Nutrient acquisition in four Mediterranean gorgonian species

MARINE ECOLOGY PROGRESS SERIESMar Ecol Prog Ser

Vol. 473: 179–188, 2013doi: 10.3354/meps10037

Published January 21

INTRODUCTION

Gorgonians are among the most abundant and im -portant suspension feeders of Mediterranean benthiccommunities (Weinberg 1976). Large gorgoniansform dense aggregations that reduce water flow andaffect sediment deposition, and species like Para-muricea clavata are responsible for a large share(~40%) of the biomass of the benthic community.

They play a paramount role in shaping these commu-nities by providing shelter for other animals andalgae (Gili & Coma 1998). They significantly affectthe carbon and nitrogen cycles of the water columneither by capturing planktonic prey and filteringlarge amounts of particulate and dissolved organicmatter, or by releasing mucus and other metabolicwastes into seawater (Gili & Coma 1998, Wild et al.2004). Knowledge of their feeding ecology is there-

© Inter-Research 2013 · www.int-res.com*Corresponding author. Email: [email protected]

Nutrient acquisition in four Mediterranean gorgonian species

Silvia Cocito1, Christine Ferrier-Pagès2,*, Roberta Cupido1, Cecile Rottier2, Wolfram Meier-Augenstein3, Helen Kemp3, Stephanie Reynaud2, Andrea Peirano1

1Marine Environment Research Center, Italian National Agency for New Technologies, Energy and Sustainable Economic Development (ENEA), 19100 La Spezia, Italy

2Centre Scientifique de Monaco, Avenue Saint-Martin, 98000 Monaco3Stable Isotope Forensics Laboratory, James Hutton Institute, Invergowrie, Dundee DD2 5DA, UK

ABSTRACT: Carbon and nitrogen isotope abundance values (δ13C and δ15N, respectively)were measured for the first time in the soft tissue, axial skeleton, and spicules of 4 Mediter-ranean gorgonians, 3 asymbiotic (Leptogorgia sarmentosa, Paramuricea clavata, and Eunicellaverrucosa) and 1 symbiotic with autotrophic dinoflagellates (Eunicella singularis). The isotopiccomposition of their diet, i.e. zooplankton, particulate organic matter (POM), and sedimentaryorganic matter (SOM), was also measured to understand gorgonian feeding ecology. (1) Car-bon and nitrogen signatures of the symbiotic E. singularis tissue in summer differed signifi-cantly from the signatures of the other species; (2) carbon and nitrogen signatures of the axialskeleton were similar to those of the tissue, because the skeleton is primarily made ofgorgonin secreted by the tissue; and (3) spicules had a high δ13C signature because they aremade by a combination of 60 to 76% of respiratory CO2 and of external CO2, with a high δ13Csignature. Comparison of the isotopic signatures of the gorgonian tissues and the food sourcesindicated that E. singularis and P. clavata had the same diet in both winter and summer,either zooplankton for E. singularis or POM and SOM for P. clavata. Conversely, L.sarmentosa and E. verrucosa shifted from zooplankton in winter to SOM in summer. Theseresults bring insights into the feeding ecology of temperate gorgonians and explain their dis-tribution, abundance, and role in the flow of particulate matter between the water columnand the benthos.

KEY WORDS: Temperate gorgonian · Feeding ecology · Symbiosis · Isotope · Spicule

Resale or republication not permitted without written consent of the publisher

This authors' personal copy may not be publicly or systematically copied or distributed, or posted on the Open Web, except with written permission of the copyright holder(s). It may be distributed to interested individuals on request.

Mar Ecol Prog Ser 473: 179–188, 2013

fore essential at different levels. It is important tounderstand the biology and distribution of gorgoni-ans, as well as their patterns of environmental adap-tation, because acquisition of energy is a determinantfactor for growth, fecundity, and thus the demo-graphic distribution of the species. Distribution andabundance in turn determine the ecological role ofgorgonians in littoral ecosystems. Finally, uptake andrelease of nutrients by gorgonian communities influ-ence by energy flow between the pelagic and thebenthic systems (Gili & Coma 1998, Coma et al.2001). Despite much evidence suggesting that thefeeding ecology of benthic suspension feeders maybe important in understanding the functioning of littoral ecosystems, especially in areas with majorenvironmental changes, natural diets and feedingrates of benthic suspension feeders are still poorlyknown.

It is difficult to study the trophic ecology of temper-ate gorgonians because they feed on a wide spec-trum of food types, from pico- and nanoplankton tomicroplankton including particulate organic matter(POM) (Coma et al. 1994, Ribes et al. 1999, 2003).Like many anthozoans, they can also feed on organicmatter contained in re-suspended sediment (An thony& Fabricius 2000, Orejas et al. 2001, Mills & Sebens2004). Food abundance and composition may changeconsiderably with depth, season, or among locations(Garrabou et al. 2002, Coma & Ribes 2003, Rossi et al.2004). Although pico- and nanoplankton show lowseasonality (Coma & Ribes 2003), detrital POM andmicroplankton exhibit marked seasonal patterns,with high values in winter and spring and low valuesin summer. Therefore, food may occasionally becomea constraining factor for gorgonians, which are con-tinuously subject to random pulses of food availabil-ity (Gili & Ros 1985, Coma & Ribes 2003, Rossi et al.2004, 2006).

Previously, feeding rates have mainly been studiedin 2 species, Paramuricea clavata and Leptogorgiasarmentosa, by analyzing polyp contents or byenclosing specimens in chambers and measuring thedepletion of natural particles (Coma et al. 1994, Ribeset al. 1999, 2003). Another approach to assess thephysiology and nutritional ecology of gorgoniansconsists of measuring the stable carbon and nitrogenisotopic signatures (δ13C and δ15N, respectively) oftheir soft tissue. Indeed, stable isotopes have beenincreasingly used to provide time-integrated infor-mation on trophic relationships (Riera 2007), espe-cially in coastal environments characterized by manylocal and imported food sources. Carbon isotopic signatures of the consumers are usually similar, or

increased by 1‰ compared to their diet, whereasnitrogen signatures are enriched by 2.5 to 3.4‰depending on tissue composition, nutritional status,and mode of nitrogen excretion (Minagawa & Wada1984, Vander Zanden & Rasmussen 2001, Gollety etal. 2010). In anthozoans containing symbiotic dino -flagellates, nitrogen fractionation can be differentcompared to heterotrophic species, as the nitrogenisotopic signature of the host is either similar orincreased by a few permille compared to its sym-bionts (Reynaud et al. 2009, Ferrier-Pagès et al.2011). Although this isotopic approach has beenapplied to tropical gorgonians to monitor seawaterpollution (Ward-Paige et al. 2005, Risk et al. 2009,Baker et al. 2010, Sherwood et al. 2010) and to deep-sea gorgonians to understand their trophic patterns(Carlier et al. 2009, Sherwood et al. 2011), the iso-topic signature of Mediterranean gorgonians fromshallow waters has been investigated only in Euni-cella singularis (Gori et al. 2012).

We sampled 4 Mediterranean gorgonian species,Paramuricea clavata, Leptogorgia sarmentosa, Euni-cella verrucosa, and E. singularis, in both winter andsummer, from the Gulf of La Spezia, where the spe-cies are particularly abundant and live together(Cocito et al. 2002). The isotopic signatures of thegorgonian soft tissue, axial skeleton, and spiculeswere compared to the signature of their potentialdiet, i.e. zooplankton, POM in seawater, and sedi-mentary organic matter (SOM). We aimed at linkingthe isotopic signature of the gorgonian tissue to aspecific food source and to highlight trophic differ-ences among species or be tween seasons. E. singu-laris, in particular, hosts symbiotic dinoflagellates,which are ex pected to transfer a large part of theirphotosynthetic carbon to their host. We thereforehypothesized that the carbon signature of the hostshould have been largely affected by this autotrophiccarbon and different from the signature of the otherspecies, which are fully heterotrophic. The secondaim was to assess the carbon signature of the spiculesto understand the origin of the carbon depositedwithin them. In scleractinian tropical corals, 70% ofthe carbon deposited in the skeleton comes frominternal respiration, while 30% has an external ori-gin (Furla et al. 2000). In gorgonians, this point hasyet to be clarified since Allemand & Grillo (1992) con-cluded that metabolic CO2 may serve as a mainsource of carbon in the red coral Corallium rubrum,whereas Lucas & Knapp (1997) found the contrary.Finally, the gorgonians’ trophic ecology is discussedto explain their particular abundance in the Gulf ofLa Spezia.

180A

utho

r cop

y

Cocito et al.: Nutrient acquisition in Mediterranean gorgonians

MATERIALS AND METHODS

Sampling and treatment of gorgonian species

Three asymbiotic gorgonian species were investi-gated: Paramuricea clavata (Risso, 1826), Leptogor-gia sarmentosa (Esper, 1789), Eunicella verrucosa(Pallas, 1766), and the zooxanthellate E. singularis(Esper, 1791). These species are among the most rep-resentative gorgonians of the western Mediterran-ean sublittoral communities (Carpine & Grasshoff1975). P. clavata and E. verrucosa are typically foundbetween 10 and 110 m depth on shaded rocky wallsand rocky bottoms, respectively; L. sarmentosa is anubiquitous species living between 10 and 200 mdepth on pebbles and boulders buried in sediment(Weinberg 1976); and E. singularis is abundant onhorizontal and gently sloping rocky bottoms but at amore shallow depth (5 to 60 m) because its symbioticdinoflagellates require light.

Sampling was performed at Tinetto Island (44°01’ N, 09° 50’ E) in the western part of the Gulf of LaSpezia (eastern Ligurian Sea). The area is character-ized by high turbidity due to terrestrial runoff fromthe Magra River and sewage discharge from the cityof La Spezia (Cocito et al. 2002). Sedimentation ishigh and constant over the year, with total suspendedsolids ranging between 1.61 and 2.65 mg l−1 and arate (mean + SD) on the vertical cliff ranging be -tween 188 ± 40 and 375 ± 61 g m−2 yr−1 (Cupido et al.2009). The visual range is very low in autumn andwinter (4.5 m) and increases in spring (8 m) and sum-mer (10.5 m). All gorgonian species were sampledbetween 22 and 25 m at the western side of theisland, where the walls descend steeply to a depth of26 m on a flat, muddy bottom. P. cla vata formscanopies on the vertical cliffs between 17 and 25 m,whereas E. singularis, E. verrucosa, and Leptogorgiasarmentosa live on boulders on the muddy bottom(Cocito et al. 2002). Gorgonians were collected bySCUBA diving in winter (March 2) and summer(August 18) 2009. From each of 3 colonies per gor-gonian species, 15 apical fragments (15 cm long)were cut en closed in plastic bags containing the sur-rounding seawater and brought to the laboratory.

For the analysis of carbon and nitrogen isotopic sig-natures of the soft tissue, 2 different protocols werecompared, each allocating 5 fragments per colony × 3colonies per species × 2 seasons (Fig. 1). The first pro-tocol envisaged incubating the fragments in sea waterenriched with magnesium chloride to relax the polyps.About 200 polyps per fragment were detached fromthe gorgonian axis using a dissecting needle and

scalpel under a binocular microscope, frozen in liquidnitrogen and freeze-dried prior to subsequent analy-sis. The second protocol (applied on 5 fragments percolony × 3 colonies per species × 2 seasons) arrangedfor precisely separating the soft tissue from the axialskeleton of frozen fragments with a scalpel. The iso-lated soft tissue was then ground in a mortar underliquid nitrogen, homogenized with a Potter blender,and centrifuged at 500 × g and 4°C for 10 min. Cen-trifugation allowed separation of soft tissue fromspicules (remaining at the bottom of the tube). Tissuesamples were then frozen in liquid nitrogen andfreeze-dried. All materials used for these extractionswere previously heat-treated at 450°C for 5 h or rinsedin distilled water. All samples were analysed for their15N and 13C isotopic composition, in triplicate, as de-scribed below. Stable isotope analyses of tissue sam-ples prepared by either method yielded comparable re-sults (ANOVA, no significant difference, p = 0.18) andwere therefore pooled for subsequent data ana lysis (n= 30 per species and season).

For the determination of 15N and 13C isotopic signa-tures of the axial skeleton and spicules, 5 fragmentsper colony × 3 colonies per species × 2 seasons wereconsidered. Each fragment was cut into a few piecesusing clean scissors and macerated at 80°C for 10 minin a 50 ml Falcon tube containing a solution of so -dium hydroxide (NaOH, 1 N), leading to a cleanskeleton (Houlbrèque et al. 2011). Pieces of axialskeleton were then isolated with clean forceps,

181

(A) winter and (B) summer

Species 1 to 4

C1 C2 C33 colonies

4 species

2 seasons

15 fragments C1-1 toC1-15

C2-1 toC2-15

C3-1 toC3-15

Treatment 1Tissue

Treatment 2Tissue, spicules,and skeleton

C1-1 toC1-5

C2-1 toC2-5

C3-1 toC3-5

C1-6 toC1-10

C2-6 toC2-10

C3-6 toC3-10

Treatment 3 Spicules andskeleton

C1-11 toC1-15

C2-11 toC2-15

C3-11 toC3-15

Fig. 1. Matrix of the sampling and treatment procedures. Foreach of the 4 gorgonian species, 3 colonies (C1 to C3) werecollected in (A) winter and (B) summer. From each colony,15 fragments were taken and distributed to the Treatments1, 2, and 3. Each fragment was analyzed in triplicate for its

δ13C and δ15N signature

Aut

hor c

opy

Mar Ecol Prog Ser 473: 179–188, 2013

rinsed with distilled water, ground under liquid nitro-gen, and freeze-dried. The remaining solution ofNaOH, containing the spicules and solubilized softtissue, was centrifuged at 1000 × g and 4°C for10 min. The supernatant was discarded, and theremaining spicules were rinsed several times withdistilled water, frozen in liquid nitrogen, and freeze-dried prior to subsequent analysis. Results for skele-ton and spicule samples obtained using this tech-nique were compared to those obtained using thescalpel technique. All samples were analysed fortheir 15N and 13C isotopic composition, in triplicate, asdescribed below. There was no significant differencebetween the 2 techniques for axial skeleton samples(ANOVA, p = 0.12), but data for spicules were differ-ent (ANOVA, p < 0.001), pointing towards a contam-ination of spicule samples with soft tissue materialwhen prepared using the scalpel. For this reason,the latter data were not included in any subsequentdata analysis.

Sampling and treatment of particulate organic matter

To characterize the isotopic signature of POM suspended in seawater, in each season, 3 separatevolumes of 10 l of seawater were sampled near thegorgonians using a Niskin® bottle. Back at the labo-ratory, at least 5 l of seawater was filtered in triplicatethrough previously heat-treated (450°C, for 5 h)25 mm Whatman GF/F filters under low vacuum. Toremove carbonates, filters were quickly acidified(1 mol l−1 HCl) and rinsed with distilled water. Theywere then freeze-dried and kept frozen until isotopicanalyses. Due to their low abundance in seawater,only a few zooplankters were collected on the filters.So, to obtain reliable isotope abundance data for zoo-plankton, concentrated samples were collected usinga plankton net (type WP2; 5 min hauls). Plankton werethen collected on filters and treated as describedabove.

Sampling and treatment of suspended organic matter

In each season, replicate sediment samples fromthe muddy bottom near the gorgonians were col-lected using plastic jars and by scraping the upper1 cm of the surface (Riera et al. 1996). Samples werethen sieved to a grain size of <50 µm to separate sandgrains from sedimentary POM. Three sub-samples of

~5 g were freeze-dried and ground using a pestleand mortar. A 200 mg aliquot from each sub-samplewas acidified with 1 N HCl to remove inorganic car-bon. To prevent any loss of dissolved organics, thesesamples were not rinsed, but acidification was performed on a hot plate to quickly evaporate theacid. Samples were dried overnight at 50°C in a fumehood (Riera 1998). Once dried, the sediment wasmixed with Milli-Q water, freeze-dried, groundagain to a fine powder, and kept frozen (−80°C) untilanalysis.

Stable isotope analysis

Samples were analysed for their 15N and 13C iso-topic composition using an elemental analyser(Flash EA-1112) coupled to a Delta V Advantageisotope ratio mass spectrometer (both Thermo-Fisher). International reference materials (USGS40and IAEA-CH6, International Atomic EnergyAgency) were analysed with each sample batchand used for scale calibration of results of 13C iso-tope analyses to Vienna PeeDee Belemnite (VPDB).Similarly, results of 15N isotope analyses werescale-calibrated to N2 in air using USGS40 and anin-house standard (leucine, δ15N = 10.77‰; Fluka),whose δ15N had been independently validated (Iso-Analytical). Two different analytical control sampleswere also analysed with each batch for quality con-trol. Precision as determined by repeat analysis ofthe reference materials and quality controls wasbetter than ±0.20 and ±0.15‰ for measured δ15Nand δ13C values, respectively. Data are expressedin the standard δ-unit notation for the heavier iso-tope h of a given element X:

δhX = [(Rsample/Rreference) −1] × 103 (1)

where R = 13C/12C for X = carbon and 15N/14N for X =nitrogen, with δ values reported relative to VPDBand air for 13C and 15N, respectively.

Statistics and models

The isotopic signatures of gorgonian soft tissue,axial skeleton, and spicules (4 species × 3 coloniesper species × 5 to 10 samples per colony × 2 seasons)were compared by a 2-way ANOVA using the soft-ware package Statistica® with 2 independent fac-tors, species (4 levels) and seasons (2 levels). Vari-ance homogeneity was checked via Cochran’s test. ATukey test was used for post-hoc comparison of

182A

utho

r cop

y

Cocito et al.: Nutrient acquisition in Mediterranean gorgonians

levels. The isotopic signatures of POM, SOM, andzooplankton, sampled at the 2 seasons, were com-pared with a 1-way ANOVA using the softwarepackage Statistica®. To give an estimation of the isotopic contribution of each determined food sourceto the isotopic signal of the gorgonian tissue, a com-puter program (IsoSource) developed by Phillips &Gregg (2003) was used.

RESULTS

For all gorgonian species, spicules showed higherδ13C values compared to the axial skeleton and tissuesignatures, ranging from −6.81 to −7.97‰, while thetissue and axial skeleton signatures were compara-ble, with lower δ13C values ranging from −18 to−24‰ (Fig. 2). Soft tissue and axial skeleton δ13C signatures (Fig. 2) were not significantly different(Tukey test, p = 0.99 and 0.78, respectively) betweenParamuricea clavata, Eunicella verrucosa, and Lep-togorgia sarmentosa. However, soft tissues and theaxial skeleton of E. singularis had distinct signaturesof their own (Tukey test, p < 0.0001 and p < 0.001,respectively), with δ13C values 1 to 2‰ more positivethan the values of the other species. As discussedlater, this difference could be due to the presence ofsymbiotic algae in the tissue. Finally, seasonal differ-ences were small, with soft tissue δ13C signatures ofL. sarmentosa and E. singularis slightly moredepleted in summer than in winter.

Concerning the carbon isotopic signature of thefood (Fig. 3), POM and SOM were significantly moredepleted than the zooplankton, in both winter andsummer (1-way ANOVA, p < 0.0001), but there wasno seasonal difference in the signature of each foodsource (ANOVA, p = 0.04).

As for δ13C, the δ15N signature wasalso different between species andseasons (ANOVA, p < 0.001). The tis-sue of Paramuricea clavata (Fig. 4) hadthe lowest δ15N values of the gorgon-ian species, in both summer and win-ter (Tukey test, p < 0.01, from 5.07 to5.61‰). The isotopic signature of thespicules of Eunicella verrucosa wasalso higher in winter compared to thesignatures of the other gorgonians(Tukey test, p < 0.04). Conversely, theδ15N value of the axial skeleton wasconsistent between species and sea-sons and ranged between 7.48 and8.20‰ (2-way ANOVA; Tukey test, p >

0.05). Concerning seasonal differences, δ15N values ofthe tissues of Leptogorgia sarmentosa and E. singu-laris were significantly higher in winter than in sum-mer (Tukey test, p < 0.01). Concerning the δ15N sig-nature of the food (Fig. 3), only POM showed asignificant seasonal effect, with δ15N values higher insummer than in winter (p < 0.01).

183

Leptogorgia sarmentosa

Eunicellasingularis

Eunicellaverrucosa

Paramuriceaclavata

0

–5

–10

–15

–20

–25

0

–5

–10

–15

–20

–25

A

B

δ13C

(‰)

δ13C

(‰)

TissueAxial skeletonSpicules

Fig. 2. δ13C values of the tissue, axial skeleton, and spiculesof the 4 gorgonian species, sampled in (A) winter and (B)

summer. Mean + SD of 5 samples

SOM POM Zooplankton0

2

4

6

8

10

δ15N

(‰)

SOMA B

POM Zooplankton0

–5

–10

–15

–20

–25

–30

δ13C

(‰)

Fig. 3. (A) δ13C and (B) δ15N values of the sedimentary organic matter (SOM),particulate organic matter (POM) and zooplankton in winter (dark) and

summer (pale). Means + SD of 5 samples

Aut

hor c

opy

Mar Ecol Prog Ser 473: 179–188, 2013

Carbon and nitrogen signatures of the differentsamples were simultaneously plotted in Fig. 5 tohighlight the preferentially exploited food sources ofgorgonians in each season. Such plots are usuallyused in complex food chains to determine the foodsource of a specific predator (Riera et al. 1999, 2009),assuming a mean trophic enrichment of 1‰ in δ13C(Rau et al. 1990) and 2.5 to 3.4‰ in δ15N (Minagawa& Wada 1984, Gollety et al. 2010) for the predatorbecause of the assimilation of food. The exact valueof the isotopic fractionation of carbon and nitrogenbetween the gorgonians and their prey is not known.We therefore applied the usual mean value of 3.4‰(represented by d in Fig. 5), considering that a valueof 2.5‰ resulted in the same conclusions. Applica-tion of the above model to the gorgonians suggestsdifferent feeding behaviours, depending on the gor-gonian species and the season. In winter, the isotopicsignatures of the tissues of Leptogorgia sarmentosa,Eunicella singularis, and E. verrucosa were close tothe zooplankton signature, while Paramuricea clavatatissue was closer to the POM and SOM signatures. Insummer, the isotopic signatures of L. sarmentosa, P.clavata, and E. verrucosa were close to the SOM sig-nature, while E. singularis tissue was closer to thezooplankton signature. According to the model of

Phillips & Gregg (2003), SOM indeed contributed ca.58% of the tissue signature of L. sarmentosa, P.clavata, and E. verrucosa, while zooplankton andPOM contributed 34 and 8%, respectively. Con-versely, there was a 94% contribution of the zoo-plankton to the tissue signature of E. singularis.

DISCUSSION

We demonstrated the acquisition of food and nutri-ents by 4 gorgonian species of the MediterraneanSea at 2 different seasons. Results will contribute tothe establishment of the role of this important compo-nent of the benthic Mediterranean community to bio-geochemical fluxes in sublittoral ecosystems and, ingeneral, to an understanding of trophodynamics inthese communities. In addition to addressing primaryresearch questions, the study may help in under-

184

Leptogorgiasarmentosa

Eunicellasingularis

Eunicellaverrucosa

Paramuriceaclavata

A

B

δ15N

(‰)

δ15N

(‰)

TissueAxial skeletonSpicules

0

2

12

4

6

8

10

0

2

4

6

8

10

Fig. 4. δ15N values of the tissue, axial skeleton, and spiculesof the 4 gorgonian species, sampled in (A) winter and (B)

summer. Mean + SD of 5 samples

2

0

4

6

8

10

12

ES LS

EV

Zooplankton PC

POM

SOM

A

2

0

4

6

8

10

–28 –26 –24 –22 –20 –18 –14

ES

EV

LS

SOM

POM

PC Zooplankton

B δ15N

(‰)

δ13C (‰)

Fig. 5. δ13C versus δ15N (mean ± SD) for the gorgonians (ES =Eunicella singularis, EV = E. verrucosa, LS = Leptogorgiasarmentosa, PC = Paramuricea clavata), sedimentary or-ganic matter (SOM), particulate organic matter (POM), andzooplankton for (A) winter and (B) summer. The lines, termi-nated by (d) correspond to the theoretical food source ofthe corals, taking into account the trophic enrichment of

1 and 3.5‰ for δ13C and δ15N, respectively

Aut

hor c

opy

Cocito et al.: Nutrient acquisition in Mediterranean gorgonians

standing the mechanisms which sustain suspensionor affect nutrient dynamics. Overall, results haveshown that each gorgonian species has its own iso-topic signature, which depends on its diet and itsrelationship with the external environment as well asits symbiotic status. In summary, Eunicella singularisclearly had a higher δ13C signature than the otherspecies, independent of the season, due to its sym-biosis with photosynthesizing dinoflagellates. Allgorgonian species shifted from one diet to anotherdepending on the season and food availability. Threeout of the 4 gorgonian species (Leptogorgia sar -mentosa, E. verrucosa, and E. singularis) consumedplankton in winter, when densities were relativelyabundant, whereas they relied on detritic matter insummer.

Spicules of the 4 gorgonians had a very high car-bon isotopic signature compared to the axial skeletonor the soft tissue due to a different carbon origin. Inanthozoans, carbon used for calcium carbonate pro-duction can either come from the external dissolvedinorganic carbon (DIC) in seawater (CO2 or bicarbon-ate, HCO3

−) or from respiratory (or metabolic) CO2

(Erez 1978, Furla et al. 2000). Although it is clear nowthat metabolic CO2 contributes to at least 70% of theskeletal carbon in scleractinian corals (Furla et al.2000, Hughes et al. 2010), only 2 attempts have beenmade to address which source of inorganic carbon ispredominantly used in spicule production. Thesestudies led to opposite conclusions but also used dif-ferent methods to constrain carbon partitioning, andeach has its own relative strengths and weaknesses.While Allemand & Grillo (1992), using 45Ca labellingof the skeleton, concluded that metabolic CO2 mayserve as a main source of carbon in the red coralCorallium rubrum, Lucas & Knapp (1997) used 14C-labelling and found the contrary (70% of externalDIC) in the gorgonian Leptogorgia virgulata. Ourδ13C analyses suggest that the carbon origin inspicules is species dependent. According to Swart(1983), there is a 7‰ enrichment between the inter-nal pool of CO2 and the pool of HCO3

− and again~1‰ enrichment between the HCO3

− and the cal-cium carbonate of the skeleton or spicules. From amean δ13C value of the spicules equal to −8‰ in theasymbiotic gorgonians, the back-calculated value forthe internal CO2 pool at the origin of spicule forma-tion is −16‰ [(−8‰) + (−1‰) + (−7‰)]. Consideringthat the respired CO2 has the same isotopic signatureas the tissue (ca. −21‰ in this study), the carbon inthe spicules is a mix of 76% of respired CO2 and 24%of external seawater CO2 (76% of respired CO2 at−21‰ + 24% of external CO2 at 0.8‰; Reynaud et al.

2002). So, the major part of the carbon used forspicule formation seems to come from respired car-bon. In Eunicella singularis, showing a mean δ13Cvalue of −4.8‰ for the spicules, the same calculationssuggest a contribution of ~60 and 40% of respiredand seawater CO2, respectively. In this case, a largerfraction of the carbon used for spicule formationseems to come from external CO2. This differencemight be due to the presence of photosynthetic sym-bionts in E. singularis, which actively take up DICfrom seawater for their own needs of photosynthesis,and a part of this CO2 (with a higher δ13C signaturethan that of respiration) might be diverted to spiculeformation. Overall, since the above techniques aredifferent, with all their uncertainties, more measure-ments are necessary to completely understand theorigin of the carbon used for spicule formation.

The axial skeleton of the 4 gorgonian species wasmuch more depleted in 13C than were the spicules.This is explained by the fact that the skeleton isformed both by calcite and by a horn-like structuralprotein called gorgonin. Gorgonin is secreted by thesoft tissues, and the ingested food serves as thesource of amino acids and other molecules used inits synthesis (Roche et al. 1960, Heikoop et al. 2002,Sherwood 2006). Such skeleton is very enriched incarbon and nitrogen (35 and 12% of the total dryweight, respectively) and has therefore carbon andnitrogen signatures close to the signatures of the tis-sue influenced by the food source (Heikoop et al.2002, Sherwood 2006) or by the external levels ofdissolved inorganic nitrogen (Ward-Paige et al.2005, Risk et al. 2009, Sherwood et al. 2010). Theexternal influences are discussed below with the tis-sue signatures.

The soft tissue isotopic signatures have highlightedseasonal and species-specific differences. Eunicellasingularis had slightly but significantly higher δ13Cand δ15N signatures compared to the other non-sym-biotic gorgonians, due to the presence of symbiontsin its tissue. Indeed, symbiont photosynthetic activityuses dissolved HCO3

− as a precursor pool for carbon,which in the case of seawater HCO3

− has a naturallyhigh 13C signature of 0.8‰ (Reynaud et al. 2002). Theresulting photosynthetic compounds therefore exhi -bit a higher 13C signature of typically −12‰ (Rey-naud et al. 2002) compared to δ13C values of −22‰typically observed in planktonic food or POM (Fry2007). Also, in symbiotic organisms, nitrogen result-ing from metabolic waste products is not excretedinto seawater as in heterotrophic organisms. Instead,it is re-absorbed by the symbionts and thereforerecycled within the symbiosis, a process which

185A

utho

r cop

y

Mar Ecol Prog Ser 473: 179–188, 2013

increases the nitrogen isotopic signature. However,the carbon autotrophic input in E. singularis did notsignificantly increase from winter to summer, asexpected if we consider that rates of photosynthesisare enhanced in summer by higher light and temper-ature levels (Ferrier-Pagès et al. 2011). This suggeststhat either E. singularis keep a constant rate of photo -synthesis under the varying light intensities of the 2seasons or that the depth at which they live, coupledwith the turbidity of the water, induces a constantlight environment. Conversely, the δ15N signature ofE. singularis decreased from 10 to 8‰, suggesting amore important input of dissolved inorganic nitrogenin summer, presumably taken up through the activityof the zooxanthellae.

The seasonal difference in the isotopic compositionof the gorgonian tissue was linked to the seasonalchanges in food availability in the surrounding sea-water, as suggested by the δ13C-δ15N plots generallyused in complex food chains to determine the foodsource of a specific predator (Riera et al. 1999, 2009).In the Mediterranean Sea, large seasonal changes inthe abundance and composition of phyto- and zoo-plankton are indeed observed, with high concentra-tions in winter due to the upwelling of nutrients tothe surface and very low concentrations in summerafter the establishment of the thermocline. Summeris therefore generally considered a very un favour -able season for all benthic suspension feeders (Comaet al. 1998), because most of these species are entirelydependent on food supply to maintain their metabo-lism (Coma et al. 2000). The expected mean winterisotopic ratios for the preferentially exploited foodresources were close to the ratio of zooplankton forEunicella singularis, E. verrucosa, and Leptogorgiasarmentosa and to the ratios of POM and SOM, i.e.organic matter from seawater and sediment, for Para-muricea clavata. These conclusions are corroboratedby the results obtained using the model of Phillips &Gregg (2003), with zooplankton contributing >90%of the diet of the first 3 gorgonian species and POMand SOM contributing 75% of the diet of P. clavata.Summer results suggested 2 different feeding strate-gies depending on the gorgonian species. Indeed, L.sarmentosa and E. verrucosa showed seasonality intheir feeding habits, as already observed with otherbenthic suspension feeders (Coma et al. 2000). Theirisotopic signature shifted from zooplankton to SOM,suggesting that in summer, zooplankton concentra-tion was no longer sufficient to sustain the metabo-lism of these species (Coma et al. 2000), which reliedon the dominant food source available, i.e. detritalorganic matter from re-suspended sediment. SOM

can provide significant amounts of nutrients, such asnitrogen to anthozoans, and even sustain their meta-bolic needs in some conditions (Mills & Sebens 2004,Anthony & Fabricius 2000). This detrital material isabundant all year in the Gulf of La Spezia, becausethe Magra River continuously discharges a highquantity of sediment and particles on the vertical cliffwhere gorgonians thrive (Cupido et al. 2009). Con-versely, the 2 other species showed little seasonalityin their food sources. The P. clavata signature re -mained close to the SOM signature, as in winter, andzooplankton seemed to remain the preferential foodsource for E. singularis. In contrast to the other 2 gor-gonian species relying on zooplankton in winter, thesymbiotic status of E. singularis perhaps allowed it tosurvive with this sole heterotrophic food source insummer, despite its scarcity.

Only 2 out of the 4 gorgonians of this study hadbeen previously investigated for their feeding habits,using punctual measurements of prey disappearancein incubation chambers (Ribes et al. 1999, 2003).These previous studies confirm the validity of ourapproach using stable isotopes. Ribes et al. (1999)observed that Paramuricea clavata from the MedesIslands mainly relied on detrital POC (86% of thetotal diet), especially during winter, when particlesare re-suspended in the water column. ConcerningLeptogorgia sarmentosa, Ribes et al. (2003) con-cluded that detritus and zooplankton were the 2 mainfood sources, which also confirms our findings thatzooplankton account for the major δ13C signal of thetissue of this species.

In conclusion, this is one of the first studies on thefeeding habits of 4 predominant gorgonian species ofthe Mediterranean Sea under natural conditions. Ithas shown a high predation rate on zooplankton inwinter by 3 out of the 4 species and a shift to detritalmaterial in summer, when plankton is no longer suf-ficiently abundant to sustain gorgonian metabolism.The Gulf of La Spezia has a high detrital load all year,and this might explain the high gorgonian abun-dance in this zone, since they can rely on this foodsource during nutrient shortage. Because the Gulf isalso under terrestrial influence, zooplankton is par-ticularly abundant in winter and therefore constitutesa high energetic food source at this season. The sym-biotic Eunicella singularis acquires additional carbonand nitrogen from its symbionts. Finally, the carbonused to build the spicules, is mainly produced frommetabolic respiration, except again for E. singularis,which derives a larger fraction of carbon from theexternal medium via the activity of its symbiotic zooxanthellae.

186A

utho

r cop

y

Cocito et al.: Nutrient acquisition in Mediterranean gorgonians

Acknowledgments. We thank R. Delfanti, Head of the Mar-ine Environment Research Centre, ENEA, for hosting C.F.-P.,S.R., and C.R. in La Spezia; D. Allemand, Director of the Sci-entific Centre of Monaco; and all those who helped us in thelaboratories and at sea, including Dr. M. Abbate, Dr. G. Cer-rati, A. Bordone, and S. Sikorski. Financial support was pro-vided by the Centre Scientifique de Monaco and ENEA.

LITERATURE CITED

Allemand D, Grillo MC (1992) Biocalcification mechanism ingorgonians: 45Ca uptake and deposition by the Mediter-ranean red coral Corallium rubrum. J Exp Zool 262: 237−246

Anthony KRN, Fabricius KE (2000) Shifting roles of het-erotrophy and autotrophy in coral energetics under varying turbidity. J Exp Mar Biol Ecol 252: 221−253

Baker DM, Webster KL, Kim K (2010) Caribbean octocoralsrecord changing carbon and nitrogen sources from 1862to 2005. Glob Change Biol 16: 2701−2710.

Carlier A, Le Guilloux E, Olu K, Sarrazin J, Mastrototaro F,Taviani M, Clavier J (2009) Trophic relationship in adeep Mediterranean cold-water bank (Santa Maria diLeuca, Ionian Sea). Mar Ecol Prog Ser 397: 125−137

Carpine C, Grasshoff M (1975) Les Gorgonaires de laMéditerranée. Bull Inst Océanogr Monaco 71: 1−140

Cocito S, Bedulli D, Sgorbini S (2002) Distribution patternsof the sublittoral epibenthic assemblages on a rockyshoal in the Ligurian Sea (NW Mediterranean). Sci Mar66: 175−181

Coma R, Ribes M (2003) Seasonal energetic constraints inMediterranean benthic suspension feeders: effects at different levels of ecological organization. Oikos 101: 205−215

Coma R, Gili JM, Zabala M, Riera T (1994) Feeding and preycapture cycles in the aposymbiotic gorgonian Para-muricea clavata. Mar Ecol Prog Ser 115: 257−270

Coma R, Ribes M, Gili JM, Zabala M (1998) An energeticapproach to the study of life-history traits of two modularcolonial benthic invertebrates. Mar Ecol Prog Ser 162: 89−103

Coma R, Ribes M, Gili JM, Zabala M (2000) Seasonality incoastal benthic ecosystems. Trends Ecol Evol 15: 448−453

Coma R, Ribes M, Gili JM, Hughes RN (2001) The ultimateopportunists: consumers of seston. Mar Ecol Prog Ser219: 305−308

Cupido R, Cocito S, Barsanti M, Sgorbini S, Peirano A, San-tangelo G (2009) Unexpected long-term populationdynamics in a canopy-forming gorgonian following massmortality. Mar Ecol Prog Ser 394: 195−200

Erez J (1978) Vital effect on stable-isotope composition seenin foraminifera and coral skeletons. Nature 273: 199−202

Ferrier-Pagès C, Peirano A, Abbate M, Cocito S and 5 others(2011) Summer autotrophy and winter heterotrophy inthe temperate symbiotic coral Cladocora caespitosa.Limnol Oceanogr 56: 1429−1438

Fry B (2007) Coupled N, C and S stable isotope measure-ments using a dual-column gas chromatography system.Rapid Commun Mass Spectrom 21:750–756

Furla P, Galgani I, Durand I, Allemand D (2000) Sources andmechanisms of inorganic carbon transport for coral calci-fication and photosynthesis. J Exp Biol 203: 3445−3457

Garrabou J, Ballesteros E, Zabala M (2002) Structure anddynamics of north-western Mediterranean rocky benthic

communities along a depth gradient. Estuar Coast ShelfSci 55:493–508

Gili JM, Coma R (1998) Benthic suspension feeders: theirparamount role in littoral marine food webs. Trends EcolEvol 13: 316−321

Gili JM, Ros JD (1985) Study and cartography of benthiccommunities of Medes Islands (NE Spain). PSZN I: MarEcol 6: 219−238

Gollety C, Riera P, Davoult D (2010) Complexity of thefood web structure of the Ascophyllum nodosum zoneevidenced by a δ15N and δ13C study. J Sea Res 64: 304−312

Gori A, Viladrich N, Gili JM, Kotta M, Cucio C, Magni L,Rossi S (2012) Reproductive cycle and trophic ecology indeep versus shallow populations of the Mediterraneangorgonian Eunicella singularis (Cap de Creus, north-western Mediterranean Sea). Coral Reefs 31: 823−837

Heikoop JM, Hickmott DD, Risk MJ, Shearer CK, AtudoreiV (2002) Potential climate signals from the deep-sea gor-gonian coral Primnoa resedaeformis. Hydrobiologia 471: 117−124

Houlbrèque F, Rodolfo-Metalpa R, Jeffree R, Oberhänsli F,Teyssié J.L., Boisson F., Al-Trabeen K., Ferrier-PagèsC.(2011) Effect of increased pCO2 on zinc uptake andcalcification in the tropical coral Stylophora pistillata.Coral Reefs 31: 101−109

Hughes AD, Grottoli AG, Pease TK, Matsui Y (2010) Acqui-sition and assimilation of carbon in non-bleached andbleached corals. Mar Ecol Prog Ser 420: 91−101

Lucas JM, Knapp LW (1997) A physiological evaluation ofcarbon sources for calcification in the octocoral Lepto -gorgia virgulata (Lamarck). J Exp Biol 200: 2653−2662

Mills MM, Sebens KP (2004) Ingestion and assimilation ofnitrogen from benthic sediments by three species ofcoral. Mar Biol 145: 1097−1106

Minagawa M, Wada E (1984) Stepwise enrichment of 15Nalong food chain: further evidence and the relationbetween δ15N and animal age. Geochim CosmochimActa 48: 1135−1140

Orejas C, Gili JM, Lopez-Gonzalez PJ, Arntz WE (2001)Feeding strategies and diet composition of four Antarcticcnidarian species. Polar Biol 24: 620−627

Phillips DL, Gregg JW (2003) Source partitioning using sta-ble isotopes: coping with too many sources. Oecologia136: 261−269

Rau GH, Teyssie JL, Rassoulzdegan F, Fowler SW (1990)13C/12C and 15N/14N variations among size-fractionedmarine particles: implications for their origin and trophicrelationship. Mar Ecol Prog Ser 59: 33−38

Reynaud S, Ferrier-Pagès C, Sambrotto R, Juillet-Leclerc A,Jaubert J, Gattuso JP (2002) Effect of feeding on the car-bon and oxygen isotopic composition in the tissues andskeleton of the zooxanthellate coral Stylophora pistillata.Mar Ecol Prog Ser 238: 81−89

Reynaud S, Martinez P, Houlbrèque F, Billy I, Allemand D,Ferrier-Pagès C (2009) Effect of light and feeding on thenitrogen isotopic composition of a zooxanthellate coral: role of nitrogen recycling. Mar Ecol Prog Ser 392: 103−110

Ribes M, Coma R, Gili JM (1999) Heterogeneous feeding inbenthic suspension feeders: the natural diet and grazingrate of the temperate gorgonian Paramuricea clavata(Cnidaria: Octocorallia) over a year cycle. Mar Ecol ProgSer 183: 125−137

Ribes M, Coma R, Rossi S (2003) Natural feeding of the tem-

187A

utho

r cop

y

Mar Ecol Prog Ser 473: 179–188, 2013

perate asymbiotic octocoral-gorgonian Leptogorgia sar-mentosa (Cnidaria: Octocorallia). Mar Ecol Prog Ser 254: 141−150

Riera P (1998) δ15N of organic matter sources and benthicinvertebrates along an estuarine gradient in Marennes-Oleron bay (France): implications for the study of trophicstructure. Mar Ecol Prog Ser 166: 143−150

Riera P (2007) Trophic subsidies of Crassostrea gigas,Mytilus edulis, and Crepidula fornicata in the Bay ofMont Saint Michel (France): a δ15N and δ13C investiga-tion. Estuar Coast Shelf Sci 72: 33−41

Riera P, Richard P, Grémare A, Blanchard G (1996) Foodsource of intertidal nematodes in the Bay of Marennes-Oleron (France), as determined by dual stable isotopeanalysis. Mar Ecol Prog Ser 142:303–309

Riera P, Stahl LJ, Nieuwenhuize J, Richard P, Blanchard G,Gentil F (1999) Determination of food sources for benthicinvertebrates in a salt marsh (Aiguillon Bay, France) bybenthic carbon and nitrogen stable isotopes: importanceof locally produced sources. Mar Ecol Prog Ser 187: 301−307

Riera P, Escaravage C, Leroux C (2009) Trophic ecology ofthe rocky shore community associated with the Asco-phyllum nodosum zone (Roscoff, France): a δ15N and δ13Cinvestigation. Estuar Coast Shelf Sci 81: 143−148

Risk MJ, Lapointe BE, Sherwood OA, Bedford BJ (2009) Theuse of δ15N in assessing sewage stress on coral reefs. MarPollut Bull 58: 793−802

Roche J, André S, Salvatore G (1960) Métabolisme de l’iode(131I) et formation de la scléroprotéine iodée (gorgonine)du squelette corné chez Eunicella verrucosa (Pallas).Comp Biochem Physiol 1: 286−288

Rossi S, Ribes M, Coma R, Gili JM (2004) Temporal variabil-ity in zooplankton prey capture of the passive suspensionfeeder Leptogorgia sarmentosa (Cnidaria: Octocorallia),

a case study. Mar Biol 144: 89−99Rossi S, Gili JM, Coma R, Linares C, Gori A, Vert N (2006)

Temporal variation in protein, carbohydrate, and lipidconcentrations in Paramuricea clavata (Anthozoa, Octo-corallia): evidence for summer−autumn feeding con-straints. Mar Biol 149: 643−651

Sherwood OA (2006) Deep-sea octocorals: dating methods,stable isotopic composition, and proxy records of theslopewaters off Nova Scotia. Dalhousie University, Halifax

Sherwood OA, Lapointe BE, Risk BE, Jamieson MJ, RobynRE (2010) Nitrogen isotopic records of terrestrial pollu-tion encoded in Floridian and Bahamian gorgoniancorals. Environ Sci Technol 44: 874−880

Sherwood OA, Lehmann MF, Schubert CJ, Scott DB,McCarthy MD (2011) Nutrient regime shift in the west-ern North Atlantic indicated by compound-specific δ15Nof deep-sea gorgonian corals. Proc Natl Acad Sci USA108: 1011−1015

Swart PK (1983) Carbon and oxygen isotope fractionation inscleractinian corals: a review. Earth Sci Rev 19: 5l−80

Vander Zanden MJ, Rasmussen JB (2001) Variation in δ15Nand δ13C trophic fractionation: implications for aquaticfood web studies. Limnol Oceanogr 46: 2061−2066

Ward-Paige CA, Risk MJ, Sherwood OA (2005) Reconstruc-tion of nitrogen sources on coral reefs: δ15N and δ13C ingorgonians from Florida Reef Tract. Mar Ecol Prog Ser296: 155−163

Weinberg S (1976) Revision of the common Octocorallia ofthe Mediterranean circalittoral. I. Gorgonacea. Beau -fortia 24: 63−104

Wild C, Huettel M, Klueter A, Kremb SG, Rasheed MYM,Irgensen BBJ (2004) Coral mucus functions as an energycarrier and particle trap in the reef ecosystem. Nature428: 66−70

188

Editorial responsibility: Christine Paetzold, Oldendorf/Luhe, Germany

Submitted: June 11, 2012; Accepted: September 7, 2012Proofs received from author(s): January 3, 2013

Aut

hor c

opy