Diel `tuning' of coral metabolism: physiological responses to light cues

The symbiotic association between zooxanthellae and corals responds to the surrounding environment by both fast acclimation and long-term adaptation of each partner in this association. Long-term adaptations involve changes in genotypes and even replacement of the zooxanthellae species(Baker, 2003). Daily changes in light intensity and spectrum, temperature, and nutrient and food availability cause changes in respiration and photosynthesis, enzyme activity and pigment composition and concentration. Increasing irradiance results in an increase in photosynthesis rates (Falkowski and Raven,1997). Under high irradiance levels in shallow reefs, corals experience elevated oxygen partial pressures (P_O2) in their tissues as a result of the oxygen produced by their zooxanthellae(Dykens and Shick, 1982),apparently without any cellular damage. High light levels, elevated seawater temperature and UV radiation lead to an increase in reactive oxygen species(ROS) within the algae as well as in the host (Lesser and Shick, 1989a,b; Dykens et al., 1992), which is harmful to both (Lesser and Shick, 1989a,b).

Superoxide dismutase (SOD), catalase (CAT) and ascorbate peroxidase function together to inactivate the harmful oxygen free radical and hydrogen peroxide (H₂O₂), thereby preventing subsequent cellular damage (Fridovich, 1986; Asada and Takahashi, 1987). Protective mechanisms against active oxygen species in both the animal host and its zooxanthellae (Lesser and Shick, 1989a,b)have been described. Responses include changes in pigment content and in the activities of SOD, CAT and ascorbate peroxidase(Shick et al., 1995).

The photosynthesis-irradiance relationship in corals changes in response to ambient light (Falkowski et al.,1990). Zooxanthellae, like most algae, photoacclimate to changes in irradiance by adjusting light-harvesting and utilization capability (for reviews, see Falkowski and Owens,1980; Stambler and Dubinsky,2004), by changing cellular pigment concentrations(Dubinsky et al., 1984; Porter et al., 1984). Photoacclimation also includes changes in the respiration of the zooxanthellae, the quantum yield and the light-saturated rate of photosynthesis (Porter et al.,1984). Various photoacclimation processes differ in their kinetics(Fisher et al., 1996). While most of these changes take longer than a day, some, such as polyp expansion and contraction, occur within minutes(Levy et al., 2003). Changes in the density of the zooxanthellae and reshuffling of their species occur within days, whereas changes in growth form of the coral colony may take several months to years (Graus and Macintyre, 1982; Miller,1994; Todd et al.,2004), and genetic selection takes thousands of years(Falkowski et al., 1990).

In the past few decades, coral reefs have been exposed to natural and anthropogenic stresses, such as increases in temperature and UV radiation,which cause major bleaching events over entire coral reef systems (e.g. Glynn, 1993, 1996; Hoegh-Guldberg, 1999; Fitt et al., 2001; Coles and Brown, 2003; Jokiel, 2004). Eutrophication and pollution are likely to act synergistically with temperature in rendering corals more susceptible to bleaching and disease(Dubinsky and Stambler, 1996). Hypotheses for the mechanism of bleaching include (1) elevated oxygen tension in the symbionts and the host, leading to internal damage by ROS(Shick et al., 1996), (2)reduction of the photosynthesis/respiration ratio, producing an energetic imbalance (Jokiel and Coles,1990), and (3) free-radical-induced damage to Photosystem II(PSII) of the zooxanthellae (Lesser,1996), triggering a membrane lipid mediated apoptosis cascade(Tchernov et al., 2004) and in some cases pathogen infection (see Rosenberg and Loya, 2004).

The aim of the present study was to document some responses of corals and their symbionts to the diel patterns of solar radiation. The study of concomitant changes in metabolism, oxygen evolution, fluorescence yields,pigments, and protective enzymes (SOD, CAT) correlated or tracking the course of sunlight, will afford a better understanding of the processes occurring in the host and in the zooxanthellae, allowing corals to optimize the benefits of photosynthesis while minimizing the inevitable risks of concurrent oxidative damage.

High-light-adapted colonies (Falkowski and Dubinsky, 1981) of Favia favus Forskal, Plerogyra sinuosa Dana and Goniopora lobata Milne Edwards and Haime of similar size (∼10 cm diameter) were collected from a depth of 4-5 m. These corals were chosen because we have information on some aspects of their energetics, such as prey capture, photosynthesis and their expansion/contraction behavior (see Levy et al., 2003, 2004, 2005a,b). The daily metabolism (photosynthesis and respiration) and chlorophyll fluorescence measurements were performed during July-August 2001. From December 2001 until March 2002, additional colonies of the same species, ∼30 cm in diameter,were collected from 4-5 m-depth next to the Interuniversity Institute for Marine Sciences of Eilat (the H. Steinitz Marine Biology Laboratory), Israel. Three mother colonies from each species were divided into six subcolonies of∼5 cm diameter each and placed at 4-5 m-depth for 75 days for `healing'and acclimation, before experimentation. One subcolony from each species was placed in the respirometer (see below) for 24-h measurements. After measurement, these subcolonies were frozen in liquid nitrogen. Parallel to the respirometer experiment, every two hours during daytime (08.00-16.00 h) and before dawn (05.00 h), three subcolonies from each species were sampled and immediately frozen in liquid nitrogen. The samples were stored for 1-2 weeks at -70°C before analysis of the biomass parameters and enzyme activity.

The light spectrum was measured at 5 m-depth (in the Gulf of Eilat, Red Sea, in front of the Interuniversity Institute for Marine Science). Measurements were conducted on several cloudless days during 15-19 February 2001. A spectral scan was performed using a Li-Cor LI1800 scanning spectroradiometer (Lincoln, NE, USA). Readings were conducted every 30 min at 2 nm interval spans between 300 and 750 nm (units are W m^-2nm^-1). Only data from 15 February are presented since that day was considered representative for a cloudless day. The data represent only UV measurements between 300 and 400 nm, which include UVA + UVB.

The in situ oxygen flux data were obtained using twin,three-chamber submersible respirometers (AIMS, Townsville, QLD, Australia). This instrument is equipped with UV-transparent chambers, each with an oxygen sensor (Kent EIL galvanic type ABB, Stonehouse, Gloucestershire, UK), one light meter (Li-Cor 4π underwater quantum sensor), a temperature probe and a data logger. A centrifugal pump flushes the water in the chambers at programmable intervals. Twenty-minute intervals were used for these experiments (Fabricius and Klumpp,1995). Prior to the incubation period, the colony surfaces were carefully cleaned of epiphytes and other debris using a small toothbrush. The respirometers were deployed at 4-5 m-depth in front of the H. Steinitz Marine Biology Laboratory (near the incubation site). Data processing was performed using the AIMS `Respiro' program for calibrating and normalizing the data. Respiration was measured as oxygen uptake in the dark. Parameters for the photosynthesis (P) versus energy (E) curves were calculated from a non-linear curve-fitting, including α (initial slope). P_max (light-saturated photosynthesis rate and compensation light levels), E_opt (optimum irradiance), E_com (compensation intensity) and E_k(saturating intensity) were calculated from a non-linear curve-fitting based on theoretical models of a hyperbolic tangent equation(Ben-Zion and Dubinsky,1988).

Diel changes in chlorophyll fluorescence and fluorescence characteristics of the three symbiotic corals were monitored for three consecutive days for a diurnal cycle from 31 July to 2 August 2001, using the SCUBA-based Fast Repeating Rate Fluorometer (FRRF; Gorbunov et al., 2000). The FRRF measures the chlorophyll fluorescence yield using a sequence of flashlets that gradually close the PSII reaction centers, leading to an increase in chlorophyll fluorescence(Kolber and Falkowski, 1993). The maximum quantum yield of photochemistry in PSII was determined in a dark-adapted state as the ratio of F_v/F_m=(F_m-F_o)/F_m,according to Butler (1972),where F_v is variable fluorescence, F_mand F_o are maximum and minimum yields of chlorophyll a fluorescence measured in a dark-adapted state (relative units). The steady-state quantum yield was measured under illumination by ambient light and defined byΔ F′/F_m′=(F_m′-F′)/F_m′(see Gorbunov et al., 1999, 2000; Lesser and Gorbunov, 2001). The quantum yield of irradiance-stimulated thermal dissipation [i.e. non-photochemical quenching (NPQ)] is assessed from the ratio of(F_m-F_m′)/F_m(Gorbunov et al., 2001). The instrument was placed on a tripod at 5 m-depth, and measurements were taken automatically from the top part of the coral colony(Lesser and Gorbunov, 2001) at 10-min intervals. Cross section was calculated(Gorbunov et al., 2001). The data presented show mean measurements for each hour.

Tissue homogenates were prepared by removing all tissue by an airbrush method, which is similar to the Water-Pik method(Johannes and Wiebe, 1970) but uses airjet to strip the tissue of the coral skeleton into a few ml of 100 mmol l^-1 phosphate buffer (pH 7.0). The volume of the homogenate was measured and subsamples taken for determination of: (1) zooxanthellae density, from direct counts on a Neubauer hemacytometer; (2) chlorophyll a concentration, measured spectrophotometrically on a Cary spectrophotometer (Varian, Palo Alto, CA, USA) in 90% acetone(Jeffrey and Humphrey, 1975)and (3) enzyme activity, quantified using the methods described below. The surface area was calculated based on a digital photograph using the program Image tool (Uthscsa Image Tool for Windows, v. 2.00, University of Texas, San Antonio, TX, USA).

The homogenate was centrifuged twice at 1500 g for 15 min. The supernatant was used for analysis of animal protein and enzyme activity. The pellets containing the zooxanthellae were suspended in 2 ml 100 mmol l^-1 phosphate buffer (pH 7.0). One milliliter was taken for HPLC analysis, and an additional 1 ml was dissolved by sonication (Misonix 3000;4×15 s pulses at 35 W). The sonicated suspension was replaced by a 0.05%Triton X-100 solution in 1 ml 100 mmol l^-1 phosphate buffer (pH 7.0). After incubation for 10 min, the suspension was centrifuged at 14 000 g for 30 min and used for protein and enzyme analyses. The enzymes were examined in the cytosoluble fraction of protein extractions. SOD activity was assayed spectrophotometrically as described by Elstner and Heupel(1976) and Oyanagui(1984). Standards were prepared using bovine erythrocyte SOD (Sigma) for each set of samples. Catalase activity was assayed spectrophotometrically by monitoring H₂O₂ depletion at 240 nm. All assays were conducted at 25°C and expressed as enzyme activity units (U) per mg protein. SOD and CAT activities were averaged and presented as a percentage of the total maximum activity. Protein content was determined by the Bradford assay(Bradford, 1976).

For pigment analysis, frozen subsamples were extracted in 1 ml cold 100%acetone, vortexed and centrifuged. The supernatant was transferred to a test tube and kept in the dark at 4°C for approximately 1 h. The acetone extract was then filtered onto a GF/F filter, and 0.3 ml 1 mol l^-1ammonium acetate was added in order to facilitate the separation of the hydrophilic components of the extract(Zapata et al., 1987). The reverse-phase HPLC system consisted of a multiple solvent delivery pump(CM4000 LDC-Milton-Roy, Riviera Beach, FL, USA), an injector (Rheodyne)equipped with a 100 μl loop and a C-18, 25×4.6 mm (Alltech,Deerfield, IL, USA) Spherisorb column. The pigments were detected with a variable-wavelength spectrophotometer (LDC-Milton-Roy) set at 436 nm. Data were recorded and processed by a digital-analog converter and software(Jasco-Borwin, La Fontanil, France). Two solvents were used in the system:solvent A consisted of 30% 1 mol l^-1 ammonium acetate (Sigma) in double-distilled deionized water and 70% methanol (HPLC grade; Bio Lab,Jerusalem, Israel); solvent B consisted of 30% ethylacetate (HPLC grade, Bio Lab, Israel) and 70% methanol. The solvent program exhibited a linear increase in solvent B from 20 to 60% in 7 min, a plateau at 60% for 5 min, a linear increase in solvent B from 60 to 100% from 12 to 20 min and was then maintained at 100% solvent B for 20 min.

Pigment identification was facilitated by the use of ChromaScope (BarSpec,Rehovot, Israel), a spectral peak analyzer that allows scanning of pigments separated by the HPLC system in the range of 360-700 nm. The spectral data of the separated peaks and their retention times were used as the parameters for peak identification using published data(Rowan, 1989; Jeffrey et al., 1997) and were compared to standards (Yacobi et al.,1996). The major pigments found in this study were isolated online; their concentrations were determined and subsequently used for quantification. Quantification of the chromatograms was facilitated by injection of known pigment concentrations into the HPLC system and calculation of the response factor based on the area under the peak. All pigment concentration calculations are presented as the mean of duplicate measurements. Individual measurements did not differ by more than 10%. The pigment data collected by HPLC were expressed in relation to moles of chlorophyll a (Chl a).

Differences between SOD and CAT activities were estimated using one-way analysis of variance (ANOVA) and the Student-Newman-Keuls (SNK) test at the 5%significance level. The means of enzyme activity for each coral species were converted using the arcsine function. The Pearson correlation factor was used in order to determine the relationship between light intensity and enzyme activity (Sokal and Rohlf,1981). Values are represented as means ± s.d.

Light intensity at the 4-5 m-depth water on the reef in Eilat (Gulf of Aqaba) on 3-5 March 2002 reached a maximum of 1100 μmol quanta m^-2 s^-1 (Fig. 1A), while in July-August it reached a value of ∼2000 μmol quanta m^-2 s^-1. The intensity values of UVA + UVB radiation in February 2001 at the 5-m depth is presented in Fig. 1B. The net oxygen evolution rate tracked the diurnal course of solar energy flux(Fig. 1A), while respiration rates did not change significantly during the night. Typical photosynthesis versus irradiation curves for colonies of Favia favus, Plerogyra sinuosa and Goniopora lobata are presented in Fig. 1C-E. The optimum light intensity (E_opt) at which photosynthesis is at a maximum without exhibiting photoinhibition was similar for all three species,approximately 1100 μmol quanta m^-2 s^-1, which is close to the maximum radiation intensity at this depth. The compensation point(E_com) at which photosynthesis and respiration rates are equal was reached at approximately 10-15% of the optimum light intensity,between 119 and 155 μmol quanta m^-2 s^-1. Favia favus and P. sinuosa were light-saturated near 700 μmol quanta m^-2 s^-1, while G. lobata reached almost 950 μmol quanta m^-2 s^-1. The maximum photosynthesis rate of F. favus colonies (0.11±0.03 μmol O₂cm^-2 min^-1) was lower than that of both G. lobata and P. sinuosa (0.22±0.05 and 0.22±0.13μmol O₂ cm^-2 min^-1, respectively). The highest P_g/R_24h ratio was found in F. favus (4.37±0.68), while P. sinuosa(3.78±0.18) and G. lobata (3.48±0.23) exhibited lower values (one-way ANOVA; P<0.05; Table 1). Net oxygen production followed ambient light intensity at the same depth of 5 m, and the peak of oxygen evolution was reached at midday. Photoinhibition was not observed even at noon (Fig. 2).

In situ measurements of chlorophyll fluorescence yields revealed a diel variation in response to ambient illumination. The variable fluorescenceΔ F′/F_m′ declined from ∼0.40(arbitrary units, a.u.) in the early morning to ∼0.07 at midday for P. sinuosa, ∼0.25 for F. favus and ∼0.1 for G. lobata (Fig. 2). Decreases in ΔF′/F_m′ values were found during the midday hours (between 10.00 and 14.00 h) in F. favus and G. lobata (Fig. 2A,C),while in P. sinuosa the values were low until late afternoon (16.00 h; Fig. 2B). In all three species from early afternoon, there was an increase inΔ F′/F_m′ values as light intensity decreased. In all three species,Δ F′/F_m′ values were inversely correlated to the net oxygen production(Fig. 2A-C). The functional absorption cross section of PSII (σ_PSII′) decreased slowly with increasing irradiance, with minimum values being observed between 11.00 and 14.00 h in all coral species examined(Fig. 3; data not presented for G. lobata). The recovery in σ_PSII′ occurred in late afternoon, when light intensity decreased down to less than ∼500μmol quanta m^-2 s^-1(Fig. 3). Theσ _PSII′ values did not change significantly during the night and remained constant (data not presented). The irradiance-induced decrease in ΔF′/F_m′ was accompanied by an increase in the quantum yield of NPQ in all corals (Figs 2, 3). In F. favus(Fig. 2A) and G. lobata (not presented),Δ F′/F_m′ decreased to a level of∼0.1-0.25 at a maximum light intensity of 2000 μmol quanta m^-2 s^-1, while in P. sinuosaΔ F′/F_m′ values were less than∼0.1 (Fig. 2B). In F. favus, NPQ maximum values reached ∼0.4, whereas P. sinuosaexhibited higher NPQ values of ∼0.65 when light intensity was above∼1800 μmol quanta m^-2 s^-1(Fig. 3).

Superoxide dismutase (SOD) activity in the animal tissue of the three species exhibited a similar trend (Fig. 4) ofincreasing activity with irradiance and net photosynthesis productivity (one-way ANOVA for F. favus, P<0.0001; for G. lobata P<0.05; for P. sinuosa P<0.001). In tissues of P. sinuosa and F. favus, SOD activity values observed between 10.00 and 14.00 h were significantly higher than during the rest of the day and night (SNK, P=0.006 and P=0.009, respectively). SOD activity in G. lobata coral tissue was significantly higher only in the afternoon (16.00 h) and during the night (05.00 h) (SNK, P<0.05; Fig. 4B). SOD activity in the isolated zooxanthellae of the three coral species did not differ significantly during day or night (one-way ANOVA, P>0.05),although a clear trend of increasing activity with increased light intensity could be observed (Fig. 4A-C).

Catalase (CAT) activity revealed a similar pattern, i.e. increased activity in the coral tissue with increasing irradiance and oxygen evolution(Fig. 5A-C). The highest value was observed during midday in F. favus (SNK, P<0.0001)and in G. lobata (SNK, P<0.05), while in P. sinuosa the highest activity values were observed in the morning and midday. The morning values were significantly higher than night-time/dawn values (SNK, P=0.034). CAT activity in the symbiotic zooxanthellae of P. sinuosa peaked during the morning hours(Fig. 5C) compared with midday and night-time (SNK, P=0.037). Although light intensity and CAT activity were correlated in F. favus zooxanthellae, the results were not significant (P=0.07; Fig. 5A). G. lobata did not reveal a significant increase in CAT activity in any of the measurements performed in this study(P=0.34). Mean diel values of SOD and CAT activity in the host and in the zooxanthellae per species are presented in Table 2. A high correlation was found between SOD activity and light intensity in both the animal tissue and algal cells (Table 2). A correlation was found between CAT activity and light intensity in the tissue of F. favus (Pearson's correlation, r=0.811) and G. lobata (r=0.969), while P. sinuosa exhibited a low correlation (r=0.594). CAT activity of the zooxanthellae exhibited a high correlation with light intensity only in F. favus(r=0.951), whereas in G. lobata (r=0.441) and P. sinuosa (r=-0.055)(Table 2), the correlation was low.

The pigment suite of zooxanthellae revealed in the HPLC chromatograms demonstrated the typical components of dinoflagellates(Table 3). The major pigments were Chl c₂, peridinin, dinoxanthin, diadinoxanthin (DD),Chl a and β-carotene. Chlorophyllide a (Chld a) was present in all chromatograms, and its concentrations were pooled with those of Chl a. Even though the samples were preserved in liquid N₂ immediately after collection of the coral, we were not able to detect diatoxanthin, which is a signature pigment for the photoprotective response. It is known that this pigment vanishes within minutes if algal cells are allowed to relax in low light(Goericke and Welschmeyer,1992). The statistical analyses did not detect any significant differences (one-way between pigment compositions ANOVA, P>0.05)(Table 3). The three coral species were almost uniform in their pigment concentration relative to Chl a throughout the day (Table 3).

Diel changes in response to the course of sunlight are among the most interesting phenomena of coral physiology. This includes patterns of calcification and photoacclimation. The latter comprises changes in photosynthesis and respiration rates, P_g/R_24h as well as pigment content(reviewed by Falkowski et al.,1990; Barnes and Chalker,1990). Changes in enzymatic activity and behavioral responses,such as variations in tentacle expansion and contraction(Levy et al., 2003), are all light dependent and have been studied for many years(Brown et al., 2002a; Stambler and Dubinsky, 2004). Hoegh-Guldberg and Jones(1999), Lesser and Gorbunov(2001) and Gorbunov et al.(2001) monitored diurnal in situ chlorophyll fluorescence yields. Brown et al.(1999) followed the diurnal changes in photochemical efficiency and xanthophyll concentrations.

The present study highlights the dynamics of the diurnal cycle of corals in situ and their mechanism of continuous `tuning' in relation to the course of light and UV radiation and their penetration into shallow water. The three corals studied, Favia favus, Plerogyra sinuosa and Goniopora lobata, are common reef-building corals in the Red Sea,from the surface to a depth of 30 m. The distributed three species exhibit a diurnal course in oxygen evolution parallel to that of light intensity. Different P_g/R_24h ratios were shown to occur as a result of polyp expansion/contraction, as described by Fabricius and Klumpp (1995) for soft corals. Their data reveal that colonies with expanded tentacles have higher P_g/R_24h values (1.3±0.02) than colonies with contracted tentacles (1.14±0.02). The P_g/R_24h ratios presented in Table 1 for the three species range from 3.48 to 4.37. Generally, P_g/R_24(h) ratios in hard corals range between 2 and 4. Scleractinian corals have much higher photosynthetic rates than soft corals (Fabricius and Klump,1995), which may be related to a far higher density of zooxanthellae cell in host tissue in scleractinians (usually in the range of 0.5-1×10⁶ zooxanthellae cm^-2) than in most octocorallians (with the possible exception of Tubipora musica),which is reflected in the higher host respiration/symbiont photosynthesis ratio. The stony corals F. favus and P. sinuosa expand their tentacles nocturnally and remain open until sunrise, whereas G. lobata remains expanded continuously(Levy et al., 2003). P. sinuosa can also expand and contract its photosynthetic vesicles in response to varying light intensities, as described by Vareschi and Fricke(1986). Therefore, the different ratios of oxygen evolution and respiration observed by us are likely to reflect polyp behavior such as expansion or contraction of the tentacles(Levy et al., 2003).

Differences in metabolism, especially in respiration rate, can also be influenced by feeding behavior. Porter(1976) suggested that corals with large polyps depend on heterotrophy more than those with small polyps(see Lesser et al., 2000);however, no support for this hypothesis was found in subsequent studies(Haramati et al., 1997). Goniopora lobata polyps are small (∼5 mm) compared with those of F. favus (∼11 mm) and P. sinuosa (up to ∼20 mm). However, G. lobata has a greater tentacular surface for feeding than is found in other small polyp corals(Sebens et al., 1997). From the P_g/R_24(h) ratios measured, it seems that G. lobata exhibits a lower ratio with a small polyp size. This can be related to the changes in tentacle state. The tentacles can be expanded or contracted during daytime, which can influence photosynthetic rate and reduce the ratio of P_g/R_24h (Levy et al., 2005a). Plerogyra sinuosa, with the largest polyp size, had similar P_g/R_24h values(Table 1) to those of G. lobata. This results from daytime expansion of the photosynthetic vesicles.

Respirometer data on coral metabolism in shallow waters do not indicate any photoinhibition during the midday hours of July-August 2001(Fig. 2A-C) and March 2002(Fig. 1A-E). A similar finding was noted by Goiran et al.(1996) for the symbiotic coral Galaxea fascicularis and by Muller-Parker(1984) for the sea anemone Aiptasia pulchella. Absence of photoinhibition in both intact coral colonies and whole reefs is common and predictable: while branch tips and surface zooxanthellae are exposed to the full, measured irradiance, most of the symbionts in branched colonies and deep-seated algae in massive species are still light limited. The result is that photoinhibition of high-light-exposed cells is masked by the status of most of the zooxanthellae,a situation akin to that of entire reefs, where surfaces may be photoinhibited, while light in reef crevices and shaded parts of colonies never reaches saturation levels.

Despite the fact that oxygen measurements did not show photoinhibition, the FRRF measurements demonstrate a clear pattern of midday depression inΔ F′/F_m′ associated with increasing light intensity accompanied by an increase in non-photochemical quenching. As ambient light intensity increased, both steady-state(F′) and maximum (F_m′) fluorescence yields decreased due to non-photochemical quenching (NPQ) of chlorophyll fluorescence (Lesser and Gorbunov,2001; Levy et al.,2004). Again, this difference between the data integrated from whole colonies, measured by the respirometer, and those of the FRRF, which observes only a small section of the colony at its exposed surface, is to be expected and underscores the benefits inherent in combining both methodologies concomitantly. NPQ is one of the protective mechanisms against oxidative damage that is associated with the ability to dissipate excess energy as heat(Falkowski and Raven, 1997). The heat from the NPQ process is partly dissipated via the xanthophyll cycle, in which diadinoxanthin is converted into diatoxanthin; in this way, corals can keep excess light energy from damaging components of the photosynthetic pathway (Ambarsari et al.,1997; Brown et al., 1999, 2002b). The diurnal changes in the chlorophyll fluorescence cycle (Fig. 2) are similar to those obtained in earlier studies by Brown et al. (1999), Hoegh-Guldberg and Jones (1999), Lesser and Gorbunov (2001) and Gorbunov et al. (2001). The higher values of NPQ found in P. sinuosa as compared with F. favus(Fig. 3) indicate that this species is capable of dissipating excess excitation energy more efficiently through the xanthophyll cycle, and by doing so, this coral might be less prone to bleaching events (see Warner et al.,1999). Diel periodicity may lead to differences in the pigment concentrations due to circadian rhythm, as reported for phytoplankton(Marra, 1980; Owens et al., 1980; Post et al., 1984). In the present study, photosynthetic pigments did not exhibit temporal patterns in concentration (Table 3). Again,it may well be that such changes did take place but went unnoticed because of the different exposure of the zooxanthellae to light on a mm-to-cm scale,which could not be resolved by us.

High oxygen levels within the host tissues of symbiotic cnidarians are a result of the high rate of oxygen production in zooxanthellae exposed to high irradiance (Lesser et al., 1989a,b). Photosynthetic oxygen evolution by the zooxanthellae in hospice can raise tissue oxygen concentrations to ≥200% saturation(Dykens and Shick, 1982; Kuhl et al., 1995; De Beer et al., 2000). Such levels of photosynthetically produced molecular oxygen act synergistically with sunlight, especially UV radiation, in the presence of photosensitizing agents (e.g. flavins and chlorophylls) to produce active harmful forms of oxygen (Foote, 1976; Asada and Takahashi, 1987). Increased SOD and CAT activity is associated with increased light and UV intensity (Fig. 3; Table 2), suggesting higher O₂^- radical and H₂O₂ production(Levy et al., 2005b). Therefore, a diurnal change in SOD/CAT activity in the host corresponds to the diurnal patterns of solar radiation both in the visible and the UV domains, acting as a protective mechanism against oxygen radicals. It is likely that the animal host is more susceptible to oxidative damage than a plant, where high oxygen tension is the norm; indeed, SOD activity in the symbiotic algae did not show a significant diurnal pattern,although it did correlate with light intensity (Figs 3, 4; Table 2). The fact that SOD/CAT activity was higher in the morning suggests that, as light increases ahead of any photoacclimation processes, the algae are susceptible at that time to the increasing irradiance and UV penetration. Oxidative stress and apoptosis occur in both the animal and the zooxanthellae(Dunn et al., 2002; Lesser and Farrell, 2004). However, we still believe that oxidative stress is primarily an animal response, aggravated by the presence of symbiotic algae that can cause hyperoxia (Nii and Muscatine,1997). The work done by Nii and Muscatine(1997) on sea anemones suggested that exposure to chronic stress may cause symbionts to release oxidants directly and that oxidative stress in the sea anemone A. pulchella is primarily an animal response. Downs et al.(2002) proposed that algae-generated H₂O₂ could diffuse from the algal symbionts into the coral cytoplasm. Once inside the coral cytosol,H₂O₂ can be neutralized by protective enzymes or converted into the hydroxyl radical, which can eventually lead to coral bleaching (Levy et al., 2005b). Therefore, the rise in free-radical-scavenger enzymes in the host is likely to be a defense reaction to the accumulation of free radicals within the host tissue. The difference in enzymatic activities between species indicates that although the acclimation process depends on environmental conditions, genetic variation among species controls the function of these enzymes and, thus, plays a role in the survival and the depth distribution of different corals in the coral reef.

Excessively high photosynthetically available radiation (PAR) is potentially detrimental to photosynthetic organisms. The photosynthetic apparatus saturates rapidly at relatively low flux levels(Hoegh-Guldberg and Jones,1999). Above these levels, light energy damages the photosynthetic machinery (Walker, 1992; Foyer et al., 1994; Osmond, 1994). Short wavelengths (290-400 nm) and elevated temperature may act synergistically with intense radiation, causing destructive effects in both partners (host and symbiotic algae; Jokiel, 1980; Shick et al., 1996). These environmental stresses may lead to photoinhibition, a reduction in photosynthesis and an increase in respiration, and to a change in pigment content. The outcome of these environmental effects leads to the evolution of several protective mechanisms. These include the production of mycosporine-like amino acids for which a UV-absorbing function has been inferred (Dunlap and Chalker,1986; Lesser and Shick,1989b; Lesser,2000; Shick and Dunlap,2002) and the development of active oxygen scavenging systems(Shick et al., 1996). The present study indicates that shallow water corals respond rapidly to the prevailing internal and external conditions and accomplish fine-tuned adjustments to ensure that they continue to function effectively.

In this fine-tuning, the coral host plays the role of an important guardian, as can be observed from the increase in the activity of two free-radical scavenging enzymes, SOD and CAT. The adjustments occur within a few minutes to several hours, while changes in the pigment content may occur on a scale of days. In addition, increasing NPQ while reducing the cross section of PSII are important mechanisms, protecting both algae and host from harmful excess energy. We suggest that the symbiotic dinoflagellates within the host live in an environment that protects them from chronic photoinhibition (Gorbunov et al.,2001). This symbiotic association owes its success partially to the effectiveness of the host's anti-oxidative defenses and partially to the photoacclimative plasticity of the zooxanthellae. Since corals have evolved in stable tropical environments, we assume that even slight changes in the environment, such as rising sea-surface temperatures, increased UV load,lowering of ambient pH and bacterial infection, occurring on time scales too short to allow the necessary genetic adaptation and extending outside the phenotypic ranges covered by acclimation, can upset this unique system.

We thank the staff of the Interuniversity Institute of Marine Sciences in Eilat for their support during our fieldwork. We acknowledge the help of E. Perel, N. Schwalb, M. Ohavia and A. Rivlin for technical assistance, E. Levin for helping in the fieldwork, M. Ben-Zion for helping us with computer programming, and D. Iluz for helping us with light measurements. We also thank A. Perelman for his suggestions on the enzyme work.