Feeding in deep-sea demosponges: Influence of abiotic and biotic factors

Author’s Accepted Manuscript Feeding in deep-sea demosponges: influence of abiotic and biotic factors Leah M. Robertson, Jean-François Hamel, Annie Mercier www.elsevier.com

PII: DOI: Reference:

S0967-0637(17)30119-X http://dx.doi.org/10.1016/j.dsr.2017.07.006 DSRI2815

To appear in: Deep-Sea Research Part I Received date: 29 March 2017 Revised date: 27 June 2017 Accepted date: 28 July 2017 Cite this article as: Leah M. Robertson, Jean-François Hamel and Annie Mercier, Feeding in deep-sea demosponges: influence of abiotic and biotic factors, DeepSea Research Part I, http://dx.doi.org/10.1016/j.dsr.2017.07.006 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Feeding in deep-sea demosponges: influence of abiotic and biotic factors Leah M. Robertsona,b, Jean-François Hamelc, Annie Merciera* a

Department of Ocean Sciences, Memorial University, St. John’s, NL Canada b

c

Department of Biology, Memorial University, St. John’s, NL, Canada

Society for the Exploration and Valuing of the Environment (SEVE), Portugal Cove-St. Philips, NL, Canada *

Corresponding author: [email protected]

Abstract In shallow benthic communities, sponges are widely recognized for their ability to contribute to food webs by cycling nutrients and mediating carbon fluxes through filter feeding. In comparison, little is known about filter feeding in deep-sea species and how it may be modulated by environmental conditions. Here, a rare opportunity to maintain live healthy deep-sea sponges for an extended period led to a preliminary experimental study of their feeding metrics. This work focused on demosponges collected from the continental slope of eastern Canada at ~1000 m depth. Filtration rates (as clearance of phytoplankton cells) at holding temperature (6°C) were positively correlated with food particle concentration, ranging on average from 18.8 to 160.6 cells ml-1 h-1 at nominal concentrations of 10 000 to 40 000 cells ml-1. Cell clearance was not significantly affected by decreasing seawater temperature, from 6°C to 3°C or 0°C, although two of the sponges showed decreased filtration rates. Low pH (~7.5) and the presence of a predatory sea star markedly depressed or inhibited feeding activity in all sponges tested. While performed under laboratory conditions on a limited number of specimens, this work highlights the possible sensitivity of deep-sea demosponges to various types and

levels of biotic and abiotic factors, inferring a consequent vulnerability to natural and anthropogenic disturbances. Keywords: Porifera; sponge; feeding; filtration; deep-sea; cold-water

1 Introduction Sponges play a unique role in marine communities, due to their ability to process and cycle dissolved and particulate compounds through their metabolic processes and via the microbial populations they host (Maldonado et al., 2012). They use a mechanism called collar sieving under the filtering action of flagellated choanocytes to capture food particles down to ∼0.1 μm diameter (Riisgård and Larsen, 2010). The contribution of sponges has been particularly well explored in shallow-water habitats from temperate and tropical areas, including the importance of the “sponge loop” as a major regulator of dissolved organic matter in oligotrophic environments such as coral reefs (De Goeij et al., 2013). While several studies have examined the feeding biology of demosponges from shallow-water habitats (e.g. Reiswig, 1971a, b, 1975; Riisgård et al., 1993), the feeding ecology and contribution of deep-water counterparts to food webs below 200 m remain virtually unexplored (Witte et al., 1997; Yahel et al., 2007), although they are ascribed a wide spectrum of feeding strategies, from filter feeding to carnivory (Chevaldonné et al., 2015). Sponges are known to dominate some deep-sea benthic communities and constitute important biodiversity hotspots (Cathalot et al., 2015; Kahn et al., 2015). However, on account of challenges associated with the collection of live sponges from the deep sea, experimental laboratory studies on these sponges remain scarce. A few

2

species of deep-water/cold-water glass sponges have been studied where they occur at more accessible depths, i.e. above 200 m in the Northeast Pacific (Chu and Reiswig, 2014; Kahn et al., 2015; Leys et al., 2011; Yahel et al., 2006; 2007). It has been estimated that glass sponge reefs can remove 90-99% of the most abundant bacteria and eukaryotic algae from the water (Kahn et al., 2015; Yahel et al., 2006; 2007). Work with the deepwater sponges Thenea abyssorum and T. muricata has revealed their size-specific food preferences for ultraplankton (Witte et al., 1997). A recent study by Kutti et al. (2015) examined the metabolic response of sponges collected at 200 m to suspended materials in the laboratory, and highlighted mechanisms deployed to resist sediment stress. Yet, how changes in biotic and environmental variables might impact the feeding of sponges from deep-sea environments, with implications for other life functions, has not previously been explored. In the deep sea, conditions are often stable, with minimal temperature fluctuations, yet periodicities in phytodetritus influx have been documented, which may modulate growth (Leys and Lauzon, 1998) and biogeographic patterns (Ruhl et al., 2008). Furthermore, temperature can still be a factor in some regions of the deep sea. The Northwest Atlantic, off eastern Canada, is under the influence of the Labrador Current and the North Atlantic Oscillation (DFO, 2009). Temperatures at depths ~500 m can range from 0°C in winter months to 5°C in late summer (DFO, 2009), including in regions where deep-sea sponges are common on hard rock substrate (or seafloor), either occurring sparsely or as dense beds (Beazley et al., 2013; J.-F. Hamel, personal observation). It can thus be assumed that thermotolerance could play a role in the adaptability of sponges on the eastern Canadian continental slope. Temperature has

3

already been shown to affect the growth and reproduction of several deep-sea species from the same geographic area, including the octocoral Drifa glomerata (Sun et al., 2010) and the scleractian coral Flabellum alabastrum (Hamel et al., 2010). Lesser and Slattery (2013) reported that predation is one of the primary determinant of sponge community structure. Accordingly, sponges have developed defenses, which have been shown to ward off fish and other potential predators (Sheild and Witman, 1993; Slattery et al., 2016). These defenses may rely on physical components, such as morphological traits and the presence of spicules, while others may revolve around chemical deterrents (Hill et al., 2005) or an ability to regenerate lost tissues rapidly (Sheild and Witman, 1993). When shallow-water sponges are in the presence of a predator, their feeding is altered, generally as a result of energy expenditure being focused on chemical defenses, metabolites or other forms of predator deterrence (Haber et al., 2011). In deep-sea environments, sponge predators include sea stars, such as Henricia lisa (Gale et al., 2013). However, it is currently unclear whether the presence of a predator could exert any influence on filter feeding in deep-sea sponges. The impact of climate change and ocean acidification on deep-sea species is not well known (Hofmann et al., 2010); however, due to the general stability of the deep-sea environment, it is believed that deep-sea taxa may be more sensitive to change and will be impacted negatively by ocean acidification (Barry et al., 2014). Because members of Porifera have a poor ability to acid-base regulate, they could be particularly vulnerable to shifts in pH and, more generally, to ocean acidification (Goodwin et al., 2014). Despite preliminary suggestions that sponges may benefit from reduced competition from corals following climate shifts (Bell et al., 2013), one of the few direct studies of their responses

4

to ocean acidification highlighted the greater susceptibility of heterotrophic species (Bennett et al., 2017), such as those found in the deep sea. The present work focused on deep-sea demosponges belonging to the Polymastiidae family. Approximately 63% of the polymastiid sponges belong to the Polymastia genus (Plotkin et al., 2012), leading to a vast distribution that ranges from tropical shallow waters to temperate and deep-sea areas (Boury-Esnault et al., 1994). Two species of this genus have been documented at 1177 m depth on the Flemish Cap, off the Grand Banks of Newfoundland (Murillo et al., 2012). The other species in the present study, Radiella hemisphaerica, is typically found in the North Atlantic Ocean and Norwegian Sea and has also been recorded from the Flemish Cap, down to 1460 m ue et al.

01 .

This study took advantage of a unique opportunity to access live deep-sea sponges to gather baseline data on their feeding ecology. To our knowledge, these are among the rare live deep-sea sponges collected >200 m to have been experimentally studied under laboratory conditions. The objective was first to assess if and how their filtration rates (measured as clearance of phytoplankton cells) would correlate with phytoplankton concentrations, and then to examine how filtration metrics varied under different conditions of temperature and pH, and in the presence of a predator. The findings will hopefully provide insightful information on the biology and resilience of deep-sea sponges, which are recognized as essential habitat-forming components of deep-water ecosystems that need to be managed and preserved (Campbell and Simms, 2009; NOAA, 2010).

5

2 Materials and methods 2.1 Collection and maintenance

Sponges were collected in December 2013 as by-catch during a routine multispecies research survey conducted on the CCGS Teleost by the federal Department of Fisheries and Oceans (DFO). Intact live specimens were collected at approximately 1000 m depth on the continental slope of Northeast Newfoundland, Canada (around 54o 59’ 1 ” N: 50o 46’ 0” W . Based on morphology, spicules and known occurrences, two sponges were identified as Radiella hemisphaerica (hereafter referred to as Radiella 1 and Radiella 2) and two others determined to belong to be the same species in the genus Polymastia (dubbed Polymastia 1 and Polymastia 2). Collections took place one year prior to the study, and specimens used for this work were visually healthy, with none showing signs of tissue damage or necrosis. The sponges were kept in a 20-L tank in a darkened flow-through system of unfiltered seawater (~72 L h-1) at ambient temperatures that fluctuated seasonally from ~1°C to 6°C. This range is consistent with temperatures occurring in the region where the sponges were collected (DFO, 2009). 2.2 Experimental design

Three weeks before the first trial in October 2014 (10 months post collection), the sponges were transferred from the main holding tank into separate experimental vessels, which consisted of glass beakers. A volume of 1000 ml was used for the largest of the two Polymastia (Polymastia 1) and 250 ml for all the other smaller specimens, maintaining biomass in the range of 0.05-0.15 g ml-1. Prior to and between experiments, the beakers with the sponges remained fully submerged into a larger holding tank (20 L)

6

submitted to flow-through conditions and natural input of dissolved and particulate organic matter (as mentioned for holding conditions above). Great care was taken to ensure the sponges remained constantly immersed at all time, received minimal light exposure during manipulations, and were given time to acclimate to the various experimental conditions. Height (measured from base to highest point, excluding papillae), width (the largest diameter of the base, excluding crest), as well as crest and papillae length (when present) were measured while the sponges were fully immersed. Volumetric water displacement and total wet weight of each sponge were only determined after the completion of the study (Table 1); these measurements required emersion, which could have impacted the health and responsiveness of the sponges. 2.3 Standardized experimental procedures

Importantly, all sponges had been adapted to laboratory flow-through conditions for several months. Before each trial, the water flow was interrupted and detritic particles present in the vessels around the sponges were siphoned gently, without touching the sponges. Vessels were slightly raised so that their rims emerged above the water line of the holding tank, enabling static conditions to be maintained for the 2-h trial duration. All trials (n=3 replicates per condition) were conducted at the same time of day (early afternoon, with 48 h between replicate runs). Preliminary experiments had confirmed that phytoplankton cells remained suspended for at least 2 h (without any aeration required) as per Peterson et al. (2006). The food was prepared at the beginning of each trial using commercial phytoplankton (Phyto-Feast® Live) diluted with seawater to the desired nominal concentration (see below for the description of the various concentrations used). 7

This mixture includes six species: Pavlova sp., Isochrysis sp., Thalassiosira weissflogii, Tetraselmis sp., Nannochlorpsis sp., and Synechococcus sp. with cell sizes ranging from 1 to 15 µm in diameter. Thalassiosira and Synechococus are two phytoplankton genera commonly found in North Atlantic waters (Johns et al., 2003; Li, 2002). Preliminary tests showed acceptance of this food by all sponges under study. Moreover, particles of <50 µm in size were determined to be ideal for sponge filtering (Duckworth et al., 2003). The cell sizes considered for this experiment were between 1-3 µm. While larger particles were found in the mixture, the smaller phytoplankton cells were the closest to the size of ultraplankton that the sponges would be feeding on at >1000 m (Pile and Young, 2006). Once the phytoplankton had been added and gently mixed (time 0), three 1-ml aliquot of seawater was removed from each experimental vessel to confirm the effective concentration around each sponge (see below for various concentrations tested). Controls with only phytoplankton cells and treatment conditions (no sponge) were used to determine cell loss outside the standard deviation, i.e. background loss, which was accounted for when calculating cell clearance from the difference between initial and final phytoplankton concentrations. At the end of the 2-h trial, three 1-ml aliquot of water were sampled from each vessel and the number of phytoplankton cells was determined by counting cells as previously described. Given the fragility and uniqueness of the study species, cell count was chosen over more intrusive methods, e.g. the use of latex beads to determine feeding flow rate (Francis and Poirrier, 1986). From phytoplankton cell clearance, filtration rates were calculated using the equation outlined in Peterson et al. (2006):

8

Filtration rate (ml h-1) = (Vsw × t-1) × ln(Co× C2-1)

Where Vsw= volume of seawater in the container, t = time feeding, C0 = cells ml-1 of at time 0-h, and C2= cells ml-1 at the end of the trial. Mean filtration rates were calculated using the averaged cell concentrations at time 0-h and 2-h. Any phytoplankton cell loss determined in the control was subtracted to obtain true average filtration rate. 2.4 Effect of phytoplankton concentration, temperature, predator and pH

Feeding trials were first conducted at ambient temperature (6 ± 0.8°C) and pH (7.95 ±0.03) corresponding to conditions occurring in the flow-through holding tanks at the onset of the study. Filtrations rates were tested under three nominal concentrations of phytoplankton: 10 000, 20 000, and 40 000 cells ml-1. Replication and sampling followed the designs described earlier. Two additional temperatures (0°C and 3°C) were tested using the intermediate phytoplankton concentration (20 000 cells ml-1). This concentration was used as it yielded high filtration rates in the first series of experiments. The range of temperatures tested represent non-lethal limits for the sponges, consistent with natural fluctuations in the deep-water environment of the Northwest Atlantic, where averages recorded fluctuate around 1-3°C in the fall months (DFO, 2009). The desired temperature was maintained constant using an electronically-controlled Boekel Grant GD100 water bath. To minimize shock and allow acclimation, the water temperature in the vessels was slowly equilibrated to the experimental temperature over a 1-h period. The range of temperature tested and the acclimation allowed were chosen to parallel natural conditions that the species are 9

likely exposed to in nature, i.e. shifts of 1-6 oC inside a few hours (DFO 2009). Replication and sampling followed the designs described earlier. The predator trials involved the addition of a predatory sea star to the vessels holding sponges. They were conducted at 3°C and a phytoplankton concentration of 20 000 cells ml-1. The predator used was the deep-sea asteroid Henricia lisa which has already been described as a sponge feeder (Gale et al., 2013). For this study, one individual of H. lisa was placed in a perforated tube (to avoid physical contact and direct predation), which was then inserted in the vessel above the sponge (one predator for each sponge). A 1-h acclimation period was allowed before the feeding trial to acclimate the sponges to the chemical signature of the predator, as per previously published preypredator interaction in marine invertebrates, e.g. (Bullock, 1953). Replication and sampling followed the designs described earlier. If a response to the treatment was determined (i.e. decreased feeding rates), a second set of trials was devised to ensure that the presence of H. lisa rather than presence of the tube was the cause of the response. Triplicate trials of the control were run using the same methods and conditions. At the end of all treatments, predation on the deep-sea sponges by H. lisa was confirmed by placing the sea stars in an open container in the flow-through holding tank (described above) at a distance of 15 cm from the sponges. They were monitored for 3 h. Any contact or attempt to feed on the sponges were documented and scored as a predator-prey relationship as per Gale et al. (2013). The pH trials were conducted at 3°C (± 0.4) with the intermediate phytoplankton concentration of 20 000 cells ml-1 using Radiella. The pH was adjusted to 0.3-0.4 units below ambient level to be consistent with predicted scenarios of ocean acidification by

10

2100 (Gehlen et al., 2014). A CO2 gas cylinder and electronically-controlled solenoid were used to decrease the pH to the desired level by bubbling CO2 into a seawater container (1 L) and the pH of the seawater was monitored using a calibrated AquaticLife™ pH controller. Methods followed previous studies (Verkaik et al., 2016; 2017), which can be consulted for estimates of the carbonate chemistry. The pH within the experimental vessels was decreased gradually over 1.5 h by slowly trickling the freshly prepared low-pH seawater. While temporal constraints cannot allow for climaterelevant variations in pH to be achieved, the acclimation period was chosen to be consistent with the previous trials and the main goal was to assess response to shifts in pH. Moreover, pH fluctuations in the order of 0.1 within a day are not uncommon in shallow waters of eastern Canada (Verkaik et al., 2016). The experimental pH was measured both at the start and end of the trials to determine any change over the 2-h period. Replication, sampling and control followed the designs described earlier. 2.5 Data analysis

Analysis of variance (ANOVA) was used to test for any significant effect of time (0 h vs 2 h) and individual on phytoplankton cell concentrations in each treatment. It was also used to explore the influence of individual and experimental factors (initial phytoplankton concentration, temperature, pH level, and presence/absence of a predator) on the calculated filtration rates. In cases where equal variance was violated, data were either natural-log transformed or ranked. Where appropriate, post-hoc multiple comparisons tests were conducted using the Holm-Sidak method. When interactions terms were significant, one-way ANOVAs or t-tests were independently conducted at

11

each level of each factor. A Spearman test was used to examine relationships between the various factors and either net/relative cell clearance or calculated filtration rates.

3 Results and Discussion Experimental studies of deep-sea species remain limited in general, and laboratory investigations of Porifera from the deep are perhaps among the rarest. Few sponges collected from the deep sea have been kept alive long enough for meaningful work to be carried out, with the exception of specimens of the genus Thenea (North Atlantic) that were used to conduct short-term experiments (Witte and Graf, 1996; Witte et al., 1997). There has also been work carried out on species collected from marginally deep areas ~150-200 m (e.g. Leys et al., 2011) and on specimens from shallow waters above 30 m that are presumed to have eurybathic distributions (e.g. Chu and Reiswig, 2014). The importance of gathering basic information on deep-water sponge communities is compounded by the fact that sponges are known to play key roles in other oligotrophic environments (De Goeij et al., 2013). The demosponges examined here were collected around 1000 m depth and survived more than 2 years under laboratory conditions. While the sample sizes available were admittedly low, this is not unusual for experimental studies on fragile deep-sea taxa (e.g. Kutti et al., 2015; Lartaud et al., 2014), and work on live deep-sea organisms in general is still rare. Moreover, evidence of growth was detected over the study period (Supplementary material, Figure S1), which lends support to the reliability and usefulness of the feeding metrics reported herein. The quantity of phytoplankton available to the sponges was an overall significant factor in the feeding responses of the demosponges under study (Figures 1, 2). Globally, all the feeding metrics measured were positively correlated with the initial phytoplankton 12

concentration (Figure 1). A three-way ANOVA (time*initial concentration*individual) showed a significant decrease of phytoplankton concentrations over time (F1,71=23.4, p<0.001) and a significant effect of initial concentration (F2,71=105.9, p<0.001) but not of sponge individual (F3,71=1.6, p=0.191). A one-way ANOVA confirmed that clearance rates were significantly greater at initial phytoplankton concentrations of 40 000 and 20 000 cells ml-1 than 10 000 cells ml-1 (p<0.001). Pairwise comparisons showed that differences in filtration rates among individual sponges were significant only at 20 000 cells ml-1 (p<0.021), in the following order: Polymastia 1 > Polymastia 2 = Radiella 2 = Radiella 1. Results were similar to those found in tropical shallow-water sponges (Axinella corrugate), where feeding typically increased with concentration of phytoplankton cells, and a concentration of 20 000 cells ml-1 of 4-10 µm sized phytoplankton was deemed to meet metabolic demands (Duckworth et al., 2003). A consistent trend was also reported by Kahn et al. (2015) in their field study of the cloud sponge Aphrocallistes vastus from ~175 m depth, which showed that the number of bacteria removed by individuals increased linearly with bacteria concentrations. When the weight of the sponges was factored into the comparisons, Polymastia 2 exhibited the highest feeding metrics (Table 1). In general, Polymastia was more responsive and had higher maximum filtration rates (9-12 ml h-1 g-1) than Radiella (1-5 ml h-1 g-1), which could be due to their morphology: Polymastia had many predominant branching papillae (which visibly developed over the study period; see Supplementary material, Figure S1) to potentially aid with pumping mechanisms under low flow or static conditions, whereas Radiella had fewer and much shorter papillae. With the caveat mentioned earlier, the present study suggests that the deep-sea demosponges under study

13

have lower filtration rates than shallow-water counterparts (Peterson et al., 2006; Riisgård et al., 1993), which exhibit weight-standardized filtration rates of 30-211 ml h-1 g-1. While direct comparisons are not possible due to morphological differences that cannot be accounted for (i.e. size data unavailable), feeding of deep-water species from the North Pacific estimated in situ revealed processing rates of ~7.9 ml sponge-1 s-1 for the hexactinellid Sericolophus hawaiicus (Pile and Young, 2006) and benthic grazing rates of up to 165 m3 m-2 d-1 for Aphrocallistes glass sponge reefs (Kahn et al., 2015). It is possible that the results obtained here could be an underestimation of filtration rates. The food used during the trials was perhaps not optimally recognized or retained. Although the phytoplankton mixture had a cell size deemed appropriate for some sponges (i.e. ultraplankton; Witte et al., 1997) and feeding was documented, it cannot be excluded that deep-sea demosponges mainly target other particle sizes or food sources, such as bacteria (Kahn et al., 2015; Verhoeven et al., 2017). It is also acknowledged that results were obtained under laboratory conditions. No significant linear correlation was detected between temperature and feeding metrics (rs=0.289, p=0.087 for of net clearance rate; rs=0.204, p=0.231 for filtration rate). The response was variable, with significant interactions detected between temperature, individual and time (F6,71=2.4, p=0.039; Figure 3), and between time and temperature (F2,71=3.3, p=0.047) on phytoplankton concentrations. At 3°C, there was also a significant interaction between time and individual (F2,23=6.6, p=0.004) and the independent tests indicated that Radiella 1 was the only one for which a statistically significant cell clearance was detected over the trial period (p<0.001; Figure 3B). At 0°C; there was an overall significant decrease in phytoplankton concentration over time (F1,23=4.8,

14

p=0.044). Corresponding filtration rates showed no significant difference among temperatures (F2,35=2.7, p=0.088) but a difference among individuals (F3,35=3.6, p=0.027). Specifically, Polymastia 1 had the highest filtration rates and Radiella 1 the lowest (i.e. Polymastia 1>Polymastia 2 = Radiella 2> Radiella 1; p=0.04-0.956). In temperature-specific analyses, filtration rates among individuals were only significantly different at 6°C (F3,11=4.5, p=0.040), whereas mean filtration rates of given individuals appeared to drop between 6°C and lower temperatures, but only in Polymastia and not in a statistically significant manner. Another study showed that a temperature decrease from 10 to <7°C arrested pumping in hexactinellid glass sponges (Leys and Meech, 2006), which have a different morphology and pumping mechanism (Leys et al., 2011). Here, a decrease in phytoplankton cell clearance was not expected to occur at 3°C, since this represents a typical temperature at 1000 m depth where the sponges were collected. On the other hand, temperatures around 0°C are less common at great depths, even if nearfreezing values can be recorded over the Grand Banks during the spring, when the Labrador Current is strong (Colbourne, 2004). Overall, the present results are consistent with the fact that important but normal temperature variations do not appear to significantly impact sponges. According to Kelmo et al. (2013), sponge diversity was unaltered and sponge density increased in all habitats examined during periods of ElNino from 1995 to 2011 (when habitats were subjected to large temperature fluctuations), suggesting that sponges can tolerate shifts in thermal regime fairly well. The presence of the predatory sea star Henricia lisa markedly decreased the ability of all sponges to clear phytoplankton cells (Figure 4). With the predatory sea star present, no significant decrease in phytoplankton cell concentrations occurred over the 2-

15

h feeding period (F1,23= 0.117, p=0.737) and there was no difference among the four individual sponges (F3,23= 0.753, p=0.536), and no interaction (F3,23=0.14, p=0.935). In the controls, the sponges showed significant phytoplankton cell clearance over the 2-h feeding period (F1,11= 24.7, p=0.001). Corresponding filtration rates decreased when H. lisa was present (mean of 0-53 ml h-1); they were significantly lower than in the control treatments (F1,11=6.7, p=0.032). When H. lisa was placed in a tank with the sponges at the end of all trials, the sea star made contact with all except the smallest Radiella, and generally had its stomach partially or fully extruded confirming the predatory nature of this sea star as demonstrated by Gale et al. (2013). The decreased feeding response in the presence of H. lisa was recorded without direct contact between prey and predator, indicating that the chemical signature/odour of the sea star was presumably detected by the sponges, activating a defense mechanism. As feeding was noted in control sponges, it is possible that exposure to potential predators resulted in closure of the ostia and oscula, in turn arresting filter feeding and masking the chemical signature of the sponge itself from the approaching predator (minimizing olfactory detection), similar to observations in shallow-water sponges (Haber et al., 2011). When the ambient seawater was acidified by 0.3-0.4 units, following predicted climate change scenarios for the deep North Atlantic by year 2100 (Gehlen et al., 2014), an inhibitory effect on feeding was recorded. There was no significant clearance of phytoplankton cells over the feeding period under low pH (Figure 5), and no statistical difference amongst individuals (F1,23=3.3, p=0.085). A two-way ANOVA on filtration rates confirmed the significant difference between ambient versus low pH treatments (F1,11=7.8, p=0.023). This could indicate that, in the short initial period following pH

16

change, sponges responded by closing ostia and oscula, therefore arresting filter feeding. However, it does not preclude adaptation to low pH over longer periods. While effects of ocean acidification on marine invertebrates are often reported (Pörtner, 2008), studies on deep-sea taxa are scarce (Hofmann et al., 2010; Taylor et al., 2014; Verkaik et al., 2017), and the responses of sponges have not been well studied. It has been predicted that deep-water scleractinian corals will show signs of altered capacity to calcify by 2100, and that ocean acidification might seriously damage these habitats in the long term (Turley et al., 2007). Out of the ~5000 species of sponge, only ~600 calcify using calcium carbonate, whereas most have siliceous spicules (Smith et al., 2013), suggesting that, unlike corals, most sponges will not be affected structurally by ocean acidification. It has even been proposed that climate change may cause sponges to become dominant, if they are able to cope with temperature and pH-related stresses over longer periods, by allowing them to spatially outcompete corals and other more sensitive species (Bell et al., 2013). Growth and metabolite synthesis of coral reef sponges were minimally affected by the mean values of water temperature and pH predicted for the end of the century (Duckworth et al., 2012). On the other hand, ocean acidification might have a negative effect on sponges due to limited physical protection, poor acid-base regulation and low metabolism, with deep-sea demosponges potentially being more vulnerable (Goodwin et al., 2014). A recent experimental study on the combined effects of ocean warming and acidification found that heterotrophic sponges (such as those studied here) may be the most susceptible to elevated pCO2 (Bennett et al., 2017). Overall, the present study indicates that environmental factors have the potential to modulate filter-feeding rates in the deep-sea demosponges, although it remains unclear

17

whether longer acclimation periods to certain conditions (e.g. lower pH levels or chemical signature of predators) would have allowed feeding efficiency to resume. Understanding the biology of sponges is vital to enhance our knowledge of deep-sea habitats where, despite ongoing efforts, Porifera continue to be less studied than other structural taxa such as corals. While admittedly still limited in scope, the findings outlined here will add to our baseline understanding of deep-sea sponge feeding metrics and acute reactions to abiotic and biotic stimuli, with potential contributions to estimates of carbon and energy flows in the deep ocean.

Acknowledgements Warm thanks are extended to E. Montgomery, J. Ammendolia, the Ocean Sciences Field Services at Memorial University, and the Department of Fisheries and Oceans (DFO) for assistance at various stages of the study. This work was funded by grants from the Natural Science and Engineering Research Council (NSERC) and the Canadian Foundation for Innovation (CFI) to A. Mercier.

References Barry, J.P., Lovera, C., Buck, K.R., Peltzer, E.T., Taylor, J.R., Walz, P., Whaling, P.J., Brewer, P.G., 2014. Use of a free ocean CO₂ enrichment (FOCE) system to evaluate the effects of ocean acidification on the foraging behavior of a deep-sea urchin. Environ. Sci. Technol. 48 (16), 9890. Beazley, L.I., Kenchington, E.L., Murillo, F.J., del Mar Sacau, M., 2013. Deep-sea sponge grounds enhance diversity and abundance of epibenthic megafauna in the Northwest Atlantic. ICES J. Mar. Sci. 70 (7), 1471-1490.

18

Bell, J.J., Davy, S.K., Jones, T., Taylor, M.W., Webster, N.S., 2013. Could some coral reefs become sponge reefs as our climate changes? Glob. Change Biol. 19 (9), 26132624. Bennett, H.M., Altenrath, C., Woods, L., Davy, S.K., Webster, N.S., Bell, J.J., 2017. Interactive effects of temperature and pCO2 on sponges: from the cradle to the grave. Glob. Change Biol. 23 (5), 2031-2046. Boury-Esnault N. Hajdu E. Borojević R. Custodio M. Klautau M. 1994. The value of cytological criteria in distinguishing sponges at the species level: the example of the genus Polymastia. Can. J. Zool. 72 (5), 795-804. Bullock, T.H., 1953. Predator recognition and escape responses of some intertidal gastropods in presence of starfish. Behaviour 5 (1), 130-130. Campbell, J.S., Simms, J.M., 2009. Status report on coral and sponge conservation in Canada. DFO, St. John's, NL (Canada). Cathalot, C., Van Oevelen, D., Cox, T.J.S., Kutti, T., Lavaleye, M., Duineveld, G., Meysman, F.J.R., 2015. Cold-water coral reefs and adjacent sponge grounds: hotspots of benthic respiration and organic carbon cycling in the deep sea. Frontiers in Marine Science 2, 37. Chevaldonné, P., Pérez, T., Crouzet, J.M., Bay‐Nouailhat, W., Bay‐Nouailhat, A., Fourt, M., Almón, B., Pérez, J., Aguilar, R., Vacelet, J., 2015. Unexpected records of 'deep-sea' carnivorous sponges Asbestopluma hypogea in the shallow NE Atlantic shed light on new conservation issues. Mar. Ecol. 36 (3), 475-484. Chu, J.W.F., Reiswig, H.M., 2014. Mechanisms of propagule release in the carnivorous sponge Asbestopluma occidentalis. Invertebr. Biol. 133 (2), 109-120. Colbourne, E.B., 2004. Decadal changes in the ocean climate in Newfoundland and Labrador waters from the 1950s to the 1990s. J. Northwest Atl. Fish. Sci. 34, 43-61. De Goeij, J.M., Van Oevelen, D., Vermeij, M.J.A., Osinga, R., Middelburg, J.J., de Goeij, A.F.P.M., Admiraal, W., 2013. Surviving in a marine desert: the sponge loop retains resources within coral reefs. Science 342 (6154), 108-110. DFO, 2009. 2008 State of the ocean: physical oceanographic conditions in the Newfoundland and Labrador region. Can. Sci. Advis. Sec. Sci. Advis. Rep. 2009/057. Duckworth, A., Samples, G., Wright, A., Pomponi, S., 2003. In vitro cuIture of the tropical sponge Axinella corrugata (Demospongiae): effect of food cell concentration on growth, clearance rate, and biosynthesis of stevensine. Mar. Biotechnol. 5 (6), 519-527. Duckworth, A.R., West, L., Vansach, T., Stubler, A., Hardt, M., 2012. Effects of water temperature and pH on growth and metabolite biosynthesis of coral reef sponges. Mar. Ecol. Prog. Ser. 462, 67-77. 19

Francis, J.C., Poirrier, M.A., 1986. Particle uptake in two fresh-water sponge species, Ephydatia fluviatilis and Spongilla alba (Porifera: Spongillidae). Trans. Am. Microsc. Soc. 105 (1), 11-20. Gale, K.S.P., Hamel, J.-F., Mercier, A., 2013. Trophic ecology of deep-sea Asteroidea (Echinodermata) from eastern Canada. Deep-Sea Res. I 80, 25-36. Gehlen, M., Séférian, R., Jones, D.O.B., Roy, T., Roth, R., Barry, J.P., Bopp, L., Doney, S.C., Dunne, J.P., Heinze, C., 2014. Projected pH reductions by 2100 might put deep North Atlantic biodiversity at risk. Biogeosciences 11, 6955-6967. Goodwin, C., Rodolfo-Metalpa, R., Picton, B., Hall-Spencer, J.M., 2014. Effects of ocean acidification on sponge communities. Mar. Ecol. 35 (s1), 41-49. Haber, M., Carbone, M., Mollo, E., Gavagnin, M., Ilan, M., 2011. Chemical defense against predators and bacterial fouling in the Mediterranean sponges Axinella polypoides and A. verrucosa. Mar. Ecol. Prog. Ser. 422, 113-122. Hamel, J.-F., Sun, Z., Mercier, A., 2010. Influence of size and seasonal factors on the growth of the deep-sea coral Flabellum alabastrum in mesocosm. Coral Reefs 29 (2), 521-525. Hill, M.S., Lopez, N.A., Young, K.A., 2005. Anti-predator defenses in western North Atlantic sponges with evidence of enhanced defense through interactions between spicules and chemicals. Mar. Ecol. Prog. Ser. 291, 93-102. Hofmann, G.E., Barry, J.P., Edmunds, P.J., Gates, R.D., Hutchins, D.A., Klinger, T., Sewell, M.A., 2010. The effect of ocean acidification on calcifying organisms in marine ecosystems: an organism-to-ecosystem perspective. Ann. Rev. Ecol. Evol. Syst. 41, 127147. Johns, D.G., Edwards, M., Richardson, A.J., Spicer, J.I., 2003. Increased blooms of a dinoflagellate in the NW Atlantic. Mar. Ecol. Prog. Ser. 265, 283-287. Kahn, A.S., Yahel, G., Chu, J.W.F., Tunnicliffe, V., Leys, S.P., 2015. Benthic grazing and carbon sequestration by deep‐water glass sponge reefs. Limnol. Oceanogr. 60 (1), 78-88. Kelmo, F., Bell, J.J., Attrill, M.J., 2013. Tolerance of sponge assemblages to temperature anomalies: resilience and proliferation of sponges following the 1997-8 El-Niño southern oscillation. PLoS One 8 (10), 76441. Kutti, T., Bannister, R.J., Fosså, J.H., Krogness, C.M., Tjensvoll, I., Søvik, G., 2015. Metabolic responses of the deep-water sponge Geodia barretti to suspended bottom sediment, simulated mine tailings and drill cuttings. J. Exp. Mar. Biol. Ecol. 473, 64-72. Lartaud, F., Pareige, S., de Rafelis, M., Feuillassier, L., Bideau, M., Peru, E., De la Vega, E., Nedoncelle, K., Romans, P., Le Bris, N., 2014. Temporal changes in the growth of 20

two Mediterranean cold-water coral species, in situ and in aquaria. Deep-Sea Res. II 99, 64-70. Lesser, M.P., Slattery, M., 2013. Ecology of Caribbean sponges: are top-down or bottomup processes more important? PLoS One 8 (11), e79799. Leys, S.P., Lauzon, N.R.J., 1998. Hexactinellid sponge ecology: growth rates and seasonality in deep water sponges. J. Exp. Mar. Biol. Ecol. 230 (1), 111-129. Leys, S.P., Meech, R.W., 2006. Physiology of coordination in sponges. Can. J. Zool. 84 (2), 288-306. Leys, S.P., Yahel, G., Reidenbach, M.A., Tunnicliffe, V., Shavit, U., Reiswig, H.M., 2011. The sponge pump: the role of current induced flow in the design of the sponge body plan. PLoS One 6 (12), e27787. Li, W.K.W., 2002. Macroecological patterns of phytoplankton in the northwestern North Atlantic Ocean. Nature 419 (6903), 154-157. Maldonado, M., Ribes, M., van Duyl, F.C., 2012. Nutrient fluxes through sponges: biology, budgets, and ecological implications. Adv. Mar. Biol. 62, 113. Murillo, F., Cristobo, J., Muñoz, P., Serrano, A., Ríos, P., González, C., Kenchington, E., 2012. Deep-sea sponge grounds of the Flemish Cap, Flemish Pass and the Grand Banks of Newfoundland (Northwest Atlantic Ocean): Distribution and species composition. Mar. Biol. Res. 8 (9), 842. NOAA, 2010. NOAA Strategic plan for deep-sea coral and sponge ecosystems: Research, management, and international cooperation. National Oceanic and Atmospheric Administration, Coral Reef Conservation Program. NOAA Technical Memorandum CRCP 11, Silver Spring. Peterson, B.J., Chester, C.M., Jochem, F.J., Fourqurean, J.W., 2006. Potential role of sponge communities in controlling phytoplankton blooms in Florida Bay. Mar. Ecol. Prog. Ser. 328, 93-103. Pile, A.J., Young, C.M., 2006. The natural diet of a hexactinellid sponge: Benthic– pelagic coupling in a deep-sea microbial food web. Deep-Sea Res. I 53 (7), 1148-1156. Plotkin, A., Gerasimova, E., Rapp, H.T., 2012. Phylogenetic reconstruction of Polymastiidae (Demospongiae: Hadromerida) based on morphology. Hydrobiologia 687 (1), 21-41. Pörtner, H., 2008. Ecosystem effects of ocean acidification in times of ocean warming: a physiologist’s view. Mar. Ecol. Prog. Ser. 7 0 -217. Reiswig, H.M., 1971a. Particle feeding in natural populations of three marine demosponges. Biol. Bull. 141 (3), 568-591.

21

Reiswig, H.M., 1971b. In situ pumping activities of tropical Demospongiae. Mar. Biol. 9 (1), 38-50. Reiswig, H.M., 1975. Bacteria as food for temperate-water marine sponges. Can. J. Zool. 53 (5), 582-589. Riisgård, H.U., Larsen, P.S., 2010. Particle capture mechanisms in suspension-feeding invertebrates. Mar. Ecol. Prog. Ser. 418, 255-293. Riisgård, H.U., Thomassen, S., Jakobsen, H., Weeks, J.M., Larsen, P.S., 1993. Suspension-feeding in marine sponges Halichondria panicea and Haliclona urceoluseffects of temperature on filtration-rate and energy-cost of pumping. Mar. Ecol. Prog. Ser. 96 (2), 177-188. Ruhl, H.A., Ellena, J.A., Smith, K.L., 2008. Connections between climate, food limitation, and carbon cycling in abyssal sediment communities. Proc. Natl. Acad. Sci. U.S.A. 105 (44), 17006-17011. Sheild, C.J., Witman, J.D., 1993. The impact of Henricia sanguinolenta (O.F. Müller) (Echinodermata:Asteroidea) predation on the finger sponges, Isodictya spp. J. Exp. Mar. Biol. Ecol. 166 (1), 107-133. Slattery, M., Gochfeld, D.J., Diaz, M.C., Thacker, R.W., Lesser, M.P., 2016. Variability in chemical defense across a shallow to mesophotic depth gradient in the Caribbean sponge Plakortis angulospiculatus. Coral Reefs 35 (1), 11-22. Smith, A.M., Berman, J., Key, M.M., Winter, D.J., 2013. Not all sponges will thrive in a high-CO2 ocean: review of the mineralogy of calcifying sponges. Palaeogeogr. Palaeoclimatol. Palaeoecol. 392, 463-472. Sun, Z., Hamel, J.-F., Edinger, E., Mercier, A., 2010. Reproductive biology of the deepsea octocoral Drifa glomerata in the Northwest Atlantic. Mar. Biol. 157 (4), 863-873. Taylor, J.R., Lovera, C., Whaling, P.J., Buck, K.R., Pane, E.F., Barry, J.P., 2014. Physiological effects of environmental acidification in the deep-sea urchin Strongylocentrotus fragilis. Biogeosciences 11 (5), 1413. Turley, C.M., Roberts, J.M., Guinotte, J.M., 2007. Corals in deep-water: will the unseen hand of ocean acidification destroy cold-water ecosystems? Coral Reefs 26 (3), 445-448. ue . Casas .M. Brodie W.B. Murillo . . Mandado M.n. ago . lpoim R. Ba n R. rmesto .n. 01 . ist of Species as recorded y Canadian and E Bottom Trawl Surveys in Flemish Cap. Northwest Atlantic Fisheries Organization 14 (005). Verhoeven, J.T.P., Kavanagh, A.N., Dufour, S.C., 2017. Microbiome analysis shows enrichment for specific bacteria in separate anatomical regions of the deep-sea carnivorous sponge Chondrocladia grandis. FEMS Microbiol. Ecol. 93 (1), fiw214.

22

Verkaik, K., Hamel, J.-F., Mercier, A., 2016. Carry-over effects of ocean acidification in a cold-water lecithotrophic holothuroid. Mar. Ecol. Prog. Ser. 557, 189-206. Verkaik, K., Hamel, J.-F., Mercier, A., 2017. Impact of ocean acidification on reproductive output in the deep-sea annelid Ophryotrocha sp. (Polychaeta: Dorvelleidae). Deep-Sea Res. II 137, 368-376. Witte, U., Graf, G., 1996. Metabolism of deep-sea sponges in the Greenland-Norwegian Sea. J. Exp. Mar. Biol. Ecol. 198 (2), 223-235. Witte, U., Brattegard, T., Graf, G., Springer, B., 1997. Particle capture and deposition by deep-sea sponges from the Norwegian-Greenland Sea. Mar. Ecol. Prog. Ser. 154, 241252. Yahel, G., Eerkes-Medrano, D.I., Leys, S.P., 2006. Size independent selective filtration of ultraplankton by hexactinellid glass sponges. Aquat. Microb. Ecol. 45 (2), 181-194. Yahel, G., Whitney, F., Reiswig, H.M., Eerkes-Medrano, D.I., Leys, S.P., 2007. In situ feeding and metabolism of glass sponges (Hexactinellida, Porifera) studied in a deep temperate fjord with a remotely operated submersible. Limnol. Oceanogr. 52 (1), 428440.

23

3.4 1.3

Radiella 1

Radiella 2 1.7

4.5

3.7

8.5

8.0

33.0

15.0

50.0

volume (ml)

Metrics

diameter** (cm)

**Excluding crest

2.0

Polymastia 2

*Excluding papillae

2.0

height* (cm)

Polymastia 1

Individuals

5.2

30.6

11.1

50.3

wet weight (g)

0.8

0.65

None

0.45

Crest length (cm)

0.2

0.05

0.6

0.3

Longest papilla length (cm)

filtration rates measured at 6°C and a nominal phytoplankton concentration of 20 000 cells ml-1.

12

12

24

114

Number of papillae

25.8

36.8

128.2

433.9

(ml h-1)

5.0

1.2

11.6

8.6

(ml h-1 g-1)

Maximum filtration rate

Table 1. Morphometric data of deep-sea demosponges under study (Radiella hemisphaerica and Polymastia sp.), and maximum

Fig. 1. Correlation between initial phytoplankton cell concentration (cells ml-1) and A) net cell clearance rates (cells ml-1), B) percent clearance and C) filtration rates (ml h-1) in deep-sea demosponges (Radiella hemisphaerica and Polymastia sp.) at 6°C. Solid lines represent linear regression fit and dashed lines represent 95% confidence interval. Fig. 2. Change in phytoplankton concentration effected by deep-sea demosponges (Radiella hemisphaerica and Polymastia sp.) over 2 h of exposure to A) 10 000 cells ml-1, B) 20 000 cells ml-1, and C) 40 000 cells ml-1. Data (mean ± SD, N=3) gathered at 6°C. Different letters indicate significant (p<0.05) pairwise differences between initial (0 h) and final (2 h) cell concentrations for each individual. Fig. 3. Change in phytoplankton concentration effected by deep-sea demosponges (Radiella hemisphaerica and Polymastia sp.) over 2 h at A) 6°C, B) 3°C, and C) 0°C. Data (mean ± SD; N=3) gathered during exposure to a nominal concentration of 20 000 cells ml-1. Different letters indicate significant (p<0.05) pairwise differences between initial (0 h) and final (2 h) cell concentrations for each individual. Fig 4. Change in phytoplankton concentration (mean ± SD, N=3) effected by deep-sea demosponges (Radiella hemisphaerica and Polymastia sp.) over 2 h exposure to 20 000 cells ml-1 at 3°C. Results obtained A) in the presence of the sea star predator Henricia lisa, and B) under control conditions (without predator). ND = not determined. Different letters indicate significant (p<0.05) pairwise differences between initial (0 h) and final (2 h) cell concentrations for each individual. Fig 5. Change in phytoplankton concentration (mean ± SD, N=3) effected by the demosponge Polymastia sp. over 2 h at a nominal phytoplankton concentration of 20 000

cells ml-1. Results obtained at A) ambient pH (7.95 ± 0.03) at 6°C, and B) low pH (7.54 ± 0.02) at 3°C. Different letters indicate significant (p<0.05) pairwise differences between initial (0 h) and final (2 h) cell concentrations for each individual.

26

Figures All figures to be printed in black and white

Figure 1 27

Figure 2

28

Figure 3

29

Figure 4

30

Figure 5

31

Highlights  Filter-feeding metrics of live deep-sea sponges were experimentally studied  Feeding was positively correlated to increased food concentration  Feeding decreased when pH was lowered or a predator was added  Most sponges grew during trials at optimal temperature and food concentrations  Some sponges shrank when exposed to the abiotic and biotic stressors

32