Introduction
Parasites are found in all ecosystems throughout the globe (Jorge and Poulin, Reference Jorge and Poulin2018). They comprise up to 40% of all described species (Dobson et al., Reference Dobson, Lafferty, Kuris, Hechinger and Jetz2008), feature in up to 70% of the links within food webs (Dunne et al., Reference Dunne, Lafferty, Dobson, Hechinger, Kuris, Martinez, McLaughlin, Mouritsen, Poulin, Reise, Stouffer, Thieltges, Williams and Zander2013), and contribute significantly to the biomass of many ecosystems (Kuris et al., Reference Kuris, Hechinger, Shaw, Whitney, Aguirre-Macedo, Boch, Dobson, Dunham, Fredensborg, Huspeni, Lorda, Mababa, Mancini, Mora, Pickering, Talhouk, Torchin and Lafferty2008). Their presence has important – though still remarkably underappreciated – implications for the structure, functioning and dynamics of entire ecosystems (Amundsen et al., Reference Amundsen, Lafferty, Knudsen, Primicerio, Klemetsen and Kuris2009; Dunne et al., Reference Dunne, Lafferty, Dobson, Hechinger, Kuris, Martinez, McLaughlin, Mouritsen, Poulin, Reise, Stouffer, Thieltges, Williams and Zander2013). The influence of parasites on how ecosystems respond to environmental change, however, particularly a warming climate (Kutz et al., Reference Kutz, Hoberg, Polley and Jenkins2005; Hoberg and Brooks, Reference Hoberg and Brooks2007), remains largely unknown. Climate warming will likely modify rates of parasite transmission (Mouritsen and Jensen, Reference Mouritsen and Jensen1997), as temperature is known to influence both parasite infectivity (Studer et al., Reference Studer, Thieltges and Poulin2010) and host immunocompetence in invertebrates (Mydlarz et al., Reference Mydlarz, Jones and Harvell2006). Moreover, by moderating host behaviour (Issartel et al., Reference Issartel, Hervant, Voituron, Renault and Vernon2005; Abram et al., Reference Abram, Boivin, Moiroux and Brodeur2017), warming influences the susceptibility of organisms to parasites (Morley and Lewis, Reference Morley and Lewis2014) and overall host functioning (O'Gorman et al., Reference O'Gorman, Pichler, Adams, Benstead, Cohen, Craig, Cross, Demars, Friberg, Gíslason, Gudmundsdóttir, Hawczak, Hood, Hudson, Johansson, Johansson, Junker, Laurila, Manson, Mavromati, Nelson, Ólafsson, Perkins, Petchey, Plebani, Reuman, Rall, Stewart, Thompson and Woodward2012).
Bioturbation – the mixing of sediment by mobile organisms – is an important ecosystem function that occurs in both terrestrial and aquatic environments. It comprises a key non-trophic mechanism through which organisms physically, chemically, and biologically structure ecosystems (Grant and Daborn, Reference Grant and Daborn1994; Jones et al., Reference Jones, Lawton, Shachak, Samson and Knopf1996; Baranov et al., Reference Baranov, Lewandowski and Krause2016; Wohlgemuth et al., Reference Wohlgemuth, Solan and Godbold2017). In aquatic ecosystems, bioturbation influences the flow of nutrients (Mermillod-Blondin et al., Reference Mermillod-Blondin, Gaudet, Gerino, Desrosiers, Jose and Châtelliers2004), oxygenation of sediments (Baranov et al., Reference Baranov, Lewandowski and Krause2016), turbidity of the water (Croel and Kneitel, Reference Croel and Kneitel2011) and sediment erosion rates (Grant and Daborn, Reference Grant and Daborn1994). Moreover, the rate of bioturbation has been shown, in a limited number of studies, to increase with warming (Baranov et al., Reference Baranov, Lewandowski and Krause2016). There is, however, little information about the influence of parasites on rates of bioturbation (Vannatta and Minchella, Reference Vannatta and Minchella2018) and whether this effect is, in turn, modified by warming. Though parasitism has been shown to modify burrowing behaviour in intertidal cockles (Mouritsen and Poulin, Reference Mouritsen and Poulin2005), and reduce their digging into the sediments, there have been no studies of which we are aware that found that parasites increase bioturbation rates of their hosts.
Gammarid amphipods contribute significantly to bioturbation in aquatic ecosystems globally (Mermillod-Blondin et al., Reference Mermillod-Blondin, Gaudet, Gerino, Desrosiers, Jose and Châtelliers2004; Hunting et al., Reference Hunting, Whatley, van der Geest, Mulder, Kraak, Breure and Admiraal2012; De Nadaï-Monoury et al., Reference De Nadaï-Monoury, Lecerf, Canal, Buisson, Laffaille and Gilbert2013; Vadher et al., Reference Vadher, Stubbington and Wood2015), primarily by reworking the uppermost layer (i.e. 2–3 cm) of sediment. In freshwaters, gammarids are also frequently infected with an acanthocephalan parasite, Polymorphus minutus, which modifies both the movement of their hosts in the water column and the rates at which they shred detritus (Bauer et al., Reference Bauer, Haine, Perrot-Minnot and Rigaud2005; Labaude et al., Reference Labaude, Rigaud and Cézilly2016). Two life-stages of the acanthocephalan – the acanthella and the cystacanth – utilize the amphipod intermediate host. The cystacanth is the life-stage associated most strongly with behavioural changes (Bailly et al., Reference Bailly, Cézilly and Rigaud2018), as it is the stage at which the parasite is infective to its definitive (that is, final) host, in this case water fowl.
We explored whether (1) parasitic infection and warming, individually or in combination, modify rates of sediment surface reworking (our measure of bioturbation) by host organisms and, if so, (2) the combined effects of parasitic infection and temperature on host bioturbation are additive, antagonistic, or synergistic. To address these questions, we quantified the bioturbation activity of Gammarus duebeni experimentally in the laboratory across the broad range of temperatures encountered in their native range. Gammarus are used frequently as a model system to examine the impacts of parasites on intermediate host behaviour (Bakker et al., Reference Bakker, Mazzi and Zala1997; Agatz and Brown, Reference Agatz and Brown2014; Perrot-Minnot et al., Reference Perrot-Minnot, Sanchez-Thirion and Cézilly2014; Perrot-Minnot et al., Reference Perrot-Minnot, Maddaleno, Cézilly and Woods2016) and G. duebeni comprise important components of the benthos throughout their native range (Reid, Reference Reid1938; Donohue et al., Reference Donohue, Donohue, Ní Ainín and Irvine2009; MacNeil and Briffa, Reference MacNeil and Briffa2009), playing a crucial role in ecosystem functioning by processing detritus (Kelly et al., Reference Kelly, Dick and Montgomery2002). As the amphipods are ectothermic (Baranov et al., Reference Baranov, Lewandowski and Krause2016), we expect warming to increase rates of bioturbation by increasing movement capacity (Dell et al., Reference Dell, Pawar and Savage2011). We also predict that parasitic infection will reduce rates of bioturbation due to reduced interaction between the gammarids and the benthos, as gammarid hosts infected with P. minutus display enhanced phototaxis and are more likely to move upward in the water column (Perrot-Minnot et al., Reference Perrot-Minnot, Maddaleno, Cézilly and Woods2016). As temperature and parasitic infection have been shown to additively impact similar gammarid behaviours (Labaude et al., Reference Labaude, Rigaud and Cézilly2016), we expect warming and parasites to additively impact bioturbation.
Methods
Experimental design
We quantified the rate of sediment surface re-working by adult Gammarus duebeni var. celticus at two levels of infection (i.e infected or uninfected by P. minutus cystacanths) and at four temperatures (4 °C, 9 °C, 14 °C, and 19 °C), encompassing the majority of the temperature range experienced by G. duebeni in their native range in Ireland, in a full-factorial experiment. Each experimental treatment combination was replicated 20 times.
Amphipods, benthic lake sediments and lake water used in the experiment were collected from Lough Lene, Co. Westmeath, Ireland (53.6625°N, 7.2340°W) on 22 January 2018. Surficial (i.e. less than 3 cm depth) benthic lake sediments were collected, homogenized, passed through a 1 mm sieve to remove macrofauna and rocks, and allowed to settle in lake water for one day before use.
Bioturbation was quantified based upon methods developed by De Nadaï-Monoury et al. (Reference De Nadaï-Monoury, Lecerf, Canal, Buisson, Laffaille and Gilbert2013) and Wohlgemuth et al. (Reference Wohlgemuth, Solan and Godbold2017). Eight 10 L buckets (28.5 cm diameter, 20 cm in height) were filled with lake sediments to a depth of 5 cm. Sterile centrifuge tubes (8.5 cm long with an internal diameter of 2.7 cm) with their tops and bottoms removed were placed into the buckets (25 pipes per bucket). Tracer sand (pink luminophores <125 µm; Brianclegg Ltd., UK) was then added to a depth of 0.2 cm within each tube. Filtered aerated lake water was then added slowly to the bucket to a depth of 13 cm above the sediment. A single G. duebeni adult (>0.02 g fresh weight) was added to individual tubes, which were then covered with mesh (1 mm aperture) to retain the organisms within the tubes whilst allowing the circulation of aerated water. Presence of the mesh also enabled clinging behaviour by Gammarus, thus allowing them to mimic their propensity to cling to floating debris in the water column. Each 10 L bucket contained ten tubes containing infected G. duebeni, ten tubes with uninfected G. duebeni, and five tubes containing no G. duebeni. The latter acted as procedural controls. Fresh mass of G. duebeni individuals at the commencement of the experiment was similar across all experimental treatment combinations (ANOVA; F 7,135 = 1.7, P = 0.12). Sediment disturbance in the procedural controls was negligible (Fig. S1), and did not vary with temperature (ANOVA, F 3,36 = 0.73, P = 0.54). Two 10 L buckets were kept at each of the four temperatures analysed. The buckets were aerated continually and kept in a 12 h:12 h light:dark cycle. After 28 days, G. duebeni individuals were removed and dissected to ensure infection status. Only organisms with single, cystacanth-stage infections were designated as infected – any hosts with multiple-infections or acanthellae-stage infections were omitted from analyses. Parasites were then examined microscopically to confirm their identity morphologically (following McDonald, Reference McDonald1988) after cystacanths were first placed in a 0.25 mm solution of sodium taurocholate, a type of bile salt which encourages extension of the proboscis, and left overnight at 37 °C.
Data analyses
Photographs of the sediment surface of each experimental tube were taken with a Canon EOS 550D (Aperature: f/4.5; Pixels: 5184 × 3456) and saved as RGB-coloured JPEGs. Images were captured under UV light (395 nm wavelength, UV LED flashlight, LightingEVER, Las Vegas, USA) to optimise fluorophore detection. Images were then processed using ImageJ (version 1.43u; US National Institutes of Health, https://imagej.nih.gov/ij/). Images were cropped, then split into red, green and blue colour channels. The red channel was selected for analyses, as it allowed for clearest distinction between the pink fluorophores and the black lake sediments. Images were then thresholded in order to colour the fluorescent particles white and the sediment particles black. The photo was then analysed and the proportion of black pixels, representing the lake sediments brought up from below the fluorophores, recorded. The total area of surface sediment reworked was then quantified in cm2.
Data were analysed in R (version 3.4.1; R Core Team, 2017). The extent of sediment surface reworking was log10-transformed prior to analyses to meet assumptions of normality and homoscedasticity. A linear mixed-effects model was constructed using lme4::lmer (Bates et al., Reference Bates, Mächler, Bolker and Walker2015), with the log10-transformed area reworked as the response variable, bucket as a random effect, and temperature and infection status as fixed effects. Model selection was done with model.sel:MuMIn (Barton, Reference Barton2016). We report the results of fixed effects on the model with the lowest AIC of the candidate models. To determine the magnitude of the effect of infection across the range of temperatures examined, we calculated the Cohen's d effect size with 95% confidence intervals using effsize:cohen.d (Torchiano, Reference Torchiano2018).
Results
Infected individuals of G. duebeni reworked significantly more sediment surface area than uninfected individuals (linear mixed-effects model, F 1,137 = 7.38, P < 0.01; Fig. 1A). Rates of sediment reworking also increased significantly with warming (F 8,137 = 5.3, P = 0.05; Fig. 1A). Combined effects of parasitic infection and warming were, however, additive, as temperature did not interact with parasitic infection in moderating bioturbation, and the magnitude of the effect of parasitism was consistent across the range of temperatures examined (Fig. 1B).
(A) Rates (mean ± s.e.) of bioturbation by G. duebeni infected (open circles) and uninfected (closed circles) with P. minutus across a temperature range of 4 °C to 19 °C. Bioturbation was measured as the area of benthic surface sediments in our experimental microcosms that were reworked over the course of the experiment. (B) Effect size of P. minutus infection on bioturbation rates by G. duebeni across the range of temperatures examined.
Discussion
We found that both parasitic infection and warming increased bioturbation by G. duebeni in our experimental microcosms. Moreover, infection and temperature moderated bioturbation additively and did not interact. This comprises the first evidence of which we are aware of parasites enhancing the bioturbation activity of their hosts. Given the importance of gammarid amphipods as key drivers of detritivory and bioturbation in freshwater ecosystems (Hunting et al., Reference Hunting, Whatley, van der Geest, Mulder, Kraak, Breure and Admiraal2012), coupled with predicted increases in the prevalence of parasites in a warmer world (Galaktionov, Reference Galaktionov2017), our findings have important implications for the structure and functioning of freshwater ecosystems under global change.
The observed enhanced bioturbation caused by infection with P. minutus contrasts with our a priori predictions. As infection with P. minutus increases movement upwards in the water column (Jacquin et al., Reference Jacquin, Mori, Pause, Steffen and Medoc2014; Perrot-Minnot et al., Reference Perrot-Minnot, Maddaleno, Cézilly and Woods2016; Bailly et al., Reference Bailly, Cézilly and Rigaud2018), we anticipated that infected hosts would interact less with the benthos, leading to a decrease in the rates of surface sediment reworking. However, parasites often do not have fine-scale control when manipulating their hosts. The manipulation of crickets by nematomorph worms provides a clear example. The worms alter the behaviour of crickets to increase their chances of entering the water. However, the manipulation is not a specific push towards the water, but rather results in an increase in erratic jumping (Thomas et al., Reference Thomas, Schmidt-Rhaesa, Martin, Manu, Durand and Renaud2002). It has been suggested previously that the behavioural manipulation of our model parasite, P. minutus, is non-specific and does not drive the intermediate host directly to the exact, preferred definitive host (Jacquin et al., Reference Jacquin, Mori, Pause, Steffen and Medoc2014). The mechanism of manipulation by the parasite is possibly related to hypoxia in the water column and anaerobic metabolism within the host (Perrot-Minnot et al., Reference Perrot-Minnot, Maddaleno, Cézilly and Woods2016). The mechanisms underlying the manipulation are not yet fully understood, though it is possible that an accumulation of lactate in the brain of the amphipods may drive the reversal in geotaxis seen with P. minutus infection (Perrot-Minnot et al., Reference Perrot-Minnot, Maddaleno, Cézilly and Woods2016). The increased digging we observed in infected G. duebeni may therefore reflect an additional impact of lactate accumulation in the brain and a consequential increase in movement, rather than a mechanism for directly increasing transmission of the parasite to its definitive host. However, further work is needed to determine whether or not enhanced bioturbation activity is adaptive for the parasite.
A wide range of animal behaviours exhibit thermal dependence, many of which can be explained by metabolic theory (Kordas et al., Reference Kordas, Harley and O'Connor2011; Dell et al., Reference Dell, Pawar and Savage2014). Higher temperatures have been linked previously to enhanced bioturbation rates in non-amphipod aquatic species (Ouellette et al., Reference Ouellette, Gaston, Gagne, Gilbert, Poggiale, Blier and Stora2004), though the extent to which temperature enhances or supresses bioturbation likely varies across species (Maire et al., Reference Maire, Lecroart, Meysman, Rosenberg, Duchêne and Grémare2010). Our results are consistent with those from previous studies (Labaude et al., Reference Labaude, Rigaud and Cézilly2016) that found additive, rather than interactive, effects of temperature and parasitic infection on a range of behaviours in Gammarus. As the climate continues to warm, alterations in the prevalence of parasites and associated shifts in the behaviour and functioning of Gammarus have the potential to moderate the impact of many of the stressors of aquatic systems associated with global environmental change (Baranov et al., Reference Baranov, Lewandowski and Krause2016).
Our results demonstrate a significant influence of parasites on the key ecosystem function that is bioturbation (Vannatta and Minchella, Reference Vannatta and Minchella2018). Bioturbation has, for example, been linked to rates of community respiration, sediment transport, nutrient availability and overall community structure (Grant and Daborn, Reference Grant and Daborn1994; Ouellette et al., Reference Ouellette, Gaston, Gagne, Gilbert, Poggiale, Blier and Stora2004; Donohue and Garcia Molinos, Reference Donohue and Garcia-Molinos2009; Croel and Kneitel, Reference Croel and Kneitel2011; Baranov et al., Reference Baranov, Lewandowski and Krause2016; Wohlgemuth et al., Reference Wohlgemuth, Solan and Godbold2017). Therefore, irrespective of whether or not the altered behaviour we found is adaptive in terms of the parasite's fitness, our findings have important implications for our understanding of the roles played by parasites in the structure and functioning of aquatic systems.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S0031182019000635
Author ORCIDs
Maureen A. Williams, 0000-0001-7886-9353; Ian Donohue 0000-0002-4698-6448; Celia V. Holland 0000-0002-0550-7287
Acknowledgements
We thank Daniel Wohlgemuth and Alison Boyce for their assistance with methodology and chamber construction.
Financial support
This research was funded through a Government of Ireland Postgraduate Scholarship from the Irish Research Council (GOIPG/2014/103).
Conflict of interest
None.
Ethical approval
Not applicable.
