Introduction

Scleractinian corals, which provide the biogenic habitat for a variety of marine organisms, form the base of coral reef ecosystems. The ability of these corals to maintain and develop the framework of reefs is vital to the long-term persistence of ecosystem prosperity (Jones et al. 2004). In addition to recruitment, the shift in balance between the growth and decline of coral colonies is a critical process upon which the integrity of entire coral reef framework depends. In particular, branching corals, such as pocilloporids and acroporids, show large variations in this balance as a result of incorporating the properties of rapid degradation and regeneration into their colonies (Hall 1997). For example, a high potential for breakage appears in branching corals because their distal portions grow rapidly in length and number while they are exposed to mechanical disturbances (Highsmith et al. 1980). In contrast to the distal portions, the tissue at the proximal regions of the coral branch appears more vulnerable and less vigorous, these parts are more aged and sensitive to mortality (Meesters and Bak 1995). The abundance of branching corals that are extremely susceptible to environmental variations has previously been attributed to this aging-related deterioration (Loya et al. 2001). The intra-colony arrangement in branching corals also shows sharp contrasts in energy metabolism, several physiobiochemical characteristics, and the correlated expression of genes involved in growth and resistance (Bay et al. 2009; D’Angelo et al. 2012; Fang et al. 1989; Gladfelter et al. 1989; Oku et al. 2002).

The stress response of scleractinian corals exposed to a variety of environmental insults, such as chemical, thermal, and irradiant stresses, strongly involves oxidative stress in the underlying etiology. As previously demonstrated, stressed corals induce a series of biochemical and physiological responses related to their oxidative circumstances, such as the overproduction of reactive oxygen species (ROS); the depletion of antioxidant capacity; the accumulation of oxidative damage molecules; and an increase in the expression of antioxidant enzymes, protein chaperons, and protein turnover activity (Baird et al. 2009; Downs et al. 2002; Lesser 2006). Oxidative stress places a burden on an organism through the excess generation and accumulation of ROS, such as superoxides and hydroxyl radicals, which cause cellular damage as the result of cellular abnormalities in energy conversion processes, including photosynthesis, catabolism, or anabolism (Lesser 2006; Murphy 2009). In addition to affecting exogenous insults, these toxic reactions appear to play major roles in various degenerative processes in the organisms, including illness and aging (Cadenas and Davies 2000). As has been revealed by extensive research, ROS can cause the oxidative modification of many cellular components, such as lipids, proteins, and DNA, resulting in decreased and inhibited cellular functioning, which can initiate the processes of cellular degradation (Avery 2011). In particular, lipid peroxidation can expand oxidative damage to the cellular membranes and is one of the vital challenges in the maintenance of cellular homeostasis. Under oxidative stress, membrane perturbations, including alterations in physical properties, modification of membrane protein activity, and permeabilization to nonspecific ions, such as Ca2+, are causally linked to the occurrence of lipid peroxidation as an initiator of cellular degradation (Beranova et al. 2010; Stark 2005). Extensive lipid peroxidation ultimately disrupts membrane function and leads to the loss of cell viability.

Organisms that maintain their cellular functions by altering their membrane lipid and fatty acid compositions can remain viable even under extreme oxidative circumstances (Mykytczuka et al. 2011; Rozentsvet et al. 2012). Membrane lipids show a great diversity of molecular structures by varying the arrangement of chemical bonds that can accommodate the changes in the cellular membrane that occur due to an environmental disturbance or physiological senescence (Freikman et al. 2011; Tang et al. 2014a; Zehmer and Hazel 2005). The modified lipid structure can generate a membrane that has entirely different properties and thus exerts a potent effect on biological processes (Lee 2004; Vigh et al. 2005). Indeed, membrane lipids constitute the basic structural elements that create cellular membranes with the physicochemical properties appropriate for the physiological demands and that optimize cellular performance (van Meer et al. 2008). Even relatively small alterations in cellular membrane properties can elicit considerable changes in the efficiency of several membrane-dependent functions, leading to a chronic impairment in the fitness of organisms (Tang et al. 2014a). As a potential site of oxidative damage, the cellular membrane should be the focus of studies examining cellular stress and regulatory responses.

Most studies on coral membrane lipid composition have characterized the contradistinctions of many lipid classes and fatty acid content that were related to the different cellular physiological conditions both between and within colonies (Bachok et al. 2006; Yamashiro et al. 2001; Oku et al. 2002). These studies have provided insight into lipid metabolic responses, but they have been unable to reveal the relevance of these lipid shifts to the modulation of membrane properties and the related physiological requirements. Cellular functions must be coordinated with the appropriate membrane properties, which are largely determined by the ability of the lipid composition to adapt to the cellular situation. Therefore, to understand cellular function and regulation, it is essential to have insight into the molecular structures of altered membrane lipids. High-throughput methods in lipidomics have made this detailed analysis possible, allowing us to systematically conduct lipid profiling with an explicit description of the molecular structure. In this study, the membrane lipid profiles of a pocilloporid coral, Seriatopora caliendrum, were characterized according to the functional differentiations within the coral colony using a mass spectrometry-based lipidomics method. The lipid profiles revealed an interesting contrast between the distal and proximal portions of coral branches. The distinctions in lipid composition were related to the physiological requirements of coral growth and the cellular accommodation of aging-related physiological senescence to augment or maintain cellular functioning.

Materials and Methods

Coral Sample Collection

Considering the reproductive influence, coral sampling was done during the period when coral reproduction is inactive (Tang et al. 2014b). A branch was obtained during spring 2013 at depths of 8–10 m from each sample colony (15–20-cm diameter, N = 19) of the pocilloporid coral S. caliendrum that inhabits the Houbibu region of Kenting National Park, Taiwan. The tip (0.1–0.5 cm from the branch end) and stalk (1–2 cm from the branch end) were collected, immediately frozen in liquid nitrogen, and stored at −80 °C. Based on a growth rate of approximately 4 cm/year (Edmunds 2005), the ages of the coral tip and stalk regions can be estimated to be 0.3–1.5 and 3–6 months old, respectively. The coral sampling was conducted under a permit issued by Kenting National Park Headquarters.

Identification of Zooxanthellae

Coral tissue was ground in liquid nitrogen and processed using a modified salting-out procedure for DNA extraction (Ferrara et al. 2006; Keshavmurthy et al. 2012). The coral sample was placed into 0.2 mL lysis buffer (0.25 M Tris, 2 % sodium dodecyl sulfate, 0.05 M EDTA at pH 8.0, and 0.1 M NaCl) with 4.0 μL proteinase K (25 mg/L) and lysed overnight at 55 °C. The lysed sample was carefully mixed with 0.2 mL saturated NaCl and centrifuged for 20 min at 10,000g. The supernatant was isolated and mixed with two volumes of 99.9 % cold ethanol and then stored at −20 °C for 30 min prior to centrifugation for 30 min at 12,000g (4 °C). The collected DNA pellet was washed three times with 99.9 % cold ethanol, dried by vacuum for 10 min, and then resuspended in sterile distilled water. To identify the clade of the zooxanthellae in the corals, nuclear large-subunit ribosomal DNA was amplified using the following host-excluding primers: forward 5S 5′-GCCGACCCGCTGAATTCAAGCATAT-3′ and reverse D23zoox 5′-TGTGGCAYGTGACGCGCAAGCTAAG-3′ (Chen et al. 2005). PCR analysis was conducted in 50-μL reaction volumes containing 10 μL 5× Excel Taq premix (SMOBIO Technology, Taiwan), 1 μL (10 pmol) of each primer, and 5 μL template DNA. The PCR reactions were carried out using a GeneAmp PCR System 2400 (Applied Biosystems, USA), and the conditions were as follows: an initial denaturation at 95 °C for 5 min; 30 cycles consisting of 95 °C for 30 s, 56 °C for 1 min, and 72 °C for 2 min; and the final elongation step at 72 °C for 6 min. Following PCR, the products were sequenced using an ABI 3730 DNA Analyzer (Applied Biosystems, USA), and the obtained sequences were searched for and aligned using the NCBI GenBank Database with BLAST (http://www.ncbi.nlm.nih.gov/).

Lipid Analysis

For lipid extraction, the frozen coral fractions were ground in a liquid nitrogen-cooled mortar and then lyophilized overnight (Tang et al. 2012). A three-replicate chemical measurement was conducted for each coral sample. The homogenous dry tissue was weighed and then extracted with a chloroform/methanol (2:1) solution. The mixture was then mixed thoroughly with a 0.15-M NaCl aqueous solution, and the complete lower layer was collected for further analysis. A validated high-performance liquid chromatography (HPLC)–tandem mass spectrometry method was used to profile the composition of glycerophosphocholines (GPCs) in the lipid extracts (Tang et al. 2012). Briefly, chromatographs of the lipid extracts were created with a reversed-phase gradient elution at a constant flow rate of 0.5 mL/min in an Agilent 1100 series HPLC system (Agilent Technologies, USA). The molecular information for the lipid species that emerged from the chromatographic column (Kinetex C18, 2.6 μm, 100 × 2.1 mm; Phenomenex, USA) was acquired and recorded by a coupled API 4000 triple quadrupole mass spectrometer (Applied Biosystems, USA). An analytical strategy consisting of a two-stage tandem mass spectrometry experiment based on unique fragmentation patterns in the collision process was used to detail the GPC molecular species composition in the sample. The preliminary profile of GPCs in each sample was obtained using a precursor ion scan of m/z 184 (corresponding to the phosphorylcholine ion) in the positive mode, because this fragment signal is derived from the common feature of the molecular structure of GPCs. Subsequently, based on this phosphorylcholine-containing lipid profile, another tandem mass spectrometry experiment was conducted to acquire the product ion spectrum of each GPC species according to their respective elution times. The molecular structure of the GPC was then determined by illustrating the product ion spectra of the [M + H]+, [M + Na]+, [M + Li]+, and [M + Ac]− ions.

Data Processing and Analysis

The acquired GPC profile data were processed using the software (Analyst 1.4; Applied Biosystems Instrument Co., USA) on the mass spectrometer. For the peak detection, each molecular mass ion signal was manually determined from a total ion chromatogram of pooled coral samples during a 0.2-min time window per step; the extracted peak was recorded as m/z and retention time. After removing duplicate peaks, the remaining peaks were checked again to remove latent isotope peaks. A quantitation method was created using the final peak list to quantify a batch of samples according to the Analyst 1.4 regulation procedure. Once a table of results was obtained, all the peaks in the batch were manually reviewed to confirm the accuracy of the peak alignment and the peak area integration. Ultimately, the peak list including the peak area of the extracted ion chromatogram was exported as a Microsoft Excel file for further analysis.

After the data were processed, each detected peak signal was normalized to the sum of the total signal response in the GPC profile. A principal components analysis (PCA) and an orthogonal partial least squares to latent structures discriminant analysis (orthogonal partial least squares discriminant analysis (OPLS-DA)) were performed using SIMCA-P 13 statistical software (Umetrics AB, Umea, Sweden) to determine whether the metabolic differences in the GPC profiles from the coral tissues could be distinguished or modeled, respectively. PCA is a pattern recognition method that condenses the original set of variables into fewer variables based on their weighted averages; the new variables are called scores, and the weighted profiles are called loadings. Each GPC profile was represented as a data point in the score plot, and each point that was clustered together had a similar profile. The loading plot revealed information about the GPC molecular species that contributed to the data point separation observed in the score plot. Furthermore, the variations within and between groups can be separated using the OPLS-DA model, and this facilitates model interpretation. This supervised multivariate analysis method is useful for surveying the data and for recognizing the patterns of GPC profiles that are associated with specific sample groups. A one-way analysis of variance (ANOVA) was used to further assess the significant differences (p < 0.05) of the sifted GPC molecular species obtained from the coral tissues at the tip and stalk positions.

Lipid Nomenclature

GPCs are composed of a glycerol backbone with a phosphorylcholine head group at the sn-3 position and a fatty acid substituent at the sn-1 and sn-2 positions. The fatty acid substituent can be linked via an ester at both positions, and long-chain alkyl ether or vinyl ether linkages can also be adopted at the sn-1 position. Because of the differences in linkages at the sn-1 position, GPCs are further divided into phosphatidylcholine (1,2-diacyl-GPC), plasmanylcholine (1-O-alkyl-2-acyl-GPC), and plasmenylcholine (1-O-alk-1′-enyl-2-acyl-GPC) subclasses. The designations and abbreviations for GPC molecular species are based on the recommendations of the Lipid Metabolites and Pathways Strategy (LIPID MAPS). For example, 1,2-dihexadecanoyl-, 1-hexadecyl-2-hexadecanoyl-, and 1-(1Z-hexadecenyl)-2-hexadecanoyl- sn-glycero-3-phosphocholine are abbreviated as PC(16:0/16:0), PC(O-16:0/16:0), and PC(P-16:0/16:0), respectively; the molecular structures for these lipids are shown in Fig. 1. The notation x:y for a fatty acid chain indicates the number of carbon atoms (x) and double bonds (y). Additionally, symbols marked with “O” or “P” indicate a fatty acid chain linkage with the glycerol backbone via an alkyl or vinyl ether, respectively, generally at the sn-1 position; no marks indicate an ester linkage. In cases where both fatty acid chains cannot be assigned, the sum of their carbon atom and double-bond numbers can be derived from the deductive inference of the molecule once the lipid subclass is identified. These lipids are correspondingly abbreviated PC(32:0), PC(O-32:0), and PC(P-32:0). Molecular species of GPCs that possess a hydroxyl substituent at either the sn-1 or the sn-2 position of the glycerol backbone are named lyso-GPC; for example, 1-hexadecanoyl-2-hydroxyl-sn-glycero-3-phosphocholines can be abbreviated as PC(16:0/0:0).

Fig. 1
figure 1

Structure of three glycerophosphocholine subclasses. PC(16:0/16:0), PC(O-16:0/16:0), and PC(P-16:0/16:0) are indicated for 1,2-dihexadecanoyl-, 1-hexadecyl-2-hexadecanoyl-, and 1-(1Z-hexadecenyl)-2-hexadecanoyl-sn-glycero-3-phosphocholine, respectively

Full size image

Results

Of the 187 signal peaks detected in the GPC profile of S. caliendrum, 144 identified molecular species are listed in Table 1 along with a ratio accounting for the GPC composition in the coral tissues at both the tip and stalk positions. The identification of the zooxanthellae (accession: KM287430) indicated a high similarity to the sequence belonging to clade C, which ensures a strong specificity for the lipid composition in the coral samples. Therefore, the result should be considered as evidence that the coral S. caliendrum harbors the specific zooxanthellae identified in this study. The analysis revealed that a few of the molecular species including PC(16:0/22:6), PC(O-16:0/20:4), PC(O-16:1/20:4), PC(O-16:0/20:5), and PC(O-18:0/20:4) accounted for more than 50 % of the GPC composition in the coral. Approximately 60 % of the GPC composition in the tissues of S. caliendrum was composed of plasmanylcholine molecular species. The other GPC subclass, phosphatidylcholine and plasmanylcholine, accounted for approximately 30 and 1.6 % of the GPC composition, respectively. In terms of total proportions, there are no clear differences in these three GPC subclasses between the coral tissues at the tip and stalk positions. Nevertheless, the total proportion of lyso-GPCs (lyso-phosphatidylcholines and lyso-plasmanylcholines) in the tissues from the coral tip (0.669 %) was significantly higher (ANOVA, p < 0.05) than that from the stalk (0.279 %).

Table 1 The glycerophosphocholine molecular species identified in the coral (Seriatopora caliendrum) tissue in the tip and stalk portions
Full size table

The differences in the phosphorylcholine-containing lipid profiles between the coral samples were visualized in a PCA score plot. As shown in Fig. 2a, the cluster of data points belonging to different groups is clearly divided, distinguishing the coral tissues from the tip and stalk positions from each other based on the differences in their GPC profiles. The first principal component that contributed to the group separation in the score plot described a variation of 22.1 %. The covariance and correlation loading plots that revealed discriminatory GPC signal peaks are shown in Fig. S1 (in the electronic supplementary material). The coral GPC profiles of the tip and stalk groups were further modeled by OPLS-DA, as shown in Fig. 2b. The model resulted in one predictive and one orthogonal components (1 + 1) with a goodness-of-fit of R2Y = 0.957 and cross-validated predictive abilities of Q2Y = 0.903. The contrast in the lipid profiles between the two groups is presented in an S-line plot where the key features can be directly observed. As shown in Fig. 2c, distinct ratios of saturated, monounsaturated, and lyso- GPC are highlighted by high levels of correlation despite their small weights. In addition, there is a key difference showing the favoring of the two coral positions by polyunsaturated GPCs with unsaturation levels that differ by one or two degrees. As shown, the covariance direction of PC(16:0/22:6), PC(18:0/22:6), PC(O-16:0/20:5), and PC(O-18:0/20:4) is opposite to that of PC(16:0/22:5), PC(18:0/22:5), PC(O-16:0/20:3), and PC(O-18:0/20:3), and such lipid species accounted for the majority of the weighting to distinguish between the two lipid profiles of the coral tissues. Likewise, the phosphatidylcholine species PC(16:0/22:6) and PC(18:0/22:6) also showed a covariance that contrasts with their corresponding plasmanylcholine species PC(O-16:0/22:6) and PC(O-18:0/22:6), which allowed us to differentiate between the lipid profiles.

Fig. 2
figure 2

Glycerophosphocholine profile analysis of coral tissues. a The score plots of the principal components analysis (PCA) and b the orthogonal partial least squares to latent structures discriminant analysis (OPLS-DA), which show a distinguishable separation between the tip and stalk tissues of the coral (Seriatopora caliendrum). In the PCA score plot, the first principal component (t[1]) describes a variation of 22.1 % that contributes to the group separation. The OPLS-DA model resulted in one predictive (t[1]) and one orthogonal (to[1]) component, with a goodness-of-fit of R2Y = 0.957 and cross-validated predictive abilities of Q2Y = 0.903. c The S-line plot of the OPLS-DA model showing the glycerophosphocholine differences responsible for the discrimination between the coral tip and stalk tissues. The S-line plot visualizes the modeled covariation (p(ctr) [1] as a magnitude eigenvalue directed toward the coral stalk), colored according to the modulus of the modeled correlation (p(corr) [1], representing the reliability)

Full size image

The GPC species with significant differences (ANOVA, p < 0.05) in the ratio between the tip and stalk positions of the coral tissue are shown in Fig. 3. The primary feature of tissue from the coral tip was a greater proportion of lyso-GPCs; the ratio of the 28 molecules was 51–581 % greater than in the stalk tissue. In contrast, the ratio of seven phosphatidylcholines and seven plasmanylcholines that are either saturated or monounsaturated was 0.5–50 % higher in the stalk tissue than in the tip tissue. There were ten molecular species of polyunsaturated plasmanylcholines that showed greater proportions in the stalk tissue, while the seven other species that possess a 20-carbon chain that is unsaturated by 4–5 degrees, including three predominant GPC species, were more highly expressed in the tip tissues. Similarly, ten polyunsaturated phosphatidylcholines, including two predominant GPC species, showed a greater proportion in the tip tissue, while five others showed greater ratios in the stalk tissue.

Fig. 3
figure 3

Significant differences in lipid species between the coral tip and stalk tissues. The difference in the proportion of glycerophosphocholine (ANOVA, p < 0.05) in Seriatopora caliendrum tissue from the tip portion compared to the stalk portion

Full size image

Discussion

Differences in Growth Rate Between the Coral Tip and Stalk

Among diverse coral forms, branching corals have been known to develop the most separation of function within colonies (Soong and Lang 1992). One of the most visible examples of this is that the coral branch grows faster at the distal position compared to the proximal portion (Elahi and Edmunds 2007). The higher levels of lyso-GPCs and polyunsaturated GPCs in the tip likely reflect the cellular physiology requirements for rapid coral growth. In particular, the majority of the favored plasmanylcholines in the tip tissue possessed arachidonic acid (20:4), and the polyunsaturated GPCs in those tissues tended to be more unsaturated.

Scleractinian coral growth involves the formation of a great quantity of skeleton that requires the transport of many materials through calicoblastic cells (Tambutte et al. 2011). Transporting calcium through calicoblastic cells into the underlying space is one of the fundamental actions during the process of coral calcification. By directly altering membrane characteristics, lyso-GPCs can increase the activity of L-type calcium ion channels and calcium ATPase and thus enhance calcium ion currents (Grosman 2001; Kraichely et al. 2009). Lyso-GPCs also potentiate the activation of protein kinases that can phosphorylate calcium ATPase, calcium channels, and their regulatory proteins to stimulate calcium ion flux activity (Edes and Kranias 2001; Sasaki et al. 1993; Shier et al. 1976). Another cellular action that is required for coral calcification is the vesicle-mediated secretion of “organic matrix” from calicoblastic cells (Clode and Marshall 2002). During this process, calicoblastic cells are extensively involved in membrane vesiculation and fusion, where the membrane adopts a highly curved architecture. The inverted cone molecular shape of lyso-GPCs allows the membrane to construct a positive curvature during vesiculation (Fuller and Rand 2001). In the membrane fusion process, saturated lyso-GPCs can induce a fully interdigitated membrane structure that enlarges the surface area and diminishes inter-membrane repulsion, which facilitates membrane contact prior to the merging step (Komatsu and Okada 1995; Lu et al. 2001). The increase in polyunsaturated phosphatidylcholines in the contacted membranes facilitates the formation of the inverted hexagonal phase and promotes the mixing of lipids during the membrane-merging step (Chernomordik and Kozlov 2008; Szule et al. 2002). Moreover, polyunsaturated phosphatidylcholines facilitate the membrane dynamics involved in organizing lipid rafts, which is an important factor in the initiation of membrane fusion (Baenziger et al. 1992; Lingwood and Simons 2010).

In addition to skeleton formation, coral growth involves cell growth that must fit the appropriate conditions of cellular physiology. Plasmanylcholine levels have been shown to positively correlate with accelerated cell growth and the levels of cellular differentiation (Albert and Anderson 1977; Chabot et al. 1990). Under conditions of cell growth, an abundance of arachidonic acid is also required for the production of eicosanoids (Sellmayer et al. 1996). Concurrently, higher levels of plasmanylcholines possessing arachidonic acid in the coral tips are also required for cellular growth. In addition, the regulation of desaturases that are involved in the modulation of cellular growth results in differences in the ratio of polyunsaturated GPCs with dissimilar unsaturation between the tips and stalks of the coral (Kates and Paradis 1973). Enhanced lyso-GPC production results in a state of accelerated cell growth (Okita et al. 1997), which is also found in the coral tip tissues relative to the stalk tissues. This metabolic distinction in lyso-GPC production is consistent with the variations in gene expression that are associated with lysosome lipase activity in coral cells (Bay et al. 2009).

Aging Gradient of Coral Polyps from the Distal to Proximal Region

In contrast to the coral tip, coral stalk tissue appears less vigorous and exhibits a reduced capability for growth and regeneration that is related to the aging of polyps (Meesters and Bak 1995). As previously reported for other organisms, the accrual of oxidative stress has been postulated to be responsible for the physiological senescence of coral polyps during the aging process (Sohal and Orr 2012). Additionally, in natural habitats, there are various environmental parameters that cyclically augment the physiological burdens of coral, such as photosynthesis-derived oxidative stress (Mayfield et al. 2010; Yellowlees et al. 2008). Once oxidative stress is enhanced to challenge the remaining anti-oxidant capacity, the cells in older coral polyps likely become more vulnerable. Membrane permeabilization has been established to be one of the underlying mechanisms that cause functional cell losses following oxidative insults that result in a decrease in molecular membrane density (Smith et al. 2009). Therefore, it is vital for the coral cells to safeguard the membrane from oxidative damage, permeabilization, and subsequent cell/organelle swelling-induced tension stress (Trump and Berezesky 1996). At the coral stalk, the older polyps appeared to decrease their lyso-GPC levels, exchange polyunsaturated phosphatidylcholines for corresponding species of plasmanylcholines, adopt higher levels of saturated and monounsaturated GPCs, and down-regulate the desaturation of GPCs. The shift in these lipids provides clues about the requirements for cellular accommodation to oxidative conditions for the purposes of maintaining the barrier property and structural integrity of the membrane.

Oxidative degradation leads to the formation of membrane pores as a result of the rebalancing of molecular interactions facilitated by the geometric packing of a specific molecule, such as lyso-GPCs (Hull et al. 2004; Smith et al. 2009). Therefore, a decrease in the levels of lyso-GPCs in the stalk plays a beneficial role in keeping coral cells from forming membrane pores, which helps to avoid effective permeabilization derived from the increase in oxidative stress. The exchange of polyunsaturated phosphatidylcholines for corresponding plasmanylcholines and the increase in saturated and monounsaturated plasmanylcholines are also beneficial to the cells of coral stalk in maintaining the properties of the membrane barrier. Due to the replacement of the carbonyl group with a methylene, plasmanylcholines can enhance the membrane barriers against ion and water permeability, regulating the oxidative assault-caused decrease of molecular membrane density (Chen and Gross 1994; Pandey and Roy 2011). The increase in monounsaturated GPCs can strengthen the mechanical stability of the membrane, allowing it to resist tension-induced leakage due to cell/organelle swelling when cells suffer from potent oxidative damage (Shoemaker and Vanderlick 2002). The decrease in lyso-GPCs minimizes the elastic free energy of the membrane during cell/organelle swelling and thus maintains architectural stability (Fuller and Rand 2001).

Polyunsaturated phosphatidylcholines are most likely to be the targets of oxidative stress due to their effects on membrane permeability and their highly reactive properties (Olbrich et al. 2000; Ollila et al. 2007). By exchanging polyunsaturated phosphatidylcholines for the corresponding plasmanylcholines, the stalk cells can further alleviate lipid oxidation by decreasing access to the polyunsaturated chain. Similarly, lowering the degree of fatty acid chain unsaturation is a protective mechanism against aging-related oxidative damage through the decrease in sensitivity to peroxidation (Pamplona et al. 2002). Saturated and monounsaturated GPCs have also been shown to increase membrane density and to be resistant to oxidative attack (Rozentsvet et al. 2012; Smith et al. 2009). Through such lipid modulation, coral cells could solidify their membrane rigidity to safeguard against oxidative degradation. As found in previous studies, a higher ratio of saturated and monounsaturated chains accompanying a lower level of polyunsaturated chains is adopted in coral polyps with relatively weak cellular physiology (Bachok et al. 2006; Yamashiro et al. 2001). Clearly, retarding oxidative progress is important for accommodating the cellular membrane in response to oxidative conditions.

In the present study, the metabolic differentiation of membrane lipid within a S. caliendrum colony was related to the physiological requirements for growth and cellular accommodation and to the oxidative effects accrued by aging. The coral cells compulsorily shifted their membrane lipid compositions during the aging process, accommodating the oxidative situation at the cost of vitality. Indeed, the shift in lipid composition includes several important signaling molecules, such as lyso-GPCs and polyunsaturated fatty acid chains, which govern basic cellular activities and actions (Meyer zu Heringdorf and Jakobs 2007). Therefore, this intra-colony shift in lipid metabolism is also likely to compromise cellular performance and ultimately lead to physiological senescence.