CAP elicited similar trends in the absorption cross section of PSII in the light (\( \sigma_{\text{PSII}}{^\prime } \)) as observed in \( F_{q}{^\prime } /F_{m}{^\prime } \) (Fig. 1d–f). At HL, both species exposed to CAP decreased \( \sigma_{\text{PSII}}{^\prime } \) over the 2-h incubation period compared with the controls, which did not significantly differ from each other or over the course of the experiment. Similarly, dark-acclimated σPSII did not differ between species at the end of the experiment in either control or CAP-treated fragments (Table 2). The PSII absorption cross section (σPSII) was also compared with maximum photochemical efficiency (Fv/Fm) in both species following 2 h of incubation (both with and without CAP addition) to assess whether an inverse relationship was observed as noted in previous studies on low-light cultured Symbiodinium (Ragni et al. 2010). However, no correlation was observed in either in P. astreoides or M. faveolata fragments (Pearson’s correlation = 0.135, r = 0018 and 0.074, r = 0.006, respectively).

Both species were subject to similar gross photoinhibition (GPiR), at LL, ML and HL (Fig. 2a; Table 2). However, M. faveolata displayed significantly higher net photoinhibition (NPiR) than P. astreoides at the medium and high light levels (Fig. 2b; Table 2). This equated to a significantly higher repair rate (RR) in P. astreoides than M. faveolata (Fig. 2; Table 2). Differences between species in NPiR or RR were not evident at LL or ML (Fig. 2b; c).

Fig. 2
figure 2

Gross photoinhibition (GPiR), net photoinhibition (NPiR) and repair rates (RR) (min−1) in M. faveolata and P. astreoides coral fragments under low (100) medium (200) and high light (589 μmols photons m−2 s−1) (n = 5 ± SE)

Full size image

Photophysiology

Both species exhibited typical responses when exposed to sequentially higher light levels, such that non-photochemical quenching (NPQ), photochemical quenching (1 − qP), antennae-bed quenching \( ((1 - \, (\sigma_{\text{PSII}}{^\prime } /\sigma_{\text{PSII}} )) \) and Electron Transport Rate (ETR), all increased with increasing light until the highest light level, at which point, data began to asymptote or decline (Figs. 3, 4). However, differences were observed between species. NPQ was significantly higher in P. astreoides at the highest light curve intensity, 2,200 μmols photons m−2 s−1 (t8 = 4.01, P = 0.00). At light levels comparable to growth irradiance (i.e. HL), there was no difference in NPQ between species (Fig. 3a). For antennae-bed non-photochemical quenching \( ((1 - \, (\sigma_{\text{PSII}}{^\prime } /\sigma_{\text{PSII}} )), \) Fig. 3b), which contributes to NPQ, the only difference between species was at irradiances below 400 μmols photons m−2 s−1, where M. faveolata had significantly higher quenching than P. astreoides.

Fig. 3
figure 3

Light response (PFD μmols photons m−2 s−1) of M. faveolata and P. astreoides (n = 5 ± SE) of a Non-photochemical quenching [NPQ] \( ((F_{m} - F_{m} {^\prime }) /F_{m}{^\prime } ) \) and b photochemical quenching \( (1 - qP{ (1} - F_{q}{^\prime } /F_{v}{^\prime } )) \). Panels (c and d) are light responses of M. faveolata and P. astreoides (n = 5 ± SE) standardised to E/Ek

Full size image
Fig. 4
figure 4

Light response (PFD μmols photons m−2 s−1) of M. faveolata and P. astreoides (n = 5 ± SE) of (a) antennae-based quenching \( (1 - (\sigma_{{{\text{PSII}}}}{^\prime } / \, \sigma_{{{\text{PSII}}}} )) \) and (b) RCII-specific electron transport rate (ETRRCII mol e− mol RCII−1 h−1). The last data point is omitted from (b) due to large variability

Full size image

However, it is important to note that the light curve data displayed here (Fig. 3a, b) do not account for differences in species-specific minimum saturating irradiances (Ek) calculated using least-squares non-linear regression (see Hennige et al. 2008a), which was significantly higher in P. astreoides than M. faveolata; 346 (±46.1) μmols photons m−2 s−1 and 222 (±37.2) μmols photons m−2 s−1, respectively (t8 = 2.09, P = 0.03). When the minimum saturating irradiance was taken into account by plotting light curves against irradiance relative to Ek (E/Ek) following Hennige et al. (2008a), NPQ differed to that plotted against irradiance alone (Fig. 3a, c). Importantly, NPQ at E/Ek = 1 and at E/Ek > 1 (where irradiance for photosynthesis is optimal and light saturating, respectively) was higher in P. astreoides than in M. faveolata, and the higher antennae-bed quenching observed in M. faveolata at low irradiances did not equate to higher NPQ than P. astreoides. It is worth noting that electron transport and antennae-bed quenching were not plotted relative to E/Ek, as the absorption cross section for PSII partially accounts for different absorption properties between both species. The trend of increased NPQ in P. astreoides was not matched in \( 1 { }- \, (\sigma_{\text{PSII}}{^\prime } /\sigma_{\text{PSII}} ) \) data versus E (Fig. 4). This result was similar in the independent assessment of antennae-based non-photochemical quenching following CAP incubation under sunlight. Although antennae-bed quenching was higher by ca. 30% in P. astreoides in both data sets, results were not significant as variability was high.

Photochemical quenching, calculated as 1 − qP, was significantly higher in M. faveolata fragments than P. astreoides at irradiances below 1,200 μmols photons m−2 s−1 (Fig. 3b). The opposite was observed in ETRRCII, with P. astreoides fragments having greater specific electron transport rates than M. faveolata, but only below 400 μmols photons m−2 s−1 (Fig. 4b). Higher ETR in P. astreoides was driven by a slower decrease in Fm and to a lesser degree Fo in response to increased light intensities (data not shown). In contrast to NPQ data, when expressed relative to E/Ek, 1 − qP plots from both species fell together (apart from values at very low E/Ek) rather than diverge from 10 to 600 μmols photons m−2 s−1 (Fig. 3b, d).

Symbiodinium identification

Both coral species harboured different in hospite Symbiodinium types according to ITS2 identification. M. faveolata fragments harboured B1, B17 and C7, either singly or as a mixed assemblage. P. astreoides fragments were dominated by A4 symbionts, and only one fragment was identified with an additional background symbiont, A3.

Discussion

Photoinhibition and photorepair

The extent of gross photoinhibition was directly influenced by the light availability of the corals, consistent with previous studies (Takahashi et al. 2009; Ragni et al. 2010) (Fig. 2). However, gross photoinhibition did not differ between P. astreoides or M. faveolata corals treated with chloramphenicol at all light levels (Fig. 2a). Since gross photoinhibition is calculated in the presence of an inhibitor of de novo protein synthesis, it represents the extent of damage that can occur in the absence of photorepair. Net photoinhibition data (Fig. 2b) demonstrated that under experimental conditions, P. astreoides exhibited substantially less photodamage than did M. faveolata at ML and HL. The increased variability in M. faveolata data compared with P. astreoides was possibly the result of the mixed in hospite Symbiodinium community, as opposed to the single symbiont identified in P. astreoides. The higher net photoinhibition in M. faveolata at ML and HL was driven by lower rates of PSII repair relative to P. astreoides and, to a lesser extent, lower NPQ than P. astreoides relative to E/Ek (Figs. 2c, 3c). This was also evident in the reduced photosynthetic efficiency of PSII, \( F_{q}{^\prime } /F_{m}{^\prime } \), in M. faveolata relative to P. astreoides at time zero (Fig. 1). Repair rates in both species increased from LL to HL. This implies that repair mechanisms were not damaged by increased light availability, but rather, that the increasing net photoinhibition at higher light intensities is due to increases in repair rate being insufficient relative to increased damage. This result corroborates recent evidence by Hill et al. (2011), who demonstrated that increases in gross photoinhibition, rather than decreases in repair rates, are a key trigger for net photoinhibition under bleaching conditions in P. damicornis. The high PSII repair rates observed in a Symbiodinium type considered ‘thermally tolerant’ by Ragni et al. (2010), complements work presented here, as P. astreoides (which had high PSII repair capacity relative to M. faveolata) is bleaching tolerant relative to M. faveolata (Warner and Berry-Lowe 2006).

Interestingly, the rate of gross photoinhibition in both coral species here was approximately 150% greater compared with the results by Ragni et al. (2010) for two different Symbiodinium clade A cultures (approximately 150% increase). Despite different types of Symbiodinium being used in previous studies, the current study provides preliminary evidence of substantially higher chloroplast protein turnover while the symbiont is in hospite.

Implications for coral bleaching

The high repair rate observed in P. astreoides relative to M. faveolata (Fig. 2c) corresponds well with speculation that photosynthetic repair rates may underlie bleaching susceptibility in corals in general (Warner et al. 1999; Takahashi et al. 2004; Yakovleva and Hidaka 2004; Ragni et al. 2010), and a central pathway to thermal tolerance in the Symbiodinium within P. astreoides in particular. This complements recent evidence which suggests that P. astreoides is increasing in dominance relative to other corals in the Caribbean and is poised to potentially undergo sustained population growth (Edmunds 2010). It is well established that high light can exacerbate the thermal stress response of Symbiodinium to induce coral bleaching (Lesser and Farrell 2004; Robison and Warner 2006), as both light and temperature act to inhibit or damage PSII, which can lead to an increase in ROS production (Smith et al. 2005; Suggett et al. 2008). Larger photosynthetic repair rates would potentially act to reduce (not eliminate) ROS production by Symbiodinium under bleaching conditions, which would ultimately confer a higher degree of resistance in coral holobionts to bleaching. This complements recent work where higher repair rates were observed in thermally tolerant Symbiodinium types than in thermally sensitive types in vitro under non-stressful conditions (Ragni et al. 2010). Although the thermal stability of repair mechanisms may remain an important factor to consider in any bleaching study (Takahashi et al. 2009), the different rates observed here between the bleaching tolerant P. astreoides and the bleaching sensitive M. faveolata under non-bleaching conditions indicate that initial repair rates and stimulated repair rates during stress, in addition to thermal tolerance and non-photochemical efficiency, may be key to mitigating the effects of bleaching events.

This raises interesting questions as to the host-Symbiodinium specificity in the corals studied here, and whether M. faveolata, may be constrained to harbouring Symbiodinium types that have lower repair rates than those in P. astreoides. Given that M. faveolata in this region may harbour as many as three different Symbiodinium spp. (Kemp et al. 2008) and the intraspecific and temporal variability in symbiont populations within this species is prevalent (Thornhill et al. 2006), rates of photodamage and repair may be quite different if other Symbiodinium species dominate in this coral.

Photophysiology

\( F_{q}{^\prime } /F_{m}{^\prime } \) values after 2 h of CAP incubation were not significantly different between P. astreoides and M. favaolata fragments (Fig. 1a, b). Dark-acclimated Fv/Fm of both species following incubation also matched this trend (Table 2). However, the extent of \( F_{q}{^\prime } /F_{m}{^\prime } \) decrease was greater in P. astreoides than in M. faveolata. This difference indicates that Symbiodinium in M. faveolata fragments were photoinactivated to some extent compared with P. astreoides. Since this difference persisted in samples recorded at night, this photoinhibition is likely chronic, i.e., does not recover in a matter or hours, and recovery may instead take days or longer (Krause 1988; Gorbunov et al. 2001). This is supported by the large difference between species Fv/Fm in dark-acclimated samples at midday, as there was very little difference between M. faveolata \( F_{q}{^\prime } /F_{m}{^\prime } \) and Fv/Fm. This means that M. faveolata had greater damage within PSII reaction centres prior to and during this experiment, as photoprotective mechanisms and repair mechanisms have been inadequate to prevent PSII degradation (Gorbunov et al. 2001).

The greatest decrease in \( F_{q}{^\prime } /F_{m}{^\prime } \) for both species during CAP incubation was at HL levels (Fig. 1c), which is a result of light exposure and damage to D1 protein within PSII. LL fragments would not have experienced as much damage to D1 proteins as HL fragments and hence the addition of CAP to LL fragments elicited less of a response. With regard to the light acclimation state, LL control fragments acclimatised during the 2-h experiment time to lower than ambient (i.e., the light level that fragments had been previously acclimatised to) light levels and hence can be considered ‘low-light’ acclimatised relative to time zero, which explains increased \( F_{q}{^\prime } /F_{m}{^\prime } \) values.

Both light-acclimated and dark-acclimated values of σPSII decreased for CAP-treated fragments of both species, whereas in control, HL fragments σPSII did not differ significantly between species (Table 2), due in part to the relatively large variability observed in M. faveolata data, and \( \sigma_{\text{PSII}}{^\prime } \) did not differ significantly over the 2-h experimental time period. The lack of any change temporally and between species at a time that arguably should elicit the largest change (midday) indicates that rapid changes in absorption cross section are not the dominant acclimatisation strategy of Symbiodinium in both coral species. This is consistent with previous research that concluded that changes in reaction centre content and not antennae bed changes are the dominant acclimation strategy for cultured Symbiodinium (Iglesias-Prieto and Trench 1997; Hennige et al. 2009). The decrease in σPSII in CAP-treated samples is opposite to the trend observed in a recently published study that used an experimental design similar to this one with cultured Symbiodinium (Ragni et al. 2010). This difference likely results from the specific photosynthetic pigment/protein architecture used in Symbiodinium spp. and the possible differences therein between algal populations in vitro versus in vivo. Ragni et al. (2010) hypothesised that the cultured Symbiodinium exhibited a ‘lake’ model of light-harvesting complex organisation (i.e. light harvesting and energy transfer is shared among several PSII reaction centres, see Bernhardt and Trissl 1999). This would explain an increase in σPSII when PSII translational repair was chemically blocked, and inactive or photoinhibited Reaction Centres (RC’s) transferred absorbed exciton energy to remaining functional RC’s. However, the data here provide better support for a ‘puddle’ model of light harvesting, wherein there may have been greater isolation and lower connectivity (ρ) between PSII reaction centres (Bernhardt and Trissl 1999) in the Symbiodinium investigated in hospite here. Thus, once RC’s became inactivated, the energy absorbed through the attached antennae beds was no longer available to neighbouring active RC’s. Considering that different types of Symbiodinium exhibit different biophysical and bio-optical characteristics when photoacclimated to various light levels in culture (Hennige et al. 2009), it is not yet known whether this is a ubiquitous response of all Symbiodinium spp. Nevertheless, the assumption of limited photosynthetic unit connectivity is in agreement with earlier work that also noted substantially lower ρ in Symbiodinium in hospite as compared to other aquatic photosynthetic organisms (Gorbunov et al. 2001).

Light response curves (Figs. 3, 4) were broadly similar for both species when plotted against applied E with slight differences in photosynthetic parameters between species as may be expected (Hennige et al. 2009). However, typical light curves do not account for ‘light’ differences between coral species such as host absorption (Hennige et al. 2008b), skeleton scattering (Enriquez et al. 2005), Symbiodinium density or long-term light history of the coral. Likewise, the light received by the in hospite Symbiodinium, i.e., the photosynthetically useable light [PUR (Kirk 1994; Hennige et al. 2008a)] may differ dramatically in different coral species. To circumvent such limitations, the minimum saturating irradiance of PSII electron transport, Ek, was calculated for each coral. The significantly higher Ek in P. astreoides indicates that P. astreoides Symbiodinium were able to process more excitation energy through PSII before saturation than Symbiodinium in M. faveolata. When applied light (E) was expressed relative to Ek (which gives a more accurate representation of the efficiency of Symbiodinium photosynthesis relative to its acclimatisation state), NPQ was higher in P. astreoides than in M. faveolata, and photochemical quenching, qP, did not differ between species (Fig. 3c, d). This is consistent with work by Hennige et al. (2008a), where different coral species were found to modify NPQ to optimise photosynthesis towards Ek. The slightly higher NPQ in P. astreoides from E/Ek = 1 (where applied light is equal to the minimum saturating irradiance) to E/Ek = 6 (at saturating light intensities) indicates that P. astreoides has more capacity than M. faveolata at that acclimatisation state to dissipate light energy through certain pathways. However, while NPQ was higher in P. astreoides than in M. faveolata above growth irradiances (Fig. 3), it is likely that antennae-bed quenching was not the driving factor. Another possible pathway for NPQ is quenching through thermal dissipation in the reaction centres themselves, which would be recorded through NPQ but not through changes in σPSII (Gorbunov et al. 2001). Increased NPQ in P. astreoides where E/Ek ~1–2, contributed to the increased electron transport rate of PSII, ETRRCII at ca. 200–400 μmols photons m−2 s−1, which indicates potentially increased productivity through higher electron flow at optimal light utilisation. Increased electron flow through functional reaction centres in these two species may be interesting for future studies, as the ‘compensatory effect’ (Behrenfeld et al. 1998; Kana et al. 2002), where electron turnover rates are increased and is potentially beneficial under conditions likely to cause net photoinhibition (Hennige et al. 2009), would complement increased electron transport rates in P. astreoides.

Conclusions

Under high light conditions and ambient temperature, P. astreoides exhibited less net photoinhibition than M. faveolata. Since gross photoinhibition was comparable in both species, the reduced net photoinhibition in P. astreoides was enabled through higher PSII repair rates relative to M. faveolata. NPQ at and above light-saturating intensities may confer some additional advantage to P. astreoides with regard to mitigating photodamage, but further study is required to expand upon this. The increased repair rate observed here in the bleaching tolerant P. astreoides coral highlights that repair rates are an interesting area for future study with regard to the bleaching susceptibility of coral, as Symbiodinium types with relatively high repair rates could potentially act to minimise net photoinhibition under thermal bleaching conditions. Further study will determine to what extent PSII repair capacity underlies bleaching susceptibility and whether repair capacity under non-bleaching conditions can be used as an indicator of coral bleaching susceptibility.