Physiological Outperformance at the Morphologically-Transformed Edge of the Cyanobacteriosponge Terpios hoshinota (Suberitidae: Hadromerida) when Confronting Opponent Corals

Terpios hoshinota, an encrusting cyanosponge, is known as a strong substrate competitor of reef-building corals that kills encountered coral by overgrowth. Terpios outbreaks cause significant declines in living coral cover in Indo-Pacific coral reefs, with the damage usually lasting for decades. Recent studies show that there are morphological transformations at a sponge’s growth front when confronting corals. Whether these morphological transformations at coral contacts are involved with physiological outperformance (e.g., higher metabolic activity or nutritional status) over other portions of Terpios remains equivocal. In this study, we compared the indicators of photosynthetic capability and nitrogen status of a sponge-cyanobacteria association at proximal, middle, and distal portions of opponent corals. Terpios tissues in contact with corals displayed significant increases in photosynthetic oxygen production (ca. 61%), the δ13C value (ca. 4%), free proteinogenic amino acid content (ca. 85%), and Gln/Glu ratio (ca. 115%) compared to middle and distal parts of the sponge. In contrast, the maximum quantum yield (Fv/Fm), which is the indicator usually used to represent the integrity of photosystem II, of cyanobacteria photosynthesis was low (0.256~0.319) and showed an inverse trend of higher values in the distal portion of the sponge that might be due to high and variable levels of cyanobacterial phycocyanin. The inconsistent results between photosynthetic oxygen production and Fv/Fm values indicated that maximum quantum yields might not be a suitable indicator to represent the photosynthetic function of the Terpios-cyanobacteria association. Our data conclusively suggest that Terpios hoshinota competes with opponent corals not only by the morphological transformation of the sponge-cyanobacteria association but also by physiological outperformance in accumulating resources for the battle.


Introduction
Competition among sessile organisms is one of the major ecological processes of coral reefs and can result in altering coral species diversity and abundance at spatial and temporal scales (reviewed in [1,2]). As the impacts of climate change and human disturbances increase, including anomalous seawater temperatures, overfishing of herbivores, increased levels of dissolved nutrients, sedimentation, and physical disturbances (e.g., typhoons), the competitive capacity of coral opponents like algae, sponges, ascidians, and skeleton-less cnidarians can be enhanced if disturbances favor the growth of sessile organisms. As a consequence, large-scale phase shifts in reefs from coral dominance to alternative steady states dominated by algae or sponges is occurring in tropical coral reefs [3,4,5].
Cyanosponges are fast growing and aggressively competitive in disturbed areas, overgrowing and killing live corals [6][7][8]9]. In the Indo-West Pacific, Terpios hoshinota is reported as a cyanosponge species that outcompetes scleractinian corals and causes significant declines in living coral cover during outbreaks over the last four decades [7,[10][11][12][13][14][15][16][17][18]. Terpios is a very thin (<1 mm) and black sheet-like sponge that firmly encrusts corals, with sponge tissue penetrating deeply into the steeply undulating coral skeleton [19]. The surface of the sponge is usually free from epibionts and sometimes contains calcareous grains underneath [19]. Terpios can be categorized as a "phototrophic" cyanosponge based on its flattened morphology [20,21] and the presence of extremely dense populations of photosynthetic cyanobacteria in the mesohyl [7,[9][10][11][12]. Histological examinations indicate that endosymbiotic cyanobacteria in the Terpios mesohyl occupy > 50% of total cell volume and may contribute to Terpios coloration [7,19]. Molecular phylogenetic analysis using the 16 rRNA gene demonstrates that the dominant cyanobacteria taxon is an isolated group that is closely related to Prochloron spp. [22], suggesting that this strain of cyanobacteria also contains phycocyanin [19] and probably chlorophyll b [9]. Terpios prefers to grow on living coral rather than other substrates [13]. When encountering coral opponents, Terpios develops four different types of morphological reactions [23]. "Hairy" tips packed with cyanobacteria, sponge tissues, and spicules were the most frequently observed reaction of sponges to corals (Fig 1). Terpios can also transform from an encrusting sheet-like structure into a thread-like tissue in order to move across a shaded area or reach new territory [23,24]. Occasionally, Terpios exhibits negative growth and can even be overgrown by certain coral species or red calcareous algae [11,23].
Although diverse morphological responses were observed at the contact between Terpios hoshinota and opponent corals, the physiological performance of the sponge-cyanobacteria association in this interaction remains unknown. As cyanobacteria occupy a significant proportion of the biomass within sponges [22], measuring the physiological performance of cyanobacteria may represent the overall status of Terpios. In this study, we examined maximum quantum yield, photosynthetic oxygen production, and chlorophyll contents to represent photosynthetic performance, free amino acid content to represent nitrogen status, andδ 13 C andδ 15 N stable isotope ratios to represent metabolic activity in the sponge-cyanobacteria association at proximal, middle, and distal regions of opponent corals. These physiological and metabolic indices measured from different parts of the sponge suggest that Terpios hoshinota exhibits physiological outperformance where it contacts coral and that this might benefit the sponge-cyanobacteria association in accumulating the resources to compete with coral.

Ethics statement
No specific permissions were required for the collection of Terpios specimens at Gong-guan, Green Island (22°39'N, 121°29'E)", according to local government regulations. We also confirmed that our field studies did not involve endangered or protected species.

Study site, field experiments, and sample collection
Field experiments and T. hoshinota sample collections were conducted in May, July, and November, 2010, at Gong-guan, Green Island (22°39'N, 121°29'E), in 2007 where Taiwan's first outbreak of Terpios was reported [12]. Colonies of T. hoshinota (approximately 50 cm in diameter) attacking catch bowl coral, Isopora palifera, and showing "hairy" tip structures during attacks on corals (Fig 1) were selected for in situ photosynthetic measurements and followon physiological and metabolic examination in the laboratory. Different parts of Terpios colonies (the growth front contacting the living coral "CF", mid-body "MB", and no-contact front "NCF"), were selected for photosynthetic measurements and collections for laboratory experiments (Fig 1).

Maximum quantum yield determination in situ
In order to evaluate the in situ photosynthetic functions of Terpios cyanobacteria, maximum quantum yields of PSII (F v /F m = (F m -F o )/ F m ) were determined using a DIVING-PAM fluorometer (Walz, Germany) following the method developed for coral and photosynthetic sponge studies [25][26][27]. Data for Terpios MB were pooled from two measurements taken from each sponge colony. Nontransparent, custom-made DIVING-PAM probe holders were affixed to Terpios CF, MB, and NCF for 30 min dark adaption (Fig 1), after which minimum fluorescence (F o ) was measured at a measuring-light setting of 8. This was followed immediately by a saturating flash of actinic light (DIVING-PAM setting 8) and the measurement of maximum fluorescence (F m ).

Laboratory measurement of photosynthetic capability and dark respiration rate
Representative samples of Terpios approximately 5 cm in diameter for Diving-PAM measurements were chiseled from sponges to determine photosynthetic oxygen evolution and respiration rates. Each sample was sealed in an individual plastic bag underwater for delivery to the maintenance tank to avoid desiccation. Terpios samples were temporarily maintained in an aerated tank (180 × 60 × 30 cm) under low light (< 20 μmol photons m -2 s -1 ) with two underwater pumps (1800 L h -1 ) to create strong water flow. Sponge samples were used for experiments or isolation of cyanobacteria within a day of collection.
Oxygen flux due to Terpios photosynthesis and respiration symbiosis was measured in a custom-made respiration chamber (400 ml) with a BOD probe (YSI 5905, Yellow Springs, OH, USA) and dissolved oxygen (DO) meter (YSI 52). Oxygen concentration changes in the respiration chamber were continuously measured for 15 min by connecting the meters to a personal computer using manufacturer software. The Terpios symbiosis dark respiration rate was measured by covering the respiration chamber with a black sheet. Terpios symbiosis photosynthesis was induced with a UV-block halogen light bulb. Photosynthesis irradiance was checked with a quantum photometer (LI-COR, model LI-250 Light Meter, Lincoln, NE, USA) attached to a 2π sensor (LI-COR, model: QUANTUM) and expressed as μmol photons m -2 s -1 P.A.R. Light intensity was increased to determine the photosynthesis-irradiation curve by moving the light source closer to the respiration chamber, which did not result in elevating the temperature of the chamber during the experiment. After determining the changes in oxygen concentration, respiration and photosynthesis rates were standardized to the area of the dried coral substrate using the aluminum foil method [28].

Isolation of cyanobacteria and physiological index determination
Live tissue from each Terpios sample (ca. 25 cm 2 ) was first scraped with a spatula and stripped of remaining tissue with a pressurized air-seawater blast, modified from a method used for coral [29]. Artificial seawater (ASW) for air blasts was prepared from sea salt (Instant Ocean, Aquarium Systems, Sarrebourg Cedex, France) and pre-cooled on ice. The tissue slurry (ca. 10 ml) was then homogenized in a hand-held tissue grinder in an ice bath. Tissues were homogenized and left to stand for 3 min to precipitate heavy debris such as coral skeleton particles and spicules, and the resulting supernatant was filtered through 15 μm nylon cloth to collect the cyanobacteria suspension. Cyanobacteria were washed with ASW three times by centrifugation (2000 rpm, 4°C, 6 min) and re-suspended in ASW to isolate them to a concentration of 10 6 −10 7 cells ml -1 .
For free amino acid (FAA) analysis, 500 μl of freshly isolated cyanobacteria were first mixed with an equal amount of ultrapure water, sonicated in ice for 3 min, and centrifuged at 15,000 rpm at 4°C for 10 min to collect the cellular extract. Extracts were precipitated in 70% ethanol to remove proteins before being subjected to amino acid analysis as described by Wang et al. [29]. Finally, the FAA content was standardized to the protein content in the extract, which was determined using a protein assay kit (Bio-Rad, CA, USA) with bovine serum albumin as the standard.
Each of the 500 μl of freshly isolated cyanobacteria samples was further lyophilized and weighed to determine chlorophyll (Chl a, b, c1+c2) content and δ 13 C and δ 15 N values. Chlorophyll in the dried cyanobacteria was extracted with 90% cold acetone and the levels of Chl a, b, and c1+c2 were determined spectrophotometrically following Jeffery and Haxo [30]. Prior to determining δ 13 C and δ 15 N values, cyanobacteria samples were acidified with 500 μl 1.0 N HCl and shaken overnight to remove carbonate. Acidified samples (1-2 mg) were combusted in an elemental analyzer (Flash EA-1100 NC, Thermo Finnigan, Waltham, MA, USA) to produce CO 2 and N 2 for separately determining isotopic compositions [31]. Isotope ratios (δX) were expressed as the difference in parts per thousand (‰) from standard reference material as indicated below: where X is 13 C or 15 N, R is the corresponding ratio of 13 C: 12 C or 15 N: 14 N, and δ is a measure of heavy to light isotopes in the sample. The international standards of Pee Dee Belemnite (PDB) for carbon and atmospheric N 2 for nitrogen were used as references.

Statistical analysis
Comparisons of the differences in oxygen concentration during photosynthesis and respiration, the contents of chlorophyll pigments, amino acid contents, δ 13 C and δ 15 N values between sponge CF, MP and NCF were made using a one-way analysis of variance (ANOVA) followed by Tukey's honest significant difference (HSD) test for multiple comparisons at a significance level of 0.05.

Laboratory measurement of photosynthetic capability and dark respiration rate
Physiological differences in the three Terpios body locations were detected during photosynthetic oxygen production and dark respiration. The photosynthesis-irradiance (P-I) curve of the sponge CF was determined before examining photosynthetic performance. Terpios reached its maximum photosynthetic rate at an irradiance of > 400 μmol photons m -2 s -1 and no photoinhibition occurred below 1,000 μmol photons m -2 s -1 (Fig 2). The data obtained at irradiances >1,000 were not used in this study because they caused respiration chamber temperatures to rise. The variables shown in Fig 2 fitted well into a non-photoinhibition equation model (P = P max (I)/ KI + (I); r 2 = 0.910). With this equation, the calculated P max (maximum photosynthetic rate) was 36 ± 3 pmol O 2 min -1 cm -2 , and KI (half saturation constant) was 93 ± 28 μmol photons m -2 s -1 . When comparing photosynthetic performance between Terpios from the CF, MB, and NCF (n = 17 for each parts) at an irradiance of 400 μmol photons m -2 s -1 , sponges displayed significantly higher photosynthetic rates in the CF samples (33 ± 3 pmol min -1 cm -2 ) than in MB (20 ± 2 pmol min -1 cm -2 ) and NCF samples (21 ± 2 pmol min -1 cm -2 ) (F (2,48) = Table 1. Changes in the maximum quantum yield obtained from Terpios-cyanobacteria symbiosis and photosynthetic pigment content of freshly isolated cyanobacteria at the coral contact front, mid-body, and no-contact front. The content of pigments is standardized to dry weight of sample and expressed as μg g -1 dry wt. Data are means ± S.E., followed by the number of replicates in parentheses. Means followed by the same letter are not significantly different at P = 0.05 (Tukey's honest significant difference test).

Discussion
Terpios hoshinota was first discovered in Guam four decades ago [10], and since then the occurrences of coral overgrowth outbreaks have been continuously reported from the Pacific Ocean and Indian Ocean [7,[10][11][12][13][15][16][17][18]. However, the interactions between T. hoshinota and opponent corals were not explored until recently, with results suggesting a significant change in Terpios morphology when combating corals [23]. By measuring the physiological characteristics of cyanobacteria from different parts of the sponge, our study represents the first to demonstrate that T. hoshinota competes with opponent corals not only by morphological transformation of the sponge-cyanobacteria association but also by physiological outperformance (e.g., higher photosynthetic rate and nitrogen status) in order to accumulate resources for fighting with corals.

Physiological characteristics of cyanobacteria in different parts of Terpios hoshinota
As a sponge species containing a high density of photosynthetic cyanobacteria, the physiological performance of Terpios is expected to closely link to the photosynthetic activity of the cyanobacteria. To evaluate the photosynthetic performance of the Terpios-cyanobacteria association, darkness adapted maximum quantum yield (F v /F m ) and oxygen production rate were measured within the three regions of the sponge. Though maximum quantum yield was commonly used to examine the health status of algal symbiotic coral and cyanobacteria [25][26][27], F v /F m values of the cyanobacteria symbiotic with Terpios hoshinota were relatively low (0.256-0.319) compared to measurements from most corals, which have F v /F m values of > 0.6 in non-stressed situations. The Campbell et al. [32] review of the chlorophyll fluorescence of cyanobacterial photosynthesis concludes that cyanobacteria containing high concentrations of phycocyanin usually show low F v /F m values. The cyanobacteria association in T. hoshinota is dominated by an isolated group closely related to Prochloron spp. [22], and microscopic anatomy and pigment characterization show that it also contains a high concentration of phycocyanin [19], suggesting that a low F v /F m ratio is an expected pattern in cyanobacteria associated with T. hoshinota. In addition, higher maximum quantum yield in the Terpios MB and NCF might be attributed to their higher Chl b in the photosynthetic pigment assembly of this strain of cyanobacteria (Table 1), because Chl b is able to amplify photosystem-II (PSII) functional Chl antenna size about 3.4 times higher than with only Chl a [33]. Furthermore, low F v /F m in healthy cells can be a measurement artefact when the light source does not provide sufficient intensity to saturate PSII [34], and modification to the light source is needed to obtain accurate F v /F m values [35]. Because the many potential interferences might affect different parts of Terpios, in situ maximum quantum yield might not be a suitable indicator to determine the photosynthetic performance of T. hoshinota.
The higher photosynthetic performance in Terpios CF was supported by the ca. 61% increase in photosynthetic oxygen production that was measured in the laboratory. Comparatively lower photosynthetic performance within Terpios MB and NCF were not due to aging or damage of the Terpios-cyanobacteria symbiosis because dark respiration rates in the sponge and the contents of major photosynthetic pigment (Chl a and c1+c2) in the cyanobacteria were comparable among Terpios' three different body parts. In addition, the maximum quantum yield of sponges measured in the field at the MB and NCF were even higher than at the CF, supporting the observation in the laboratory that low photosynthetic performance in the MB and NCF was not due to aging or damage to cyanobacteria.

Accumulation of resources to fight with opponent corals?
Why does T. hoshinota from the CF have a significantly higher photosynthetic performance than the MB and NCF? One hypothesis is that the Terpios CF needs to accumulate resources including nutrients for cell proliferation to fight opponent corals. Our previous study showed that when encountering coral opponents, Terpios develops morphological modifications in preparation for the battle by packing itself with cyanobacteria, sponge tissues, and spicules [23]. This change might require sponge cell and spicule regeneration and particularly high cyanobacteria proliferation resulting in a higher density of germs per unit area of the Terpios CF than MB and NCF. The higher performance of these physiological parameters might result from the higher cell density at the Terpios CF. This hypothesis is supported by scanning electron microscope imaging in which "hairy" tips at the Terpios edge are packed with a high density of cyanobacteria and spicules [22,23]. However, cyanobacteria density in Terpios is nearly impossible to determine accurately since Terpios tissue penetrates deeply into the undulating coral skeleton matrix [19], making it difficult to recover all cyanobacteria from the coral skeleton substrate. Even with painstaking air/seawater brushing, the coral skeleton substrate still contained blackish cyanobacteria. Nevertheless, a higher proliferation rate of cyanobacteria in the Terpios CF was indirectly indicated by higher δ 13 C values than in the MB and NCF (Fig  4A), since a high growth rate would reduce stable carbon isotope fractionation in many marine phytoplankton taxa including cyanobacteria [36,37]. Direct evidence to demonstrate the increase in proliferation rate will be further determined by more precise techniques such as the BrdU staining and cell cycle analysis [38,39].
Does coral provide a nitrogen source for Terpios hoshinota?
The second scenario for explaining the elevation of photosynthesis at the Terpios CF might be ammonium enrichment derived from the disintegration of coral tissues overgrown by the sponge. Seawater at the Green Island sampling site is very low in nitrogen (with non-detectable NH 3 , NO 2 -1 , and NO 3 -1 from P-J Meng unpub. data). The oligotrophic environment at Green Island might create ammonium-limiting conditions for symbiotic cyanobacteria in Terpios as described for coral symbiosis by Cook et al. [40]. It is also well documented that prolonged exposure to ammonium enrichment by nitrogen-starved phytoplankton elevates algal photosynthetic rates [41][42][43]. Therefore, abundant ammonium provided by disintegrating coral tissue at the CF might elevate cyanobacteria photosynthesis in the sponge compared to the MB and NCF. A higher ammonium supply could also explain the slightly higher discrimination of 15 N in CF cyanobacteria because an increase in dissolved inorganic nitrogen would elevate stable nitrogen isotope fractionation [44][45]. Furthermore, restricted ammonium enrichment at the coral contact front might be attributed to a strong dilution effect in the water column and the limited internal transportation efficiency in a Terpios body cavity having a crowded cyanobacteria population. This possibility was also supported by total and proteinogenic FAA contents being higher in CF cyanobacteria than in MB and NCF tissues ( Table 2). It has been suggested that internal FAA levels in phytoplankton and photosynthetic cyanobacteria are mediated by the external ammonium supply [46][47].
In conclusion, our data on photosynthetic capability, nitrogen status, and stable carbon isotope data imply that Terpios might have to accumulate resources at the sponge edge to fight opponent corals for substrates. Such resources might be derived from photosynthates released by cyanobacteria or supplied by inorganic nitrogen during the process of overgrowing encountered corals, which were higher than in distal parts of sponge tissues not in contact with corals. However, the question of whether Terpios kills coral for nutrients or only competes for substrate and obtains nutrients as byproducts is still unanswered. More studies are needed to resolve this puzzle, especially since there are more than four types of boundary interactions between Terpios and corals [23] and Terpios does not always win the competition.