Seasonal variation in the antifouling defence of the temperate brown alga Fucus vesiculosus

The important role of marine epibiotic biofilms in the interactions of the host with its environment has been acknowledged recently. Previous studies with the temperate brown macroalga Fucus vesiculosus have identified polar and non-polar compounds recovered from the algal surface that have the potential to control such biofilms. Furthermore, both the fouling pressure and the composition of the epibiotic bacterial communities on this macroalga varied seasonally. The extent to which this reflects a seasonal fluctuation of the fouling control mechanisms of the host is, however, unexplored in an ecological context. The present study investigated seasonal variation in the anti-settlement activity of surface extracts of F. vesiculosus against eight biofilm-forming bacteria isolated from rockweed-dominated habitats, including replication of two populations from two geographically distant sites. The anti-settlement activity at both sites was found to vary temporally, reaching a peak in summer/autumn. Anti-settlement activity also showed a consistent and strong difference between sites throughout the year. This study is the first to report temporal variation of antifouling defence originating from ecologically relevant surface-associated compounds.


Introduction
Marine microbes exert continuous fouling pressure on any submerged surface, whether dead or living. An average ml of seawater contains up to 10 7 viruses, 10 6 bacteria, 10 3 fungi and 10 3 microalgae (Cole 1982;Jennings & Fenical 1994). Most of them strive to form biofilms facilitating interaction and proliferation (Grossart 2010). The establishment of a biofilm on a living surface, such as an algal thallus, has multiple and important consequences for the interactions between the host and the environment, eg access to nutrients and light, and modulating the capacity of the alga for further recruitment of microfoulers or macrofoulers (reviewed in Wahl et al. 2012).
Marine macroalgae may bear dense and diverse biofilms (Lachnit et al. 2009). While certain alga-microbe interactions are beneficial (Egan et al. 2001), other associations have been found to be detrimental (Dobretsov et al. 2011). The epibiotic microbial film (Costerton et al. 1995) formed, for instance, on the thallus of a macroalga replaces the original interface between the host organism and its environment by a new epibiotic interface with often strikingly different physical, chemical, mechanical, topographical and biological properties (Wahl 2008). Bacterial colonisers can enhance the colonization of the surface by larvae and spores of macrofoulers (Unabia & Hadfield 1999;Dobretsov et al. 2009Dobretsov et al. , 2011Mieszkin et al. 2012) along with all the associated consequences, eg hindered transepidermal exchanges, increased weight and friction of the host (Prescott 1990;Dougherty & Russell 2005), or they may hinder colonisation (Nasrolahi et al. 2012). Microbial pathogens may cause extensive tissue damage (Sawabe et al. 1998). Furthermore, the epibiotic biofilm may impede gaseous exchange, reduce the intensity of incoming radiation and thereby reduce the photosynthetic activity of the alga (Wahl 2008). Given the potential consequences of being fouled (de Nys & Steinberg 1999); some control by the host over the type and/or abundance of epibionts is generally expected and should be of selective advantage .
It has been demonstrated that macroalgae possess chemical defences to control fouling of their surfaces (Lau & Qian 1997;Sneed & Pohnert 2011). Although a number of studies have reported antifouling properties of algal metabolites, until now such roles have only been rigorously demonstrated for 5 species in an ecologically relevant manner (Schmitt et al. 1995;Paul et al. 2006;Persson et al. 2011;Saha et al. 2011), ie at concentrations that roughly correspond to natural 'near-surface' concentrations and against naturally co-occurring fouling species.
Temperate macroalgal species have been reported to exhibit seasonal variation in their antifouling activity in (presumed) phase with fouling pressure and abiotic factors such as light intensity and water temperature Maréchal et al. 2004;Wahl et al. 2010). These studies, however, investigated total rather than surface extracts, and thus, metabolites deployed at the surface and those stored within the thallus could not be distinguished.
The rockweed Fucus vesiculosus (Phaeophyceae), which has a temperate to arctic distribution is known to up-regulate or down-regulate its anti-herbivory defence in response to variations in grazing pressure (Rohde & Wahl 2008). However, no information is available for ecologically relevant, ie surface-deployed, variation in the antifouling defence of F. vesiculosus or any other macroalga.
Previous studies (Saha et al. 2011(Saha et al. , 2012 demonstrated that F. vesiculosus possesses surface-associated non-polar (fucoxanthin) and polar (DMSP and proline) metabolites, with anti-settlement activity against a wide range of ecologically relevant fouling bacteria. A first indication for the existence of defence regulation would be if the level of these metabolites was lower when light was limiting (resource-driven regulation) and/or when fouling pressure was low (demand-driven regulation). Since both of these conditions coincide in the winter months, it would be expected that for F. vesiculosus in the Western Baltic, the deployment of chemical antifouling defences at the thallus surface would be lower in winter than in summer. To test this hypothesis, the antimicrobial activity of surface-bound metabolites was quantified every month over a period of 1 year.

Materials and methods
Sample collection and sample sites F. vesiculosus was collected monthly between March 2010 and February 2011 from 2 geographically distinct locations separated by ca 300 km of coast line: Gelting, Germany (54°48′ N/9°44′ E) and Poel, Germany (54°01′ N/11°28′ E). Upon collection, the wet algae were individually sealed in zip-lock bags (filled with ambient water) and transported back to the laboratory in a cooler box.
To gain a general overview of the abiotic regime at Gelting and Poel, the temperature and salinity regime between 1 and 3 m depth was extracted from the GEO-MAR Kiel Baltic Sea Ice Ocean model with a daily resolution (courtesy Andreas Lehmann, GEOMAR Kiel) with regular ground-truthing by, inter alia, in situ measurements using loggers (Star Oddi, Reykyavik, Iceland). Gelting featured a higher salinity, but a similar temperature (yearly mean of 15.7 psu (SD 1.1), and 9.4°C (SD 7.8)) compared to Poel (yearly mean of 13.0 psu (SE 2.1), and 9.9°C (SD 7.9); Figure S1a, b in the Supplementary information). [Supplementary material is available via a multimedia link on the online article webpage.] The salinity difference was most pronounced from May through to September ( Figure S2).

Surface-specific extraction of algae
Surface extraction was performed in less than 12 h after collection. Macroscopic epibionts such as filamentous green algae and mussels were gently removed using tweezers. Subsequently, the algal branches were first spin-dried, weighed and then surface-extracted by dipping them for 10 s into a mixture of methanol (MeOH): hexane (1:1 v/v) as described in Saha et al. (2011). This method has been shown previously to be non-destructive for the epithelial cells of F. vesiculosus (Saha et al. 2011, Figure S4). Larger thalli had to be cut just prior to spindrying and extraction. Care was taken to avoid contact between the cut ends and the solvent mixture to avoid contamination by intracellular compounds. The resulting extract was filtered through a GF/A filter (Whatman, Ø = 15 mm) and the solvent was reduced under vacuum at 30°C using a rotary evaporator. Replication was fivefold for each month and each location, respectively, except for Poel in March and April 2010 and from Gelting in April and November 2010 when only 4 replicates were available.
Calculating the volume of the extracted surface-associated boundary layer Pilot studies had shown that algal growth was more or less isomorphic in the distal thallus parts used for extraction and that 1 g of algal wet weight corresponded to a surface area of 25.57 cm 2 (both faces added, SD ± 1.88, T. Lachnit (pers. comm.)). Thus, algal surface area was determined by multiplying algal wet weight by 25.57 cm 2 g À1 . Subsequently, the extracted algal surface volume was calculated by multiplying the calculated algal surface area by 30 μm (estimated thickness of the water film adhering to the thallus surface after spindrying; Wahl et al. 2010).

Bacteria
Eight common strains of bacteria belonging to the regional pool of microbial foulers were used in the settlement assays: Cytophaga sp. was isolated from the brown macroalga Halidrys siliquosa and also occurred on the brown macroalga Saccharina latissima; Bacillus aquimaris (Yoon et al. 2003) was isolated from the brown macroalga H. siliquosa and also occurred on the brown macroalga Desmarestia aculeate and the red alga Ahnfeltia plicata; Cobetia marina (Arahal et al. 2002) was isolated from seawater; Ulvibacter litoralis (Nedashkovskaya et al. 2003), Alpha proteobacterium DG1293 (Submitted to the International Nucleotide Sequence Database Collection (INSDC) by Green et al. in 2006, unpublished data) and Vibrio sp. SIGA198 (Hardwick et al. 2003) were all isolated from the brown macroalga Fucus serratus; Pseudoalteromonas BSw20057 (Submitted to the INSDC by Ji et al. in 2008, unpublished data) and Alteromonadaceae E1 bacterium (Patel et al. 2003) were isolated from the red alga Polysiphonia stricta. Strains were isolated from the named macroalgae by first rinsing the algal surface twice with sterile seawater (to remove non-attached bacteria), then vortexing fragments of the thallus for 20 s in sterile seawater. The resulting bacterial suspension was log-diluted in 6 steps (down to a dilution of 10 À6 ). From each dilution step, 100 μl were spread onto marine nutrient agar (2.5 g peptone, 0.5 g yeast extract and 15 g agar 1 À1 seawater) and incubated overnight at 28°C. The next day, single colonies were transferred into liquid culture medium (the same composition as before omitting the agar) and grown over night at 28°C on a shaker. From these cultures, a dilution spread on nutrient agar was performed triplicate, and grown overnight at 28°C. Subsequently, a single colony was grown overnight in liquid medium at 28°C, before being transferred to a Rotistore® cryo tube (Carl Roth GmbH, Karlsruhe, Germany) and stored at À80°C. For sequencing, DNA was extracted using an Aqua Pure Genomic DNA isolation kit (Biorad). Bacterial strains were identified by sequencing of their cDNA as described earlier . Strains were isolated and identified by F. Symanowski, GEOMAR.

Anti-settlement assay
Anti-settlement activity was used as the most relevant measure of microfouling control because it combines both repellent and/or toxic effects (Wahl et al. 1994). Bacterial strains were grown in nutrient enriched medium (5 g peptone, 3 g yeast l À1 filtered seawater) for 18-20 h. Prior to the assays, the optical density (OD) of the bacterial cultures was determined with a Beckman Du ® 650 spectrophotometer (λ 600 nm). Bacterial cultures with an OD range of 0.5-0.8 were used. Settlement assays were conducted in multi-well plates (96 polystyrene wells, flat bottom, Greiner ® ). Aliquots (96 μl) of bacterial suspension were added to the wells. Four μl of extract dissolved in DMSO (dimethylsulphoxide) at 25-fold natural concentration were added to the 96 μl of bacterial suspension so that the tested extract in the final mixture were diluted to their natural concentration (ie corresponding to the concentration in the boundary layer covering the algal thallus; Lachnit et al. 2010). Bacteria were not exposed to concentrations of DMSO higher than 5% in order to prevent toxic effects . Wells with DMSO and bacterial suspension served as controls (n = 8). The well plates were incubated for 1 h on a shaking table (100 rpm) at 20°C (Saha et al. 2011). After that, the bacterial suspension was removed from the wells by overturning the plates and unattached cells were eliminated by gently rinsing (Â2) with 100 μl of sterilised filtered seawater. The attached cells were quantified by staining with the fluorescent DNA-binding dye, Syto 9 (0.005 mM) for 10 min in the dark (Invitrogen, GmbH). A washing step was not necessary since the defence strength of F. vesiculosus extracts was calculated by dividing the fluorescence values from the test wells (settled bacteria in the presence of Fucus extract in DMSO plus stain) by the fluorescence value in the control well (settled bacteria in the presence of DMSO plus stain). To account for the autofluorescence of the extracts themselves, extract-only samples (extract in DMSO without bacteria) were measured using the same methods. Thus, the fluorescence attributable to the settled bacteria in the test well was obtained by subtracting extract-only fluorescence from the fluorescence values obtained in the test well containing extract, bacteria and stain. Fluorescence, as a proxy for the relative abundance of settled bacteria, was measured at excitation/emission wavelengths of 477-491/540 nm using a plate reader (Hidex Chameleon IV, Turku, Finland). The 8 strains were tested individually in anti-settlement assays. The extract of each of the 5 individuals of F. vesiculosus was tested in triplicate to account for variability in bacterial settlement rates. The fluorescence values from the 3 replicates were averaged and represented as 1 true replicate. For certain strains of bacteria replicate testing of individual extracts was not possible due to shortage of material. The defence strength was expressed as the 'log effect ratio', ie the logarithm of the fluorescence attributable to settled bacteria of strain Y in the presence of an extract divided by the fluorescence attributable to settled bacteria of strain Y in the absence of that extract (after the appropriate corrections mentioned above). A log effect ratio value of 0 (ie an equal number of bacteria in wells with and without extract) indicates that the tested extract had no effect on settlement, whereas a negative log effect ratio value indicates an inhibitory effect and a positive log effect ratio value indicates an attractive effect. Thus, a log effect ratio of À1 represents a 10-fold reduction whereas a value of +1 represents a 10-fold enhancement of bacterial settlement caused by the presence of the extract.

Fouling regime
Since most planktonic bacteria prefer to recruit onto solid/ liquid interfaces (Grossart 2010), the cell density of bacteria in the plankton may serve as a rough proxy for fouling pressure on macroalgae. Thus, the monthly bacterial density at 1 m depth at Boknis Eck (54°45′N/9°83′E), which is an open-water habitat with a maximum depth of 27 m situated in between the 2 F. vesiculosus collection sites and was assessed in a separate project from 2005 to 2008 (data courtesy of H.G. Hoppe & R. Koppe, GEOMAR) was used as an approximate proxy for microfouling pressure.

Statistical analysis
The strength of anti-settlement activity of extracts on bacterial settlement was expressed as the log effect ratio (see above). Mean effect strength values were obtained by averaging the effect of each replicated extract on each of the 8 bacterial strains. A 2-way ANOVA was used to analyse the effect of seasonality and site on the mean antifouling activity of the macroalgae. A Shapiro-Wilks test was used to determine the normality of the distribution while Levene's test was used to test for homogeneity of variance. Only true replicates (ie extracts of individual algae) were used for the purpose of analysis. Statistica software (Stat Soft, Tulsa, OK, USA) was used to conduct all statistical tests.

Temporal variation in anti-microfouling defence
The anti-settlement defence strength of F. vesiculosus varied among seasons, sites and with regard to target bacterial strains (Figures 1-4). The sensitivity of the various bacterial strains differed by a factor of up to 10, with B. aquimaris and Cytophaga sp. being more sensitive than the other epiphytic bacteria (Figure 4). These relative strain sensitivities were similar at the two sites. Anti-setttlement activities were approximately 1.4 times higher in the population of F. vesiculosus from Poel than in the population from Gelting (2-way ANOVA, F = 18, p < 0.001, Figures 1 and 2, Table 1). The pattern of seasonal variability, however, was similar in the 2 populations of F. vesiculosus.
The mean anti-settlement activity of surface extracts (averaged over all 8 target strains) showed a clear seasonal pattern (2-way ANOVA, F = 5, p < 0.001) with stronger repellence in summer and autumn, and weaker repellence in winter and spring. This pattern was consistent between the 2 locations (Figures 1 and 2, Table 1). Anti-settlement activity in summer/autumn was ca 1.45 and 1.3 times stronger than in winter/spring for Poel and Gelting, repectively. There was no interaction between the two factors, site and season (2-way ANOVA, F = 1.7, p > 0.05) illustrating that defence strength in the 2 sites fluctuated synchronously.   . Seasonal defence strength variation pooled over strains and sites (n = 10, n = 8 for April and November (Gelting) and for March and April (Poel), respectively). Error bars are ±SE. Anti-settlement defence is expressed as the log effect ratio, ie the more negative the value, the stronger the defence.

Microfouling pressure
The density of bacterial cells in the plankton of Boknis Eck follows a quadratic seasonal curve peaking in summer and autumn ( Figure 5) and was about 4 times higher in summer than in winter.

Discussion
Anti-microfouling activity of F. vesiculosus varied among target bacterial strains, sites and seasons. Cytophaga sp. and B. aquimaris were the most sensitive, while C. marina along with strains isolated from the brown alga F. serratus and the red alga P. stricta were relatively less sensitive to the total surface extract (Figure 4). Specific knowledge is insufficient to explain the causes of the differences in sensitivity among strains. However, the fact that the bacterial strains react differently to the surface metabolites of F. vesiculosus may contribute to the commonly observed non-random distribution of epibiotic bacteria. Indeed, F. vesiculosus is known to possess a biofilm that differs conspicuously from the biofilm on the surfaces of other neighbouring macroalgal species, non-living substrata and from the coloniser pool in the surrounding water column (Lachnit et al. 2009). Such an observation, along with the strainspecific sensitivities observed in the present study, is indicative of selective recruitment of bacterial foulers on the algal surface. This is explainable by either strain-specific preferences for certain traits of the algal surface (eg wettability, surface free energy, exudates suitable as an energy source; Wahl et al. 2010) or by the strain specific stimulation or antifouling activities of compounds on the surface of the alga (Saha et al. 2012). Epibacteria associated with basibionts are also known to modulate further microfouling and macrofouling (Holmström et al. 1992;Dobretsov & Qian 2004). Bacterial strains isolated from F. serratus inhibited fouling by diatoms (S. Alpert, GEO-MAR, pers. comm.) as well as by cypris larvae of the barnacle Amphibalanus improvises (Nasrolahi et al. 2012). Similarly, monospecific bacterial films of Shewanella baltica and Pseudoalteromonas arctica, isolated from the red alga P. stricta, reduced the attachment of barnacle cypris larvae (Nasrolahi et al. 2012). Biofilm communities which differed in their composition and were from different environments (Stratil et al. 2013) also differed in their potential to modulate fouling, while all biofilms isolated from the surfaces of Fucus spp. repelled barnacle larvae. The latter effect was stronger by 39 and 23% for biofilms from algae reared at 5 and 15°C, respectively, cf. to the repellency observed for biofilms from algae reared at 20° (Nasrolahi et al. 2012).  While the set of strains used in this study is not identical to that in previous monostrain or multistrain assays, it is apparent that the selective activity of secondary metabolites against microfouling may play an important role in moderating the epibiotic biofilm of the alga Wahl et al. 2012).
The observation that defences in the Poel population were stronger than those in the Gelting population may be due to (1) genetic differences between the two populations, (2) small-scale differences in fouling pressure between the 2 sites (if fouling pressure drives defence strength, which is not yet proven) and/or (3) site differences in salinity ( Figure S2, Supplementary information) or (4) possible differences between the sites regarding grazing, nutrients or hydrodynamic conditions, which have been shown to have the potential to affect chemical defence production in brown seaweeds (Hemmi et al. 2004;Macaya et al. 2005;Yun et al. 2007). In addition, fluctuations in water level at the Poel site often result in the algae being temporarily exposed to air (K. Maczassek and S. Stratil, GEOMAR, pers. comm.). Transient exposure to air facilitates the accumulation of antifouling products at the surface of the thallus (Brock et al. 2007). However, despite this geographical heterogeneity, a similar seasonal pattern of defence strength was found in the 2 populations.
It was postulated in an earlier study , that if the production of antifouling defence compounds competes with other metabolic budgets for limited resources, eg growth rate or reproductive output (Dworjanyn, Wright et al. 2006), defence might be jeopardised when energy input in winter is low (Lehvo 2001) and/or when growth (mainly in early summer) and reproduction (mainly in spring) consume stored or newly produced energy. This investigation showed a seasonal defence pattern with a maximum in late winter/early spring. The apparent contradiction is explainable by the different extraction techniques used. The earlier investigation ) used whole tissue extracts, a method that has been state-of-the-art for decades. The present investigation used a surface-extraction technique that has been optimised in recent years. One interpretation of the contrasting temporal pattern of defence could be that the whole extract mainly contained metabolites responsible for 'internal' defence against pathogens and which was different in structure and seasonal dynamics from the surface-bound metabolites. However, it is considered more likely that the distribution of defence metabolites on the thallus, vs inside the thallus, differs among seasons. The results presented here are more relevant ecologically than those published earlier, since they quantify the effect of the metabolites that potential foulers encounter physically on the surface of the macroalga. However, the earlier results are also interesting, and can be interpreted as indicating the potential for defence, ie the amount of defensive metabolites produced and stored within the alga 'waiting' to be deployed when the need arises at the onset of spring and summer; this scenario assumes that antifouling defence is regulated. If defence metabolites that are produced exude through the thallus epithelium in an uncontrolled manner, the seasonal patterns detected by the 2 extraction procedures should coincide. The fact that they do not may be interpreted as indicating a decoupling between production and deployment of anti-microfouling defences, ie their regulation. This interpretation of the apparently conflicting seasonal defence patterns within and on the surface of algae requires futher research in order to address a number of key questions. (1) Whether the chemical antifouling defences in F. vesiculosus and other macroalgae are costly to an extent that they compete for resources with other metabolic processes.
(2) Can antifouling defence be regulated as suggested by the seasonal pattern described in the present study and if so what is the mechanistic basis for such regulation? (3) Are the drivers for such regulation the availability of free energy and resources (ie that not appropriated for growth or reproduction), and/or the number and nature of fouling organisms contacting the algal surface? If the latter driver is of most consequence how does the alga perceive fouling on the surface of the thallus? The robustness of the seasonal pattern in defence strength at the thallus surface suggests that the fluctuation is not random. The drivers remain to be identified.