Finding of 132, 173‐cyclopheophorbide a enol as a degradation product of chlorophyll in shrunk zooxanthellae of the coral Montipora digitata

We examined the morphology and pigment composition of zooxanthellae in corals subjected to normal temperature (27°C) and thermal stress (32°C). We observed several normal and abnormal morphological types of zooxanthellar cells. Normal cells were intact and their chloroplasts were unbroken (healthy); abnormal cells were shrunken and had partially degraded or broken chloroplasts, or they were bleached and without chloroplasts. At 27°C, most healthy zooxanthellar cells were retained in the coral tissue, whereas shrunken zooxanthellae were expelled. Under thermal stress, the abundance of healthy zooxanthellae declined and the proportion of shrunken/abnormal cells increased in coral tissues. The rate of algal cell expulsion was reduced under thermal stress. Within the shrunken cells, we detected the presence of a chl‐like pigment that is not ordinarily found in healthy zooxanthellae. Analysis of the absorption spectrum, absorption maxima, and retention time (by HPLC) indicated that this pigment was 132, 173‐cyclopheophorbide a enol (cPPB‐aE), which is frequently found in marine and lacustrine sediments, and in protozoans that graze on phytoplankton. The production of cPPB‐aE in shrunken zooxanthellae suggests that the chls have been degraded to cPPB‐aE, a compound that is not fluorescent. The lack of a fluorescence function precludes the formation of reactive oxygen species. We therefore consider the formation of cPPB‐aE in shrunken zooxanthellae to be a mechanism for avoiding oxidative stress.

Reef-building corals harbor symbiotic algae (zooxanthellae) within their endodermal cells. Zooxanthellae produce photosynthate, a proportion of which is translocated to the host corals and used in their metabolic processes. The zooxanthellae are essential for coral growth and persistence. In the last two decades, the world's climate has changed rapidly; these changes have negatively impacted reef-building corals, particularly those inhabiting warm waters. Coral bleaching is an indicator of such negative impacts. The color change in coral results from the expulsion of zooxanthellae by the host (Hoegh-Guldberg and Smith 1989, Gates 1990, Brown et al. 1995, Jones 1997, or from the degradation of photosynthetic pigments in zooxanthellar cells Warner 1995, Fitt et al. 2001). The loss of zooxanthellae has been linked to environmental stressors, including high light intensity and UV radiation (Dustan 1979), elevated seawater temperature (Hoegh-Guldberg and Smith 1989), cold stress (Saxby et al. 2003, Hern andez et al. 2010, Lirman et al. 2011, Paz-Garc ıa et al. 2012, low salinity (Coles andJokiel 1978, van Woesik et al. 1995), low food availability (plankton; Titlyanov et al. 1996), and bacterial infection (Kushmaro et al. 1996). Zooxanthellae are especially susceptible to high water temperature, which damages cells by changing chloroplast morphology and function (Bhagooli and Hidaka 2002. Bacteria accelerate the bleaching process in damaged corals (Higuchi et al. 2013). During the massive Okinawan bleaching event in 1998, zooxanthellae retained in coral tissues changed morphologically and lost pigmentation (Kuroki and van Woesik 1999). Diverse forms of zooxanthellae were observed in tissues of naturally bleached coral during summer (Mise andHidaka 2003, Reimer et al. 2007). Thermal stress inflicts damage in chloroplast thylakoid membrane structures by altering their lipid composition (Tchernov et al. 2004) and inducing the production of reactive oxygen species (ROS; Smith et al. 2005), which ultimately destroys chloroplast organization (Salih et al. 1998). Thermal stress impacts coral in combination with high light stress, but it also has negative effects in darkness that degrade the zooxanthellar photosynthetic system (Suwa and Hidaka 2006).
The details of the bleaching process are not yet well understood. The mechanism of algal cell loss is unclear and there is a paucity of information on algal cell physiological status in stressed corals. Our aim in this study was to take steps toward filling some of the knowledge gaps on these issues through experimental analyses of the coral Montipora digitata. We applied temperature stresses and examined the morphology and abundance of expelled and retained zooxanthellae. We collected expelled algal cells for classification, enumeration and analysis of their pigments (by HPLC). We compared parameters between retained and expelled cells, paying particular attention to pigment composition shifts in response to thermal stress. In addition we examined impacts on coral hosts, including their responses to changes in algal cell physiology.

MATERIALS AND METHODS
Coral samples, aquaria, and incubations. Branches of M. digitata were collected from a single colony growing in the waters off Bise, Motobu, Okinawa, Japan (26°42 0 N and 127°52 0 E) on May 2011. Collected corals were transported to the laboratory in the Tropical Biosphere Research Center (University of the Ryukyus) on Sesoko Island and kept in the aquarium with natural seawater for physiological adaptation over 10 d.
Three branches~5 cm long were placed in each of two glass bottles containing 800 mL of seawater that had been passed through a cartridge filter (pore size 0.2 lm; Advantec Mfs, Dublin, CA, USA). Incubation vessels were maintained in water baths at 27°C (control) and 32°C. Filtered seawater was continually supplied to each incubation vessel at a flow rate of 10 mL Á min À1 and mixed with a stirrer bar. To make observations on expelled zooxanthellae, outlet water was collected in 10-L polycarbonate bottles. These bottles were changed twice daily. Water was collected separately over the following daily time periods: 6:00 a.m. to 6:00 p.m. and 6:00 p.m. to 6:00 a.m. Two liters of the collected water were vacuum filtered through 2.0-lm Nucleopore polycarbonate membranes (Whatman, GE Healthcare, Springfield Mill, UK) on which we observed and counted expelled zooxanthellae; the remainder of the water was passed through GF/F filters (Whatman) for pigment analyses. Light was provided above the incubation vessels by metal halide lamps providing a photon flux density of 400 lmol photons Á m À2 Á s À1 ; the daily photoperiod was 12 h.
Collection of zooxanthellae and pigment measurements. Coral branches were first washed with 3.5% NaCl solution to remove loosely attached plankton and other organisms. Subsequently, we collected zooxanthellae by removing the coral tissue from skeleton by using a Waterpik water flosser (Johannes and Wiebe 1970), which pumped a 3.5% NaCl solution.
Cells were homogenized with a glass homogenizer. The coral tissue solution was centrifuged at 2,810g for 15 min, after which we removed the supernatant. Zooxanthellar pellets were resuspended in fresh NaCl solution. This treatment was performed a second time to remove remaining coral tissue. From the final zooxanthellar suspension, we used 5 mL for pigment analysis and 1 mL for zooxanthellar counts. Mixtures for pigment analysis were passed through GF/F filters (Whatman) using a plastic syringe and filter holder. Three replicate samples were prepared. Data were normalized to the surface area of each coral branch (cm 2 ) as described in the section below (Surface area). Expelled zooxanthellae were collected from the seawater in which the incubated coral had been immersed.
All filters for pigment analysis were stored for a maximum of 1 week at À30°C and then extracted. Cell observations and counts were made within 1 h of collection.
Surface area. The surface areas of coral branches were determined using the aluminum foil method of Marsh (1970). Coral skeletons were carefully wrapped in single layer pieces of aluminum foil, which were weighed. We separately determined the mass/area relationship of foil samples (27 mg Á cm À2 , R 2 = 0.9885, N = 12) and used this value to back-calculate the surface areas of aluminum pieces wrapped around each coral sample.
Observations and counts on zooxanthellae. We classified zooxanthellae into three categories: (i) healthy morphology with normally expanded chloroplasts; (ii) shrunken cells of reduced size with partially fragmented, darkened and shrunken chloroplasts; and (iii) bleached cells with pale or colorless chloroplasts (Fig. 1). Algal cells were counted using an ECLIPSE 80i microscope (Nikon, Tokyo, Japan). Zooxanthellae in coral tissue were counted using a Neubauer-line hemocytometer (Erma Inc., Tokyo, Japan). Filters bearing captured zooxanthellae were mounted on glass slides for microscopic cell counts. Counts were made in 10 visual fields and normalized to coral surface area as (cells Á cm À2 coral) to estimate rates of expulsion over 12 h time intervals. Zooxanthellae and their fluorescence images were photographed under an Olympus IX-72 fluorescence microscope (Olympus Corp., Tokyo, Japan).
Photosynthetic activity of symbiotic zooxanthellae. We measured the maximum fluorescence of symbiotic zooxanthellae (as an index of their photosynthetic activity) using a portable pulse amplitude modulated fluorometer (DIVING-PAM, Walz, Effeltrich, Germany) following the method of Schreiber et al. (1998). Optimal quantum yield was calculated as is the initial fluorescence after dark adaptation, and F m is the maximal fluorescence after dark adaptation (Krause and Weis 1991); accordingly, coral branches were placed in darkness for 15-30 min before fluorescence measurements. Fluorescence data were collect three or more times from different parts of each coral branch.
Pigment analysis. Pigment analysis was performed by HPLC following Zapata et al. (2000). Filters containing coral tissue were cut into small pieces and homogenized in a mill with 3 mL of cold 95% (v/v) methanol. Pigments were extracted by applying 5 min of sonic treatment. Extracts were filtered through a syringe filter (0.2 lm, Millex-LG, Millipore, Bedford, MA, USA) to remove cell debris. To avoid shape distortion by previous elution peaks, the methanol extract (1.0 mL) was mixed with 0.2 mL of deionized water (Milli-Q water, Millipore) immediately prior to injection (Zapata et al. 2000). These extracted samples (200 lL) were immediately injected into the HPLC system (LC-10A; Shimadzu, Kyoto, Japan). All samples were prepared under subdued light and subjected to HPLC analysis within 5 min of extraction to avoid pigment degradation. The HPLC system was equipped 38 TOSHIYUKI SUZUKI ET AL. with a Symmetry C8 column (4.6 9 150 mm; Waters, Milford, MA, USA). Pigments were eluted at a flow rate of 1.0 mL Á min À1 at 25°C using a programmed binary gradient elution system. We used the following solvents: (A) methanol: acetonitrile: 0.25 M aqueous pyridine solution (50:25:25, vol/ vol/vol), and (B) methanol: acetonitrile: acetone (20:60:20, vol/vol/vol). Separated pigments were detected spectrophotometrically using a photodiode array detector (SPD-M10A; Shimadzu) with an optical resolution of 1.2 nm across the 320-720 nm bandwidth (with monitoring of five channels of representative wavelengths). Each peak was identified by comparing HPLC retention times with the absorption spectra of standards and data obtained from photodiode array detection. Photosynthetic pigment concentrations of zooxanthellae in coral tissue were normalized to coral surface area, and those of expelled zooxanthellae were calculated by coral surface area and duration of collection (12 h).

RESULTS
Color of coral surfaces and maximum quantum yield (F v /F m ). After 4 d of incubation, coral branches incubated at both 27°C and 32°C were still brown and there were no marked differences between temperatures (Fig. 2). Maximum quantum yield (F v /F m ) values at 27°C and 32°C were 0.67 AE 0.13 and 0.62 AE 0.07 (mean AE SD), respectively, and did not differ from the initial value (0.64 AE 0.15).
Zooxanthellar count. We observed three zooxanthellar morphologies in coral tissue ( Fig. 1), but bleached cells were rare (0.39% of totals at 27°C and 1.97% at 32°C). As indicated in Table 1, the density of zooxanthellae in coral branches after 4 d of incubation at 27°C was similar to the initial value, as were the proportions of shrunken and healthy zooxanthellae. At 32°C, zooxanthellar density declined significantly (t-test: t 6 = 0.00175, P = 0.002) to 42% of the initial value, and the number of shrunken cells increased from 3.78 9 10 4 to 4.25 9 10 5 cells Á cm À2 , accounting for~18% of total density. The zooxanthellar mitotic indices at 27°C and 32°C were 1.3% and 0.76%, respectively.
Zooxanthellar expulsion rates at the two temperatures are detailed in Table 2 and Figure 3. At 27°C, expelled cell numbers ranged from 3.78 9 10 2 to 2.39 9 10 3 cells Á cm À2 coral during 12 h of darkness, and from 3.06 9 10 3 to 1.82 9 10 4 cells Á cm À2 coral during 12 h of illumination (Fig. 3). More zooxanthellae were expelled during illumination and most were shrunken. At 32°C, expelled cell numbers ranged from 2.27 9 10 2 to 1.41 9 10 3 cells Á cm À2 coral during 12 h of darkness, and from 5.47 9 10 2 to 9.87 9 10 2 cells Á cm À2 coral during 12 h of illumination. Total numbers of expelled cells over 4 d were 4.39 9 10 4 at 27°C, and 6.00 9 10 3 at 32°C, accounting for 1% of the total zooxanthellae contained in coral tissue at the outset. Table 2 lists the expulsion rates of zooxanthellae during illuminated and dark periods of the day. The expulsion rate of shrunken zooxanthellae was 22-fold higher at 27°C than at 32°C. At 27°C, the daytime expulsion rate of zooxanthellae was higher than the rate in darkness (13-fold higher in the case of shrunken cells). Day and night expulsion rates were similar at 32°C. Pigment analysis. We collected shrunken zooxanthellae from the outlet seawater draining from the  Figure 4. Our HPLC analysis separated >30 peaks. We identified 17 pigment species among these. Ten pigment species were not fully identifiable, but we describe their characteristics here. A full summary is provided in Table 3. Seven peaks appeared only in samples dominated by shrunken zooxanthellae: pheophorbide a and pigments similar to it (17.05min retention time), a pigment similar to peridinin (22.33 and 22.69 min), a pigment similar to diadinochrome (27.76 min), a type of chl (31.03 min), a pigment similar to alloxanthin (32.13 min) and pyropheophytin a (39.93 min). A noticeable pigment peak at 31.03 min retention time had a maximum absorption peak at 686 nm (red band), matching a report by Goericke et al. (2000). The absorption spectrum of this pigment (extracted from zooxanthellae) is compared with an authentic standard of cPPB-aE in Figure 5. The match is almost perfect, and we therefore identify the extracted pigment as cPPB-aE (on the basis of retention time and absorption spectrum).
Concentrations of most pigments in zooxanthellae retained within coral tissue increased over 4 d of incubation at both 27°C and 32°C. Concentrations of chl a, peridinin and chl c 2 after 4 d were closely similar at the two temperatures (Fig. 6), however, the concentration of cPPB-aE was much higher at 32°C (Fig. 6), even though cell numbers had declined at this higher temperature (Table 1). Therefore, the concentration of cPPB-aE per unit surface area of coral had increased at 32°C.

DISCUSSION
The number of zooxanthellae expelled from coral tissues during the experimental period amounted to less than 1% of the total algal cell content; the number of cells expelled at 27°C was seven-fold higher than that at 32°C (Tables 1 and 2). It was reported that zooxanthellae are expelled at rates similar to those we measured from Pocillopora damicornis (Stimson and Kinzie 1991) and other corals (Acropora muricata, Pocillopora eydouxi, Porites lutea, Acropora cf. grandis, Favites abdita, Cyphastrea serailia, and Acropora nobilis; Yamashita et al. 2011). Since the number of cells expelled was very low in our incubations, we propose that the expulsion of zooxanthellae from coral is a natural physiological phenomenon and may not be the main mechanism underlying coral bleaching.
Although zooxanthellae were expelled at a low rate from coral held at 32°C, the algal cell density greatly decreased inside coral tissues. We also observed a decrease in the zooxanthellar mitotic index in corals held at 32°C (compared to 27°C). Therefore, algal cell division was reduced by thermal stress, but this does not explain the decrease in cell density inside the coral tissue (Table 1): there must be a process of zooxanthellar degradation inside the host. Titlyanov et al. (1996) reported that coral hosts commonly digest their algal symbionts, a phenomenon that has also been observed in the sea anemone Phyllactis flosculi (Steele and Goreau 1977), in giant clams (Fankboner 1971) and in the marine hydroid Myrionema ambionense (Fitt and Cook 1990). High temperature also results in a significant decline in maximum electron transport rate (ETR max) without any change in F v /F m (Bhagooli and Hidaka 2006), a process that generates increased levels of oxidative stress and oxidative damage in larvae of Acropora intermedia (Yakovleva et al. 2009). Hydrogen peroxide is formed in zooxanthellar cells under thermal stress; this may be a mechanism that triggers coral bleaching (Smith et al. 2005). Downs et al. (2002) reported that zooxanthellae are digested by coral under oxidative stress and removed by symbiophagy, a xenophagic-like process (Downs et al. 2009(Downs et al. , 2013.
We found a large number of shrunken zooxanthellae in coral tissue and in the outlet water of our experimental system. Zooxanthellae with shrunken cytoplasm and reduced chloroplasts have been observed previously in corals under thermal stress (Fukabori 1998). Similar observations have been made for several coral species; e.g., M. digitata (Titlyanov et al. 1996, Papina et al. 2007, Stylophora pistillata (Titlyanov et al. 1996, Kuroki and van Woesik 1999, Titlyanov et al. 2001, Galaxea fascicularis (Bhagooli and Hidaka 2002), and Zoanthus sansibaricus (Reimer et al. 2007), among others (Acropora selago, Acropora muricata, Heliofungia actiniformis, Ctenactis echinata, Oxypora lacera, and Pocillopora eydouxi; Fujise et al. 2013). Deformed zooxanthellae have also been observed in planulae of S. pistillata (Titlyanov et al. 1998). Although these zooxanthellae have been classified as degraded (Titlyanov et al. 1998, Downs et al. 2009, the mechanism responsible for the formation of these shrunken cells is poorly understood. We demonstrated that shrunken zooxanthellar cells accumulate in coral tissue under thermal stress. In contrast, these deformed cells are expelled at normal temperatures, especially during daytime. Titlyanov et al. (1996) reported that most zooxanthellar cells expelled under normal conditions had degraded shapes. We detected the presence of cPPB-aE and small amounts chl a and peridinin when shrunken zooxanthellae were abundant. We also detected pheophorbide a, pheophytin a, and pyropheophytin a (which are the degradation products of chl a) and (13 2 R/S)-hydroxychlorophyllones a, which are the products of biotic processing (Aydin et al. 2003, Mawson andKeely 2008) and/or abiotic oxidation products (Louda et al. 2000) of cPPB-aE in expelled zooxanthellae. Healthy zooxanthellae extracted from unstressed coral had no cPPB-aE or degraded pigment products. cPPB-aE has been reported as a degradation product of chl a in phytoplankton; it is commonly present in aquatic  (Table 3).
DEGRADATION PRODUCT OF CHLOROPHYLL environments (Kashiyama et al. 2012). cPPB-aE is a chl-like pigment frequently found in marine and lacustrine sediments (Chillier et al. 1993, Harris et al. 1995, Ocampo et al. 1999, Louda et al. 2000; it has also been identified in sponges (Karuso et al. 1986), bivalves (Sakata et al. 1990, Yamamoto et al. 1992, Watanabe et al. 1993, Louda et al. 2008) and protozoa (Goericke et al. 2000). The production pathway of cPPB-aE has been identified only recently. Kashiyama et al. (2012Kashiyama et al. ( , 2013 found that herbivorous protozoa produce cPPB-aE when they graze on and digest microalgae. cPPB-aE is generated from pyropheophytin a . Several types of phytoplankton are also able to generate cPPB-aE ). Yamada et al. (2013) reported the formation of Bold values and texts also indicates pigments that were detected only in shrunken zooxanthellae † (R/S)-hCPLs: (13 2 R)-and (13 2 S)-hydroxychlorophyllones a. ‡ Pigments with absorption spectra and chromatographic properties similar to known pigments (retention time differences of 1 min) are referred to as "pigment-like"; "derivatives" are pigments with absorption spectra similar to those of know pigments; however, derivatives and known pigments differed in terms of their chromatographic properties.  cPPB-aE (in small quantities) in zooxanthellae that had been extracted from coral and cultivated in flask in stationary phase. We detected cPPB-aE and pyropheophytin a in pigment extracts of shrunken zooxanthellae. Thus, cPPB-aE is likely generated from chl a through a degradation pathway leading to the formation of shrunken zooxanthellae.
The red fluorescence of chl was largely quenched in the shrunken algal cells we observed (Fig. 1). Kashiyama et al. (2012) also reported that chloroplasts of diatoms grazed on by protozoans were shrunken and had no chl fluorescence. This loss of fluorescence may indicate that ROS are not produced by chls freed from damaged chloroplasts. Free chl a released from broken chloroplasts becomes a generator of singlet oxygen when exposed to light, thereby promoting the formation of ROS (Perl-Treves and Perl 2002). Protozoans that feed on microalgae have transparent bodies, and their body contents are therefore always exposed to light during daytime. Accordingly, these organisms have developed a strategy for detoxifying free chl a through degradation to the non-fluorescing product cPPB-aE (Kashiyama et al. 2012(Kashiyama et al. , 2013. Corals also have transparent bodies and they live symbiotically with zooxanthellae. Therefore, they are always exposed to potential damage caused by the oxidative stress of ROS (Lesser et al. 1990, Dykens et al. 1992, Downs et al. 2002. Oxidative damage becomes more severe as UV radiation and water temperature increase (Lesser et al. 1990). Moreover, damaged chloroplasts are repaired with difficulty during thermal stress episodes, and ROS formation is thereby increased (Bhagooli and Hidaka 2006). We propose that corals and zooxanthellae employ the detoxification strategy used by herbivorous protists and phytoplankton, viz., degradation of chl a to cPPB-aE. Furthermore, reductions in zooxanthellar numbers within coral tissue caused by degradation of algal cells and bleaching may be an important mechanism for reducing the production of ROS. Thus, coral bleaching is a physiological mechanism used as a survival strategy by corals facing oxidative damage. Under normal conditions, corals regulate the numbers of symbiotic zooxanthellae by releasing excess cells, especially those that are unhealthy. Thermally stressed coral holobionts reduce oxidative stress through the formation of cPPB-aE, possibly in association with zooxanthellar degradation. However, although we know that zooxanthellae are degraded in coral tissues, the mechanism underlying this degradation is still unclear. Two studies have reported that corals can degrade their symbiotic zooxanthellae (Downs et al. 2009(Downs et al. , 2013, and others have indicated that zooxanthellae have an independent capacity to degrade chl a and generate cPPB-aE (Yamada et al. 2013). Nevertheless, dramatic changes in cell shape and pigment composition occur only within the host. Host-symbiont interactions may drive the process.
Coral bleaching is produced by reductions in the numbers of symbiotic zooxanthellae, but some corals under thermal stress generate bleached or pale zooxanthellae that do not have a shrunken form (Mise and Hidaka 2003). Moreover, some types of coral expel more normal zooxanthellae under thermal stress than under normal conditions (Fujise et al. 2013), and these species may possess other mechanisms to protect themselves when stressed. M. digitata, however, bleaches by degrading chl a to a non-fluorescent pigment in a strategy to combat oxidative stress. cPPB-aE was formed only in shrunken zooxanthellae whose presence may be used as an indicator of thermal stress in corals that have not reached the stage of visible bleaching. Monitoring of cPPB-aE will facilitate prediction of bleaching and promote the study of coral responses to changes in the climate.
We are grateful to Mr. Yoshikatsu Nakano of University of the Ryukyus for support in setting up the aquarium and the laboratory, and to Mr. Sunao Uehara for assisting in wild coral sampling. We are indebted to Dr. Yuichiro Kashiyama of Fukui University of Technology for providing the cPPB-aE standard and for valuable suggestions on the manuscript. We thank Dr. Mohan Prasad Niraula of Shizuoka University for valuable comments and English editing. This study was sup-