J Chem Ecol DOI 10.1007/s10886-015-0610-x
Chemoreception of the Seagrass Posidonia Oceanica by Benthic Invertebrates is Altered by Seawater Acidification Valerio Zupo 1 & Chingoileima Maibam 1 & Maria Cristina Buia 1 & Maria Cristina Gambi 1 & Francesco Paolo Patti 1 & Maria Beatrice Scipione 1 & Maurizio Lorenti 1 & Patrick Fink 2
Received: 9 December 2014 / Revised: 27 April 2015 / Accepted: 22 June 2015 # Springer Science+Business Media New York 2015
Abstract Several plants and invertebrates interact and communicate by means of volatile organic compounds (VOCs). These compounds may play the role of infochemicals, being able to carry complex information to selected species, thus mediating inter- or intra-specific communications. Volatile organic compounds derived from the wounding of marine diatoms, for example, carry information for several benthic and planktonic invertebrates. Although the ecological importance of VOCs has been demonstrated, both in terrestrial plants and in marine microalgae, their role as infochemicals has not been demonstrated in seagrasses. In addition, benthic communities, even the most complex and resilient, as those associated to seagrass meadows, are affected by ocean acidification at various levels. Therefore, the acidification of oceans could produce interference in the way seagrass-associated invertebrates recognize and choose their specific environments. We simulated the wounding of Posidonia oceanica leaves collected at two sites (a control site at normal pH, and a naturally acidified site) off the Island of Ischia (Gulf of Naples, Italy). We extracted the VOCs and tested a set of 13 species of associated invertebrates for their specific chemotactic responses in order to determine if: a) seagrasses produce VOCs playing the role of infochemicals, and b) their effects can be altered by seawater pH. Our results indicate that several invertebrates recognize the odor of wounded P. oceanica leaves, especially those strictly associated to the leaf stratum of the seagrass. Their
* Valerio Zupo
[email protected] 1
Stazione Zoologica Anton Dohrn, Center Villa Dohrn - Benthic Ecology, Punta San Pietro, 80077 Ischia, Italy
2
Cologne Biocenter, Department of Aquatic Chemical Ecology, University of Cologne, Zülpicher Straße 47b, 50674 Köln, Germany
chemotactic reactions may be modulated by the seawater pH, thus impairing the chemical communications in seagrass-associated communities in acidified conditions. In fact, 54 % of the tested species exhibited a changed behavioral response in acidified waters (pH 7.7). Furthermore, the differences observed in the abundance of invertebrates, in natural vs. acidified field conditions, are in agreement with these behavioral changes. Therefore, leaf-produced infochemicals may influence the structure of P. oceanica epifaunal communities, and their effects can be regulated by seawater acidification. Keywords Acidification . Posidonia oceanica . Wound-activated . VOC . Invertebrate . Seagrass . Odor . Infochemical
Introduction The acidification of oceans, due to increasing levels of CO2 in the atmosphere and surface oceans (Brewer 2013), may interfere with the lives of various organisms and communities, due both to chemical influences imposing physiological adaptations and to changed relationships among organisms, affecting their communications and coexistence (Fabricius et al. 2014; Kroeker et al. 2011; Wyatt et al. 2014). Most seagrasses have been demonstrated to be able to thrive at high levels of CO2 (Apostolaki et al. 2014; Garrard et al. 2014), and to survive in areas constantly characterized by low pH (Gartner et al. 2013). However, their epiphytic communities are dramatically influenced by seawater at low pH, since various calcareous algae may be selected against, even by slight pH changes (Donnarumma et al. 2014; Martin et al. 2008). The invertebrate community of the leaf stratum is associated strictly to the epiphytic community (Lebreton et al. 2009), both for shelter
J Chem Ecol
and food availability (Mazzella et al. 1992). Therefore, a slight lowering of pH, still insufficient to influence the growth of a seagrass, may have dramatic effects on the animal assemblages normally present, leading to a simplified trophic structure (Kroeker et al. 2011). In this case, the species diversity and the abundance of invertebrates may remain almost unchanged (Garrard et al 2014). We also know that associated invertebrates recognize the seagrass meadows, where they find food, shelter, and protection from predators, due to visual and chemical stimuli (Zupo and Nelson 1999). It is further known that seagrass tissue composition (e.g., the abundance of phenolic compounds) changes according to the ambient pH, and that lower amounts of deterrent compounds are produced by P. oceanica living in an acidified environment (Arnold et al. 2012; Garrard and Beaumont 2014) as well as by other seagrasses (Campbell and Fourqurean 2013). Therefore, the tissues of seagrasses living in an acidified environment may play different functional and structural roles (Jernakoff and Nielsen 1998) compared to plants growing in normal conditions (Mazzella et al. 1992), although this aspect has not been investigated. This effect is added to the direct influence of changes in the associated epiphytic cover that reduces the shelter for animals and the complexity of food webs. These changes influence as well the nutritional value and the C/N ratios of epiphytes (Ricevuto et al. 2015). The ability of algae to produce volatile organic compounds (VOCs) that trigger specific reactions in some invertebrates has been demonstrated recently (Maibam et al. 2014). Individual species of invertebrates are attracted by the odor of specific microalgae because they may represent feasible food, or be deterred by the odor of other algae because their wounding may indicate the presence of predators (Jüttner et al. 2010). Volatile organic compounds produced by wounded tissues are different from and have a different role, in respect to constitutive metabolites present in the tissues of various organisms (Thoms and Schupp 2008). In fact, constitutive compounds often are confined in the tissues, not diffusing into the environment, when the plant cells are intact. In the case of constitutive emissions, the diffusion is not linked to specific events generated by plant-animal relationships (Grote et al. 2013; Monson et al.2012; Niinemets et al. 2013), as is the case of wound-activated compounds. In contrast, the wounding of plants represents a dynamic event in the frame of plant-animal relationships, and its recognition may be crucial for some invertebrates to identify food sources or the presence of possible predators (Dicke and Sabelis 1988). Thus, several wound-activated VOCs play the role of infochemicals (Maibam et al. 2015), and they may trigger specific reactions in selected invertebrates (Fink 2007). In addition, the recognition of specific bouquets of odors depends on the ecological relationships of invertebrates with their environment (Maibam
et al. 2014), and the perceptive abilities of invertebrates are inversely correlated to the toxicity of wound-activated compounds (WACs). Volatile organic compounds are quickly delivered to target organisms, even at a long distance from the source of the Bodor^ (Kaasik et al. 2011; Lewis et al. 2012), and they often play the role of infochemicals. Chemoreception in terrestrial and aquatic complex ecosystems requires the detection of small differences in mixture composition, as opposed to the recognition of a few specialized compounds (Horner et al. 2006). Consequently, invertebrates exclusively or frequently associated to a given seagrass meadow should be able to recognize its Bodor^, since they have evolved specific chemoreceptive abilities (Jüttner et al. 2010) towards a bouquet of various VOCs produced upon wounding of the plant tissues. In fact, the wounding activity may indicate the presence of consumers or predators (Pohnert et al. 2007), requiring an active reaction by the target organisms. However, we still need investigations to establish whether invertebrates are able to recognize (Pohnert et al. 2007) the odor of a wounded seagrass, and if specific reactions may be induced by higher plant infochemicals. As well, it is still unknown whether seagrass tissue modifications influenced by the acidification of oceans may trigger changes in the reactions of associated invertebrates. It is important to stress that the behaviors triggered by chemical cues are not only due to the invertebrate’s search for food (Fink 2007), since only a few species directly graze on Posidonia leaves (Mazzella et al. 1992). Indeed, the infochemicals produced by the wounding of leaves can indicate the occurrence of larger organisms (e.g., fishes), and may represent critical signals for the presence of predators (Dolecal and Long 2014). To understand these fundamental aspects of their complex ecology, we collected Posidonia oceanica leaves at two different sites, one at normal pH (8.1) and the second at low pH (7.7), close to the ocean acidification scenario and conditions forecasted for the end of the present century (Caldeira and Wickett 2005; IPCC 2007). Subsequently, we exposed a set of 13 benthic invertebrates to the VOCs produced, after wounding, by the two types of Posidonia leaves, under normal pH conditions (8.1) and in acidified water (7.7), to understand if they recognized them by interpreting their chemokinetic reactions (Horner et al. 2008). In this way, we expected to detect any variation in respect to Bnormal^ behavior (Weissburg and Zimmer-Faust 1991) due to either i) the different composition of P. oceanica tissue grown in normal pH (8.1) and acidified (7.7) environments, and ii) any chemical modification of the VOCs due to the pH of the medium (8.1 vs. 7.7) or a modification of the invertebrate’s receptors (Wyatt et al. 2014). The research aimed at checking whether: 1) invertebrates recognize the VOCs produced by wounded Posidonia leaves,
J Chem Ecol
2) the intensity of attraction/repulsion exhibited is according to the degree of association of each species to the leaves of P. oceanica, and 3) the responses of invertebrates may explain their abundances in P. oceanica either in Bnormal^ or in acidified waters.
Methods and Material Study Sites Posidonia oceanica leaves were collected at Castello Aragonese, located in the island of Ischia (Gulf of Naples, Tyrrhenian Sea, Italy). It is represented by a volcanic islet located off the northeast coast of the island (40°43.853′N, 13°57.698′E) and is characterized by the presence of volcanic CO2 emissions (Tedesco 1996). Due to the gas emissions, the pH of the seawater drops to exceptionally low values, from ambient levels (ca. 8.1) in the control station, that is far from the venting areas, down to 6.4 in the areas of intense venting (Kroeker et al. 2011). Thus, the site shows a permanent pH gradient, from high venting areas to absence of venting, approx. 200 m long, and occurring both off the north and south sides of the islet. It represents a Bnatural laboratory^ simulating the pH conditions forecasted for the future of oceans (Hall-Spencer et al. 2008). We collected Posidonia leaves in control areas (pH 8.1), and in the acidified area characterized by a pH close to 7.7 (Garrard et al. 2014). Invertebrate specimens were collected mainly at Lacco Ameno (Gambi et al. 1992), located on the north-west coast of Ischia (40°45.432′N, 13°53.135′E), approx. 6 km apart from Castello vents. It hosts a well studied P. oceanica meadow that extends continuously from 1 to 32 m (Buia et al. 1992; Zupo et al. 2006). The pH of the seawater is stable around 8.1–8.2 (Garrard et al. 2014; Ricevuto et al. 2015), in accordance with the average values for the Mediterranean sea. Collection of Posidonia oceanica Collections of P. oceanica leaves were conducted by SCUBA divers at Castello Aragonese, in two different areas, i.e., in the acidified meadow (pH about 7.7), and in the meadow at normal pH (8.1). Five shoots for each location were selected randomly, collected over a surface area of about 20 m2, and immediately transferred to the laboratory. Shoots collected in the acidified area are morphologically different from those living at normal pH, due to a strong grazing activity by herbivorous fish (e.g., Sarpa salpa). Therefore, all leaves collected in the acidified area are short cut, and they lack apical tips (Donnarumma et al. 2014). For our experiments, only the central parts of leaves were considered, in order to avoid comparisons of different portions of the shoots due to the above-mentioned morphological differences. In particular, the lowest 10 cm of intermediate and adult leaves were discarded; the next 20 cm were used for VOC extractions; the remaining tips, when present, were discarded.
Leaves from each shoot were separated, cleaned from epiphytes by means of a steel blade, and then gently cleaned with soft paper to remove any trace of microalgae still present on their surface. Subsequently, the middle sections, as above specified, were cut into small pieces, mixed, and weighed (fresh weight). Three small samples of leaves were wrapped in aluminum foil and dried in an oven (65 °C) until constant weight (dry weight). The fresh weight/dry weight ratio was calculated, to be used for evaluation of the actual biomass tested in our bioassays. The collected leaf portions were frozen immediately (−20 °C) and used for the extraction of VOCs, a few hours prior to the choice tests. Collection of Invertebrates Collections of invertebrates were performed both at Lacco Ameno, at normal pH conditions, and at Castello Argonese, in the sector characterized by normal pH. Our collections were not quantitative (see Garrard et al. 2014, for a quantitative estimation of invertebrate associations present at Castello Aragonese and Gambi et al. 1992 for invertebrates at Lacco Ameno) since they were used only to provide living specimens for choice experiments. A circular plankton net (1 m frame diam; mesh size 100 μm) was gently trawled horizontally by a research boat over the Posidonia oceanica meadow, and invertebrates were collected in a glass jar attached to the end of the net. First sorting of the collected animals was performed on board, and specimens of various taxa were pooled into plastic bags containing clean seawater, and immediately transferred to the laboratory. Additional invertebrates were collected by SCUBA divers, by towing a rectangular framed net over the leaf canopy (Buia et al. 2004). All invertebrate specimens collected were identified by specialists, in vivo, under a stereomicroscope, moved into aerated vessels, and kept in a thermostatic chamber, in the presence of small pieces of Posidonia leaves (for shelter and food) up to the day before the experiment. They were starved for 24 h prior to being tested. A species was considered for choice tests when at least 30 individuals were collected and available. Thirteen species were selected, belonging to the main mesofaunal taxa present in the seagrass environment, i.e., polychaetes, gastropod mollusks, isopods, amphipods, and decapod crustaceans. Specimens assayed at low pH conditions were adapted slowly to acidified water (pH 7.7), starting the night before the experiment. After the experiments in acidified water, each specimen was kept at pH 7.7, in glass vessels, prior to being slowly adapted to normal pH (8.1). All specimens still alive were returned to the sea after the completion of the experimental procedures. Extraction of VOCs Volatile organic compounds (VOCs) were extracted twice from frozen leaves (2×4 g DW) of Posidonia oceanica after grinding them in a mortar under
J Chem Ecol
liquid nitrogen. The same weights of leaves were used for extraction from the normal pH site and the acidified site. Thus, the VOCs incorporated into the agarose blocks were proportional to the original concentration of infochemicals in the wounded leaves. Volatile organic compounds were concentrated by closed-loop stripping (Jüttner 1988) performed at 22 °C for 45 min. For this purpose, 8 g of ground and sonicated leaves were suspended into 40 ml of filtered (Millipore 0.22 μm) seawater and transferred to a 100 ml round bottom flask. After addition of 10 g NaCl, the VOCs were extracted and absorbed onto a Tenax TA cartridge (Fink et al. 2006 a, b). The cartridge was removed and eluted with 6 ml diethyl ether. The ether was gradually evaporated using nitrogen gas (N2, grade 5.0), and the residue was re-dissolved in 300 μl of pure ethanol. Controls were prepared according to the same procedure, but stripping was performed on filtered and sterilized seawater without the addition of leaves. All VOC samples and controls were stored at −80 °C until the choice experiments were conducted. Preparation of Gels Bioassays were conducted as reported below, in static chambers (Jüttner et al. 2010), after the inclusion of VOCs into small blocks of jellified agarose. To this end, we prepared agarose gels added with VOCs and controls. To prepare a 0.06 % agarose gel, 1.2 g of agarose (Sigma A-9045) were dissolved in 200 ml of filtered and sterilized seawater, heated (80 °C), and stirred until completely transparent. The pH of the solution was adjusted to a value close to 8.2 by adding 3.3 ml of 0.1 M NaOH. Controls were prepared by incorporating 250, 25, and 2.5 μl of the above described control extract into liquid (but close to room temperature) agarose, just before gelling, to obtain three concentrations of control gels. The agarose solution then was poured into a Petri dish and allowed to gel in a refrigerator at 5 °C, 1 h prior to the start of assays. To prepare VOC agarose blocks, 250, 25, and 2.5 μl of the ethanolic VOC extract were incorporated, respectively, into the still liquid (but close to room temperature) agarose, just before gelling. In this way, we obtained three different concentrations, namely, Blow^, Bmedium^, and Bhigh^. The low concentration simulates the VOCs released by 5 mm2 of Posidonia leaf wounded by a grazer. The medium concentration corresponds to 50 mm2 of Posidonia leaf wounded by a large grazer. The high concentration corresponds to 500 mm2 of Posidonia leaf wounded by a herbivorous fish or a similar grazer. Finally, the agarose gel disks were cut (using clean glass coverslips) into small blocks, each measuring 0.5 cm3 and used for the choice tests on invertebrates. Choice Tests on Invertebrates Thirteen invertebrate species were selected from our samples and identified. The set is representative of different feeding habits and various levels of associations to Posidonia oceanica leaves (Table 1).
Therefore, we expected that sensitive organisms would orient, according to their preferences, along the VOC gradient (Chase 1982). The assays were conducted in 14 cm (diam) Petri dishes set over circular experimental arenas printed on paper sheets, according to the protocols suggested by Jüttner et al. (2010). Each arena consisted of five sectors, (−2, −1, 0, 1, 2) indicating the rate of repulsion or attraction according to the invertebrate movements (Fig. 1). In fact, these annotations refer to the distance from the positive target, being sector +2, the one containing the agarose added with VOCs and the −2, the one containing the control agarose; B0^ is the central sector, intermediate between the positive target and the negative control. For the experiments in acidified conditions, Millipore (0.45 μm) filtered seawater was added with CO2 using a Ferplast CO2 Energy® reactor, which allows for acidification of seawater by avoiding any bubbling. The water was checked for its pH (7.7) prior to being used for filling the experimental arenas. Five individuals of each species were released at the center (marked as a circle) of each arena, and they were allowed to perceive the odor of the Posidonia leaves, diffusing from the B+2^ target. The number of individuals present in each sector of the arena was recorded at four time intervals (5, 10, 15, and 20 min) from the start of each test. Precautions were taken to minimize any external factor possibly influencing the movement of animals during the experiment, such as light, temperature, magnetism, etc. In particular, experiments were conducted at 18 °C under a well-lit and diffused light, and each of the replicated two arenas were positioned in such a way that the positive targets opposed each other. Six replicates were obtained for each invertebrate species for each treatment (Table 2). Four treatments were considered: i) VOCs from Posidonia collected at ambient pH tested in Petri dishes containing seawater at normal pH (8.1); ii) VOCs from Posidonia collected at pH 7.7 tested in Petri dishes containing seawater at ambient pH (8.1); iii) VOCs from Posidonia collected at ambient pH tested in Petri dishes containing acidified seawater (pH 7.7); iv) VOCs from Posidonia collected at pH 7.7 tested in Petri dishes containing acidified seawater (pH 7.7). For each treatment, three concentrations were tested (low, medium, and high), corresponding to 5, 50, and 500 mm2 of P. oceanica leaves, respectively. In total, 6 replicates of 12 treatments were tested on 13 invertebrate species (Table 2). When some invertebrates had to be tested again under different treatments, they were allowed to rest overnight in a thermostatic chamber (18 °C 12:12 h L/D) prior to start the new experiment. The assemblages of 5 individuals used for each replicate were randomly re-assigned every time. Calibration of Experimental Vessels The choice experiments were performed based on the assumption that VOCs added to an agarose gel may diffuse producing a continuous
J Chem Ecol Table 1 Nr
Thirteen species of macroinvertebrates tested for their responses to volatile compounds produced by three benthic diatoms Taxon
Species
Trophic habits
Assoc.
1
Polychaete
Platynereis dumerilii (Audouin & Milne Edwards, 1834)
herbivore
1
2
Polychaete
Kefersteinia cirrata (Keferstein, 1862)
omnivore/carnivore
1
3 4
Isopod Amphipod
Dynamene bifida Torelli, 1930 Caprella acanthifera Leach, 1814
herbivore omnivore
1 3
5 6
Amphipod Decapod
Gammarella fucicola (Leach, 1814) Hippolyte inermis Leach, 1815
herbivore herbivore/omnivore
1 4
7
Decapod
Cestopagurus timidus (Roux, 1830)
omnivore/carnivore
4
8 9
Decapod Gastropod
Calcinus tubularis (Roux, 1830) Rissoa italiensis Verduin, 1985
omnivore/carnivore herbivore
1 3
10 11
Gastropod Gastropod
Rissoa variabilis (Von Mühlfeldt, 1824) Rissoa violacea Desmarest, 1814
herbivore herbivore
2 2
12
Gastropod
Bittium latreilli (Payraudeau, 1826)
detritus feeder
1
13
Gastropod
Gibbula umbilicaris (Linnaeus, 1758)
herbivore
3
Nr.: reference to the invertebrates as used also for cluster analysis. The level of association to Posidonia oceanica meadows is categorized in the last column (assoc.): 1, present; 2, typical; 3, generally abundant; 4, almost exclusive
gradient, as suggested by Steinke et al. (2002). Nevertheless, the behavioral response of animals within each test may or may not justify this assumption. In order to confirm the presence of a VOC gradient, we performed a calibration experiment prior to the start of the choice tests, by using a wellknown VOC that is easily quantifiable by means of spectrophotometric analyses. For this purpose, agarose blocks were loaded with the volatile compound 2-trans-4-trans-decadienal (Sigma-Aldrich),
an aldehyde produced by some planktonic diatoms upon wounding (Wichard et al. 2005). In particular, 150 μl of decadienal, previously dissolved in 4 ml of methanol, were added to 200 ml of freshly prepared agarose, in order to obtain a final concentration of 240 μg/ml. The agarose gels were refrigerated and cut into 0.5 cm3 blocks, as reported for the experimental VOCs. The blocks then were positioned in the (+) targets of the experimental arena, while negative blocks (agarose gels) were positioned in the (−) targets. Four sets of 3 replicated arenas were prepared contemporaneously. The vessels were positioned on a squared matrix, containing the indication of 30 sampling points. Every 5 min, a set of 3 replicates was sampled by means of an automatic pipette, and 1 ml of the medium was collected, from each sampling point on the grid. These collections were repeated at 5, 10, 15, and 20 min. The collected solution then was spectrophotometrically analyzed (Hewlett Packard 8453 spectrophotometer) at a wavelength of 282 nm, and the concentration of decadienal was calculated Table 2
Experimental plan
pH of P. oceanica collection site
Treatments (performed at 3 concentrations)
7.7
pH 8.1 (A – C) pH 8.1 (C - C)
8.1
Fig. 1 Experimental arenas used for choice experiments. Animals were deployed in the central circle at the start of the experiment. The VOCadded block of agarose was fixed into the B+^ square. The control agarose was fixed into the B−^ square. Test invertebrates were counted every 5 min into the sectors, according to their ranking, from −2 to +2
pH 7.7 (A - A) pH 7.7 (C - A)
Posidonia oceanica leaves were collected at two stations at Castello Aragonese, characterized by different pH (first column), and choice experiments (treatments) were conducted, for both collections, in experimental vessels filled with seawater at two different pH (last two columns). Each treatment was repeated at three concentrations of VOCs (low, medium and high) A acidified, C control
J Chem Ecol
(Pippen and Nonaka 1958) according to its molar epsilon 31, 000 m−1 cm−1 of optical density (OD) referred to a lambda max of 274 nm in methanol. All concentrations obtained were recorded into a matrix indicating the position of samples, and the average of three replicated samples was calculated. These values were computed using the Kriging technique (Matheron 1969, 1970) that allows a spatial representation of the concentration, considered as a stationary phenomenon. In this way, we obtained a map of the concentrations measured in each point of the experimental arenas, described by isolines, suitable to confirm the exactness of our hypotheses and to track the diffusion gradient from the agarose gels in the Petri dishes during the experiment. Statistical Analyses To compare the reactions of invertebrates towards the odor of Posidonia leaves collected at two different sites (normal and low pH), under two experimental conditions (normal pH and acidified water), we calculated and plotted the average Preference Index (P.I., as proposed by Jüttner et al. 2010) exhibited by each tested invertebrate and the standard error. The latter was chosen according to previous investigations (James et al. 2008) because the individual variability characterizing ethological responses is well known, and we aimed at showing the average level of preference of each species, not the obvious scattering of results due to the natural random invertebrate movements. The significance of the results provided by each choice test was evaluated by means of the Z-test on proportions, by comparing the proportion of individuals present in the B+^ sectors at the end of the experiment to the proportion present in the B−^ sectors. The test indicates whether the observed distribution is significantly different from a normal distribution of individuals (Sprinthall 2011). In addition, a matrix Btreatments vs. arena sectors^ was filled and submitted to cluster analysis (by means of STATISTICA 10, StatSoft) to allow grouping of the species tested according to their patterns of reactions in the various experimental trials.
Results Calibration of the experimental vessels confirmed the production of a VOC gradient initially concentrated into the B+2^ sector (Fig. 2a), slowly diffusing towards the B−2^ sector (Fig. 2b, c) according to the time of sampling. After 20 min (Fig. 2d), the gradient was still present, but the differences between the two extremes of the experimental vessel were reduced, as compared to the start of the assay. The choice tests yielded complex patterns of reactions, variable in each invertebrate species according to the concentration, the type of Posidonia tissues, and the pH of the medium.
Posidonia oceanica leaves sampled at normal pH prompted some evident and significant reactions, indicating that several invertebrates recognize their odor (Fig. 3). In particular, Platynereis dumerilii, Kefersteinia cirrata, Dynamene bifida, and Gibbula umbilicaris exhibited a significant reaction of attraction towards the odor of P. oceanica at low concentration, while Rissoa italiensis and other gastropods showed slight repulsion (Fig. 3a). At the medium concentration, the leaves of P. oceanica produced more repulsion, and, limiting our examination to significant results, we observed (Fig. 3b) that Gammarella fucicola, Cestopagurus timidus, Calcinus tubularis, and Bittium latreilli were repelled by P. oceanica VOCs, while Rissoa violacea was attracted. The choice tests conducted at high concentration yielded the largest number of significant reactions, but some invertebrates exhibited contrasting reactions in respect to those reported for the low concentration. Besides polychaetes, isopods and amphipods, which elicited a positive chemotactic reaction at all concentrations, mollusks and decapods showed a repulsive reaction (significant in the case of G. umbilicaris, B. latreilli, R. italiensis, C. tubularis, C. timidus, and H. inermis) at different concentrations (Fig. 3b, c). In the case of B. latreilli, the repulsive effects of P. oceanica VOCs were proportional to the concentration: apparent low repulsion (P.I. −0.33, but not significant) at low concentration, higher repulsive reaction (P.I. −0.66, significant) at the medium concentration, and the highest repulsive reaction (P.I. −1.2, significant as well), at high concentration. Similar relationships were observed for species attracted by P. oceanica VOCs, although the results obtained at medium and high concentration are generally in better agreement. For example, in the case of Kefersteinia cirrata, attraction was demonstrated by a preference index of 1.16, 0.62, 0.81, respectively, at low, medium, and high concentration, exhibiting a decrease at medium and high concentrations in respect to the low concentration. The same trends were shown by other polychaetes, amphipods, isopods, and decapods. In the case of H. inermis, for example, low concentration of VOCs triggered slight attraction (Fig. 3a), medium concentration triggered a lesser attraction (Fig. 3b), and high concentration triggered an escape reaction (Fig. 3c). Consistently, a decreasing trend of attraction was observed moving from low towards high concentrations, when P. oceanica leaves collected at normal pH were tested in a medium at pH 7.7, although some invertebrates reversed their reactions (Fig. 4). In fact, most invertebrates exhibited a significant positive reaction at high concentration (Fig. 4c): the responses of polychaetes, amphipods, isopods, decapods, and most mollusks (with the exception of Rissoa violacea) were in agreement with the results obtained for the low concentration tested at normal pH (Fig. 3a). In contrast, low and medium concentrations of VOCs triggered less obvious reactions (Fig. 4a, b).
J Chem Ecol
Fig. 2 Calibration of the experimental vessels. The diffusion of 2-trans-4-trans-decadienal is indicated by isolines as calculated by the Kriging technique, at various sampling times. a, 5 min; b, 10 min; c, 15 min; d, 20 min
A quite different pattern of chemotactic reactions was prompted by VOCs extracted from P. oceanica leaves collected in acidified conditions (pH 7.7) and tested at the same pH (Fig. 5a). Most species exhibited a low-level response at low concentration, and only Gammarella fucicola and Hippolyte inermis were significantly attracted by the VOCs, while Cestopagurus timidus was significantly repelled. A larger number of invertebrates exhibited a significant chemotactic response at Bmedium^ concentration of VOCs (Fig. 5b). In fact, the two amphipods both exhibited a low but significant repulsion, along with the mollusk Rissoa variabilis. Hippolyte inermis and Calcinus tubularis showed a clear positive chemotaxis. The high concentration disoriented most invertebrates: only three significant responses were exhibited (Fig. 5c) by Cestopagurus timidus, Calcinus tubularis (repulsion), and Rissoa italiensis (attraction). The same tests, performed in normal seawater, yielded a complex pattern of results. Three species (Gammarella fucicola, Cestopagurus timidus, and Gibbula umbilicaris)
showed significant attraction, while Rissoa violacea showed rejection at low concentration (Fig. 6a). Consistently, a larger number of species exhibited significant reactions at medium concentration (Fig. 6b). The two polychaetes showed attraction along with Gammarella fucicola, Calcinus tubularis, and Rissoa italiensis. Cestopagurus timidus, in contrast, was repelled by these VOCs tested at normal pH. Only four species reacted at high VOC concentrations (Fig. 6c). Hippolyte inermis, Cestopagurus timidus, and Rissoa variabilis were attracted by acidified Posidonia leaves, while Calcinus tubularis was repelled in these conditions. The chemotactic responses observed are, overall, in accordance with the distribution of the considered species at ambient pH and in the acidified area (Table 3). The cluster analysis performed on the matrix of preference indices shows two main groups of invertebrates, separated based upon their specific reactions. The first group contains P. dumerilii, C. tubularis, G. fucicola, K. cirrata, and D. bifida (Fig. 7). These species are all characterized by a low level of
Fig. 3 Results of the choice experiments performed on volatile organic compounds (VOCs) extracted from Posidonia oceanica leaves collected at normal pH and tested in normal seawater (pH 8.1). Bars indicate attraction (positive bars) or repulsion (negative bars) according to the preference index proposed by Jüttner et al. (2010). a. tests performed at
low VOC concentration. b. Tests performed at medium VOC concentration. c. Tests performed at high VOC concentration. * indicates significant results at P0.05). Standard errors (N=6) are indicated by error bars
J Chem Ecol
Fig. 4 Results of the choice experiments performed on volatile organic compounds (VOCs) extracted from Posidonia oceanica leaves collected at normal pH (pH 8.1) and tested in acidified seawater (pH 7.7). Bars indicate attraction (positive bars) or repulsion (negative bars) according to the preference index proposed by Jüttner et al. (2010). a. Tests
performed at low VOC concentration. b. Tests performed at medium VOC concentration. c. Tests performed at high VOC concentration. * indicates significant results at P0.05). Standard errors (N=6) are indicated by error bars
association to P. oceanica (Table 1). The second group contains all the other species, mostly characterized by a higher level of association with P. oceanica (Table 1).
chemotaxis (Fink 2007; Zhou and Rebach 1999). Several superimposed factors affect the behavior of invertebrates and contribute to the results reported above (Briffa et al. 2012). In addition, chemical cues may be influenced by local ecological conditions that affect the quality and quantity of VOCs produced by plants, as well as the perceptive abilities of individual species of invertebrates (Wyatt et al. 2014). Nevertheless, some clear and significant trends were observed and, given the experimental procedures followed, we are able to discriminate the effects attributable to changes in
Discussion Invertebrate Reactions The analysis of invertebrate behavior in response to chemical stimuli is complex, due to individual variability and to stochastic factors that influence the animal
Fig. 5 Results of the choice experiments performed on volatile organic compounds (VOCs) extracted from Posidonia oceanica leaves collected in the acidified site (at pH 7.7) and tested in acidified seawater (pH 7.7). Bars indicate attraction (positive bars) or repulsion (negative bars) according to the preference index proposed by Jüttner et al. (2010). a. Tests performed at low VOC concentration. b. Tests performed at
medium VOC concentration. c. Tests performed at high VOC concentration. * indicates significant results at P
