Zooxanthellae density and Chlorophyll a.

Coral samples (∼3–5 cm long) for the analysis of zooxanthellae density (per cm²) and Chl a concentration in the tissue (µg cm⁻²) were stored at −20°C prior to analysis. Methods for zooxanthellae density and Chl a in coral tissue were adapted from Siebeck et al. [39]. Briefly, a jet of pressurized air in filtered seawater was used to remove coral tissue from the skeleton and the slurry homogenised (45 s) after measuring exact volume of the homogenate. Homogenate samples (10 ml) were taken, fixed with formaldehyde (2 ml formaldehyde per 10 ml solution) and a Neubauer haemocytometer used to count total number of zooxanthellae under a microscope (four replicate counts on four plates). Calculations of the total number of symbiotic zooxanthellae per area were based on the volume of the homogenate and the coral surface area. Coral surface area was measured using paraffin wax method [40].

Chl a concentration was determined following the methods of Jeffrey and Humphrey [41]. Briefly, three aliquots of the homogenate sample (10 ml) (as above) were centrifuged (3500×g) for 15 min. The supernatant was discarded and the pellet re-suspended in 10 ml of acetone (100%) and Chl a extracted for 24 h in the dark (−20°C). After centrifugation at 3500×g for 10 min fluorescence was measured in a T700 fluorometer (Turner Designs) and standardized to coral surface area.

RNA/DNA and protein analysis.

Samples for RNA/DNA and protein analysis were stored at −80°C prior to analysis. RNA content varies with metabolic demand and is correlated with new protein synthesis, while DNA content is largely stable [14], [15]. Methods for analysis of RNA/DNA followed Humphrey [42]. Briefly, coral samples (∼1 cm long) were crushed in liquid nitrogen and TE extraction buffer (Tris-EDTA with 1% sarcosyl added (10 ml)). Samples were sonicated in an ice bath, centrifuged (1200×g for 3 min) and the supernatant (100 µl) as well as TE buffer (900 µl) transferred into deep well plates. The same solution was used for protein analysis. Samples (75 µl, triplicates), nucleic acid standards (0–2.5 µg ml⁻¹ for DNA and RNA, duplicates) and control homogenates were added to three 96 deep-well microplates. Plate one was treated with TE buffer (15 µl) while plate two with TE buffer (7.5 µl) and RNAse (7.5 µl) and plate three with RNAse (7.5 µl) and DNAse (7.5 µl). Plates were incubated (35 min for plate one and two and 60 min for plate three) and Ribogreen added to each well. Plates were read in a microplate reader (485 nm excitation and 528 nm emission). RNA was calculated by subtracting the fluorescence reading of plate two from plate one while DNA was calculated by subtracting the fluorescence reading of plate three from plate two. Calculating of RNA and DNA concentrations was based on the standard curves of each plate. Finally the ratio between DNA and RNA concentration was determined.

For protein analysis, a standard DC Protein Kit was used [43]. Sample solution (5 µl, as above (triplicates)) was added together with protein standards (0.2 mg ml⁻¹ to 1.5 mg ml⁻¹ protein duplicates) in a 96 deep-well microplate. After adding 25 µl of reagent A and 200 µl of reagent B to each well, absorbance was read after 15 min at 750 nm. Protein content was calculated based on standard curve of each plate and all protein concentrations standardized afterwards to DNA values.

Effective quantum yield and relative electron transport rate.

Variations in the effective quantum yield (yield) of photosystem II (PSII), a way to assess the photosynthetic performance of PSII, were measured using an underwater pulse-amplitude modulated (PAM) fluorometer (Diving-PAM, Walz) on 10 sections of the tagged coral colony. The fiber was placed at a fixed distance (1 cm) in front of the coral tissue. The yield (ΔF/Fm′) was measured by exposing 10 sections of the colony separately to a 0.8 s period of saturating light (ca. 8000 µmol m⁻² s⁻¹) [11]. Yield measurements were converted to relative eletron transport rate (rETR) with the formula rETR = ΔF/Fm′×PAR×0.5 with PAR as the immediate radiance and 0.5 as the factor that accounts for the distribution of electrons between photosystem I and photosystem II [44]. Relative measure was used for the rate of electron transport since light absorption characteristics of tissue are unknown for these species.

Abiotic factors.

Light and temperature were measured using temperature loggers (Hobo Pendant Data Logger) that were deployed between 3 to 7 days before the sampling.

Short-term current speed was estimated using drifters. Two crucifix design drifters [34] containing a GPS were simultaneously deployed for approximately 10 minutes and the GPS location and time of their deployment and collection recorded. The distance and time of each deployment allowed a simple estimate of surface flow speed.

Biotic factors.

Water quality samples were taken for the analysis of dissolved nutrients, total nitrogen (TN), Chl a, and picoplankton (Synechoccocus, Prochloroccocus and picoeukaryotes).

For the analysis of dissolved inorganic nitrogen (DIN) (NO_x (NO₃+ NO₂) and NH₄), water samples (40 ml) were filtered through 0.45 µm filters and stored at −20°C, before flow injection analysis (FIA) with detection by absorbance at specific wavelengths for nitrate (NO₃)/nitrite (NO₂) (QuikChem FIA+Lachat 8000 series). Briefly, Nitrate was reduced to nitrite through a copperized cadmium column that than reacted with sulphanilamide under acid conditions to form a diazonium ion. The diazonium ion is coupled with N- (1-naphthyl) ehthylenediamine dihydrogenchloride that has an absorbance maximum at 520 nm (Quikchem Method 31-107-04-1-A) [45]. Ammonium was measured by fluorescence (Global FIA high sensitivity gas diffusion unit- HPMSD, Shimadzu RF-10Axl Fluorescence detector) [46]. Briefly, Ammonium was liberated as ammonia by sodium hydroxide and passed subsequently through a porous PTFE membrane (HPMSD), reacted with ortho-phthaldiadehyde and sulphite and formed a fluorescent derivative. The derivative was measured with an excitation wavelength of 310 nm and an emission wavelength of 390 nm. Dissolved inorganic nitrogen (DIN) is the sum of ammonium and nitrate and nitrite.

For total nitrogen (TN), unfiltered water samples (50 ml) were stored at −20°C until analysis. TN was determined from autoclave digests with potassium persulphate (Lachat Quick-Chem 8500 Automated Flow Injection Analyser) [47].

Organic nitrogen (ON) was calculated by subtracting DIN from TN.

Seawater samples (1 L) for Chl a was filtered onto 0.7 µm filters (Whatman GF/F) and the filters stored at −20°C in the dark until fluorometric analysis of duplicated 90% acetone extracts was carried out [48].

To determine concentrations of autotrophic picoplankton groups (Synechoccocus, Prochlorococcus and picoeukaryotes), 1.5 ml seawater was fixed with gluteraldehyde (0.5% final concentration) for 10 minutes and then snap frozen in liquid nitrogen. Samples were analysed using flow cytometry following the technique of Patten et al. [49]; samples were thawed at 37°C, 1 µm fluorescent beads (Molecular Probes) added as an internal standard and samples were analysed using a FACSCANTO II (Becton-Dickinson) flow cytometer fitted with a 488 nm laser on high throughput mode at a flow rate of 60 µl min⁻¹ for 100 s. Nitrogen (N) and carbon (C) content of picoplankton was calculated based on fixed factors for picoeukaryotes (39.2 fg N and 836 fg C ) [50], [51], Prochloroccocus (4 fg N and 46 fg C) and Synecoccocus (30 fg N and 213 fg C) [52], [53]. Total nitrogen and carbon content for picoplankton was then calculated by the sum of nitrogen and carbon concentrations for Prochloroccocus, Synechoccocus and picoeukaryotes.

Zooplankton was sampled with a plankton net (90 cm diameter, 10 min in the water with a speed of 2.5 km h⁻¹ for 50 m around the station) and preserved with formaldehyde (ca. 5%). Dry weight analyses were carried out for zooplankton between 100 and 1000 µm. Wet zooplankton samples were added to a pre-weighed container after cleaning samples of salt with DI water (30 s) and dried in an oven (60°C for 24 hours). Samples were then cooled down in a desiccator and weighted for subsequent calculation of dry weight (DW) [54].

Protein concentration.

Our results suggest that protein concentration is relatively stable on a diel basis, suggesting that this index could be used primarily to determine changes in seasonal and long-term coral health. This is in agreement with previous work which showed that protein concentration only shows detectable changes after weeks or months of perturbation, generally associated with changes in particulate food availability [16], [57], [58] or extreme temperatures [59], [60]. The precise time it takes to build-up and use protein reserves seems to be species-specific and depends on individual metabolic rates [59], [60]. While metabolic rates of A. spicifera were generally higher than of A. digitifera [32], no recurrent diel or daily patterns were observed for either species. Significant correlations between protein concentration and short-term light exposure (seconds – minutes), as well as (slightly weaker) correlations between integrated light exposure over the previous several days suggest that protein content does in fact respond to light, but that corals effectively integrate energy across diel fluctuations in light intensity, such that heterotrophic processes including protein synthesis, remain relatively stable. For example, coral’s protein synthesis is often dependent on DIN uptake, a light dependent process, which has been shown to express diel patterns [61]. In fact, overall in our study, nitrogen was a primary physico-chemical predictor for protein concentration in A. digitifera on a diel basis and for A. spicifera protein synthesis was correlated with light across seasons [38]. Positive correlations of protein concentration with light (seconds to days) might also be related to changes in the translocation rates of photosynthetic products from the zooxanthellae to the coral host which can be as rapid as 15 minutes, but might take 48 hours, at which time the products are finally integrated into coral tissue [31]. Overall, it is likely that a variety of physico-chemical factors varying on a diel, daily and seasonal basis light, nitrogen concentrations and zooplankton concentrations [38] will influence changes in protein concentrations in corals; however these changes are integrated via the longer metabolic time lines for protein synthesis. This index therefore emerges as a robust choice for the measurement of medium to long-term changes in coral health.

RNA/DNA ratio.

RNA/DNA ratio showed significant differences between species, between seasons, between days and times of the day. This index is therefore an appropriate indicator for short as well as long-term changes in metabolic rates of corals. Because daily and diel variations are smaller than seasonal differences, the RNA/DNA ratio also provides an index of coral health that integrates appropriately across time scales. Seasonal changes in RNA/DNA ratio have been seen for corals [14] as well as changes within days when transplanted to different depths [42]. Diel patterns in our study were only recorded when data between species was pooled perhaps due a too small sample size given the intraspecific variation. In general, there is a clear response of RNA/DNA ratio to diel physiological fluctuations [30] as described for other species such as molluscs [23] and fish [22]. In our study, both species showed a winter pattern of lowest RNA/DNA ratio at noon and highest at midnight, while summer patterns were more variable, with highest RNA/DNA ratio in the day time, possibly since physico-chemical factors showed stronger diel fluctuations in summer then winter. Given the apparent importance of light, plankton concentrations and nitrogen concentrations in driving RNA/DNA ratios [14], [38], we suggest that this reflects corals’ reliance on photosynthates transferred from the zooxanthellae to the coral host, as well as on energy gained through heterotrophic feeding. Short-term (seconds to minutes) light variation showed the strongest correlation with RNA/DNA ratio for both Acropora species but also integrated light level from up to three previous days were correlated with the RNA/DNA ratio. However, the lack of correlation between RNA/DNA ratio and rETR, an indicator for photosynthetic activity [12], is contradictory to the hypothesis that diel variations in RNA/DNA ratios are driven by photosynthetically derived carbon. However, even though no clear correlations exist, diel variations in photosynthesis are likely to be at least partly responsible for diel changes in RNA/DNA ratio potentially due to a time lag effect of the integration of translocation products into the tissue [31] or an overestimation of rETR [12]. Plankton and nitrogen concentration can also trigger diel and daily variations in coral health, since diel cycles are often set by the availability of demersal plankton food [17], [18] and alternate sources of nitrogen [19] which can change on a daily as well as seasonal basis. This time-lag in feeding can be observed in the dynamics of coral food vacuoles, which appear in digestive cells within 2 hours of feeding and subsequently decrease in number per digestive cell and as the percentage of digestive cells with food vacuoles, only after 5–7 hours post-feeding [9]. Overall, RNA/DNA ratio can be used as a short-term as well as long-term indicator even though further studies are needed. When used in monitoring projects it has to be kept in mind though that RNA/DNA measurements should be taken at the same time each day to make data comparable.

Zooxanthellae densities and pigments.

In general, zooxanthellae densities in our study fell within the range of previous studies, ranging between 0.5×10⁶ and 5×10⁶ cells per cm² [29], [62], [63]. Zooxanthellae densities in corals in our study were surprisingly resistant to diel and daily changes in irradiance. Diel changes in zooxanthellae densities have been observed for some species due to synchronous division and degradation over a diel cycle [8], [9]. However, this diel cycle in corals seems highly species specific and not always present [40], [64], [65]. The results in our study suggest that if such cycles occur in zooxanthellae division, other processes regulating cell concentrations are likely to be operating at the same time, such as degradation [66], digestion and extrusion [65], resulting in no significant change in zooxanthellae density.

The strongest correlation of incident light with zooxanthellae density occurred more than 24 hours prior to sampling in A. digitifera. Previous studies have shown that it takes 24 hours for zooxanthellae densities to be modified [67], since algal division rate is set to more than 24 hours prior to actual cytokinesis [9]. Rhythmical changes in proliferation of zooxanthellae with periods of 3–6 days can certainly occur in addition to daily proliferation periodicity [68] due to changes in light [9], as well as local supply of zooplankton and nutrient concentrations [9], [69]. The fact that for A. spicifera, changes in zooxanthellae density do not seem to be directly driven by light history within either hours or days, suggests that zooxanthellae density is driven primarily by other factors, such as nitrogen uptake [38]. Our results suggest that for certain species such as A. digitifera zooxanthellae densities can be used to describe long-term (seasonal) changes in coral health. This is in agreement with other studies which also observed seasonal changes in zooxanthellae densities [28], [29].

Daily changes in physico-chemical factors such as temperature, light and nutrient concentrations might have been not large enough to result in measurable changes in zooxanthellae density. Seasonal environmental variation is much greater, resulting in larger variation in symbiont densities. Temperature, nutrients and light were all observed to be important in driving measurable variation in coral health indices [38].

We found no seasonal or diel difference for Chl a per unit coral surface area, or for Chl a per zooxanthellae. Our measured daily variations suggest Chl a can be a good indicator for short-term (days) changes in physico-chemical impacts on coral health. Daily changes in light intensity have also been known to impact coral Chl a concentrations [70], but no significant diel patterns were observed in our study or a previous study [5]. Our results do indicate that short term light levels (within hours) have the strongest impact on Chl a concentrations, suggesting there are no real time-lag effects. In regard to seasonal changes no significant variation was detected in this study in contrast with previous studies that have shown significant variations in Chl a concentration throughout the year in the studied species [28], [29], [71]. Thus Chl a might be suitable to describe seasonal changes in physico-chemical factors in case these changes are big enough, however it should be kept in mind that daily changes might also influence these patterns.

Effective quantum yield and relative ETR.

Diel patterns in photosynthetic yield of A. spicifera are in accordance with previous studies showing highest values at night, a decrease in the morning, low values around noon and an increase during afternoon towards evening [10], [72], [73]. Down-regulation is attributed to dynamic photo-inhibition mediated by non-photochemical quenching in the reaction centre and antenna pigment bed [44], [73]. Yield depends on daily occurrence of peak solar radiation and is reversible and photo-protective [74]. The slightly different pattern for A. digitifera, with lowest values in the morning might be due to different genetic types of Symbiodinium [75], [76], photo-protection trait in the tissue [77] and tissue thickness [78]. In general yield values between 0.3 and 0.6 were in the range of previous studies [11], [79]. The fact that overall no species-specific differences occurred, suggest that both Acropora species’ photosystems react in similar ways to variations in light levels. However, yield values showed only a low correlation with short-term light history (R = 0.24 and 0.26) suggesting that other physico-chemical factors are more important for changes in yield. Despite previous studies which showed seasonal changes in yield values due to temperature [27], [80] our study did not observe seasonal variations even though temperature difference varied between 23 and 31°C between winter and summer. Overall for the studied Acropora species yield does not seem to be a good indicator for determining any seasonal differences, however yield appears more appropriate indicator for short-term changes since it reacts quickly to variations in light intensity. However since our study fell within a La Niña year and light values during summer were lower than normal [38] further studies are needed to determine seasonal changes. In addition when used for short-term monitoring projects light intensities are needed and samples should be taken at the same time to make data comparable.

Seasonal, daily and diel changes in rETR were clearly observed in our study, suggesting this measure was a good short-term and long-term indicator of coral health. Higher values at noon than in the morning are in the same range as those seen in previous studies [12], [72] for both species. Higher values during summer than in winter are also in accordance with previous work [27]. Short-term light history had an overwhelming correlation with changes in rETR while also light from previous day showed a smaller impact. Our values are in the range of other studies for morning (50–150) and midday (100–200) [12] as well as with seasonal studies [27]. Rates are influenced by phylotype and physiology (colour and tissue thickness of the coral species) of resident Symbiodinium sp., colony morphology, animal behaviour (polyp extension and contraction) as well as the light history of a particular location from which fluorescence measurements are taken [72], [81]. However, photosynthetic productivity can be overestimated when interpreted only through rETR [12] since rETR is non-linearly correlated with primary productivity and the relationship between biochemical and energetic assays of photosynthesis is influenced by photo-acclimatisation state of individual colonies [12]. Thus the photosynthetic rates we estimated at noon when PAR was highest might be an overestimation.