Home - Bleaching Part 1 - References

---

REVIEW ON CORAL REEF BLEACHING (Part 2)

BY

MARTIN PECHEUX (1997)

(martinpecheux@minitel.net)

 

 

TABLE OF CONTENTS

SECOND PART : BIO-PHYSIOLOGICAL KNOWLEDGE

A) BLEACHING EXPERIMENTS

Warming

Cold

Light

Carbon dioxide

Oxygen

UV

Salinity

Miscellaneous (darkness, inhibitors, starvation, burial, others)

Expulsion of "surpernumerary" symbionts

B) SOME CONSIDERATIONS ON THE PROCESS OF BLEACHING

Time dynamic of bleaching process

Immediate reactions

Mid- and long-term reactions

Oxygen tension or low photosynthesis/respiration ratio ?

Temperature effects

Mechanisms of expulsion

Are expulsed symbionts healthy ?

C) THEORETICAL BIOLOGICAL BACKGROUND

Photoinhibition and photooxidation

Fluorescence quenching

Light influence

Temperature influence

Carotenoid pigments and the xanthophyll cycle

Carbon fixation and photorespiration

The carboxysome pist

Carbon uptake or carbon limitation ?

Environmental conditions and carbon fixation

Relations between carbon fixation and photoinhibition

Physiological effects of water agitation

Lipid phases and permeability

Stress proteins

Warm-induced bleaching of Euglena, a model ?

THE CO2-PHOTOINHIBITION HYPOTHESIS

CONCLUSION

Annex I: A preliminary CO2 coral bleaching experiment

Annex II: CO2 effects on PS II in symbiotic foraminifers

Annex III: Formation of dense, hot, hypersaline surface water

Annex IV : Implications for the carbon cycle

Annex V : Shell abnormalities in large foraminifers, Mauritius, 1989

Annex VI : Some recommendations

SECOND PART : BIO-PHYSIOLOGICAL KNOWLEDGE

 

Few hypothesis on the mechanism of bleaching have been so far proposed:

- elevated oxygen tension in the symbionts and the host, leading to internal damages by oxygen radicals, hypothesis first advocated by Sandeman (1988);

- some emphasis has been put on the observed fall of the photosynthesis/respiration ratio, producing an energetical imbalance (Jokiel and Coles, 1990), but the primary reason of this reaction remains unexplained;

- a direct effect of warm temperature on membranes, perhaps through failure of ions pumps, elevated intracellular calcium concentration and cell adhesion dysfunction (Muscatine, 1991, Muscatine et al., 1991, 1992, Gates et al., 1992);

- deleterious effects of UV, perhaps becoming harmful only at high temperature (Lesser et al., 1990);

- breakdown of the mechanism of symbiont recognition by the host was suggested by Davies (1992), but the basic mechanisms of signal exchange, recognition, specificity, and regulation in endosymbiotic systems are almost unidentified (see Reisser, 1992);

- recently, symbiont photoinhibition under elevated temperature and light was designated as the first mechanism of bleaching (Pêcheux, 1992 widely distributed first version of this report, 1993, 1994, Iglesias-Prieto et al., 1992, Iglesias-Prieto, 1995, in press, Warner et al., 1996, B.E. Brown, Panama 1996 oral com., Strasser et al., in press) caused by CO2 rise (our works). This early event might be followed by energy inbalance, disruption of recognition (Iglesias-Prieto, 1995, in press) or perhaps by symbiont toxic stress signals.

 

 

A) BLEACHING EXPERIMENTS

Bleaching can be induced in a reproductible manner in laboratories. The most studied factors were temperature, salinity, light and UV. Physiological responses such as symbiont expulsion, photosynthesis and respiration are now more or less known. Characterization of bleaching at the biochemical and enzymatical levels is still very poor.

 

Warming

Isolated symbionts

The first important learning is that freshly isolated zooxanthellae (FIZ) are viable and photosynthetizing at 35°C (Sandeman, 1988), a temperature which provokes the rupture of the symbiotic association. This remains however to be confirmed at the time scale of days to weeks at which bleaching occurs. Comparison between FIZ of Acropora palmata, which has exhibited little bleaching in situ, and FIZ of Porites porites, moderatly bleached, showed that i) the photosynthetic rate between 20°C and 35°C of the former ones is relatively constant whereas it increases rapidly in the latter; ii) the first ones are in addition photoinhibited at high irradiance level; iii) O2 inhibits photosynthesis of Acropora FIZ more than that of Porites FIZ. Nonetheless, a threefold reduction of photosynthesis when O2 saturation is raised from 90% to 120% (at 25°C and saturation light) appears absolutely unusual for FIZ, and for photosynthesis in general, and suggests other factor involvement (such as CO2 change with O2 bubbling). Sandeman concluded that symbionts with low level of regulation of photosynthesis induce high internal oxygen pressure and toxicity, bringing in bleaching. Two different regulations were put into evidence: photoinhibition (by light) and photorespiration (by O2).

 

Saks (1981) found that cultivated Chlorella symbionts of the Red Sea foraminifers Amphistegina lessonii and Amphisorus heimprichii are viable from 15°C to 38°C and growing between 22°C-33°C, whereas the maximum temperature in situ in Aqaba is only 27.3°C. It can be concluded that these isolated symbionts are not sensitive per see to bleaching temperature, but it must be quoted that a) no bleaching is reported in Aqaba, though drastic fall of foraminifer standing crop (J. Erez, com. pers.) ; b) Chorella are accessory symbionts, representating less than a few percent of symbionts, and absent in Amphistegina from Hawaii (Lee et al., 1980).

 

Cultivated Symbiodinium of the jellyfish Cassiopea showed a photosynthesis increase with temperature till 30°C with a normal Q10 of 2.2, but was null at 35°C, at contrast to previous data for FIZ. No measure of photosynthetic O2 was made between this two temperatures, but DCMU-induced fluorescence after 45 minutes in darkness began to decrease around 32°C (Iglesias-Prieto et al., 1992). Impairment of the photosystem II was implicated in bleaching after fast fluorescence kinetics study of cultivated symbionts of Montipora (Iglesias-Prieto, 1995, in press). Overall quantum efficiency Fv/Fm decreased at high temperature of 34-35°C. Initial fluorescence rose at 35°C in dark and maximum fluorescence as well as initial rate of photochemistry at 37-38°C in light, corresponding to thermal inactivation, probably at the PS II donor side, the O2 water-splitting complex. This indicated also only a limited light-induced thermoprotection. But more relevant to mass bleaching, at 32.5°C in light, Vj, an indicator of photoinhibition at Qa-Qb site, rises strongly with time. After few hours it saturates at a very high level. At contrast, in the dark for 20 hours, there is emergence of a very rapid K transient indicative of damage at P680 or O2 site. A remark must be made that cultures were devoid of bicarbonate whereas experiments were done with high 5mM NaHCO3.

 

Symbiotic associations

The first experiments on temperature resistance of corals were run by Mayer and Vaughan. Mortality rather than bleaching was observed after short-term temperature shocks. In darkness, with a warming rate of 2°C per hour, two dozens of Australian and Caribbean corals species resisted without apparent injury between 35.7°C and 38.2°C and died between 35.8°C and 38°C (Mayer, 1914, Table 6, Mayer, 1918a, Table 6, 10), or, without progressive warming, between 34.7°C and 38.2°C during one hour (Mayer, 1918b). Species sensitivity in these experiments bears some similarity with the bleaching sensitivity, but with important exceptions, as for example Siderastrea spp. which were ranked as resistant. The resistance was the same either in darkness or in light, or with 1.7 to 6.6 cc O2/l. He found that the rate of respiration was correlated with the sensitivity to temperature, except for Pocillopora damicornis (Mayer, 1918b, 1924, incorrectly spelled "Mayor"), and hypothetized "that acid may accumulate in the tissues, due to the increased metabolism under the influence of heat and thus poison the coral". The lethal temperature in 1 hour of 12 species of alcyonarians groups them in 3 groups: 34.5-35°C for Eunicia and Plexaura, 37-37.5°C for Gorgonia, and 38.2°C for Briareum asbestum (Cary, 1918). He found that the respiration rate increased with high temperature but was not correlated to temperature sensitivity.

 

Death of various species of corals were observed at 40°C, and death or bleaching at 36°C within 1/2 to 2 hours by Yonge and Nicholls (1931a). They noted that some wandering cells, which contain the symbionts in normal time, were devoid. Expulsed zooxanthellae appeared healthy. They suggest that starvation of CO2, nitrogen or phosphate is a likely cause of the bleaching. Low photosynthesis/respiration ratio were observed during the warm summer in (involuntary) experiments (Yonge et al., 1931), leading to the belief of a low contribution of zooxanthellae as an organic carbon source.

Bleaching of Zoanthus sp. in culture at 26.5°C occurred when submitted to a sudden warming of 30°C during 2 days (Reimer, 1971). Thereafter, extrusion took place continuously during three months. This author also stated that bleaching occurred in Palythoa sp. when culture temperature was raised by only 1°C, as well as with salinity increase from 33°% to 35°%, but without further precision. Bleaching by warming was also seen in the sea anemone Anthopleura (Buschbaum, 1968, PhD., in Muscatine 1974). Röttger (1972) reported bleaching in the large foraminifer Heterostegina depressa at 31°C.

 

Impact of thermal effluent of a power plant in Biscayne Bay, Florida, was surveyed by Thorhaug et al. (1973) during 4 years. They showed that coelenterates (probably mostly alcyonarians, cf. Roessler, 1971) are more sensitive to high temperature than fishes, echinoderms or porifers: 50% loss of coelenterate species occurred at temperature about 1-2°C lower than other taxons. Solenastrea hyades and Siderastrea sidera were pale, damaged or dead in "area of marginal influence on other biota" (Purkerson, in Johannes, 1975).

 

Around Kahe power plant effluent in Hawaii, nearly all corals died inside the + 4-5°C zone, bleached inside the + 2-3°C one, and were unaffected at the + 1-2°C one (Coles and Jokiel, 1974, Coles, 1975). Effects were correlated with absolute maximum temperature and not with the fluctuations, which were of the order of 3-4°C in minutes. The relative sensitivity among coral species followed their metabolic rate. Bleached corals sometime retained their carotenoids while loosing totally their chlorophylls, and sometime retained chlorophyll a while losing chlorophyll c. In Porites lobata, pale specimens had greater carotenoids/chlorophylls ratio than either bleached or normal ones (Coles and Jokiel, 1974).

 

Thermal plant effluents in Guam were used by Neudecker (1981) in transplant experiments. Acorpora formosa transplanted from a zone with "periodical excess of 32°C" to "36°C not uncommon" had zooxanthellae and tissue sloughing, with death within a few hours to one week. Pocillopora damicornis bleached in one week and died in one month, and some P. andrewsi survived 2 months but showed no growth, in correspondance with the growth rate.

 

The team of Coles, Jokiel and collaborators studied extensively the detrimental influence of temperature in laboratory. Coles et al. (1976) observed the first signs of bleaching for various Hawaiian corals after 90 hours at 31.3°C, and for Eniwetok ones after 60 hours at 32.5°C. Death was in both cases produced by 2°C or more above summer maximum and 3°C above summer mean. Each degree increase reduced the survival time to about a third, with a difference between death and bleaching temperature dose of 1°C. If this relation holds, bleaching in the order of one month (as observed in situ) is induced by 1°C more than the in situ maximum temperature. They addressed the question of whether the Eniwetok-Hawaiian difference has a genetic or physiological adaptation basis. In a subsequent one-month experiment with full natural light on three coral species of Hawaii (Jokiel and Coles, 1977), some paling appeared after several weeks at 29.7°C, and bleaching and death above 29.6°C (within range 30-33°C) (mean summer is 26.5-27°C with maximum of 27.5°C). No influence of colony size was noted, with the important exception of newly-settled colonies which nearly all died in one month after 7 days at 30-32°C, whereas all adults fully recovered in 2 months. There is a very sharp optima around 26-27°C for coral larval settlement (Jokiel and Guinther, 1978). Settlement was also enhanced after a short passage in thermal plant system (Coles, 1984). On the other side, thermal shock provoked planulae release, or maybe in fact abortion (Edmuson, 1929, 1946). In any case, no normal reproduction probably occurs during bleaching stress conditions.

 

Other relevant remarks (Jokiel and Coles, 1977) to mass bleaching are that, at 32°C, retraction of polyps lasted only one day; zooxanthellae probably migrated from coenosarc area, which faded, to polyps, which first darkened; and bleaching occurred first on the upper face. Growth rate, measured by buoyant weight technique, was not correlated with bleaching sensitivity: whereas Montipora verrucosa grows the fastest, the bleaching sensitivity ranked P. damicornis > M.verrucosa > Fungia scutaria. Calcification was reduced near sublethal temperature. Retrospectively, one of their conclusions must be quoted: "An increase [of temperature] as large as 2°C would not decrease net annual growth".

 

Coles and Jokiel (1977) measured the photosynthesis and the respiration with O2 evolution in two alternate 10-40 minutes light/dark cycles, first at ambient temperature (24°C for Hawaiian corals and 28°C for Eniwetok ones), then, after 15 minutes of acclimatation, at 5 temperatures between 18°C and 31°C. Corals showed an increase of photosynthesis and respiration with warming. Both processes increased more or less similarly in Hawaiian corals, whereas Eniwetok ones increase 1.6 to 5 time photosynthesis more than respiration. Differences in the responses between Hawaiian and Eniwetok P. damicornis were noted, but not for M. verrucosa. The P:R ratio decreased from 18°C to 31°C in Hawaiian corals. This decrease occured only above 25°C in Eniwetok corals. The P:R value at high temperature was correlated with bleaching sensitivity. There are some contradictions in their data because the photosynthesis almost always increased more than the respiration with temperature (text, table 1) whereas the P:R ratio is said to fall with temperature.

 

In other experiments on Montipora verrucosa from Hawaii, Cole and Jokiel (1978) observed mortality within 2-4 hours at 33°C. 10% of the colonies died, and 50%-80% bleached or paled in 22-56 day exposure at 28.1°C under natural light (in disagreement with the finding of their precedent work). Paling but not bleaching was also present at natural temperature (24°C), particularly at full light. They acclimated nubbins during 2 months at temperature ranging from 20°C to 28°C. Carbon fixation, growth and chlorophyll content showed an optimum around 26°C, but the number of symbionts was not influenced.

For Marcus and Thorhaug (1981), Porites compressa from Hawaii began to bleach in 5 days at 32°C (4-5°C above mean summer), and bleached at 33°C. P. porites from Florida began to bleach in 10 days at 33°C (2-3°C above mean summer), with total bleaching at 34°C. The low light intensity (175µE/m2.s) in experiments corresponded to the depth of collection. The percentage of bleached colonies increased logarithmically with the time of exposure, and the overall effect was correlated with the speed of bleaching. About 1°C above threshold doubles the speed or proportion of bleaching at any given time.

Incubation during one day of Porites from Okinawa by Kato (1987) produced no bleaching at 29°C, 10% at 31°C, 30% at 32°C and 100% at 33°C. The production of mucus-sheet and the bleaching response were parralell. 20% death occured at 32°C within one day, and 37% at 33°C within two days.

 

Hoegh-Guldberg and Smith (1989) worked with colonies of Stylophora pistillata and Seriatopora hystrix from Lizard Island, Great Barrier Reef. In situ mean summer water temperature of Lizard Island is around 29°C (and up to 29.8°C, Wolanski and Pickard, 1985). They provoked mortalities of 10-20% at 30°C in 4 days but without signs of bleaching, 40-50% at 32°C with paling within 2 hours, and 100% within 8 hours at 34°C. Colonies incubated at 32°C showed a 20%-40% reduction of zooxanthellae number, with no change of chlorophyll per zooxanthellae, except those incubated at 30°C. There was even more symbionts in one individual of Stylophora (cf. fig. 7). Respiration increased approximately linearly with temperature in both species. It was respectively 2.7 and 6 times greater at 32°C than at 27°C. This corresponds for Q10 value of 7.3 and 36. These increases were expressed at 32°C after 6 hours, or less. In contrast, photosynthesis per zooxanthellae stayed about the same, except in Stylophora at 30°C (+60%). Expulsion of zooxanthellae increased in 30 minutes at 30-32°C up to 100 times its normal low level, and, in a second step, after 4 hours, accelerates sharply to 500-1000 times its value, corresponding to 10-13% zooxanthellae loss per day. After 23 days at most after transfer back to normal temperature, recovery was evident, with even slightly more symbionts than at the begining. Thus, a temperature 1°C higher than mean summer one induced the first physiological effects (higher expulsion and respiration) without bleaching in 4 days, 3°C higher inducing bleaching and 5°C higher being quickly lethal.

 

Glynn and D'Croz (1990) incubated Pacific Pocillopora damicornis from the Gulf of Panama to up to 32°C and others from Chiriqui to up to 30°C (this locality has lower temperatures) during 10 weeks. Number of symbionts decreased by 3-4 order of magnitude. Only at 32°C in Panamian corals was the chlorophyll per zooxanthellae reduced (by 2-3 order of magnitude). The decrease was discreete in Chiriqui corals. Decreases of symbiont number, pigment per symbiont and protein per surface followed a logarithmic evolution with time, with no initial time lag. The histopathological deterioration were first observed in mucous cells, then general atrophy-necrosis occurred. The nuclei of necrotic zooxanthellae remained unstained in slide. Okinawan Acropora valida and Pocillopora damicornis paled only slightly momentarly the third week when exposed to 30.1°C, above 1980 in situ bleaching at 29.6°C (Glynn et al., 1992). They died completly at 31.3°C.

 

Lesser et al. (1990) tested the bleaching-sensitive zoanthid Palythoa caribaeorum from Bermuda at 29-33°C during ten days. Compared to the 26°C control, there was both a decrease of symbiont number (-31%) and of pigment content per symbiont (-50%), together with a lowering of the chlorophyll a/c2 ratio from about 2.3 to 2.

Shock warming from 16°C to 26°C for 6 days of the temperate Anemonia viridis induced paling (Suharsono et al., 1993). It was observed increase of primary and secondary lysosomes associated with membranes, more lipid in mesenterial area (but for both symbiotic and aposymbiotic anemones), space around symbionts associated with the algal membrane, and large number of degenerate zooxanthellae, but equivocal evidence of their digestion.

 

Fitt and Warner (1995) and Warner et al. (1996) studied various coral species, the sensitive Montastrea annularis, Agaricia lamarcki and the more resistant M. cavernosa, A. agaricites, Siderastrea radians under 28-36°C for 2-4 days with light of site origin (400-500µE/m2.s). In one series of experiment up to 34°C, the symbiont number stayed constant, together with a slight increase of chlorophyll per zooxanthellae; in another, after 1 day at 32-34°C or 2 days at 30°C, only sensitive species losed half of their symbionts, containing one half-one third chlorophyll content of control. To be short, O2 photosynthesis, P/R ratio and PAM fluorescence were stable at 30°C. Falls occured in sensitive species at 32°C, in the others at 34°C. Lower Fv/Fm was not due to an initial fluorescence rise (indicative of direct thermal damage). Analysis is given of the measured fluorescence quenching: adaptation mechanisms for energy dissipation (non-photochemical quenching, Qnp) rose more in S. radians than in M. annularis. But it must be said still to lower level. DCMU-induced fluorescence of freshly isolated symbionts is well curvilinear-correlated with whole M. annularis colonies Fv/Fm, a surprise for me as I found a long lasting half decrease of Fv/Fm after isolation (unpublished results). Temperature shock (1°C per 20 minutes up to 38°C) indicated a critic temperature of Fo rise above 34°C in M. annularis, not in S. radians. These works enlight the inactivation of photosystem II, supposed at O2 site or D1 protein, before any other sign of bleaching.

 

Effects of slowly rising the temperature on symbiotic large foraminifers were studied by fast fluorescence kinetics by Strasser et al. (in press). Fluorescence monitoring allowed adjustment of the stresses in order to stay near limit of adaptability. We used shallow Mauritian Amphistegina lobifera and Amphisorus heimprichii, and the Mediterranean Sorites variabilis, the first symbiotic with diatoms, the other with Symbiodinium sp.. They all showed early bleaching signs at end (browning of last chambers). They were incubated from 25°C to 30°C then 32°C in 5 days, then recovery. Light was first set low, 70µE/m2.s, and only raised to 500µE/m2.s the last afternoon, still less than can be experienced in situ. Day/night oscillations in photochemistry of PSII were more proheminent with elevation of temperature and light (both of culture or induced by instrumental protocols). Most sensitive was Sorites then Amphisorus, the less Amphistegina, with stress temperatures under low light above respectively 25°C, 28°C and 30°C. Lower trapping per absorbtion (Fv/Fm) and following electron transport (1-Vj) were compensated by increased absorbtion per reaction center, which maintained constant excitation rate per PS II ("cruise control"), the best way to avoid photodestruction, but certainly by PS II inactivation.

 

In other symbioses not yet known to bleach, Alberte et al. (1986, 1987) conduced temperature experiments on the ascidian Lissoclinum patella/Prochloron symbiosis, from Palau (with natural temperature range 26-32°C). It is worthwhile to note that it has a break in the Q10 of photosynthesis at 30°C (3.5 under, -1.5 above), both in freshly isolated symbionts and of the whole symbiosis, whereas respiration increases regularly between 15 and 45°C (normal Q10 of 1.6 to 2). Isolated symbiotic dinoflagellate of the tropical planktonic foraminifer Orbulina universa had also a sharp cessation of growth rate at 30°C, with diseappearence of the normal motility in culture after 2-3 days (Spero, 1987). Zooxanthellate planktonic foraminifers have a very strong decrease of growth rate at 30°C-32°C, but this might be first of all understood as the open ocean maximum temperature. Non-symbiotic foraminifers have a similar upper viable temperature (Bijma et al., 1990).

About acclimatation to warming, there was no clear difference between bleaching experiments conducted in Hawaii in summer or in spring, suggesting no seasonal adaptation (Jokiel and Coles, 1977). In contrast, better resistance was found for warm-acclimated corals during 2 months against 32-32.5°C during 2 days (Coles and Jokiel, 1978). They also noted that already pale specimens were less damaged.

 

In Montipora verrucosa subjected to 32°C, 3-D reconstruction of confocal microscopy revealed that chloroplast were the first affected organelles with disruption of thylakoids (Salih et al., 1996).

 

In conclusion, experimental data converge to indicate that bleaching is induced by temperatures of 2°C above in situ maximum. Exceptions come for the Jokiel and Coles (1978)'work, which is in contradiction with their results of 1976, and of the undocumented case of Palythoa of Reimer (1971). Glynn and D'Croz (1990) did found bleaching at 30°C, a temperature reached during the (very exceptional) El Niño 1983. Also, Hoegh-Guldberg and Smith (1989) corresponding temperature difference is only 1°C or less, and moreover for a short-span experiment. Of course, it is easy to argue that incubations represente already a stress, and that only the maximum temperature difference is relevant.

 

Cold

Low temperature of 21°C during 6 hours induced "considerable expulsion of zooxanthellae" in Eniwetok corals (Coles and Jokiel, 1977). Floridan Porites porites partially bleached at 16-20°C in 10 days, and totally at 15°C in 5 days (Marcus and Thorhaug, 1981). In the experiments of Jokiel and Coles (1977), mortality occurred at 18.3°C within 2-3 weeks for Hawaiian corals. In contrast to bleaching due to warming, it was characterized by 1) tissue loss and not expulsion of zooxanthellae; 2) a different time dynamic, with fast mortality starting only after a week; 3) an absence of recovery when transferred back to normal conditions.

 

Effects of low temperature shock have been well studied by Steen and Muscatine (1987), Muscatine et al. (1991). Expulsion of zooxanthellae of the sea anemone Aiptasia was provoked by transient cold shock, with a clear threshold at 16°C. Its rate increased with exposure time up to 8h, as well as with cold. The rate appeared somewhat function of exposure time multiplied by degree under threshold. It reached a saturation level of 3.3%/hour. Even after return to normal conditions, expulsion may be continuous for up to 8 days (cf. also Steen, 1986). Recovery needed 7 weeks after a 1 hour 4°C shock (Weis, 1991).

Sensitivity of corals to cold shock appeared more variable, notably with death without bleaching for the sensitive warm-bleaching species Montipora cavernosa, Millepora complanata , M. verrucosa and Palythoa. In this case, the process of bleaching is a dissociation of the endoderm inside the coelenteron, and an expulsion in the medium after rewarming, when the mouth opens. A likely explanation of cold shock effects is a thermotropic effect on membranes. Gates et al. (1992) examined these expulsed endodermic cells, containing 1 to 5 symbionts. They were viable for a around one day.

Whether mild cold-, cold shock- and warm-induced bleachings are similar is yet difficult to know (Gates et al., 1992). The "cold case" pecularity observed by Jokiel and Coles (1977) seems to vanish for "cold shock" (see also below).

 

Light

Many quantitative data on zooxanthellae number in cnidarians were collected in function of light or depth, but no general trend appears (see Battey, 1992). Zooxanthellae number ranges from 0.25 to 15.106/cm2, with most values been between 1 and 5.106/cm2. Content of chlorophyll pigments increase in general with low light. The chlorophyll a/c2 ratio ranges from 0.3 to 3.7. It often increases in response to low light or deepening.

 

Chlorophyll a per cm2 was reduced under full light after two month incubation of M. verrucosa, particularly at the extreme temperatures (20°C and 28°C). The light level in the acclimatation phase had no clear influence on warm-bleaching sensitivity (Cole and Jokiel, 1978). Pocilloporids incubated during ten days under full light had twice less chlorophyll per zooxanthellae than under 25% full light, but the same amount of the number of symbionts, though a slight elevation of expulsion in the first day was noticed (Hoegh-Guldberg and Smith, 1989). In the experiments of Lesser et al. (1990) on Palythoa, the zooxanthellae number was unchanged, whereas pigment content decreased, by about one third, at high (1700µE/m2.s) versus low (170µE/m2.s) light level, with a chlorophyll a/c2 ratio slightly lowered.

 

Synergistic effects of light level with warm-induced bleaching were clearly detrimental for Montipora verrucosa either during the treatment or the recovery phase (Cole and Jokiel, 1978). One the other side, almost none were measured by Lesser et al. (1990) in experiment with Palythoa.

 

Bleaching can be induced in situ with upward transplantations (see also below, under UV), from 14-50m depth to 1 m depth for Scolimia and Mycetophyllia (Lang, 1973), from 80m to 6m for Oculina varicosa (Reed, 1981), and from 30-40m to 15-0.5m for Montastrea annularis (Dustan, 1979). In the latter case, slight bleaching was also noted with downward transplants. In aquarium, it is current know-how that replacement of halide lamps, which might double irradiance, often induce bleaching in corals (Monaco aquarium managers, pers. com.).

 

In the case of solar bleaching after emersion of Goniastrea in Thailand, covering by neutral plexiglass, which reduced light by only 6% (in order to compare with UV-shielding), led to the surprising total protection instead of 80% bleaching of unshielded corals (B. Brown, oral com., Panama, 1996). In addition to light effect, this emphasizes spectacularly the high sensitivity of corals to very small environmental changes at stress limit.

 

Polyp retraction seems a way to avoid light by self-shading: in Acropora acuminata, light saturation level and compensation point were 25% greater when polyps were retracted than when they were partially expanded (Crossland and Barnes, 1977) and photosynthesis was reduced by 27% (Lasker, 1981). Xeniids stopped completely their O2 production when closed (Svoboda, 1977). Contraction of temperate intertidal anemon when emerged, or during peak irradiance, likewise decreased greatly their photosynthesis (Schick and Dickens, 1984). Migration of zooxanthellae at the base of polyps has probably similar results. Another consequence is surely the reduction of exchange rates, as indicated by the lowering of respiration by 15-33% in Montastrea annularis when polyps contracted (Lasker, 1981).

 

In the freshwater Hydra-Chlorella symbiosis, continuous light (600-1900 W/m2 during 3-5 days) induced partial or total bleaching, depending on host strains (Pardy, 1976). Red light was the most effective in causing bleaching. The disrupted appearance of the symbionts suggested that light acted on the algae directly. Red light induced an high degree of bleaching of Montipora verrucosa after 2 months, more than with prolonged shading (Kinzie and Hunter, 1987). This might be related to the "red drop" quantum efficiency of Photosystem II.

 

Large foraminifers appears particularly sensitive to high light : paling or mottled appearance was observed in Heterostegina in culture after 8 months at 600 lux, not 300 lux with "day-light neon tube" (Röttger and Berger, 1972). Mottling and bleaching can be induced in Amphistegina gibbosa and A. lessonii in culture at 25-26°C with fluorescent lights at intensities exceeding 60µE/m2.s and 140µE/m2.s respectively (Hallock et al., 1986).

 

Carbon dioxide

The first experiment of acidity on coral is from the famous Peyssonel (1744) which stated : "This insect closes itshelf when one throws, in the glass where it is, acid liquors" (p. 44). Later, following his acidosis hypothesis, Mayer (1918b) tested the resistance of 8 species of corals in (unprecised) CO2-enriched water. They survived from less than 1 hour to more than 4 hours. Corals showed nearly the same resistance to CO2 as they did to temperature. He verified that low pH (4.85-5.8) does not alter the rate of respiration. Mayer (1924) ran experiments to check if there were synergical effects: at 37.7°C, the coral reactions appeared similar whether at 8.22 or at 6.95 pH; the one-hour lethal temperature in CO2-enriched water at 6.65-6.9 pH was slightly lower than in normal sea water, the difference being -0.85°C to +0.1°C. The short time exposure prohibits to draw firm conclusions.

 

Very few other experiments were done. Considerable numbers of zooxanthellae were found in fecal pellets in clams preincubated overnight in carbon depleted seawater (Yellowless et al., in press). In various reefal algae, Tolbert and Garey (1976) induced bleaching within 5 days by incubation in flushed-CO2 free water (with unknow pH). The calcareous algae Halimeda was found to be the most resistant (it can be guessed that CaCO3 dissolution released some bicarbonate). It must be also mentionned that Loomis (1957) demonstrated that sexual reproduction in freshwater Hydra is elicited by high pCO2.

We made a preliminary experiment on the effect of CO2 on corals which gave good indication that increased pCO2 and lowering pH is a bleaching factor (Pêcheux, 1993, 1994, see Annex I).

 

Monitoring by PAM fluorescence of the Mediterranean large foraminifer Sorites variabilis was undertook under either 25 or 32°C, 200 or 500µE/m2.s, 7.5 or 8.5 pH for one day, with two replicates each (thanks to M. Havaux). It showed that photoinhibition arose at the high temperature or the high light. At 25°C, 500µE/m2.s, adaptation was faster at the higher pH than at the lower one. And only samples at the low pH, 32°C, 500µE/m2.s could be distinguished from all others after 4 days in recovery, by their strong Fv/Fm inhibition, slow recuperation and bleaching signs.

 

Effects of temperature and light on symbiotic foraminifers were studied by fast fluorescence kinetics by Strasser et al. (in press) (cf. above). The experiment was driven under either 8.00 or 8.33 CO2-induced pH conditions, and our subsequent analysis of this factor revealed that it became influent under highest temperature and light stresses, with several parameters involved in a complex way at 10-50% change level (unpublished results). A further test demonstrated existence of complex interactions between pH and PS II photoinhibition state, of a magnitude compatible with a CO2 rise origin of the mass bleaching (Annex II).

 

One of our more lancinating questions was : because light is so strongly involved in bleaching, why foraminifers do not move toward shade place, as they otherwise obviously move to optimal place, for example under boulder in very shallow depth ? a possible answer is : in husbandry of S. variabilis with normal temperature and light (25°C, 90µE/m2.s) at eight pHs from 7.7 to 8.8, an unexpected strong correlation was found between low pH and upward active movement (r2=0.873, p=0.0007 after up to 40 days), which suggests that, given CO2 rise, large foraminifers expose themshelf to photoinhibitory conditions (under redaction).

 

Oxygen

High oxygen concentration

High internal O2 pressure is though to be experienced by photosynthetic symbiosis, as the diffusive path is a priori important. Oxygen-derived free radicals represent a potential danger and could be responsible of bleaching, when elevated temperature and irradiance presumably increase photosynthesis or host sensitivity.

 

Yet, the only case of bleaching induced by elevated oxygen level is described by Dykens and Shick (1984) for one specimen of the temperate intertidal sea anemone Anthopleura, incubated for 12 hours at 200% O2 saturation in light, which resulted in loss of about 40% of its chlorophyll and in "observed zooxanthellae expulsion". Anemons incubated in darkness were unperturbed. Cultivation of Cassiopea symbionts at either 300%, normal or 10% O2 level had no effect on the temperature threshold of decline of DCMU-induced fluorescence (Iglesias-Prieto et al., 1992).

 

External high tension of oxygen are experienced in normal time by sea anemone: 200% O2 saturation was measured in the gastroderm after 5 minutes at 315 µE/m2.s under mild anesthesia (Dykens and Shick, 1982). Corals can withstand up to 300% O2 saturation in closed incubation chamber without apparent detrimental effect (D'Aoust et al., 1976), and 373% O2 saturation had been measured just at the surface (Kaspar, 1992). Bubbles of probably pure O2 are formed by Acropora acuminata (Crossland and Barnes, 1977). Trapped gaz in Millepora branches contained 70% O2, coming from photosynthesis of endolitic algae (Bellamy and Risk, 1982). However, Wells et al. (1973) measured only 28-32% O2 in bubbles forming in situ at 2 m depth. In respect to the oxygen (and CO2) question, the Tridacna mollusc may constitute a interesting comparative case as their symbionts are in very close proximity to the blood system. The question is the buffering capacity of the blood system. Mangum and Johansen (1982) found that during the day, the blood is normoxic or only slightly O2 sursaturated, with a maximum of 110%, but Schick and Dykens (1985) measured up to 175% O2 saturation. The first authors suggested that the blood can store the equivalent of 5-15 minutes of photosynthesis. In addition, high internal O2 pressure is surely exacerbated when reef waters are themselfes oversaturated.

 

Nocturnal morphs of Montastrea cavernosa close their polyps at day and open them in the dark. When an enclosed portion of an head is exposed at night to twice the ambient oxygen, the enclosed polyps are only partially extented while those outside are fully extended, suggesting O2 sensitivity, and perhaps regulation of photosynthesis by retraction (Wells et al., 1973).

Elevated oxygen tension increases oxygen consumption and provides a negative feedback (already in Henze 1910, in Mayer, 1918b). In dark (and probably also in light), 250% O2 saturation during 4 hours generally enhances the aerobic respiration (from -11% for Protopalythoa to +46% for Palythoa tuberculosa). This response is probably function of the adaptation to the rate of exchange with the water in normal time, and thus influenced by the morphology (Shick, 1990). Dark oxygen uptake follows Michaelis-Menten dynamic (Km of 13-38% O2 saturation) and is not completly saturated at normal O2 level (Newton and Atkinson, 1991). The short term (at most 10 minutes) increases up to 60% of the respiration following long periods of photosynthesis might be related to the same effect (Edmunds and Davies, 1988), or to ATP production for bicarbonate pumping (see below).

 

Oxygen detoxifying enzymes

The oxygen question was mostly examined through the study of the activity of the oxygen detoxifying enzymes, which are the most studied ones in cnidarian biology. First, the direct production of oxygen radicals by excitation or electron transfers at the photosynthetic site inside the symbionts (see below) must be distinguished from the overall increase of the formation of radicals by the many aerobic reactions under elevated oxygen tension, occurring also in the host. Plants and animals possess protective mechanisms, assured mainly by enzymes (see Asada and Takahashi, 1987, and below): superoxide dismutase (SOD), catalase (CAT), and ascorbate peroxidase (ASPX) associated with glutathione transferase.

 

In reef photosymbioses, high level of those enzymes are associated with oxygen and photosynthesis. They decrease with depth (Schick et al., 1995). In the sea anemone Anthopleura, the host SOD content decreased within 15 days incubation in darkness, as well as in light when the photosynthetic inhibitor DCMU is added. It increases in darkness under 300% O2 saturation (Dykens and Shick, 1982). SOD and CAT are associated with zooxanthellae area (Dykens and Shick, 1984). SOD and CAT activities are generally correlated with chlorophyll content between individuals and within various reef photosymbiotic taxons (Dykens and Shick, 1982, Dykens, 1984, Schick and Dykens, 1985) but this relationship was questioned by Tytler and Trench (1988) who found higher level of CAT in some non-symbiotic animals. High level of glutathione-S-transferase in the animal tissues of the coral Favia fragum, which exceeds those found in most invertebrates, may be also related to protection from oxygen tension (Gassman and Kennedy, 1992).

 

Cultivated zooxanthellae have about twice higher activities of SOD, CAT and ASPX than freshly isolated ones, which is probably due to shading inside the host, but the photorespiration pathway (see later) may be involved instead (Tytler and Trench, 1986, Lesser and Shick, 1989). In addition, the study of Matta and Trench (1991) on cultivated zooxanthellae at 28%, normal and 270% O2 saturation showed that up to 99% of enzyme activity can be found in the previously overlooked pellet fraction. In their results, SOD activity followed oxygen tension, but CAT was lower at the normal, intermediate atmospheric O2 level. The case of ASPX was even less clear: normalized to total proteins, it increased with O2, but on a cell basis, the reverse trend was observed.

 

Experimental conditions susceptible to promote bleaching tended to enhance oxygen defences. When incubated 2 weeks with both high light and natural UV together, SOD and CAT activities, but not ASPX, increased by 40% in cultivated zooxanthellae of Aiptasia. FIZ from incubated colonies which supported the same treatments do not show significant changes (Lesser and Shick, 1989). More interestingly, in Palythoa, an increase of zooxanthellar SOD, CAT, ASPX activities exists with temperature of 31°C instead of 26°C (respectively 24%, 174% and 47%), with high light (46%, 67% and 38%) and with UV (32% and 51%, none for ASPX). In the animal part, only CAT is affected by temperature (163%). CAT of host and symbiont are well correlated, but not SOD (Lesser et al., 1990). One must be aware that in aposymbiotic anemons too, SOD level increases with oxygen, as expected, but also with light and UV; CAT activity increases only with oxygen, being multiplied by 5 with 200% O2 saturation (Dykens and Shick, 1984).

 

In short, evidences of involvment of high oxygen tension in bleaching are rather indirect, except for one case. Adaptation clearly occurred. It will be of interest to measure oxygen pressure during warming experiments and in situ as well.

A very low O2 level is also a stress. Corals ejected zooxanthellae after in the dark within 2 to 6 days when incubated in O2 saturation less than 20% (Yonge et al., 1930, Yonge, 1966). This was also observed in 2-3 weeks with corals in sealed jars in the light, in water having a low pH and O2 content, but it was attributed to starvation.

 

Ultra-violets

There is a general concern about detrimental effects of UV on reef organisms. Most experiments were realized with UV-unshaded runs and shaded controls.

80% of "shade-loving" epifauna died within 3 days when exposed to natural UV level in aquarium, in particular the sponges, with bleaching of Zygomycale or discoloration of Mycale cecilia, whereas others had tissue loss (Jokiel, 1980). Incubation of Pocillopora damicornis with and without natural UV had no effect on its zooxanthellae number nor on the chlorophyll per zooxanthellae after 40 days; UV reduced the coral growth by a mean of 25% (ranges: 5.0-8.5 versus 5.3-9.5 mm/40 days) (Jokiel and York, 1982). No difference was noted between Okinawan corals submitted to either natural sun irradiance or with reduced UV to 10%, both in synergy with warm temperature (30.1°C, above in situ 1980 bleaching at 29.6°C); at 31.3°C, Acropora but not Pocillopora died faster with natural UV (Glynn et al., 1992).

 

In Aiptasia, treatment with UV for 2 weeks did not significantly alter neither total chlorophyll, nor chorophyll per symbiont, nor chlorophyll a/c2 ratio, but reduced zooxanthellae doubling/day; photoinhibition was slightly greater under UV, and the maximum photosynthetic rate lower (Lesser and Shick, 1989), or the threshold of photoinhibition lower. On the contrary, a reduction of symbionts number was observed in Palythoa after exposure to natural UV for 10 days (by about 20%, but with great variability), whereas neither pigment composition nor content per symbiont changed (Lesser et al., 1990).

 

Interspecific differences in UV responses were clearly apparent in the work of Shick et al. (1991). They incubated "shade-loving" octocoral Clavularia and "sun-loving" sea anemone Phyllodiscus with and without UV during 5 months. Though there was in all cases an "high mortality", the number of symbionts was unchanged and no photoinhibition was observed. Pigment content of the symbionts decreased by one half only in the "sun" species, together with a lowering of the chlorophyll a/c2 ratio. Nonetheless the photosynthesis of the freshly isolated zooxanthellae was about 75% greater, tentatively explained by self shading. The photosynthesis of whole colonies of Clavularia was a contrario 50% lower, thought to arise from an impairement of PS II, and may explained the observed reduction of host SOD content. Photosynthesis and respiration of symbionts of Clavularia was 50% lower. Schick et al. (1995) measured a one third inhibition of photosynthesis of transplants of Acropora from 20m and 30m to 1m depth compared to similar ones under UV shield, but no difference from those coming from 2 and 10m. Montipora verrucosa collected from 10 m depth and transferred to full sun aquaria bleached and died in 2 days with natural UV and in 3 weeks without. They survived in 60% sun light, either with or without UV (Scelfo, 1986). When not shield from natural UV, transplanted Montastrea annularis from 24m to 18m and to 12m in Bahamas had one third zooxanthellae density without changed chlorophyll concentration per zooxanthellae nor MAA after 3 weeks (Gleason and Wellington, 1993). Paling occurs only in UV-unshield 12m corals (see critics of Dunne, 1994). Those authors monitored in the same time in situ level of UV (Wellington and Gleason, 1993). According to their data (but see Dunne and Brown, 1996) and admitting constant UV absorbtion over the water column, the increase of mean daily irradiance between 18m and 24m is only 10% at 375nm wavelength, but almost 200% at 300nm, with mean increase of UVB by x2.5 and x6 at 18m and 12 m. The former wavelengthes are surely of very low detrimental effect (Caldwell and al., 1986, Quaite et al., 1992), and harder UV have not doubled at global scale. It is therefore hard to evaluate what in situ UV increase is necessary to elicit mass bleaching. As the spike of UVB-UVA at 24 m when high water transparency had the same energy as the bleaching dose at 12 m, it conduced to the hypothesis of increase UV by dolldrums and clear waters, nonwithstanding its negligeable effect in very shallow waters (see above). At contrast, Dunne and Brown (1996) pinpointed that with emersion, corals of Phuket naturally experiment 20 fold increase of UV, while their bleaching is not related from UV (Brown et al., 1994a).

 

In synergy with a stress temperature of 32°C, Fitt and Warner (1995) tested the response measured by PAM fluorescence of Montastrea annularis under in situ light and without blue, UVA and UVB domains. Photosynthesis efficiency (Fv/Fm) fell more strongly with wider spectrum range apart UVB. Here one may simply suspects saturating photon dose effect.

Artificial UV irradiation from 2 to 42 W/cm2 was tested on 5 coral species by Siebeck (1981, 1988). It did not induced bleaching but polyp withdrewal, production of mucus, swelling of tissue, ejection of mesenterial filaments and finally death. The half lethal dose (LD50) in the 280-380nm wavelength ranged from 100 to 250kJ/m2. The dose/effect law is of logarithmic type, with various scaling factors according to species. Corals from 1.5 m depth are only twice more resistant than those from 20 m. The resistance was 5 times greater as measured by LD50 when corals were exposed to light, particularly the blue, during the recovery phase. This unexplained protecting effect of blue light probably does not involve the photosynthetic apparatus, as it was also observed in the crustacean Daphnia.

 

Adding 1µE/m2.s UV to temperate Anthopleura and tropical Cassiopea, Banaszak and Trench (1995) did not noticed any effect on symbioses, only on cultivated Symbiodinium (reduced growth, less chlorophyll, low motility, multiple cell walls). By incubating tips of Acropora prolifera for one month at 27°C, 29°C and 31°C, crossed with normal UV 2160, 2400 and 2600 J/m2.day, Reaka-Kudla et al. (1993) found a synergy between these two factors, with roughly 20% more bleaching with 10% more UV and 50% more bleaching with 20% more UV.

 

Almost unmeasurably small intensities of UVB irradiation in addition to fluorescent lamps can induces mottling to total bleaching of the large foraminifer Amphistegina at only 23-26°C, and UVB at irradiance superior to 0.005W/m2 kills then within days (Hallock et al., 1995). After an incubation during 50 days at 25°C with a) normal fluorescent at 16µE/m2.s (Å7W/m2) ; b) with Mylar, which reduces visible light by 45% and UVB by 75% ; and c) 0.005W/m2 UVB added, the proportion of bleached specimens was respectively 40%, 23% and 67%, but with the same growth rate (Hallock et al., 1995).

Bleaching by photopigment destruction and subsequent tissue necrosis has been carried out on the kelp algae Ecklonia radiata of Marmion Reef in aquaria under natural light, while no such effects were observed in UV-shaded controls (Wood, 1987).

Species of marine dinoflagellates have variable adaptation capacity to UV (Jokiel and York, 1982). Growth of isolated zooxanthellae of the sea anemone Aiptasia and the medusan Cassiopea can be severely impaired. Cultivated symbionts of Aiptasia are very sensitive to UVA and UVB, but only when light was above 20% of surface irradiation. As it is probably the irradiance at most inside the host, UV increase should have no effect on the symbiosis (Jokiel and York, 1984). Recently, Hebling et al. (1992) have shown that photosynthesis of tropical phytoplankton, at contrast to polar one, were suprinsingly not significantly altered by UVA, UVB or UVC.

 

UV-absorbing compounds

Jokiel and York (1982) stressed that if UV are often a harmful ecological factor, they may be also utilized photosynthetically under low light (Halldal, 1968, cf. Lesser and Shick, 1989, and ref. herein). Radiations as hard as 310 nm are used for O2 evolution (Halldal, 1968). The famous cnidarian fluorescence is probably a way to use shortwave photons in certain environments such as turbid waters. In the remarkable case of Leptoseris fragilis living between 100 and 145 m depth, host pigments absorbs UV-C photons and reemitte them at the maximal absorption of the algae pigments (Schlichter et al., 1986, Schlichter and Fricke, 1990).

 

Cnidarians contain substances called "S-320" which absorb UV with maximum in the 310-330nm range, and which been identified as host water-soluble mycosporine-like amino acids or "MAA" (mycosporine-Gly, palythine and palthinol, see Dunlap and Chalker, 1986, Stochaj, Dunlap and Shick in Shick, 1991, Schick et al., 1995). In bleached corals near a thermal plant in Hawaii, S-320 substances decreased by one half in Porites and Pocillopora while they increased by one half in Montipora, but this variability was said to have resulted from loss of tissue (?) (Jokiel and Coles, 1974). Aposymbiotic Anthopleura sea anemons have 3.5 times less S-320 substances than symbiotic ones (Shick and Dickens, 1984). Host content of S-320 is higher after UV incubation of P. damicornis during 40 days (Jokiel and York, 1982). S-320 compounds can increase tremendously in less than 4 days in Montipora verrucosa when transplanted from 10m and 3m depths to aquaria, in shade or full sun, and, ironically, with or without UV (Scelfo, 1986). In Clavularia, more than 4 months is needed in UV-transparent aquarium before an increase can be observed in palythine (about 25%) and palythene (600%); in Phyllodiscus, the global level of S-320 compounds was similar, but mycosporine-Gly was mostly localized in zooxanthellae instead of being equally distributed between host and symbionts (Shick et al., 1991). No changes in S-320 concentration was observed in sea anemones after 2 and 4 weeks in the presence or absence of UV, pointing here to slow responsiveness, at least with reduction of natural level of UV (Stochaj, Dunlap and Shick in Shick, 1991). Perhaps more important, the S-320 decreased in Palythoa with higher level of temperature and light (except with unfiltered UV). This decrease is observed with S-320 normalized against dry wheight basis, not protein content (reduction of protein occurs in bleaching conditions). This suggested that warming may predispose to UV sensitivity (Lesser et al., 1990).

 

It may be signalled that with photooxidation of PS II (see below), new fluorescence peaks at 280nm appeared, presumably due to colorless isoprenoid degradation products of bleached pigments (Telfer et al., 1991).

 

To summarize, there are many data on effects of UV compared to other factors. UV have been showed to induce bleaching in sponges and probably in Amphistegina, in upward coral transplantation, and some symbiont loss in Palythoa, but their detrimental effects correspond rather to death. Dose needed to elicit bleaching responses is hard to precise in term of wavelenght and energy. Distinction between lethal UV effects and photosynthetic impairment is perhaps to be done. Adaptation to UV by production of S-320 is evident, but seemingly with great variability in time response. The unusual, often "fluorescent" colors observed during mass bleaching suggest an UV effect, but their appearance is probably only due to the disappearance of symbiont pigments. A synergy with temperature appears as a possible important factor.

 

Salinity

Coles and Jokiel (1992) wrote a very complete synthesis on effects of salinity on coral reefs, both in situ and in experiments. Salinity must greatly be lowered in order to induce lethal effects on corals: Marcus and Thorhaug (1981) observed no responses above 25°% within 20 days. First partial bleaching occurs at 20°%, with or without production of mucus. No effect on photosynthesis nor on respiration were measured on Siderastrea sidera when salinity was changed by plus or minus 10°% during 6 days (Muthiga and Szmant,1987). Kato (1987) studied the production of mucus as a stress response to low salinity : one third of colonies produced it at 27°%, and three quarters at 21°%, with a paling threshold below 25°%.

Synergistic effects between low salinity and high temperature have been put into evidence by Cole and Jokiel (1978) on Montipora verrucosa, with the very first effects at 30°% together with 32°C during 4 hours, 25°% being clearly detrimental, even at lower temperature. Under low irradiance, there were no particular influence of lower salinity (30°%) at temperature from 27°C to 34°C during 7 hours in Stylophora and Seriatoporix (Hoegh-Guldberg and Smith, 1989).

Effect of a slight elevation of salinity is more difficult to precise: Pacific and Atlantic Porites become partially bleached in 1-3 weeks at 40°% and normal temperature (Marcus and Thorhaug, 1981), whereas Hawaiian ones tolerated 40-45°% (Coles, 1992, abstract). More relevant, corals exposed to 33°C for 2 hours survived few days longer with a concomittant 40°% than those with a normal 35°% salinity (Cole and Jokiel, 1978).

 

Red and brown algae bleached in a few days at low salinity (15-20°%), while high salinity does not affect pigmentation (Arai and Miura, 1991, Yokoya, 1992). Carbon plays a role much more important than other ions in the relation salinity-photosynthesis (Ogata and Matsui, 1965, Hammer, 1968, Dawes and McIntosh, 1981). Low salinity decreases strongly CO2 and HCO3 affinity (Booth and Beardall, 1991). Calcium concentration may also play a role (King and Schramm,1982).

Miscellaneous

 

Darkness

Darkness is known since a long time to induce bleaching, from casual observations in situ (Duerden, 1902, in Yonge and Nicholls, 1931a). Vaughan (1914), Yonge and Nicholls (1931a) and Yonge et al. (1932) maintained corals in darkness, with continuous renewal of seawater. Darkness provoked expulsion of zooxanthellae within 2 weeks to 5 months. Yonge and Nicholls (1931a) described an abnormal abundance of mucus-glands and cells with refractile granules termed "wandering" cells. Zooxanthellae were in degenerating state, without clearly-marked nucleus nor pyrenoid and with large number of granules, some of them refractile. Corals remained in good health as long as 7 months when fed. Bleaching was induced by darkness in 10 days to 2 months (Goreau, 1959, Goreau and Goreau, 1960). Franzisket (1970) observed that specimens halted calcification on the first day and bleached after 10-20 days in darkness. Pocillopora died in 1 month, while after 2 months, Porites were in a very atrophied state, and Montipora and Fungia remained healthy. The intensity of the effect followed the growth rates. The extraordinary resistant Oculina patagonica, newly introduced in Mediterranean, losed its zooxanthellae when transplanted in a dark cave during 29 months, but survived well, and recovered thereafter (Zibrowius and Ramos, 1983).

The temperate coral Pleisiastrea urvillei began to expulse its symbionts and to reduce the calcification only after 50 days in darkness. Surprisingly, the normal proportion of 10% of degenerate zooxanthellae did not change during the bleaching. Expulsed zooxanthellae were both healthy or degenerated (Kevin and Hudson, 1979). In situ shading carried out by Rogers (1979) bleached all Acropora in 3 weeks, whereas Montastrea and Diplopora were more resistant. Corals losed 22% of their chlorophyll when shaded by hypersuspension of peat in only 19 hours (Dallmeyer et al., 1982). Oculina varicosa bleached when transplanted from 6m to 80m depth (Reed, 1981). Partially bleached Condylactis contain only 30% of C30 gorgosterol normally produced by symbionts (Boatwright, 1974, in Ciereszko, 1988). The tropical sea anemone Aiptasia pulchella bleached at 60% within 10 days in darkness (Steen and Muscatine, 1987), and can be cultivated in aposymbiotic state more than 4 years in darkness (Wilkerson and Muscatine, 1984, Steen, 1986) and apparently contained a residual heterotrophic population of zooxanthellae able to recover in light, according to Steen (1986). The large foraminifers Amphisorus and Amphistegina bleached in darkness and died after 2-4 months (Lee et al., 1991), as did Heterostegina (Röttger, 1972). The freshwater hydra can sustained indefinite symbiosis with its native Chlorella in darkness (if feeding is provided); but hydra infected with non-native algae does bleach in the same conditions (Muscatine and McNeil, 1989).

 

Chemical inhibitors

Bleaching of corals can be caused by long term exposition to photosynthetic inhibitors (Lasker et al., 1984, unpubl. res.). The medusan Cassiopea andromeda losed its zooxanthellae under chronic DCMU incubation (Hofmann and Kremer, 1981), and Anemonia viridis with 100µM DCMU for 2 months (Suharsono et al., 1993). The large foraminifer Amphistegina lessonii became aposymbiotic in five days with 10µM DCMU, with drop of photosynthesis to 7% (Lee, 1983, Lee et al., 1983). White specimens had normal pseudopodial activity and recovered in one week. Some symbionts remained normal, others perhaps underwent autolysis. The thylakoids became more compact and some were disrupted. Collapse of pyrenoids was often observed. Methylviolagen was not as efficient. An identical bleaching was induced in symbiotic planktonic foraminifers with 10µM DCMU in three days (Bé et al., 1982, Spero, 1987), and the zooxanthellate symbionts underwent lysis in their perialgal vacuoles. Higher DCMU dose brought out high mortality rate of the host, whereas longer exposure produced host cytoplasmic alteration including dense deposits within the mitochondrial matrix. DCMU favorized bleaching of the Hydra-Chlorella symbiosis under high light level but not ambient (Pardy, 1976) and induced symbiont digestion (Muscatine et al., 1979). DCMU is known to reduce not only photosynthesis but also calcification in corals and larger foraminifera (Vandermeulen et al., 1972, Crossland and Barnes, 1977, Dugay and Taylor, 1978, Barnes, 1985). Many commercial herbicides (phenoxy-acids, triazines, glyphosphate, acetochlor, bromacil,...) can induce bleaching in large foraminifers at concentration between 0.01-10ppm (unpubl. res.). The first effect is immobility and loss of adherence. Symbionts expulsion begin with the browning of the last chambers before complete whitening in 3-15 days. Borelis are the most sensitive, followed by Soritidae and Heterostegina, while Amphistegina spp. are quite resistant. Jaap and Wheaton (1975) provoked bleaching of corals within hours and within months by application of the fish-collecting quinaldine and rotenone. Glynn et al. (1984) showed that P. damicornis died by loss of tissue whithin 1-2 days with 0.1ppm of a phenoxy-acid herbicid (2,4D-amine Na-salt), or with a wetting agent, at 0.25ppm (Tergitol). The very surprising result is that this wetting agent had a protecting effect at 0.025ppm against bleaching at elevated temperature (from 25.7°C to 34.5°C in 4 hours, then gradual cooling) which normally causes death in 3-5 days. Dispersants were lethal at low dose according to Elgershuizen and De Kruif (1976), had no effect alone (Cook and Knap, 1983) or, together with oil, increased calcification during 2 months (Dodge et al., 1985). Bleaching was also observed 5 to 13 days after 30 seconds submersion in marine diesel (Reimer, 1975). Loss of zooxanthellae was seen with 3 months incubations in 0.1 and 0.5ppm oil (Peters et al., 1981), but it might be because temperature reached 31°C. Tentatives of bleaching of anemons and corals by algicid and streptomycin were not successfull (Muscatine, 1961, PhD., in Muscatine, 1974, unprecised).

 

Starvation

The first response to starvation in filtered sea water is expulsion of zooxanthellae, which can be rather fast, as soon as 7 days, or can take up to as long as 5 months, according to Yonge and Nicholls (1931b), but other reports did not confirm this observation (see discussion in Franzisket, 1970). The former authors also observed that, with starvation in light, 10-75% of expulsed zooxanthellae were still alive, whereas they were generally all dead when starvation occurred in darkness. In the temperate Anemonia sulcata, starvation in light or darkness induced an expulsion of symbionts, without leading to total bleaching (Taylor, 1969).

 

Burial

Mayer (1918a) observed that the coral species which survive 30 hours buried under mud were also those which resits to 37°C; those killed by 11 hours of the same treatment died at 36.5°C or lower. He suggested that temperature-induced bleaching is related to their capacity to resist asphixation. High rate of sediment loading (2-8kg/m2) provoked bleaching by covering and abrasion (Rogers, 1983). Sand deposition on the center of alcyonacean corals induced serious bleaching after 5 weeks on covered center of alcyonacean, with less chlorophyll (Riegl, 1995). In experiments, rates comparable to natural one does not usually induce bleaching (Peters and Pilson, 1985, and ref. herein) though they observed a few bleached patches. In addition to necrosis and decrease of mucus production, histopathological examination revealed an accumulation of unusual basophilic mucoid material in the calicoblastic layer, and a change in pH and in composition of mucin, as indicated by the Pentachrome stain (similar to the observation of Glynn et al., 1985b, in corals during mass bleaching, see above).

 

Others

Osmotic stress (probably low salinity) applied on Palythoa sp. provoked fast symbiont expulsion in mucus strands via the mouth. Extruded algae retained morphological integrity and photosynthetic capability (Trench, unpublished, in Trench, 1974).

The freshwater Hydra viridis expulsed its Chlorella symbionts in few days after incubation in 0.5% glycerol (Whitney, 1907). This may be not linked to photosynthesis as this treatment reduced the growth by 20% both in green and albino hydra (Muscatine and Lenhoff, 1965). Glycerol caused algae to disintegrate in situ, leaving behind empty intracellular vacuoles (Muscatine 1974).

 

Porites lutea bleached at 50% with iron concentration superior to 0.005mg/l, which is a very high dose (Harland and Brown, 1989). In aquarium, a solution of strontium, molybdemum, complexed iron, iodine and other ions was said to "prevent bleaching" out of many Actinodiscidae and Zoanthiniaria (Wilkens, 1990).

 

Expulsion of "surpernumerary" symbionts

Low rate of expulsion of zooxanthellae also occurs regularly, and is though to correspond rather to a regulation of their number. It is thus fondamentally different from bleaching, althougth both may have in common some mechanisms.

Healthy zooxanthellae in various stages of life, including zoospore, are normally released in pellets by the sea anemone Aiptasia tagetes and various others species, according to Steele (1975, 1977). They contained degenerated cysts as well, with a large accumulation body, numerous and larger cytoplasmic inclusions, and reduced chloroplasts. On the other hand, Taylor (1968) observed that there were always degenerated symbionts in the mesenteries, and that "viable zooxanthellae have never been found among excreted material" of the temperate Anemonia sulcata. In this species under normal conditions, older cells of zooxanthellae (as seen by the waste accumulation) showed first a thickening of their cell wall, latter an enlargment of the accumulation body with an oxalate crystal formation, a loss of starch and pyrenoid body, and lastly an appression of the chloroplast and the disappereance of the stroma. Those cysts were transported as intracellular inclusions in "undifferentiated amoeboid cells" to the mesenteries (Taylor, 1969). Degeneration of zooxanthellae in healthy culture of Zoanthus (Trench, 1974) was first visible with enlargement of vacuoles and their coalescence, with myelin figures and crystalline bodies. Chloroplasts and mitochondries then became disorganized, but algal periplast membranes remained intact, suggesting that the process originated within the zooxanthellae. They did not fix radioactive carbon any longer. Pulse-chase experiments and autoradiographies proved that degenerating zooxanthellae, previously marked, migrated in 5-10 days to mesenteries where most were found. Trench did not observed any wandering amoebocytes nor migration from cell to cell, and concluded that symbionts are exocytozed out of gastrodermis cells, then phagocyted in the digestive-excretory mesenterial zone before definitive ejection. Acid phosphatase activity, typical of digestion, was never observed together with vacuoles containing symbiont, even degenerated ones.

 

In giant clams, intact, often dividing zooxanthellae, as well as zooxanthellae in various stages of disorganization, are found in the digestive tract and in fecal pellets. They are surely derived from the clams themshelves, and many are photosynthetically functional and can produce viable culture (Trench et al., 1981).

 

Regular expulsion of zooxanthellae can be also experimentally enhanced in conditions which favorize multiplication of symbionts. Under continuous-light regime, Palythoa sp. (but not P. tuberculosa) regularly expulsed by the mouth healthy zooxanthellae in mesenterial filaments or in "zooxanthellar bodies" with a definite shape, probably rolled mesenterial filaments (Reimer, 1971). Immediately after being fed, starved specimen of the same species expulsed symbionts with a very high mitotic index (10%). The sea anemone Anthopleura elegantissima also normally expulses symbionts in bolus, which were found to have four fold greater mitotic index than in hospite. Such expulsion increased with light (McCloskey et al., 1996). When highly enriched in nutrients (17mM NH4), the coral Pocillopora damicornis released zooxanthellae at a rate 40% superior than in controls, but, as symbiont number was three times higher, with an equal or lower growth rate, it corresponds to a decrease of the specific release (5.10-3 instead of 11.10-3 per day, Stimson and Kinzie, 1991).

 

In the Mediterranean symbiotic sponge Cliona viridis, bleaching by lysis has also been observed concomitant with reproduction, either sexual or budding, which was considered as inducing a stress (Rosell, 1993).

 

 

B) SOME CONSIDERATIONS ON THE PROCESS OF BLEACHING

Time dynamic of bleaching process

Immediate reactions

To understand bleaching, the knowledge on the first reactions is clearly the most critical. Under normal conditions, the rate of expulsion of zooxanthellae is low, from less than 0.01% per day to a maximum of 0.1% according to Hoegh-Gulberg et al. (1987). Stimson and Kinzie (1991) found higher values, with a mean of 1% per day, with a clear diurnal variation by a factor greater than 10, with the maximum at noon. When temperature is raised by 2-3°C above the threshold, expulsion of zooxanthellae is almost immediate, at least in experiments. It is multiplied by 100 above normal level in half an hour or less (Hoegh-Guldbergh and Smith, 1989); a second marked increase to a steady level of expulsion of 10-15% symbiont population per hour occurred 3-4 hours after the application of the temperature stress. Edmunds (1994) observed a mean loss of zooxanthellae of 0.07% per day after 3.5 days at 27°C-32°C in P. porites. A slow rate of expulsion (0.6%/day), 15 times above normal, may be induced by mild stress (Xenia during 30 hours in unstirred, darkness and low O2 conditions, Hoegh-Gulberg et al., 1987). A similar rapid rate in the field is proved at least in some mass bleaching events, as well as with warm incoming tidal water, by the observations of clouds of expulsed zooxanthellae. It implies that massive expulsion can take place in less than few hours. On the other hand, most reports indicated that the bleaching process in situ takes weeks, at least at population level.

 

Dark respiration increased within 2 hours (or less) with warming (Coles and Jokiel, 1977), probably from host because it reachs level 3-6 times normal rate whereas symbionts represent usually 10-20% the mass of the host. It does not correspond to a normal temperature dependence (Q10 up to 36), but more probably to a stress reaction. The 40-60% increase of the respiration observed with elevated oxygen level, either directly tested (Schick, 1990, Newton and Atkinson, 1991) or through light and photosynthesis (Edmunds and Davies, 1988) seems too weak to be responsible. In this later case, it might be due to energy consumption for bicarbonate uptake which must reach 20% of photosynthesis (taken one ATP per bicarbonate ion uptake and 5 ATP per fixed carbon in sugar equivalent).

 

The high respiration rate may provide an explanation to the increase of oxygen-detoxifying enzymes in bleaching experiments. Oxygen radicals are produced in mitochondries in amount proportional to the respiration rate multiplied by the oxygen concentration (Fridovich, 1986). SOD activity had been correlated with this measure in various organs of Tridacna (Schick and Dykens, 1985). Higher respiration rate will thus enhanced oxygen radical formation, even taking in account the lowering of internal oxygen concentration, which is limited by diffusion.

 

Photosynthesis reactions are less well known. It increased within 2 hours in the experiment of Coles and Jokiel (1977) but was normal in Stylophora and greatly reduced in Seriatopora after 7 hours at 32°C, even on a per zooxanthellae basis (Hoegh-Guldberg and Smith, 1989). Four days later, during recovery, photosynthesis per zooxanthellae was at normal level, except for Stylophora treated by a mild warming (30°C), whose rate was roughly twice that of controls. Also note an increase of photosynthesis just in the first hour at 34°C, in the resistant species Montastrea cavernosa but not in the other ones (Fitt and Warner, 1995, fig. 1). Photosynthesis of isolated zooxanthellae responds to high temperature in a variable way, increasing or remaining the same (Sandeman, 1988) or decreasing (Iglesias-Prieto, 1992). No short-term data on pigment composition and activities are available.

 

Also, coral calcification is quickly reduced. Calcification first, not photosynthesis, began to decrease at 28°C in Acropora (Kajiwara et al., 1995). Radioactive calcium incorporation disminished by an half after at most 3 and 6 hours at 30°C, compared to the maximum rate at 25°C (Clausen, 1971), and by the same amount after at most only half an hour during an exposition at 36°C compared to the rate at 31°C for corals living at 28°C (Clausen and Roth,1975). Low and quasi constant level of inhibition at 30, 60 and 90 minutes strongly suggests that the reaction is a very fast one. Below 30°C, Q10 of 6 and 12 for calcification (Clausen, 1971) are abnormally high for an enzymatic reaction.

 

Mid- and long-term reactions

Bleaching can be observed even one month after a short-term temperature stress in experiments (Coles and Jokiel, 1978), and appears in those cases as an irreversible process. Host respiration stayed at a high level before returning to normal after 23 days in the experiment of Hoegh-Guldbergh and Smith (1989). This points to a long-lasting stress reaction. It can be related to tissue reparation, as histopathological studies reveal general necrosis, and/or to a shock-like process. A possible internal pH perturbation, one month after bleaching, was suggested by staining reaction (Glynn et al., 1985b). Physiological degradation may also precede bleaching, as observed by a slower growth rate one month at least, maybe three, before bleaching in the transplant experiment of Shinn (1966, table 3).

 

Oxygen tension or low photosynthesis/respiration ratio ?

On one hand, high internal O2 tension and consecutive tissue damages have been hypothetized as the cause of bleaching. This is corroborated indirectly by sometime O2 evolution in isolated zooxanthellae and by studies on oxygen-detoxifying enzymes. On the other hand, the sensitivity to bleaching has been correlated with the fall of the photosynthesis-respiration ratio observed during bleaching, due to an higher increase of respiration rate with temperature (Jokiel and Coles, 1974, Coles and Jokiel, 1977, see also data of Hoegh-Guldberg and Smith, 1989). This fall implies a low internal O2 concentration. These two point of vue appear contradictory. However, in experimental bleaching with sudden warming, there may be a transient burst of photosynthetic oxygen, triggering a fast protective and energetically costly process, responsible of the P/R ratio. Transient reactions of photosynthesis to warming are unknown, but would be faster than 2 hours (experiments of Coles and Jokiel, 1977).

 

Temperature effects

As expected, when either the temperture or the time exposure are increased, the bleaching is more prononced. Smith and Buddemeier (1992.) spoke of "temperature dose" as excess temperature times duration of excursion. The processes involved in bleaching follow crudly a logarithmic dynamic, as seen by the proportion of colonies which bleach or die with time, or by loss of zooxanthellae or of protein (Glynn and D'Croz, 1990). This indicates a constant rate of symbiont expulsion and/or degradation. The effects are generally multiplied by 2 or 3 for each degree increase (see data of Coles et al., 1976, Marcus and Thoraug, 1981, Glynn and D'Croz, 1990).

 

A discrepancy seems to exist between the temperature increment needed to elicit a bleaching response in the field and in laboratory : during bleaching events, warming is quite usually by no more than 1°C above "normal maximum", whereas in laboratory a 2-3°C increase above this maximum is necessary. Unresolved questions are whether bleaching in the field is consecutive to an "heat shock" or a temperature rise above a certain limit, and to which extent sudden temperature changes in laboratory experiments are similar to in situ bleaching:

- Coles (1975) analyzed the bleaching effect of thermal effluent. He concluded that bleaching is not due to "thermal shock variation", but is a response not to prevailing mean temperature variation but to its absolute value. He did not exclude that temperature variation has an influence, but only near the upper thermal limit;

- the absolute value of 30-32°C°C generally encountered in bleaching is not due to some fundamental limitation in the symbioses because the temperate anemon A. viridis bleaches when transferred from 16°C to 26°C (Suharsono and Brown, 1992) ;

- in situ, the reported rate of warming are relatively slow, during period of the order of weeks. In oceanic waters, up to 1°C in 10 days or less has been exceptionally noted (Goreau et al, 1992), and it is possible that even more rapid warming occurs with hot ring passages. In reefal areas, a record is given by Coles et al. (1976) with a warming from 25.5 to 32°C in 1 hour after a storm. Perhaps temperature can change rapidly after calm weather onset (see above, Great Barrier Reef event);

- in addition, mass bleaching occurred in inshore waters with great temperature variation, often small but probably important in some case (5°C in Indonesia during the 1983 bleaching event, Brown and Suharsono, 1990). In normal time those variations are generally less than 1°C, but reach 3°C in fringing zone and 12°C at shore (synthesis in Andrews and Pickard, 1990), thus greater than the experimental temperature "shock".

 

Mechanism of expulsion

A cytological pathway of zooxanthellae during experimental warming-induced expulsion was well described by Yonge and Nicholl (1931a), and appears as an highly controlled mechanism. Zooxanthellae are quickly carried via the mesenteries to the "absortive" zone of the filaments, from where they are ejected into the coelenteron. The observation that symbionts are transported by "wandering cells", which remain thereafter empty in the corals, bears great similarity with the description of expulsion of "surpernumerary" symbionts by Taylor and Trench (see above). Jokiel and Coles (1977) observed that coral polyps darkened as coloration faded in the coenosarc tissue, apparently due to migration of the zooxanthellae. Glynn (1989) also observed "vacuoles vacated by zooxanthellae" in bleached corals (cf. also Muscatine, 1974, Brown et al., 1995). A process which lets behind such vacuoles would constitute a special type of exocytosis.

 

In contrast, Gates et al. (1992) have examined the material released by Pocillopora damicornis and Aiptasia pulchella after both cold shock (2-4 hours at 12°C) and warming (up to 16 hours at 32°C) in both cases in darkness. The material released after heat stress was more difficult to handle than that produced by cold stress, but appeared similar. They found host cells detached from the endoderm, containing 1 to 5 symbionts. Calcium-free seawater can also provoke detachment of endoderm cells (Gates and Muscatine, 1992). Epifluorescence shows they were viable for a short term, around one day. Those authors proposed that host cell detachment is caused by adhesion dysfunction after temperature shock, either cold or warm (see ref. therein for similar shock processes after strong stress in other organisms), maybe due to failure of ion pumps and elevated intracellular calcium. They discuss potential mechanisms of zooxanthellae release after temperature stress (exocytosis, apoptosis, necrosis, pinching off and host cell detachment).

 

Carefull histopathological examination of five coral species only 5 weeks after the beginning of a bleaching event by Brown et al. (1995) revealed a more complexe figure. There was both release from excretory zones and zooxanthellae degradation inside coral cells. It existed loss of endodermal cells in some Goniapora, interpreted as an extreme response with death of coral. In addition, a migration within wandering cells is not excluded as suggested by the presence of zooxanthellae in mesoglea.

Other reef photosymbioses must have quite different mechanisms: zooxanthellae in giant clams must surely follow the digestive path, as for the slow natural expulsion. In symbiotic large foraminifers, exocytosis is obvious as they are unicellulars. No observations have been made on sponges, in which cyanobacterial symbionts are intra- perhaps also extracellular. We conclude that, although a fundamental incertitude is still present concerning the cytological mechanism of coral reef bleaching, expulsion of symbionts is the likely main process.

 

Are expulsed symbionts healthy ?

Zooxanthellae expulsed after coral starvation appeared dead at 75% (in light incubation) or at 100% (in darkness) to Yonge and Nicholl (1930b). In their high temperature experiments, they stated that they remained healthy. Suharsono and Brown (1992) studied the temperate Anemonia viridis subjected to temperature increase from 16°C to 24°C during six days. They confirmed that expulsed zooxanthellae are viable, and even more, that they have a 5 to 6 times higher rate of division, supposely because they are unrestricted from host control. It was suggested that this increase of the division rate occurred already inside the tentacules. Expelled cells from Xenia under mild stress had also a mitotic index 15 times grater than normal (Hoegh-Gulberg et al., 1987). The only remark on the state of released symbionts in situ came from Bunkley-Williams and Williams (1990b) which stated that they were "dead and dying". Histopathological observations provide ambiguous indications. Zooxanthellae are often already degenerated in the coral tissues, but in some cases, they appeared healthy inside necrotic tissue host (Glynn et al., 1985b). As seen before, works on expulsion of "surpernumerary" symbionts gave also a controversal picture.

 

 

C) THEORETICAL BIOLOGICAL BACKGROUND

The understanding of the biological basis of mass bleaching is almost non-existent. Only few indices and hypothesis are at our disposition. In contrast, the current knowledge in many fundamental biological and physiological processes is very advanced. Technics which are widely used in biological laboratories have clearly the power to identify what happens during bleaching.

What are the points which might be useful to treat ? From an up-to-down point of view, bleaching affects reef photosymbioses which have high metabolism and reduced exchanges: high temperature and light involvment, as well as preferential bleaching of fast growing corals, indicates high metabolism, notably photosynthesis. Reduced exchanges is first an inherent characteristic of symbioses, with the host barrier for the symbionts. They are also expected during dolldrum time, at the organism/environment interface, between reef and open ocean waters, and between sea and air. They would involve rather the O2/CO2 couple, as it is hard to believe that nutrients are involved in mass bleaching. A best guess is that a basic biochemical process of photosynthesis and related host-symbiont exchanges is in question. Two critical phenomena limit photosynthesis: photoinhibition, mostly controlled by light level and able to degenerate in photooxidation ("bleaching" in the botanical sense), and photorespiration, due to limitation of carbon and oxygen exchange. And the key enzymes of both, the photosystem II and the Rubisco, are said to be the most temperature sensitive ones in plants.

 

Other primary possible cause of bleaching are a direct temperature effect on enzymes or on lipidic membranes, or a relation with the stress shock reaction. Regulation mechanisms between host and symbionts are almost unknown in reef organisms. Few other indications from studies on exocytosis and biochemistry are directly relevant.

 

Photoinhibition and photooxidation

From a physiological point of vue, photoinhibition is the reduction of plant photosynthesis, observed in particular under high irradiance. Works on its biochemical basis are rapidly progressing, and the excellent review of Melis (1991) is a recent one (see also Virgin et al., 1992, Aro et al., 1993). Photoinhibition has its origin in the impairment of the function of the photosystem two (PS II), located in the thylakoid membranes of chloroplasts. PS II carries out the delicate tasks of charge separation from light excitation, splitting of water, which produces O2 on one side of the PS II, and reduction of quinone A and B (Qa-Qb) on its other side. The core of this complex, the D1 protein, is easily destroyed and degraded. In the chloroplast thylakoids, functional PS II are appressed in the grana. When D1 is degraded, unfunctional PS II (perhaps the so-called PS IIß) are disassembled from their light antenna and transported in the stroma where they are regenerated. In plants, even under the best conditions, between 20% and 40% of the PS II are in the unfunctional state, and protein production and turnover are the highest for D1 in chloroplast, though it represents less than 1% of total proteins (Critchley and Rusell, 1994). Protective mechanisms exist, primarly by disconnection of the light antenna by phosphorylation. They are though to be driven by the pH transthylakoid gradient or the redox state of NADPH, which carries away the energy. A good description of the cyanobacterial PS II can be find in Feher et al. (1989).

 

The events in D1 degradation are not completly elucided. A critical step, particularly slow, seems that quinone B must be charged twice before to be protonated. Recent works tend to indicate that, when this downward electron transfer is impaired, one of the two central carotens quickly bleach, probably by direct charge transfer from P680 chlorophyll and not through oxygen singlet radical (Telfer et al., 1991). D1 and D2 tyrosines, transferring electron between P680 chlorophyll and water, are therafter degraded (Blubaugh et al., 1991). It may be said that photoinhibition acts as a protection against further photooxidation. There is still a controversy about oxygen involvement in photoinhibition. Several mechanisms might be involved, in which O2 attacks at the Qb site. Otherwise cytochrome b559 can protect P680+ from degradation by electron cycling, and oxygen competes with it for electron. "Anaerobiose" effect on photoinhibition appears now to be due to bicarbonate depletion as in all experiments, N2 flushing is used (Sundby, 1990). Reviews on the physiological effects in photoinhibition have been done by Powles (1984), Kyle (1987), Anderson and Osmond (1987), Krause (1988) and Baker (1991).

 

The photosystem one (PS I) raises the redox potential from PS II to produce NADPH (see a description of PS I in Brettel, 1997). Most oxygen radicals in chloroplasts are produced at PS I (see reviews by Fridovitch, 1978, 1986, Asada and Takahashi, 1987, Foyer et al., 1994). Reduction of O2- to hydrogen peroxyde (H2O2) is done by the superoxide dismutase enzyme (SOD). In the chloroplast, ascorbate peroxidase (ASPX) reduces H2O2 by deshydrogenation of ascorbate, which can be regenerated by NADPH directly or through glutathione, preventing the formation of the highly reactive OH radical and singlet oxygen. In this ascorbate-glutathione cycle, re-reduction of oxygen to water dissipates the reducing power ("Mehler reaction"). In the cytoplasm and other organelles, the direct scavenging of H2O2 to oxygen and water is generally insured by catalase (CAT). In algae, catalase is mostly found in peroxisomes, where H2O2 is a by-product of the photorespiration pathway.

 

Photooxidation arises when light energy dissipation is impaired, by overloading of the system of energy transfer between light capture and CO2 reduction. Chlorophyll and carotenoid pigments are progressively destroyed ("bleached" sensu stricto), by decay of excitation state or by oxygen singlet radical attack. Almost always, bleaching occurs after photoinhibition (Powles, 1984). To simplify, PS II are more suceptible to photoinhibition and PS I to bleaching (cf. Miller and Carpentier, 1991).

Many inhibitors are used in photoinhibition studies (Powles, 1984, Krause, 1988). DCMU (and other herbicides such as triazines) binds at the Qb PS II site. It inhibits photosynthesis but prevents D1 degradation during short-term photoinhibition by shunting reduction of quinone and providing a sink to reducing power; in addition it allaviate oxidative stress (Greer et al., 1991, cf. Tsang et al., 1991).

 

The photoinhibition as defined at the biochemical level by the degradation of the core of PS II must be distinguished from the "physiological" photoinhibition as seen by the low oxygen production. This production can be also reduced by decoupling of light antennae or by their pigment bleaching, as well as by non-radiative dissipation processes by futile cycles, i.e. electron cycling through cytochrome b559, the Mehler reaction, and photorespiration. Nonetheless, photoinhibition measured by fluorescence quenching is well correlated with reduction of net photosyntesis under a wide variation of environmental conditions (Genty et al., 1989, Seaton and Walker, 1990, Ögren, 1991).

 

Fluorescence quenching measurements

Most progresses in the studies on photoinhibition came from the simple non-destructive measurement of chlorophyll fluorescence quenching, widely used in study of stress in land plants (reviews by Baker and Horton, 1987, Krause and Weiss, 1991, Geider and Osborne, 1992, Lichtenthaler, 1995). It is a different technique than remote detection of coral bleaching using pulsed-laser fluorescence spectroscopy, which gave only indications of quantities of photosynthetic pigments, though respective proportion of PSI and PSII might be calculated from their data (Hardy et al., 1992).

Fluorescence quenching techniques measure the long wavelenth light emitted under a shorter wavelenth probe light. Most of this fluorescence is the fact of the PSII, and the principal "quencher" is the Qa in its oxidized state. Hence the photochemical reaction can be easely monitored under various environmental regimes. Three techniques are of current use among many others. Measurements of change of fluorescence before and after DCMU addition (blocking electrons at Qa) provides a rapid way to measure photoinhibition (Prézelin, 1981, Neale, 1987) and was carried out in isolated zooxanthellae (Iglesias-Prieto et al., 1992, Fitt and Warner, 1995, Warner et al., 1996). The classical "pulse-amplitude modulated" (PAM) fluorescence uses modulation of a weak probe light, allowing continuous recording even under light. Most important values are the dark-adaptated initial (Fo) and maximal (Fm) fluorescences from which is calculated the variable part of the fluorescence (Fv/Fm=(Fm-Fo)/Fm), describing the quantum yield of exciton trapping. Changes in light-adapted samples give indications on photoprotecting mechanisms (cf. Warner et al., 1996). Fast fluorescence kinetic (or fluorescence induction) profits from the polyphasic rise of fluorescence at microsecond scale after light probe on, due to cascade events at PS II. In addition to Fo, Fm, it allows determination of other characteristics of PS II state, first photon absorbtion, then efficiency of electron transport at Qa-Qb site and later energy sink to PS I (as in Iglesias-Prieto, 1995, Strasser et al., in press, see annex II).

 

Unfortunately, there are few works on photoinhibition in aquatic plants, and they are mainly on freshwater or marine planktonic ones (Powles, 1984, Neale, 1987). Recent data indicate low relative variable fluorescence in oligotrophic pelagic ocean with strong diel pattern (from 0.4 in day downto 0.25 in night, Behrenfeld et al., 1996, and ref. herein).

 

In reef, in situ measurements have been performed on marine macrophytes (Hanelt, 1992, Hanelt and Nultsch, 1994). In less then one meter depth, they showed considerable level of photoinhibition: Fv/Fm ratio, normally of 0.7-0.8, was often around to 0.5, and as low as 0.3. There were fast responses (15 minutes) depending on light level, the temperature being probably not involved. Larkum et al. (1996) confirmed this strong photoinhibition with Fv/Fm downto 0.2 in Halimeda. The only value of coral Fv/Fm in reef is given by Fitt and Warner (1995, fig. 5), around (a slightly low) 0.66 in Montastrea annularis. Physiological experiments have given strong support of a synergy of temperature, light, agitation and CO2 on PS II photoinhibition as the cause of bleaching (see above and under).

 

Irradiance influence

Photoinhibition is primarly induced by high irradiance, with variations due to adaptation, acclimatation and age (Powles, 1984). High light adaptation and acclimatation were reviewed by Anderson and Osmond (1987), and are characterized by fewer and less thylakoids, lower grana/stroma ratio (perhaps not in C4 plants), higher chlorophyll a/b ratio, more electron carriers, more ATP synthesis and higher /chlorophyll. Dinoflagellates, and perhaps cyanobacteria, are more sensitive to photoinhibition than diatom or green algae (Neale, 1987). It is not known whether UV-induced photoinhibition depends only on photon flux dose (Neale, 1987), but the PSII D1 protein Qa-Qb site appears as the preferential target of UVB damage in plants (Friso et al., 1995, Lichtenthaler, 1995, and ref. herein).

 

Temperature influence

Photoinhibition is minimum around 30°C, which is a condition of mass bleaching, because chloroplastic protein biosynthesis, very sensitive to temperature, is maximal at this temperature, thus insuring a rapid turnover of the chloroplastic D1 protein (Powles, 1984, Greer et al., 1986, Greer and Laing, 1990, Baker, 1991, Ögren, 1991). Appearance of light stress and photoinhibition at high temperature is observed in terrestrial plants generally above 42°C (Ludlow and Björkman, 1984, Havaux et al. 1991, Georgieva and Yordanov, 1994). There may be less fluorescence quenching of non-photochemical origin at higher temperature (Ögren, 1991). Isolated thylakoids of plants grown at 20°C show a dramatic increase of PS IIß when heated one minute at only 35°C (Sundby et al., 1986).

 

Tobacco treated by 5000 lux and 37°C had mitochondrial and cystosolic Mn- and Cu/Zn-SOD increase but stable chloroplastic Fe-SOD level, while the reverse was observed at 4°C, from which it was concluded that heat photoinhibition, in contrast to chilling photoinhibition, was not caused by photosynthetic oxygen (Tsang et al., 1991). An interesting complex synergy of temperature and salinity is signalized by Larcher et al. (1990) of temperature and salinity in land plant : salinity stress enhances mild temperature stress, but protects against severe one.

 

There are very few data on "high-temperature-exacerbated photoinhibition" at less than 40°C (Ludlow, 1987). Reports on the interaction of light and temperature are somewhat contradictory but in general low light provides some protection. Light-induced intrathylakoid acidification may activate Rubisco and/or stabilize PS II complex (Weis, 1982, Havaux et al., 1991). Light-induced protection to temperature begins already at 31°C in potatoes (Havaux, 1994, fig. 4). This kind of mechanism corresponds to the diurnal cycle supported by land plants. However, it seems to be absent in unicellular algae so far studied (Srivastava and Strasser, 1995) or only very limited (in cultivated coral zooxanthellae, Iglesias-Prieto, 1995).

Rise of initial fluorescence in the dark (Fo) during heating at rate of 1°C per minute is a standart method to assess direct thermal damage of PS II, which temperature is well correlated with plant resistance in the field. It occurs above 36°C-49°C in land plants (Schreiber and Berry, 1977, Havaux, 1993, 1994). Thanks to M. Havaux, we measured such rises in the large foraminifer Sorites variabilis, symbiotic with Symbiodinium sp. like corals. They were collected in french Mediterranean sea, thus presumably particularly sensitive to warming as summer maximum is less than 29°C. Critic temperature (Tc) was found between 36.5°C and 38°C, whereas they bleach at 32°C in other experiments (unpublished results). We conclude that a direct thermal damage of PS II is not involved in bleaching. Albeit done with other protocols, Fo rise at similar temperature in cultivated symbionts (Iglesias-Prieto, 1995, in press) or in the bleaching-sensitive Montastrea annularis (Warner et al., 1996) confirms this view. A clear difference must be made between direct thermal damage of PS II occuring at high temperature (>35°C), seemingly at the the donor side water-splitting complex and/or because chlorophyll antenna disconnection, and a temperature-induced sensitivity of photoinhibition, put in evidence at lower, 30°C-32°C, bleaching temperature.

It is noteworthy, that, according to M.D.A. Le Tissier (com. pers., 1994) and Salih et al. (1996), one of the first event in warm-induced bleaching in situ or in experiments is desappression and degradation of thylakoids. It is a strong argument in favor of photoinhibition as the mechanism at origin of bleaching. Stroma/grana morphology is intimately link to PS II functionality, desappressed stroma being the place of D1 remplacement (see Critchley and Rusell, 1994).

 

Carotenoid pigments and the xanthophyll cycle

Carotenoids play a role both in light harvest and photoprotection (reviews by Siefermann-Harms, 1985, 1987, Koyama, 1991). The peridinin is a light-harvesting pigment, typical of dinoflagellates, and is often involved in light adaptation (Prézelin, 1976, cf. Dustan, 1982). Kleppel and al. (1989) observed in bleached Montastrea annularis a decline of peridinin relative to chlorophylls. Their data also indicated a slight increase of diadinoxanthin in bleached corals, while in the experiments of Hardy et al. (1992, fig. 9), peridinin+xanthophylls appeared to be reduced as chlorophylls. In "solar bleaching" after emersion, strong changes in xanthophylls were observed, typical of photoinhibition (B.E. Brown, Panama 1996, oral com.). Photoprotection by xanthophyll pigments came recently in evidence (reviews by Demming-Adams, 1990, Demming-Adams and Adams, 1996, also Young, 1991). Xanthophyll might also be involved in thermal protection of PS II (Havaux and Tardy, 1996). Adaptation of reef-building corals to high light intensity follows generally an increase of xanthophylls/chlorophyll a, ratio which is between 6% to 14%, and up to 34% (Tilyanov, 1981, Titlyanov et al., 1982). In cultived zooxanthellae, xanthophyll/chlorophylla ratio increased more or less regularly from 35% to 50-75% with level of light, and xanthophyll/carotenoids ratio rose from 10-20% to 30-45% (Prézelin, 1976, Chang et al., 1983).

 

Xanthophylls, located in the light-harvesting antennae, often constitute 20-30% of the carotenoids in dinoflagellates (Demers et al., 1991), 15-25% in higher plants (Young, 1991). Under inhibitory conditions, short-term de-epoxization of violaxanthin to zeaxanthin delays PS II chlorophyll desactivation in higher plants. Long-term acclimatation to photoinhibition also involves an increase of both compounds. This is also known in the marine macroalgae Ulva (Franklin et al., 1992). These yellow pigments are represented by diadinoxanthin and diatoxanthin in zooxanthellae and some other microalgae. De-epoxization of diadinoxanthin to diatoxanthin occurs under high light in a diatom (Willemoes and Monas, 1991), in a green microalgae and a dinoflagellate (Demers et al., 1991). Adaptation takes place in minutes to hours with a change of diatoxanthin/diadinoxanthin ratio from 0% to 55%, and in days with an increase of the ratio of xanthophylls/chlorophyll a from 35% to 80%.

 

Miscellaneous

Methyl violagen, known commercially as paraquat, is a potent Mehler reagent. It enhances the formation of oxygen radicals at PS I site and impedes NADP reduction (Asada and Takahashi, 1987). Vandermeulen et al. (1972) have tested this compound on Pocillopora damicornis in 1-hour incubation. Carbon incorporation (radiocarbon in NH4 digested tissue) was surprinsingly enhanced in a very regular way with concentration from 0.5µM to 0.5mM, and even in some specimens up to 50 mM (by 50-80%, r2=0.70 to 0.83, from op. cit., Tab. 4). Polyps began to retract at 50mM. At 500mM, photosynthesis was null and tissues readily sloughed off.

 

Interaction of nutrients was studied in Ulva, where nitrogen starvation reduced adaptative capacity to photoinhibition, as it slows down the biosynthesis of Rubisco and probably of D1 protein (Henley et al., 1991); as reef organisms live in oligotrophic waters, they might be more susceptible to photoinhibition.

 

Carbon uptake and fixation, and photorespiration

Yonge and Nicholls (1931a) suggested that bleaching was due to CO2 starvation. On the contrary, experiments by Sandeman (1988) tended to indicate that photorespiration, favorized by high O2/low CO2, has a protective role against bleaching. Photorespiration and its links to photoinhibition, will be examined now.

 

Carbon dioxide fixation is the subject of many reviews (Osmond, 1981, Lorimer and Andrews, 1981, Edwards and Walker, 1983, Raven, 1984, Petterson, 1990, Oliver and Kim, 1990). CO2 is fixed by the Rubisco enzyme (Ribulose 1,5-biphosphate carboxylase/oxygenase), which is slow and relatively inefficient. This enzyme can constitute up to one half of the protein (and nitrogen) content in plants. When oxygen is present at the active site of the Rubisco, there is a "parasite" reaction, by which the sugar substrate is oxidated instead of being reducted. This oxygenase activity, termed photorespiration, represents 20-40% of the carbon fixation, or carboxylase activity, and is the major limitation to photosynthesis.

 

Photorespiration is proportional to the O2/CO2 ratio. It increases with temperature, due both to the the enzyme characteristics and to differential solubility of O2 and CO2. The oxygenase/carboxylase ratio increases by 1.5 for each 10°C increase (Jordan and Ogren, 1981, 1984). The oxidized substrate (phosphoglycolate) is recycled through a complicated pathway (C2 or photorespiration cycle) with further oxidation and consumption of energy. In a side effect, H2O2 and ammoniac are released, important in oxidative damage and nutrient equilibrium (review by Lorimer and Andrews, 1981, Bowes, 1989, Oliver and Kim, 1990). The so-called C4 terrestrial plants concentrate CO2 in the chloroplasts in order to avoid photorespiration, and they are increasingly abundant toward the warm latitudes and low nutrients soils (Edwards and Walker, 1983), just like photosynthetic symbioses in seawater. Photosynthetic rate is better correlated with Rubisco activity than with chlorophyll content (Anderson and Osmond, 1987).

 

The carboxysome pist

Photorespiration is avoided when Rubisco is isolated from the oxygen-producer PS II. This is certainly the main function of the cyanobacterial carboxysome, where most of the Rubisco is (Kaplan et al., 1991). Carboxysome-like structures, containing also most of the Rubisco, have been encountred in endosymbiotic zooxanthellae of certain cnidarians (Cassiopea, Montipora verrucosa and Zoanthus socialis), which is quite exceptional for eukaryotic algae (Blank, 1986, 1987, Blank and Trench, 1988). Pyrenoids may play a similar role (Osafune et al., 1989, Kaplan et al., 1991) probably also in zooxanthellae (indirect indications in Markell et al., 1992).

 

Carbon uptake or carbon limitation ?

Active transport of inorganic carbon (always in HCO3- form as far as can be judged) is now well known in a wide range of aquatic phototrophs (reviews by Raven, 1984, 1991, Lucas and Berry, 1985, Badger, 1987, Prins and Elzenga, 1989). The goal of carbon pumps is to reduce photorespiration, which is rather critical in warm oligotrophic tropical water. The efficiency of the carbon concentrating systems depends principally of the CO2 leakage. A requisite is thus low permeability to CO2, but must be balanced by the drawback of low O2 permeability favouring photorespiration (and photooxidation). Note that symbiotic foraminifer shell is an efficient peculiarity in this respect, in imperforate and perforate (as bleaching Amphistegina) sub-orders.

 

Raven (1991) pointed simply that the presence of carbon pumps is implicated by the steep O2 gradient surrounding cnidarians (twice or more O2 saturation, see below), because a purely diffusive use of CO2 would need an equal but inverse CO2 gradient, impossible given its low concentration in seawater (also Raven, 1984, p. 207). An HCO3 active uptake was clearly put in evidence in the larger foraminifers Amphistegina and Amphisorus because: a) internal inorganic carbon concentration is 2-3 order higher than in seawater; b) by the dynamic of the uptake (ter Kuile and Erez, 1988, ter Kuile et al., 1989a, 1989b). Evidence for similar mechanisms in cnidarians are clearly emerging (Weiss et al., 1991, Al-Moghrabi, 1992, Goiran et al., 1996, Al-Moghrabi et al., 1996, Bénazet-Tambutté, 1996). Normal photosynthesis rate at pH 9 simply prooves it. In Tridacna, only CO2 might be used (at least at 25% normal photosynthesis) (Yellowless et al., in press). Aiptasia pulchella had 26% higher photosynthesis at 5mM carbon, suggesting that it is carbon limited (Weiss, 1993). Carbonic anhydrase enzyme is an indispensable piece for the conversion of HCO3 to CO2 for photosynthesis (see review of Sültemeyer et al., 1993). Lower level (35%) of this enzyme was found in bleached portion of Tridacna (Yellowless et al., in press). It is more abundant in host (Graham and Smillie, 1976). Its activity is correlated with photosynthesis and number of zooxanthellae (Weis et al., 1989, Weis, 1991), localized at or near the symbiont vacuole but perhaps not associated with the membrane (Weiss, 1993) and at surface of freshly isolated zooxanthellae (Al-Moghrabi et al., 1996). It is also probably involved in calcification (Isa and Yamazoto, 1984). Its inhibitor Diamox reduces photosynthesis (Goreau, 1977, Crossland and Barnes, 1977, ter Kuile et al., 1989b, Weis et al., 1989) and calcification (Goreau, 1959). Use of inhibitors of anion carrier, H+ATPase, carbonic anhydrase, or HCO3-, Na-, Cl-depleted seawater in coral and anemon colonies as well as in freshly isolated and cultivated zooxanthellae converged to indicate that the host pumps bicarbonate while the symbionts probably profit from an high pCO2 inside the symbiosome (Al-Moghrabi, 1992, Goiran et al., 1996, Al-Moghrabi et al., 1996, Bénazet-Tambutté, 1996, Bénazet-Tambutté et al., 1996). In well stirred condition, coral colonies had low inorganic carbon affinity, and saturated around seawater level (Al-Moghrabi, 1992, Goiran et al., 1996), probably not under low agitation (see below).

 

Trench and Fisher (1983, see also Tytler and Trench, 1986) synthetized the works made on photorespiration in isolated Symbiodinium, which appears often as very low. Contradictory results let suppose either an important variability between symbionts or an high sensitivity to slight variations of experimental conditions (see also Downton et al., 1976, Black et al., 1976). Alternatively, isolation of zooxanthellae may favour photorespiration (cf. transient glycolate release, in Trench, 1971b, Fig. 4-7). Whatever the figure is, indications of existence of both high and low photorespiration level are clear. Cultivated zooxanthellae possess bicarbonate pumps of their own (Goiran et al., 1996, Al-Moghrabi et al., 1996). In hospite, photorespiration is surely important in the jellyfish Cassiopea andromeda (Hofmann and Kremer, 1981), but low in Acropora acuminata (Crossland and Barnes, 1977). Some enzymes of the C4 "CO2 concentrating mechanism" are present in cultivated symbionts, particularly of Montipora verrucosa, as well as in Palythoa (Tytler and Trench, 1986), but the simplest explanation is that they belong to the anaplerotic pathway, i.e the usual ±10% direct carbon fixation on C3 compounds for biosynthesis of C4 ones.

 

Carbon limitation has been hypothetized at high symbiont density in Millepora dichotoma (Schonwald et al., 1987) and Stylophora pistillata (Dubinsky et al., 1990), and also because of enhancement of photosyntesis by flow (Dennison and Barnes, 1988, Patterson et al., 1991, Lesser et al. 1994). Muscatine and Weis (1992) developed arguments along this line, in particular in view of the heavy organic d13carbon (d13Corg) observed in coral tissues and symbionts which is around -13°% (range -9.6°% to -22.2°%) (Land et al., 1975, Muscatine et al., 1989, Yamamuro et al., 1995), clearly heavier than in tropical phyto- and zooplankton (-20°%, Rau et al., 1989). Shortly stated, Rubisco fractionates by -27°%; when limitation is important, it fixes all the carbon present, and thus the fractionation is not so great. As CO2 is around -7°% in tropical seawater, a value of -9.6°% (in Montastrea annularis, one with the highest photosynthetic rate), it would indicate a tremendous carbon limitation. Instead, it proves that corals use HCO3, which is around +1°% in seawater. The variation of d13Corg in carbon concentrating systems is principally determined by the ratio of CO2 leakage to CO2 fixation, in terrestrial C4 plants (Farquhar, 1983, Henderson et al., 1992) and in HCO3 pumping microalgae (review by Berry, 1989). In Bahamian corals, the d13Corg of symbionts and coral species are ordered quite similarly to the general bleaching sensitivity: begining by the heaviest ¶13Corg (and the most sensitive species), this order is : Montastrea annularis > Porites asteroides > M. complanata > Agarcia agaricites > Acropora cervicornis > A. palmata > Dendrogyra > Eusmilia > Madracis (data of Muscatine et al., 1989). In fact, the d13Corg and the annual integrated photosynthetic rate are well correlated together (at least at 1m depth, and with both decreasing with depth), so this rate might be simply the main factor of bleaching, though it is more "noisy". In this respect, head corals (Montastrea) rank higher than branched ones (Acropora), and this may alleviate the discrepency between preferential bleaching of massive "slow-growing" corals in Atlantic and of "fast-growing" ones in Pacific. Again Tridacna symbiosis stands apart, with Black and Bender (1976)'s single measurement of host and symbionts values of d13Corg -16°% and -23.3°% respectively. A weak correlation exists between the (highly variable) carbonic anhydrase activities and the interspecific bleaching sensitivity (Palythoa variabilis > Gorgonia ventalina > Montastrea cavernosa, Millepora alcicornis > M. dichotoma, Stylophora, but Agaricia fragilis and Siderastrea sidera are out of order) (data of Weis et al., 1989).

 

Environmental conditions and carbon fixation

Temperature

The Rubisco is one of the most heat-sensitive enzyme of the photosynthetic apparatus, being half-inactivated between 30°C and 40°C in tens of minutes, depending of light and pH (Weis, 1981a, 1981b, 1982).

Phytoplankton populations living in water above 15°C up to at least 27.5°C have all a rather constant temperature at which their photosynthesis is null, about 35°C (Li, 1985). This may indicate some intrinsic difficulty of marine photosynthesis to adapt to high temperature (in laboratory experiments, the maximum temperature can be quite higher, see for example Eppley, 1972). Also in phytoplankton, Rubisco activity begins to decline at about 30°C (Smith and Platt, 1985), activity which is well correlated with their photosynthesis (Rivkin, 1990). In the green microalgae Chlorella vulgaris, the induction of the CO2 concentration mechanism and carbonic anhydrase by transfer from high to low CO2 condition is accelerated by temperature increase up to 37-40°C (Shiraiwa and Miyachi, 1985).

 

Carbon dioxide

First a non-scientific observation: "...such oxygen poisoning, which I have observed in many leather corals [zooxanthellae expulsion or crumpling], (...) is even worsened with additional carbon dioxide, such as through a diffuser into the basin" (Wilkens, 1990). Effects of increasing CO2 concentration on the physiology of reef photosynthetic organisms are badly known, and the experiments of Mayer are still informative. At normal temperature, no effect is evident. Polyps of Cyphastrea lived 4 months at pH 7 (Edmonson, 1946). A reef mesocosm is maintened in an sub-healthy state notwithstanding permanent acidosis in Monaco (unpublised results). The reef water in Biosphere II with pCO2 of 1000-3500ppm during two years, was buffered at pH around 7.7 (thus with strong alkalinity rise, around three fold bicarbonate) with no mention of problem (Nelson et al., 1993). In Hawaiian aquaria, corals are growing well inspite of 2-4 higher pCO2 and 7.5-7.8 pH (Atkinson et al., 1995).

Burris et al. (1983) observed few or no responses of corals to increased concentration of bicarbonate and total carbon (as might be expected if carbon uptake is optimized, cf. ter Kuile et al., 1989a, 1989b), but did not tested CO2 influence. This influence is difficult to separate from pH change, as H+ concentration is quasi proportional to CO2 concentration around the natural range. In their experiments, addition of up to 2mM bicarbonate would have lowered the pH by only about 0.04, given the (quite high) reported initial inorganic carbon concentration and assuming a pCO2 of less than 400ppm.

 

With slight acidification, which increases both CO2 and HCO3 concentrations, larger foramifers reduce slightly their photosynthesis, and markedly their calcification (ter Kuile et al., 1989a, 1989b). Recently, Al-Moghrabi (1992), Goiran et al. (1996) showed that corals have their photosynthesis optimum at pH 8.6-8.8 in 15 minutes incubation. Extrapolated to present seawater pH change, it would mean that their photosynthetic rate have decreased by few percent compared to preindustrial one, quite important in term of growth rate. Whatever the magnitude of the effect, this quite unusual behavior for phototroph with respect to carbon may be related to indirect pH effect: the internal pH of larger foraminifers might be sensitive to external pH (ter Kuile et al., 1989b), and the temperate Anemonia viridis has a weak control of its internal pH (7.4 in the light, 6 in the dark, Rands et al., 1992). Such a low control is inferred in coccolithophorids where the growth rate closely follows the internal pH, influenced by external pH and CO2 (Nimer et al., 1994). A good model of coral photosynthesis is a dependency on HCO3 uptake with a Michaelis-Menten dynamic (in first approximation not too different than an Hill-Whittimgham one), uptake driven by internal/external proton gradient, thus quite sensitive to pH global change (see note in Goiran et al., 1996). On the other hand, the haemolymph of Tridacna, in closed proximity to symbionts, is buffered at pH 8 with an inorganic carbon concentration of 1.6mM (Yellowless et al., in press). Coelenteric fluid was said to be at pH 7.8 in normal time and 7.1 during digestion by Yonge and Nicholls (1930), but Bénazet-Tambutté (1996) measured fast light-induced alkalinization to pH 9 in temperate Anemonia viridis.

 

A preliminary experiment gave some indications that increasing pCO2 is a bleaching factor (see Annex I).

In conclusion, the carbon concentration mechanism appears as a fundamental element of reef photosymbioses, and the carbon dioxide question is worthy of further research in the perspective of recent bleaching (D'Elia et al., 1991).

 

Relations between carbon fixation and photoinhibition

The feedback regulations between photochemical and carbon-fixing reactions are quite complex. First, photorespiration provides an efficient protection against photoinhibition, not so because it reduces oxygen concentration but because it provides a gratuitous use of reduced NADPH and a sink to light energy (Osmond, 1981, Kyle, 1984, Krause and Cornick, 1987, Krause, 1988, Heber et al., 1990, Petterson, 1990, Kozaki and Takeba, 1996). Alternatively CO2 rise can increases PS II efficiency by sink energy of dark reactions (Habash et al., 1995). Photoinhibition is rapid when both photosynthesis and photorespiration are suppressed, under concentration lower than 4% O2 and 45 ppm CO2, far below natural levels (Krause and Cornic, 1987). Also, oxygen both protects by enhancing photorespiration and destructs through oxygen radical formation in the Mehler reaction (Petterson, 1991).

 

The other strong link between carbon and photoinhibition is a direct one. PS II requires HCO3 at mid concentration (Km=80µM to 180µM, with pH dependency) for electron transport between the quinones A and B at the non-heme iron site, a bottleneck in the photosynthetic chain, and site of photoinhibition (Stemler, 1989, Sundby, 1990, Blubaugh and Govindjee, 1988, 1991, see also Strasser et al., 1992, Crotty et al., 1994). It was well demonstrated in green microalgae by Demeter et al. (1995). They are supports of a physiological role of HCO3 binding to PSII as coenzyme regulating quinone redox state, particularly under CO2 depletion when "hot weather or high light intensity" (Ireland et al., 1987, Diner and Petrouleas, 1990). Low pH might also accelerate photoinhibition at O2-side of PS II (Speata et al., 1997).

 

For unknown reason, low level of CO2 does not protect terrestrial plants possessing a CO2 concentration mechanism (Krause and Cornic, 1987). Under high light and normal O2, chlorophyll destruction in maize is faster without CO2, and level of up to 3000ppm CO2 is needed to alleviate it (Lechowski and Bialczyk, 1991). There is a very high photoinhibition at CO2 compensation point in marine phototrophs with HCO3 uptake, such as the cyanobacterium Chlamydomonas (Sültemeyer et al., 1989) and the green microalgae Dunaliella salina, where photoinhibition is reversed at 3% CO2 (prelim. results in Smith et al., 1990, cf. also Kaplan, 1981).

 

Chlorophyll fluorescence differences are observed in the calcareous algae between acid and alkaline induced zones (Plieth et al., 1994). In cyanobacteria and green macrophyte Ulva, inorganic carbon uptake is highly correlated with a rapid quenching of chlorophyll a fluorescence, a yet unexplained phenomenon (Miller and Canvin, 1987, Espie et al., 1989, 1991, Osmond et al., 1993). It could be due to an energy dissipation by the carbon pump, or more probably because HCO3 binding to PSII. The effect of bicarbonate addition on thylakoid membranes is very important, and 20mM inorganic carbon protected efficiently against photoinhibition (Å10% instead of Å70% inactivated PSII after half an hour, Sundby, 1990). It must be noted that CO2 depletion increases photoinhibition, but would protect against irreversible photoinhibition, as DCMU (Demeter et al., 1995).

Indeed, we put in evidence complex interactions between the CO2/pH factor and the photoinhibition at PS II in symbiotic foraminifers under bleaching stresses (see Annex II).

 

Physiological effects of water agitation

The "dead calm wheather" often reported during mass bleaching events might have a direct effect on reef organisms through a lack of water agitation. The water motion influences the physiology because turbulence reduces the unstirred layer, or quiet water film, which surrounds all immersed solid. This film increases in thickness with the size of the object. Because transport of solutes occurs only by molecular diffusion in this layer, it acts as a barrier to exchange between organisms and the well mixed water (Smith and Walker, 1980, Raven, 1984, Lucas and Berry, 1985). This introduces a rate limitation to maximum possible fluxes. Compounds with a purely diffusive behavior, such as oxygen or CO2, and those which are actively transported through organism membrane (HCO3-, nitrogen and phosphate ions) must be distinguished. Dynamic corresponds to the simple Fick law in the first case, whereas the latter is best described by the Hill-Whittimgham equation, with a sharp transition between diffusion limitation and uptake limitation. This last equation was fitted for bicarbonate uptake in two large foraminifers (ter Kuile et al., 1989a). Large foraminifers appear to be optimized to the turning point between diffusion and uptake limitations (this can be simply seen in aquarium as they move agaisnt current till a certain agitation is reached), and this applies certainly for corals as well. Data of Al-Moghrabi (1992), Goiran et al. (1996, fig. 2) correspond rather also to an Hill-Whittimgham bicarbonate uptake.

 

It is current "know how" that corals are quickly stressed and killed when kept in still water (cf. Addey et al., 1983). Jokiel (1978) studied growth rate of 3 coral species during 70 days and found that it was correlated with water motion (approximatly linearly, op. cit., table I and II). Reproduction and mortality was also similarly influenced. These species were fitted to their environment of origin, and tissue loss occurs with too high water motion, and especially with too low one (also Rex et al., 1995). He also showed that within time of 2 weeks at most, acclimatation can be effective. Growth is enhanced by flow in the ascidian Didemnum molle/Prochloron symbiosis (Olson, 1986). At saturating light level, Acropora in 2 hour incubation in still condition had 25% reduced photosynthesis, respiration and light-enhanced calcification compared to a moderate stirred incubation (Dennison and Barnes, 1988). Maximum photosynthesis of Montastrea annularis may be multiplied by about three with water flow doubling (Patterson et al., 1991, fig. 4). Respiration rate also increased significantly (op. cit., no data). Photosynthesis in a mesocosm was reduced when flowing system was halted (Adey and Loveland, 1991). Linear growth of Pocillopora damicornis over 44 days was one third greater in fast-flowing tank than in control in oligotrophic conditions (Stambler et al., 1991). At contrast, Atkinson et al. (1994) found no influence of water velocity on calcification nor respiration of a Porites community in flume, in measures lasting only one daylight.

 

In Porites cylindrica (Rex et al., 1995), photosynthesis follows almost a Michaelis-Menten pattern in function of agitation (perhaps due to Hill-Whittimgham one), with an half level corresponding to low, still, bleaching condition (op. cit., Fig. 5, 6). A enhancement of photosynthesis and respiration was observed in P. damicornis (Lesser et al., 1994). Under experimental high flow regime, a greater activity of antioxidant enzymes was measured (following O2 production) as well as for other enzymes and particularly for free Rubisco. Activity of carbonic anhydrase was lower. No such differences were observed in situ between specimens taken from high and low flow habitats. There is thus a short-term effect, a mid-term adaptability at biochemical level, and a long-term one, presumably related to morphological plasticity. They noted no difference in PAM fluorescence quenching (unpublished). But a fast fluorescence test with large foraminifers gave us a coherent picture of 6-19% better quantum and electron transport yields of PS II with agitation ( in Amphistegina and Sorites, three replicates, either still condition or 200rpm in 5ml tubes for half an hour, unpublished results).

 

The morphology of the fast-growing and bleaching-sensitive species Agaricia agaricites is not influenced by prey capture, nor light, nor hurricanes, and control by O2/CO2 exchange was suggested (Helmuth and Sebens, 1993). Shick (1990) suggested a diffusion limitation because of enhancement of dark aerobic respiration by increasing O2 in sea anemones, corals, and zoanthids, particularly Palythoa. On the contrary, Newton and Atkinson (1991), Atkinson et al. (1994) did not find any effect of water velocity on dark oxygen uptake of P. damicornis, but the deduced kinetic parameters were highly variable, probably at least influenced by the initial O2 concentration.

 

Flow also enhanced inorganic nutrient supply: phosphate uptake rate is about five times greater of a coral reef-flat community at water velocity of 50cm/s than in still water (Atkinson and Biger, 1992), and ammonia uptake doubled (Atkinson et al., 1994). Sponges (with or without photosynthetic symbionts) have a very reduced growth in lower current speeds (Wilkinson and Vacelet, 1979). Capture of zooplankton would be greater at mid-flow (Patterson, 1991, Sebens and Johnson, 1991, Helmuth and Sebens, 1993).

 

At ecosystem level, it is obvious that corals grow better in agitated waters ("corals make reefs") but it can be due either to direct flow influence, pH/O2 influence, or less likely because of greater zooplankton supply. Addey and Steneck (1985) stated that flow rate controls the productivity gradient from fore to back reef, but not the difference of productivity between reefs.

Thickness of the unstirred layer has been estimated to be about 200µ in Amphistegina in the experimental conditions of ter Kuile et al. (1989a). It may range from 1mm to 1cm in corals, according to indirect calculations of Patterson et al. (1991, unpubl. res.). The diffusive boundary layer around massive and branched corals was measured with O2 microelectrode to be 1.5mm to 4 mm, with reduction of about one third with a current of 5cm/s (Shashar et al., 1993). Oxygen saturation ranged from 0% in dark to 373% in light. In coralline red algae, the quiet film measures 100-200µ at 1-3cm/s flow and up to 2.5mm in unstirred condition; extreme variations of O2 saturation at organism surface were 11% and 400% (Kaspar, 1992), with deduced corresponding pH of 7.9 and 9. Similar data, 100-400µm, are fournished by Israel and Beer (1992). For comparison, Dennison and Barnes (1988) found equivalent flow speeds of 5 to 39cm/s in Acropora biotopes (during a period with higher than usual waves, with winds of 2.5-10 m/s). Water agitation can be easily measured in situ by plaster dissolution (op.cit, and Petticrew and Kalff, 1991), or by fully automated electromagnetic device, and video of particle mouvements (Helmuth and Sebens, 1993). Potts and Swart (1984) used the former technique during two years at five sites of Heron Island and found a some correlation between wind speed and water agitation (r2=0.14, p<0.001) as well as with tides.

 

Bleaching in valley or pillar in Montastrea, at the center and/or the periphery of Agaricia colonies, or verrucae of P. eyouxi could be related to micro-currents, but zebra patterns are yet harder to explain. Polyp retraction and zooxanthellae migration at the polyp base must clearly exacerbate the exchange problem. Only one anedoctical report of bleaching in relation to agitation condition exists, namely in colonies of ascidians at the downstream end of a flow chamber after six days in darkness (Olson, 1986). Mayor (1924) explained the temperature resistance of Pocillopora by its "capacity to drive stagnant CO2-laden layer from its stems by its cilia" and its great "aeration area".

 

Given that photosymbioses must have more solute exchange problems than algae, and that calm wheather is generaly associated with bleaching, this effect needs further analysis.

 

Lipid phases and permeability

Change of lipid phases with warming and temperature-induced membrane thermotropism was suggested by Gates et al. (1992) as a possible origin of bleaching, perhaps through change of ions permeability. Some examples of phase changes in lipids with temperature are known (bilayer/non bilayer/gel), as well as destruction of selective permeability, or vesiculation and fusion of membranes (Quinn, 1989). Fluidity increases with temperature and lipid unsaturation, and thus the permeability of membranes to gaz and ion. In plants, tolerance to heat stress (5 hours at 37°C or more) is correlated with the degree of lipid saturation in thylakoid membranes, and change of their composition is induced by heat-shock proteins. It may allow a conservation of a constant proton permeability, permitting to the transthylakoid pH gradient to stabilize photosynthetic apparatus (Suss and Yordanov, 1986). Conversely, greater unsaturation of lipids of thylakoid membranes confers resistance to low temperature and chill-induced photoinhibition (Murata et al., 1992).

 

Bleaching in the reefs is provoked by temperature change of one degree or less, and membrane lipidic compositions are surely under adaptative mechanisms. Phase transitions are smooth for natural heterogenous lipid membranes. Membrane perturbations by temperature during bleaching events needs further informations. Destruction of membranes by oxygen radicals and lipid peroxidation is also a possibility. It could be measured by malondialdehyde analysis, or tested with an enhacer of the phenomenon, such as Rose Bengale (cf. Hodgson and Raison, 1991).

 

Preliminary results of ESR study with 12-doxyl-stearic acid showed linearity and no lipid phase transition between 15°C to 35°C in Aiptasia, and between 5°C to 40°C in Pocillopora damicornis (L. Muscatine, com. pers., 1994).

 

Exocytosis

Many works have been done on this subject, particularly in medecine, but unless concrete informations are available in bleaching, it cannot be guessed which information may be usefull. There is often internal acidification (pH -0.5) and increase of internal free calcium (from 100nM to 250nM), but "it is not mandatory nor sufficient" (Plattner, 1989, Almers, 1990, examples in Cannon et al., 1985, Madshus, 1988, Blackbourn et al., 1991). The relation between internal pH and calcium is widely known (Frelin et al., 1988, Evans et al., 1991). Use of 1µM ionomycin which raised internal free calcium pool from 50nM to 600nM did not provoke bleaching (L. Muscatine, com. pers., 1994).

 

Stress proteins

Stress proteins will surely be found in bleached corals and other photosymbioses, as "heat shock", "stress proteins" are synthesized within minutes of exposure to adverse environmental conditions and persist over time. Stress proteins are found in many taxa from prokaryotes to plants and vertebrates and are highly conserved. The heat-shock response has long been implicated in acquired thermotolerance (reviews by Lindquist, 1986, Bradley, 1990, Nover, 1991, Ang et al., 1991). High light alone, or in synergy with temperature, can induce, shift or reduce expression level of stress protein (Knack and Kloppstech, 1992). Heat shock protein can protect against photoinbition (Schuster et al., 1989). Oxidative stress response bears similarity with heat shock one (Tsang et al., 1991, and ref. herein).

 

The heat shock proteins ("Hsp") have been studied in the freshwater symbiotic cnidarian Hydra by Bosch and Praetzel (1991, and ref. herein). An Hsp60 protein was synthesized under warming (33°C durin 2 hours, instead of 18°C). This protein has no relation with ubiquitous Hsp70 family, which is also present in normal condition. One species, temperature sensitive and found only in environment with constant temperature, lacked the ability to synthesize three major groups of stress proteins. Heat shock proteins are produced at 33°C to 40°C after two hours, whereas an Hsp 70 is already present in subtidal corals at 28°C (B. Brown, oral com., Luxembourg 1994). Different Hsp were also produced by Montastrea (Hsp74) and Aiptasia (Hsp68 and 72) under short term temperature shock of two hours at 33°C and 35°C ; none were evident at 31°C for one week (Black et al., 1995, and ref. herein), casting doubt on a link with bleaching.

 

While the presence of stress proteins would not be very informative of the nature of the stress, they might be used as in situ or in vitro early indicators, as well as to track late events in the mechanisms of bleaching.

 

Warm-induced bleaching in Euglena, a model ?

Euglena is unique among higher plants and algae in that "the chloroplast can be completely gratuitous for growth, provided a utilizable source of organic carbon", in a certain sense as for bleached corals. The parallelism of some features in the heat bleaching in Euglena with those of reef symbiosis are noteworthy (see Brandt, 1988, Ortiz and Wilson, 1988, Ortiz and Kutner, 1990, Ortiz, 1990a, 1990b, 1991, 1992, Conkling et al., 1993 and ref. herein):

- Euglena gracilis bleached when transferred from 23°C to 32-34°C. It is a low temperature compared to heat-induced bleaching in all other plants and algae, which occurs only at temperature superior to 40°C;

- warm-bleaching in Euglena appears to target the chloroplast almost exclusively since heterotrophic growth remained unimpaired at warm temperature. Euglenid chloroplasts are though to be derived from symbiotic green algae, implying a relatively late acquisition of endophotosymbioses within this group (Knoll, 1992);

- light intensity is important for warm-bleaching, and light incubation is needed to eliminate non permanent bleached strains. UV (and streptomycin) can also induce bleaching;

- bleaching-sensitivity is variable among Euglena strains, and depends of culture phase and growth rate;

- loss of chlorophyll is exponential (about 50% loss in 24 hours) and becomes abruptly irreversible after 40 hours (corresponding to 4 generations). In young cultures, an increase of chlorophyll content of up to 50% is observed whithin the first 15 hours, tentatively reminiscent of the observed increase of chlorophyll in some bleached coral symbionts.

Brandt (1988) hypothetized that one of the two chloroplastic tRNA polymerase become non-functional at warm temperature, and is possibly related to heat-sensitivity of 70S ribosome assemblage, but this is not the root-cause of heat-bleaching (Conkling et al., 1993). Some other observations of the process of bleaching at biochemical level are herebelow summarized:

- there is an overall increase of nucleocytoplasmic biosynthesis, coresponding to the temperature increase, whereas chloroplast biosynthesis is reduced. Putative heat shock proteins of nuclear origin are imported in the stroma (63kD) and thylakoid (60kD);

- the increase of chlorophyll in the first phase could be a side effect of glutamate pool build-up. The level of most of the chlorophyll-binding proteins of chloroplast origin decreased at the same time. Degradation of chlorophyll is thereafter controlled by nuclear proteins. Thus the bleaching process is not determined primarly at pigment level;

- the quantity of the large unit of Rubisco (chloroplastic coded), as well as PSI and II polypeptides, decreases sharply whereas that of the small one (imported from cytoplasm) increases in the first 15 hours. This corresponds to a loss of the carbon-fixing enzymes and of their regulations;

- "heat bleaching" at 33°C and "heat shock" at 36°C appears as two different processes, at least in term of protein biosynthesis.

Perhaps the process of heat bleaching in Euglena has nothing in common with that of reef symbioses, but it may provide at least an interesting heuristic.

 

THE CO2-PHOTOINHIBITION HYPOTHESIS

 

We will now summarize our hypothesis that the recent worldwide mass bleaching is due to the CO2 rise through symbiont PS II photoinhibition. We came to this conclusion in time when CO2 and PS II were not edvoked, but after a carefull integration of all the available data (1992 first version of this report, Pêcheux, 1993, 1994). Contrary to common scientific problematic, reef mass bleaching, a global question, must be understood from connections and convergences of all science disciplines, from biochemistry to physiology, ecology, oceanography, planet systems and geology, as we tried to do in this review.

The first argument in favor of the CO2 hypothesis is that it is not other factors, warming, weather nor ultra-violets. Global Warming is easely considered as responsible, but it is all but global. No warming trend exists in Caribbean. Clear cases of bleaching when temperature was not exceptional are convincing. Long term data in place are inexistant, and when it is suspected temperatures to be above previous maxima, the increase is one or two tenth of a degree at most (apart the El Niño 1983). Peak bleaching of large foraminifers occurs at normal temperature before warmest time. There is no evidence of a hydrological change which could affects all biotopes of reefs worldwide, even given the climate shift around 1976. One can be sure there was already some severe dolldrums in tropics in the past. Water transparency can not have change significantly, certainly not in very shallow waters. We know that ultra-violets, often suspected, did not increase over reef areas. This lets CO2 rise as the last Global Change which must be carefully examined.

 

CO2 rise explains well the global distribution of bleaching, and its occurence in every environments, from island to barrier reef, from surface to great depth, from lagoon to reef slope, in pristine or polluted areas. It is the "invisible factor", affecting all them equally.

 

CO2 is a fundamental factor of photosynthesis, which is at root of bleaching. The change is of great magnitude, 30%, with an induced acidification of 21% more H+ concentration. And precisely, reef photosymbioses, at least corals and large foraminifers so far studied, react in a particular way in front of CO2 augmentation : they are inhibited, at contrast to most other phototrophs. They use bicarbonate instead, and in a pH dependent way, certainly using H+ gradient.

 

In fact we are convinced this caracteristic is at origin of the symbiotic nature of the reef primary organisms, like dominance of C4 plants in land tropics. It allows an efficient inorganic carbon pumping in an environment where photorespiration is critical due to low CO2, high temperature and oligotrophy, by multistep bicarbonate gradients and/or high CO2 around symbionts, the latter case explaining perduration of symbioses.

 

Essentially, reefs are CaCO3 producers, since the first stromatholithes more than 3.5 Gyr ago. Marine CaCO3 producers regulate Earth carbon cycle, hence temperature, through pH-controlled bicarbonate uptake for photosynthesis, this is our next work. Shutdown of biocalcification is the obligate Earth response in front of CO2 rise. CO2 level in the geological past was higher, but this did not mean low pH; long term evolution of Rubisco compensated lowering of pCO2; and it is a question of change speed.

 

At ecosystem level, there are few works but they indicate great sensitivity of reefs to CO2. Apart lagoons, which indeed have low coral cover, no natural experiment is available since CO2-rich waters are also cold and eutroph. The general reef deterioration probably represents the first effect of a general CO2 stress by loss of competitivity of corals.

 

We take CO2/pH as the fourth factor of bleaching, after temperature, irradiance included UV, and water agitation, but it is the changing one. There is strong synergy between CO2 and the others: temperature increasing photorespiration, thus inorganic carbon need, light also, whereas low agitation restrains its uptake. It is a marginal physiological factor, difficult to put in evidence especially given its slow rise. It acts strongly only at the stress limit because maximum photosynthesis. It is an evidence that reef ecosystem is at limit given bleaching, whatever the cause. Strong biotic competion has selected adaptated organisms with just the sufficient high cost resistance to extreme stress, either in lagoon with great diurnal variations or in ocean bathed open site, stable for temperature or light; and pH as well.

 

The CO2 hypothesis does not explain well why mass bleaching began so suddenly around 1980. But perhaps the similar anthropogenic 80ppm CO2 rise, the glacial/interglacial one, and maximum air-sea difference in long-residence water lagoons is not casual. Perhaps this change, modulated by the various biota, might help explaining some of the so vexating uncoherent spatial pattern of bleaching.

 

Some indices are in support: the carbon isotope of coral organic matter following the bleaching sensitivity; the most affected Palythoa which has symbionts in ectoderm, the less Tridacna; the interaction with salinity; the observed change of internal pH; perhaps a greatest effect on calcification. A strong one is foraminifer shell abnormalities, only comparable to those just after the Cretaceous/Tertiary boundary, a time of tremendous CO2 release. Also their behavior, as they move upward toward light when lowering pH. We also like the empirical knowledge of an aquariologist, Wilkens, and the "acidosis hypothesis" of Mayer, a fine naturalist early in this century.

 

Today, the only direct physiological experiment gives confirmation of CO2 as a bleaching factor (Annex I). We fully admit the need of replication.

 

But would only carbon starvation throws symbioses integrity in desequilibrium so rapidly ? Among stresses, strong irradiance is maybe the first in reefs, and obvious in bleaching. The delicate light capture threatening flash over, the other fundamental limit of photosynthesis by PS II, the photoinhibition is involved. Quantum efficiency is low in oligotrophic tropic in general, and in reefs in particular. It is not known what is happening at the 30-32°C bleaching temperature. Land plants have on contrary greatest D1 repair at this temperature, but they have other adaptation necessities. Whatever, recent experiments told us that great photoinhibition is among earliest signs of bleaching.

 

Some few arguments more, first of all thylakoid disruption. Also indicative is probable change in xanthophyll cycle, according to rarissim data; perhaps related is the orange color found in resistant corals of Jamaica and in a foraminifer. Weak links can be found with bleaching by red light or DCMU.

 

Of course a definitive answer will come from fluorescence measures during natural bleaching, nonwithstanding it probes any stress. Rather the question is what is in play in PS II ?

 

Our conviction was boosted by the many and fundamental interactions between CO2 and PS II. They are well studied in land plants and microalgae. Different mechanisms exist: by energy sink of photosynthesis or photorespiration, well understood; at the Qa-Qb bottleneck because of low bicarbonate level as many theoritical works indicate; and keeping in mind influence on O2 site or transthylakoid proton gradient. Relevance at ecosystem level remains to be established. Such connections are real albeit complex, according to our first physiological data, and enough strong for making CO2 rise critical in bleaching conditions (Annex II). Its impact is of similar magnitude than other supposed causes.

 

One can suspects other interactions between host and symbionts in the bleaching processes, such as temperature sensitivity of host bicarbonate pumps or distress signals from symbionts. A role of PS I and O2 can not be excluded. Of course there is still work to understand fully all the details.

 

We believe only a coherent picture as here presented on the processes of bleaching, involving the two key enzymes of autotrophy and the base of life, CO2, can explain the crash of an entire primary level of an earth ecosystem, insured by so diverse taxons.

 

The consequence is in deadly earnest. At contrast to warming of seawater which, perhaps, is physically limitated, anthropogenic CO2 will continue to be released at increasing rate, with future level according to best serious economic models with "strongest mitigation option" up to 470ppm CO2 (56% more H+), otherwise maybe up to 1000ppm (197% more H+). Remember that using one oil liter corresponds to ten kilograms coral less for Earth pH regulation. It is difficult not to predict complete collapse of reefs. So I can just hope this hypothesis is wrong.

 

 

CONCLUSION

 

The main knowledge on the reef mass bleaching phenomenon may be summarized as follows:

- it is a global phenomenon, observed worldwide in reef areas;

- it is very recent and sudden (since 1979);

- it is increasing in frequency and magnitude, becoming chronic;

- no coherent spatial patterns emerge (depth, reef zonation, geographic area), except maybe a relationship with groove structures and passes ;

- it affects almost all, and probably all, and quasi-exclusively animal-algae symbioses which constitute the founder of reef ecosystems;

- in very general term, it affects preferentially fast-growing, or better said high-photosynthetizing associations;

- it affects them in mass, with subsequent variable mortality.

 

After almost two decades of mass bleaching events, no consensus emerges clearly about its origin, though it is widely believed to be related to the "Global Changes":

- it is generally associated with warmest temperatures, but not with exceptional ones in a few well studied cases. It is yet difficult to relate it to global warming;

- it is frequently associated with calm weather, which might have various consequences (on hydrological patterns, water transparency, air-sea exchanges, and physiology). Changes in the last decades of other weather parameters during bleaching remain to be studied;

- light is often involved, probably in relation to photosynthesis;

- UV can not be directly, and hardly indirectly, responsible for bleaching, as they have not yet increase in tropics;

- carbon dioxide build up remains as the last serious hypothesis, but rather because its effects are badly known.

The mechanisms of bleaching have not been identified. Hypothesis of internal elevated oxygen tension is contradictory with an increase of respiration. Many fundamental questions are unanswered: decrease of photosynthetic pigments at time of bleaching, process of symbiontic rupture if there is really one, healthiness of expulsed symbionts. There is very few elements to treat the question of whether bleaching originates on host or symbionts side (or, in our opinion, both). Convergent data on a symbiont photoinhibition mechanism are promising.

 

Thus, consequences of mass bleaching is yet hardly previsible. Its continuous increasing frequency and magnitude demonstrates the possibility of a catastrophic issue. Mass bleaching is a threat on all the reef ecosystems, potentially more dangerous than every other disturbances, such as sea level rise or sea level fall (Jacobs, 1992, and ref. therein), local pollutions, or expected future UV increase. It is even more alarming than impact of ozone depletion in polar regions or acid rain on temperate forests, for which remedies are known. If indeed CO2-induced symbiont photoinhibition is the cause and mechanism of reef mass bleaching as we believe, reefs are in very great danger given future CO2 rise.

 

 

Acknowledgments : This work is an updated version of an unpublished report, December 1992, written under a contract of the Observatoire Océanologique Européen, Monaco. Thanks to its friendly members. Update was supported by French RMI n°224397K.

 

ANNEX I: A preliminary CO2 coral bleaching experiment.

 

Material and method

We used tips of 1-3 cm of Stylophora pistillata from Aqaba, 5 m depth, healthily cultivated at 75µE/m2.s (10L:14D) and 22-23°C (maximum temperature in reef of Aqaba is about 26°C).

They were subjected to raises of temperature to 23-25°C and light to 550±30µE/m2.s (12L:12D), in agitated 5-liter aquaria of filtered mediteranean seawater. CO2 bubbling was monitored by pH electrodes, with:

1) normal pCO2, a little more than 360ppm (pH=8.03±0.05); 2) intermediate pCO2, about 1000ppm (pH=7.67±0.20); 3) and high pCO2, about 5000ppm (pH=7.14±0.08). Five replicates were incubated in each aquarium for 60 hours (begining with a dark period) before transfer back to normal conditions, except for light (10µE/m2.s the first week). Photosynthesis and respiration were measured during respectively 15 and 10 minutes twice a day on one clone of each aquarium in a small incubation chamber with O2 and pH electrodes.

 

Results

Nubbins in the control aquarium appeared always healthy, with polyps fully expanded. In the intermediate pCO2 condition, there was visible paling and closure of polyps, which were still partially retracted even two weeks later; one nubbin losed its tissue 5 days later. In the high pCO2 condition, the nubbins showed the first sign of bleaching after two days, then were bleached after 60 hours, with zooxanthellae clumps sticking to bottom aquarium. Then they progressively losed their tissues before they died all within 2 weeks.

Photosynthesis and respiration of the three clones showed transient patterns during the first 24 hours. Thereafter, for the control nubbin, the photosynthesis was normal within 3-5 days after a slight depression; in the intermediate condition, it was reduced by one third, with recovery after one week; and for the high pCO2 condition, it quickly fell down to zero with bleaching.

 

Discussion

This experiment demonstrates that elevated CO2 is a stress to coral, inducing bleaching at least in synergy with setting up of summer temperature and light. The question remains if present CO2 rise is of sufficient magnitude to be responsible of mass bleaching. We used strong increase of CO2 in this experiment, but, in term of pH, which is probably the physiological relevant parameter, the 0.4-0.5 pH lowering (at 1000ppm) is to be compared with the present 0.1 pH ocean acidification. The situation is similar to most warm-induced bleaching in laboratory with 2-3°C above in situ maximum temperature compared to few tenth degree of warming if present responsible of bleaching. Moreover, temperature experiments have shown that there is an approximate logarithmic relationship between the speed of bleaching and the magnitude of the stress. As the paling and bleaching observed in our experiment were very fast, compared to weeks generaly observed, it will be not a surprise that present CO2 rise is a relevant factor in reefs.

Increasing pCO2 appears as the marginal factor throwing photosymbioses out of equilibrium when they are subjected to yearly maximum temperature and light. We nevertheless stress that this short-term experiment is preliminary, and must absolutely be confirmed by further ones.

 

 

ANNEX II : CO2 effects on PS II in symbiotic foraminifers.

 

Introduction

In view of the results of Strasser et al. (in press) and with M. Havaux (see text), we decided to run a test with several pHs and under high light. Although significance is limited by the few number of replicates and by the short duration, given the many factors involved (pH, light, temperature, day/night alternance, species difference), some of the results are judged whorthy of presentation, as they are the first ones on CO2/PS II interactions.

Material and method

Known to bleach worldwide, we used the large foraminifer Amphistegina lobifera, symbiotic with diatoms (mostly Fragilaria sp.), freshly collected from Mauritius in back reef, 1.5 m depth. Identical measures were made with the french Mediterranean Sorites variabilis, harbouring Symbiodinium sp. like corals, coming from same depth, and cultivated for 9 months in a closed aquarium at 25°C, 50-100µE/m2.s and 8.2-8.5 pH.

 

After selection and cleaning, ten specimens of Amphistegina or one of Sorites were distributed per glass-tubes of 5.5 ml Mediterranean seawater (40 S, standart alkalinity 2720µM Eq./kg). Seawater was exchanged daily with pCO2-controlled pH at either 7.50, 8.00, 8.50 or 8.90±0.03 (with parallel upward daily shift of 0.10±0.05, slightly more in low pH condition probably due to lower calcification). Light was delivered with a 12:12 hours cycle by a Life-Glo day-light fluorescent tube, at either 200µE/m2.s (two replicates per pH) or 500µE/m2.s (one replicate). Thermostated temperature was 25°C, raised to 32°C the third day light period. Then, recovery began back to culture condition (8.24 pH, 70µE/m2.s and 25°C).

At the end of test, most Amphistegina were actively moving, few were non-adhering. One third to two third within tubes showed browner last chambers contrary to normal, 10-30% even with small clump of expelled symbionts at aperture. One was truly bleached (but with orange color !), at the lowest pH and highest light. Sorites were often paling or discoloured in patch, about one third with whiter center and darker periphery, and some with few expelled symbionts at rim. They moved less, particularly under strongest light or lowest pH. None were firmly sticking as usual. Care was taken to clean the clumps of expelled symbionts each day with seawater change. A tentative to measure their fluorescence was unsuccesfull because of the low quantity.

 

Fast fluorescence kinetics was measured with a Plant Efficiency Analyser (PEA, Hansatech Ltd., UK), with 1 minute predarkening in days and 5 seconds full exciting light (about 6000µE/m2.s). The principle is simple. Strong red light (630nm) is directed onto the sample, absorbed by chlorophyll pigments, and photosystem II reemits through antenna few percent of energy in light at higher wavelength (680nm). This signal is measured with 12 bits precision from 10 microseconds to seconds after light probe on. Fluorescence is proportional to reduced quinone A (Qa-). The charge curve shows clearly the rate of energy capture by chlorophyll antenna in the first hundred microseconds, trapping by the PS II reaction center, then transported as electron charge through the Qa-Qb quinones, and finally to the plastoquinol pool (PQH2) after tenths of milliseconds, and to PS I and CO2 fixation (Fig. 1). The polyphasic fluorescence rise shows transients called O-J-I-P sequence, which was analyzed according to Strasser et al. (1995) and Strasser and Strasser (1995). The fluorescence at 50µs gives the Fo level and the maximum is the Fm level. The Fv/Fm (=(Fm-Fo)/Fm) corresponds to the quantum yield of primary photochemistry (ratio trapped/absorbed exciton, here called Trap/Abs). The probability that it is transported into the electron chain (Transp/Trap) is given by 1-Vj, Vj being the relative variable fluorescence (Vt=(Ft-Fo)/(Fm-Fo)) at the first transient J, occuring around 2 milliseconds. The specific energy flux of trapping per reaction center per millisecond (Trap/RC) is deduced from the initial slope of fluorescence rise from 50µs to 150µs normalized by Vj as before Qb reduction (dV150µs/Vj). Absorbtion and transport per reaction center (Abs/RC and Transp/RC) as well as overall ratio of transport per absorbtion (Transp/Abs) can be then deduced.

 

Results

Temperature-light effects

The pattern of the main parameters pooled by pH (n=4 at 500µE/m2.s or n=8 at 200µE/m2.s) are given for Amphistegina in Fig. 2. Oscillations are not seen the 1st day (though lowering of Fo and Fm, not shown), but clearly the 2nd and 3rd stress days. An usual stress pattern is observed, with the lowest values of efficiencies Trap/Abs and Transp/Trap in the morning. Abs/RC is greater and the Transp/RC smaller, in anti-parallel behavior. The rate of trapping per reaction center rises only slightly ("cruise control") albeit regularly during stress (by 19% at 200µE/m2.s and 29% at 500µE/m2.s, r2=0.268 and 0.474, p<0.0001). As expected, the swings are exacerbated at 500µE/m2.s compared to 200µE/m2.s for all parameters. Good recovery can be seen the 5th day.

There is no clear influence of higher temperature the 3rd day. The values are similar or better the 3rd day at 32°C than the 2nd day at 25°C, apart the Transp/Trap at 500µE/m2.s in the morning.

 

pH effects

Parameters were pooled for the stress period (Fig. 3). No effects are sure (p>0.02) for Trans/Trap (p=0.79) or at 500µE/m2.s though some high correlation coefficients. Otherwise systematic differences at p<0.02 are visible, with 7-22% difference. The most strinking feature is that high pH acts as a stress pattern for Trap/Abs and Trans/Trap, only opposed for Abs/RC at 200µE/m2.s (p=0.0001). High pH shifts values in the same direction as high light does.

Perhaps the most interesting trend is found in Trap/RC (p=0.0001, only 0.034 for the four highest light samples), lower with higher pH. The pH effect over the range 7.5-8.9 is far greater than the 200-500µE/m2.s light difference, this time low pH corresponding to high light. It is slightly less than the rise during the three stress days. Similar observations can be made taken only the shifts between end night and early morning (not shown).

One surprise was that good correlations were observed in recovery runs, two days after the pH was set equal for all specimens, and with an opposite trend than in stress period. Concerned parameters were mainly Trap/Abs (p=0.0008) and Transp/Trap (p=0.0041). In fact correlation coefficients between Transp/Trap and pH showed a regular increase with time (r2>0.347, p<0.0055 for either 200µE/m2.s, 500µE/m2.s or pooled). The best correlation was found for end day recovery Trans/Trap normalized by their initial levels (r2=0.849, p=0.0001) (Fig. 4).

 

Another clear pH effect is seen on the level of fluorescence at step I (either at 30 or 60 milliseconds), recovery values excluded. Low correlations were found but with high signifiance (r2=0.105, p=0.0001, n=120) (Fig. 5). This pH trend was more expressed in night (r2=0.374) and moreover in the 200µE/m2.s samples (r2=0.486). Parameter changes is only 3% but calculated on the more relevant scale 1-Vi, they are 30% for pooled values, 58% for 200µE/m2.s samples in night.

Measures on Sorites are in general agreement with these finding, but they were less clear, mainly because of greater variability in the day/night oscillating pattern, a surimposed temperature effect (increase of Trans/Trap at morning) and often apparent opposite pH trends between the two light levels. In general, the pH trends are better for 500µE/m2.s samples, but with only four values, confidences are not enough to be presented. As in Amphistegina, a strong negative correlation is found between pH and Trap/RC, rather at end of the days (r2=0.226, p=0.0034), or in the case of 500µE/m2.s samples all along the stress days (r2=0.392, p=0.0031). In electron transport 1-Vj, an influence is seen also mostly in recovery runs (pooled values: r2=0.392, p=0.029; 200µE/m2.s: r2=0.531 p=0.04; 500µE/m2.s: r2=0.812, p=0.09), and for Trap/Abs, only in 500µE/m2.s samples (r2=0.812, p=0.11). Vi levels are similarly affected (r2=0.090, p=0.0005), but at contrast to the former species, better in days, and even more only for high light condition (r2=0.573, p=0.0001). Time of J step was often shorter than 2 milliseconds, with one fifth below 1.5ms and as fast as 0.7 ms (as Iglesias-Prieto, 1995, found in cultivated Symbiodinium), but calculated Vj values were at worst 5% less then true values.

 

Discussion

The most clear influence is the greater stress under higher light. Yet the level was far less than experienced by those very shallow water organisms in a sunny day, which must reach about 1500µE/m2.s. Thus, strong photoinhibition is certain to exist under bleaching conditions.

 

The abscence of temperature responses in Amphistegina, and even a better electron transport at morning in Sorites, was surprising. The temperature stress was clearly seen in in a similar experiment but running a week (Strasser et al., in press). It is hard to say if it is a lag effect, as for light the first day, or a strong light-induced thermoprotection with already an adaptation or saturation to further stress. In natural medium, bleaching of Amphistegina arises because of strong light more than high temperature, which can be as low as 27°C, seemingly unlike corals (see the works of Hallock team).

Let us come now to the most interesting factor, pH. First, we are in trouble that high pH acts as a stress, contrary to our hypothesis, with lower trapping and electron transport. The difference with O2 photosynthesis response (ter Kuile et al., 1989a, 1989b) suggests that low pH promotes photorespiration but not down regulation of primary photochemistry. The surprise that pH trends were of opposite sens in recovery period points to a long lasting effect of pH, increasing with time. Alternatively it is possible that that an high pH stress induces a better adaptation at long term. The pH factor is different than temperature and moreover light which vary strongly at year and daily scale, and this must be kept in mind in future physiological experiments.

 

The influence on the trapping rate Trap/RC was clear, higher pH mitigating its regular increase during the stress period. The consequence is important as a lower rate is the best protection against overloading. Here again, at contrast to temperature and light general effects, there is no "cruise control" of trapping rate in front of acidity change. No evident interpretation can be now given on the origin of this effect, unexpected given current litterature. Thermal dissipation by xanthophylls at reaction center, or an influence on the PS II donor side (cf. Speata et al., 1997) might be suggested.

 

The lower level of fluorescence with high pH in late PS II processes as measured by Vi as an obvious explanation, sink energy by CO2 fixation (very slow Qb reducing centers or side effect of Trap/RC can not be excluded). But if it is logical to found a stronger effect during the day and under high light as in Sorites, why so in night for low light samples in Amphistegina ? One can think on a difference between diatom and dinoflagellate symbionts. I will also remark that pH regulation is quite distinct in imperforate foraminifers as Sorites, where calcification balances photosynthesis as soon, and in perforate foraminifers as Amphistegina, which buffers carbonate ion equivalent for one or two days before calcification (ter Kuile and Erez, 1988).

 

Conclusion

First, fluorescence measure and particularly fast kinetics appears as one of the best tool to monitor bleaching process.

It was seen that pH has effects at several sites of the PS II: clearly at the reaction center core (Trap/RC), some at the electron transport, and by downstream CO2 fixation. The weak effect at the Qa-Qb site was deceptive, but one may remember that during the one-minute predarkening, bicarbonate pool has time to fill up.

The magnitude of the pH effects were not very pronounced, considering the wide range of experimental pH, 7.5-8.9, compared to the global change of 0.083 lowering due to the CO2 rise. Nonetheless, we consider that the predicted few percent change in PS II activity nowadays can be critical under the strongest summer stress in nature, particularly if the several kinds of influences are additive. In every cases, the pH impact appears of similar or greater magnitude than the other usually considered causes of mass bleaching, a 0.1-0.2°C warming or a slight increase of light due to undetected lower water agitation.

Albeit its limitations, this test emphasizes the complexity of pH interactions with PS II, calling for further study given the implication, as we are programming to do unless limited ressources.

 

Acknowledgements

We are thankful to R.J. Strasser and M. Tsimilli-Michael for friendly invitation, help and laboratory support. Thanks also to N. Von Arnim for supplying large foraminifers from Mauritius.

 

 

ANNEX III: Formation of dense, hot, hypersaline surface water.

 

We present a simplified calculation on the conditions which can lead to the formation of dense, hot, hypersaline surface water, which was evoked as a possible cause of mass bleaching (see under "Water stratification"). We take into consideration that the moisture deficit increases in tropics (Flohn and others, Kumar et al., 1994, Graham, 1995), enhancing the evaporation, might be responsible for such a new phenomenon. It is an example, and has to be checked in real cases of mass bleaching.

Standard water characteristics during summer are T=30°C and S=35°%. Take a temperature increase of 1 °C, supposed sufficient to elicit the bleaching response. If water is to conserve constant density (theta=21.73), salinity must increase by 0.46°%, i.e. a fraction f=0.46/35=0.0131 (dS/dT with theta(S,T)=constant) (seawater state equation from Unesco, 1983). This requires an evaporation which, if not counterbalanced by energy input, would cool the water by :

 

ÆT=f.Lv/(1-f).Cp Å 9.5°C

 

with the latent heat coefficient Lv=2.42 106J/kg at 30°C and the specific heat capacity of seawater Cp=3.391 103J/kg.°C at the specified T and S.

Thus, during the formation of hot dense water, the energy taken away by the evaporation E must be superior to 9.5 times the net energy input, Qnet, corresponding to the warming of 1°C of the water, so the condition: {E>9.5*Qnet}.

Qnet can be calculated from the energy balance:

 

Qnet= IRup + IRdown + S + H + E

 

with IRup, outgoing longwave radiation, IRdown, downward longwave radiation, S, solar energy input, H, sensitive heat exchange, E, latent heat exchange during evaporation.

A crude evaluation of Qnet for a theoretical case is made to emphasize the importance of the most imprecise parameters: downward longwave radiation, albedo, wind speed and latent heat transfer coefficient. We suppose no advection, clear sky, and mean winds of 6.8m/s (Flohn et al., 1990). IRup is Å 480W/m2 at 30°C, with dIRup/dTÅ6.1W/m2.°C (black-body law), IRdown Å 190W/m2 to a maximum of 280W/m2 (Ramanathan and Collins, 1991). The annual mean equatorial sun input of Å424W/m2 must be reduced by the albedo, which is 6% over oceans, and is more important over reefal areas (perhaps up to 30% ?). The sensitive heat H is taken as 10-20% of E (Bowen ratio). The latent heat exchange during evaporation E is given by:

 

E = roair . CE . Lv . UZ . ÆQs

 

with air density roair = 1.15 kg/m3 at 30°C, latent heat transfer coefficient CE = 1.2 10-3 (which maybe lower in stable condition, see Walker et al., 1987), latent heat coefficient Lv=2.42 106J/kg, wind speed UZ =6.8m/s and moisture deficit ÆQs=5g/kg in 1949 and 5.6g/kg in 1979 (Flohn and others), giving E=114W/m2 and 128W/m2 in 1949 and 1979 respectively. Qnet can take a wide range of value, from -153W/m2 to 64W/m2.

Qnet can be evaluated from the rate of warming. A typical condition during bleaching is a lagoon with a water column of 5 meters which warms by 1°C in 15 days. This gives a Qnet of 13.4W/m2. Our condition {E>9.5*Qnet=9.5*13.4=127W/m2.day} was therefore not verified in 1949 {114<127} but it became true in 1979 {128>127}. The hypothesis that present change of hydrological cycle in tropics is involved in mass bleaching through hot dense water formation is thus rather sustained.

There is one strong argument against this hypothesis. Mass bleaching is often associated with dolldrum time (mean winds<1m/s), in which case E<17-19W/m2 only, excluding dense water formation, although one may consider water residence time and stratification. Also, there might be examples of bleaching in pure oceanic water with short residence time over reefal areas (as probable in the small Ile Plate, Mauritius, pers. obs.).

 

 

Annex IV : Implications for the carbon cycle.

 

Some concerns were raised about the implications of coral bleaching for the carbon cycle, with the idea that the reduced calcification in reefs and their erosion would slow down the sequestration of carbon, and therefore accelerate CO2 build up and greenhouse effect (Lang in Holling, 1988, Glynn, 1989).

First of all, calcium carbonate precipitation releases CO2. This counter-intuitive effect is due to the remouval of the two charges of alkalinity of CO32-, promoting release of H+ and acidification, and thus formation of CO2 gaz. Nowadays, precipitation of 1 mole of CaCO3 in ocean induces formation of 0.64 mole of CO2 (Ware et al., 1992). This is the well known alkalinity buffer factor. It is increasing and should be around 0.80 in the future at the expected 1000ppm CO2 atmosphere (using equations of UNESCO, 1987, and mean ocean surface water values). Conversely, increase of dissolution, or reduction of calcification, absorbs CO2, taken constant other fluxes of Earth carbon cycle. Decay of reefs is thus a CO2 regulator. In the same way, at constant pH, the exhausted CO2 from one liter of oil dissolves of 10kg of CaCO3 .

Secondly, as for the human CO2 perturbation, reef process is insignificant, at least for the next few centuries. Man releases 7±1 GTC/y (gigatons carbon/year) (IPCC, 1990, 1995), whereas reefs have a net sedimentation of only about 0.1 GTC/y in CaCO3 (Smith, 1978, Kinsey and Hopley, 1991). Organic carbon sedimentation in reefs is negligible (0.003 GTC/y, Crossland et al., 1991).

 

Runs of a standard 3-boxs model of the global carbon cycle (Pêcheux, 1993b) show that reefs are responsible at scale of thousand years of only about 17ppm CO2 in the atmosphere (between 10 and 20 ppm according to Berger and Keir, 1984, Berger, 1982, not Walker and Opdyke, 1995). Theoretically, would reefs totally stop their 0.1 GTC/y net CaCO3 precipitation, it would absorb 0.064 GTC/y of CO2. But with ocean mixing, the decrease of atmospheric pCO2 is of only 3 ppm in 100 years, with an indetectable increase of alkalinity of ocean surface layer of 12.3 µEq./liter. Given the human perturbation in the future, the maximum possible effect of total reef decay would occur in about 10000 years, when it could reduce by 15% the CO2 atmospheric content, some 440 ppm instead of 500 ppm, which will be not quite negligible (unpublished) (see also Morse and Mackenzie, 1990, Walker and Kasting, 1992).

 

However, net sedimentation may become negative by the reduction of biocalcification and the increase of dissolution (and not to speak of a direct, perhaps fundamental, CO2 dissolution effect). It must be comptabilized dissolution fluxes if gross calcification by organisms is halted whereas dissolution goes on. The ratio of gross/net calcification is very insufficiently known and only the net dissolution at night has been detected (Chave et al., 1972, Sournia et al., 1981, Barnes and Devereux, 1984, [measurements during a bleaching event], Barnes and Lazar, 1991, Copin-Montegut et al., 1992, abstract), but it would multiply our previous estimation of reef importance in the carbon cycle by 2-10 times. Ultimately, mass bleaching of symbiotic planktonic foraminifers, which are mostly tropical, must be logically envisaged (Hallock and Talge, 1993, Hallock et al., 1995, J. Erez, com. pers.), as the discoloration observed once in medusans also suggested. Given recent frequency of observations on planktonic foraminifers, such a phenomenon would certainly be missed for a long time. Those organisms normally precipitate around 0.7 GTC/y in calcite (pers. evaluation).

 

 

Annex V : Shell abnormalities in large foraminifers, Mauritius, 1989.

 

In Mauritius, November 1989, about six months after a bleaching event, we analysed faunal composition of samples, dead and living forms together, but with fresh shells recognizable (Hottinger and Pêcheux, 1991). In addition, high resolution X-ray microradiographies were realized in order to study biometric. In Heterostegina, chamberlets walls were often bifurcated, atrophied, reduced to a simple notch, even absent. Spectacular abnormalities were affecting chamber walls, which were incomplete, distorted, often given an impression of a "soft calcification". Breakage and reparation were currents. Specimens cultivated in aquarium in Basel have followed calcification in a normal way. From our limited data set, we found a correlation between the magnitude of the abnormalities and the destruction of the biotope after bleaching (algae cover on coral skeletons, etc...). Operculina also showed such abnormalities. Some deep samples had only quite regular specimens, whereas in front reef, 5-25 m depth, where bleaching was most noticeable, there was high frequency (13-31%) of slight irregularities (small chambers not reaching the margin, slightly wrinkled, spirale withinside folding, covering of ombilic) to strong abnormalities (8-47%) (very irregular chambers, angular spirales, supplementary calcification at ombilic, excrescence, supplementary counter spirale). Borelis, which can be observed only with binocular, showed in their last whorl from an abscence of some chamberlet partitions to true architectural chaos, up to 37% in one sample. Amphistegina appeared in our survey as the most resistant (hence the significance of Hallock team work !), with strong abnormalities only in one of the studied sample (5m depth, front reef), with 10-20% of deformed spire or calcification at the ombilic. It is noticeable that Elphidium (symbiotic with free chloroplasts) also show some alterations in this sample.

 

Important is that such malformations are observed in pristine areas (local pollution has been suspected at first in Mauritius), in front reef well bathed by open ocean water, in the most stable conditions.

 

 

Annex VI : Some recommendations.

 

In addition to the recommandations of quality established by Odgen and Wicklund (1988), D'Elia et al. (1991) and IOC (1992), we would like to put emphasis on some points :

- coordinated and state-of-the-art biological studies to identify rapidly the mechanisms and causes of bleaching;

- accurate and very complete studies at some few sites of all meteorological, hydrological and physico-chemical aspects, with extensive sampling and follow-up of mass bleaching as soon as it appears, and as much as possible, before, together with a multidisciplinary task group ready on alert for sampling (W90); we think in particular of Looe Key or Jamaica where it seems that mass bleaching can be foreseen a week or so with good probability;

- studies at global scale of all the weather parameters and not only temperature (winds, wind pattern, sky cover, outgoing longwave radiation, water saturation, moisture deficit). Data exist in global databases, easily accesible by dedicated programs;

In situ measurements during calm weather/mass bleaching:

 

Hydrology:

- current patterns;

- density; density flow; density difference inshore/offshore;

- stratification, with top and bottom thermometers;

- spur and grooves flows, with flow, temperature, salinity, density;

- water agitation (plaster technique at least);

- sea surface state (waves, surface slick), refraction and sun glitter,...

 

Chemical analysis:

- O2, pH and alkalinity transects, in particular during calm weather;

 

Biology:

- fluorescence studies of photoinhibition;

- short and mid term in situ 14C incubation, and later compound analysis (healthyness of symbionts, photorespiration,...);

- collection of bleached corals and other organisms as soon as possible, and conservation in dry ice and at 77K in liquid nitrogen; if possible collection at 77K of expulsed symbionts. In prevision of a) microscopy and count of symbionts; b) pigments analysis, in particular carotenoids, diadino and dinoxanthin; c) fluorescence studies of PSII state (photoinhibition ? degradation ?); d) histopathology studies (thylakoids,...);

- observation of daily growth band to know when metabolism shut down;

- organic 13C fractionation;

- perhaps collection of sea water and test if toxic.

 

Biological experiments

- first, effect of temperature on freshly isolated and cultivated symbionts;

- uses of other symbioses: Tridacna (in particular for O2/CO2 exchange and for respiration), larger foraminifers (with diatom symbionts), sponges (with cyanobacterial symbionts, and non-photosynthetic ones); alcyonarian;

- physiological reaction at the very beginning of bleaching: photosynthesis, respiration, calcification;

- chlorophyll fluorescence (and pigments, xanthophylls, ß-caroten of the core complex);

- O2 and CO2 effects on bleaching;

- 14C incubations, with temperature or other bleaching stress, for analysis of sugars, other counpond, enzymes. Reaction product of photosynthesis (glycolate) ? Where and to what is used increased respiration (pulse-chase) ?;

- test of H2O2 oxydo-reduction state. Maybe use of paraquat.

 

Long-term monitoring

In addition to standart monitoring, we propose two other long-term monitoring techniques, which we believe are quite whorthy:

A fixed automatic underwater camera, which would take one photo each day, in order to pinpoint timing of bleaching. It could be powered by solar energy, and might have a flash and a windscreen wiper. If constructed by one technical center, its cost can be minimum.

 

A memory center

In addition to data immediately useable, the coral reefs monitoring should gather samples and records which will provide informations of actual reefs state in the near or distant future. One such kind is conventional sediments samples (in particular in respect to displacement of carbon system and increased corrosivness of seawater). Another kind of record is seawater sample, at best frozed in nitrogen liquid. I was impressed by a remark in an article on isotopes of methane aimed to identify the source of this greenhouse gaz, without historical background, which stated "the best experiment in the last century would have been to sealed a bottle of air for the futur generation studies". This remains true for us. Samples can be quite small, let's say one deciliter or centiliter : carbon isotope study, from which reef global photosynthesis can be deduced, actually needs nanogramme. An other example is the study of the fractionation in the lipids C37 alkenone (produced by coccolithophorids) which is probably a record of CO2 pressure, and which analysis can be conduced today on milligramme sediments samples. DNA analysis requires nannogrammes. Analysis techniques in 20, 50 or "many" years will be somewhat powerful. Identification of all trace biochemical compounds will gave insights on trophic complexity of actual reefs. "One liter of seawater in thirty years might provide more informations than all our present work" (B. Salvat, Monaco whorkshop, Dec. 1991). We propose therefore to stock at low temperature samples collected at least once, with some seasonal, diurnal and local spatial series of :

- seawater samples of small volume;

- filtered seawater, plankton tows residus and sediments, filtered and not;

- some biological materials of principal and accessory groups.

Enough informations can be stocked in one/a few m3 for the whole reef ecosystem. The ratio cost of stockage/quantities of stocked informations is probably one of the lowest than can be imagined. Whether or not such samples are sufficient to save biodiversity is a distinct question.

If one thinks what could provides in the future critical informations on reefs and their dynamics, some fews supplementary suggestions can be made :

- organized and regular films with sound records;

- with some taste of science-fiction, frozen brains of reef fishes and octopus, juveniles and adults. Who know better what is a reef ? and these compact nanomoles of informations are surely possible to be retrieve in some not so far future.

 

 

(References)

---

MARS © Copyright 1997 - All Rights Reserved