INTRODUCTION
Marine fouling is a costly problem: $5 billion per year to the US maritime industry alone (Bhadury & Wright, Reference Bhadury and Wright2004). As a result, there is a large volume of research into commercial fouling covering several aspects of marine fouling including microbial fouling (Dempsey, Reference Dempsey1981), macrofouling (Wieczorek et al., Reference Wieczorek, Murray and Todd1996; Qian et al., Reference Qian, Rittschof, Sreedhar and Chia1999), and inhibition of fouling, both by synthetic means (Wisely, Reference Wisely1962) and using natural products (Bhadury & Wright, Reference Bhadury and Wright2004; Dobretsov et al., Reference Dobretsov, Dahms and Qian2006). There is also research on succession through the life-span of fouling assemblages (Miller et al., Reference Miller, Rapean and Whedon1948; Holmstrom & Kjelleberg, Reference Holmstrom and Kjelleberg1994; Harder et al., Reference Harder, Lam and Qian2002). However, relatively little published research has considered the fouling of biotic surfaces as opposed to abiotic ones, even though recent studies have focused on invertebrate natural products and particularly their antifouling activity (Dobretsov et al., Reference Dobretsov, Dahms and Qian2006).
Bryozoans are dominant macrofouling species within fouling assemblages (Key et al., Reference Key, Jeffries and Voris1995, Reference Key, Jeffries, Voris and Yang1996). Bryozoan colonies are modular, consisting of a large number of units, called zooids. These sessile, colonial animals are found in a diverse range of marine environments, attaching on both abiotic and biotic surfaces. Several marine species have developed defence mechanisms against bryozoan fouling, for example the marine macroalga Dilsea carnosa (Schmidel) Kuntze, which sheds its epidermis to remove the fouling bryozoans Electra pilosa (Linnaeus) and Membranipora membranacea (Linnaeus) (Nylund & Pavia, Reference Nylund and Pavia2005).
Bryozoans foul other marine surfaces, but in turn are also the subject of marine fouling, both at the macro- and micro-level. Macrofouling epizoites have been described to restrict the growth of Flustra foliacea (Stebbing, Reference Stebbing1971). A number of reports have illustrated the types of marine micro-organisms that foul bryozoan surfaces (Scholz & Krumbein, Reference Scholz, Krumbein, Gordon, Smith and Grant-Mackie1996; Palinska et al., Reference Palinska, Scholz, Sterflinger, Gerdes and Bone1999; Gerdes et al., Reference Gerdes, Kadagies, Kaselowsky, Lauer and Scholz2005). High levels of microbial diversity have been reported from four main studies on microbial interactions with bryozoans. These include the ctenostome Amathia wilsoni Kirkpatrick (Walls et al., Reference Walls, Ritz and Blackman1993) and the cheilostomes Orthoscuticella ventricosa (Busk), Cellaria pilosa (Kirchenpauer), Bugularia dissimilis (Busk) (Walls et al., Reference Walls, Ritz and Blackman1993), Bugula pacifica Robertson and Tricellaria occidentalis (Trask) (Shellenberger & Ross, Reference Shellenberger and Ross1998), Aspidelectra melolontha (Landsborough), Electra pilosa Linnaeus, Electra monstachys (Busk) and Conopeum reticulum (Linnaeus) (Kittelmann & Harder, Reference Kittelmann and Harder2005) and Flustra foliacea (Linnaeus) (Pukall et al., Reference Pukall, Kramer, Rohde and Stackebrandt2001).
Studies on Tasmanian bryozoans, A. wilsoni, O. ventricosa, C. pilosa and B. dissimilis showed differential levels of bacterial fouling between these species (Walls et al., Reference Walls, Ritz and Blackman1993). This study also noted that the distribution of bacteria throughout A. wilsoni colonies was not homogeneous, but did not attempt to measure this aspect of differential fouling.
Temporal studies on T. occidentalis and B. pacifica showed that quantitative estimates of bacterial growth on bryozoan surfaces were not constant over a one month period (Shellenberger & Ross, Reference Shellenberger and Ross1998).
The study by Kittelmann & Harder (Reference Kittelmann and Harder2005) on the encrusting bryozoan species A. melolontha, E. monostachys and E. pilosa collected from the Jade Bight, North Sea indicated that bryozoan colonies were largely free of epibiotic bacterial colonizers. However, bacterial counts of 10 pseudoreplicate samples taken from the surface of C. reticulum yielded approximately 4.8 × 107 bacterial cells cm−2.
A study of fouling on the erect, laminar species, F. foliacea collected from 3 locations approximately 9 nautical miles north-east from the island of Helgoland in the North Sea, showed that microfouling was most prevalent at the proximal end of the zooids and around the operculum area (Pukall et al., Reference Pukall, Kramer, Rohde and Stackebrandt2001). This is of interest because the presence of differential fouling suggests that there may be a control mechanism for anti-fouling allowing the manipulation of microbial films by the bryozoan colonies (Shellenberger & Ross, Reference Shellenberger and Ross1998). Flustra foliacea is also known to produce metabolites that can interfere with bacterial communication (quorum sensing) systems and this may be a mechanism by which fouling inhibition occurs (Peters et al., Reference Peters, Konig, Wright, Pukall, Stackebrandt, Eberl and Riedel2003). It is unclear whether the differential fouling observed in F. foliacea is species-specific or whether this is a general trend in Bryozoa. In the F. foliacea study (Pukall et al., Reference Pukall, Kramer, Rohde and Stackebrandt2001) evidence showed a qualitative difference in fouling within a zooid.
In our study we investigate the possibility of differential fouling on the surface of another cheilostome, Pentapora fascialis (Pallas), a sublittoral species with a distinctive, three-dimensional, approximately hemispherical structure in the Atlantic form (Figure 1). The Mediterranean form has a smaller, more reticulate colony type. Individual colonies of P. fascialis may grow up to 1 m in diameter in British waters, reaching much larger sizes in the Mediterranean where colonies often fuse together to form extensive reef-like structures 2–3 m across. Pentapora fascialis colonies are usually found below 18 m on current-swept rocky ground or boulders around the west and south coasts and into the English Channel. Pentapora fascialis may be a dominant part of the sessile benthos on rocky ground with fast currents, with colonies often reaching a frequency of 1 colony per m2 (Hayward & Ryland, Reference Hayward and Ryland1999). The geographical range of P. fascialis in the British Isles extends from Shetland in the north, around the west coast and south up into the English Channel (Jackson, Reference Jackson2006). This species is of ecological importance in providing shelter for many invertebrate species, including other bryozoans. The foliose shape of P. fascialis provides a refuge from strong currents, as well as increasing the spatial complexity of rocky reef environments. When observed in situ, colonies of P. fascialis appear distinctly more orange in the new growing tips of the colony than in older, more basal parts of the colony, which are often dull grey-orange in colour.
Fig. 1. Photograph of a single Pentapora fascialis colony in situ at Skomer Island Marine Nature Reserve. Scale bar = 35 cm.
In this study we use a fully quantitative approach, to investigate differential fouling of bryozoan surfaces using P. fascialis as a model. Two main questions are addressed: (i) are the proximal regions of individual zooids more heavily fouled than distal areas of zooids, as is the case in F. foliacea; and (ii) are newer parts of colonies less fouled than older parts, due to their shorter exposure period?
MATERIALS AND METHODS
Collection and preparation
A fragment from each of five P. fascialis colonies (Figure 1) was hand collected from the Waybench location, North Wall of Skomer Island Marine Nature Reserve (MNR), Wales, UK (OS ref: SM 724 094) by the Skomer MNR dive team in June 2005. Fragments of P. fascialis were carefully excised from living colonies and packaged in situ in local seawater in individual zip-lock bags. Sub-samples of material from fragments collected from 5 colonies were kept in seawater from the collection site for 24 hours before rinsing 3 times with sterile distilled water (SDW) for 10 minutes. Rinsed samples were immersed in 2.5% (v/v) glutaraldehyde in phosphate buffered saline (PBS) for 1 hour to fix. Fixed samples were washed in PBS twice before dehydrating using an alcohol series from 30% to 100% ethanol. The ethanol was then substituted with acetone and the samples were critical point dried using CO2. Dried samples were attached to 10 mm diameter microscopy stubs using quick-drying silver paint before sputter-coating with 4 nm platinum/palladium. Following analysis specimens were lodged at the BM NH (2007.11.6.1–9).
Image capture
Samples were examined using a Jeol 840 scanning electron microscope (SEM) with a 7 mm working distance, at an accelerating voltage of 5 kV. Digital images were captured at a magnification of ×1000 using the software PrinterFace (Energy Beam Sciences). Within each colony, 2 regions were designated: the 3–4 rows of zooids closest to the growing edge of the colony and zooids at least 10 zooid rows proximal to the growing region. Within each of these areas 5 zooids were randomly selected. Six separate images were taken of the external surface from each of these 5 zooids; 3 images from the proximal end and 3 images from the distal end of each zooid. This gave a total of 300 images for analysis.
Image analysis
Each of the 300 images was analysed manually using Microsoft Office Picture Manager (Microsoft, 2003). Each image was analysed qualitatively for overall impression of level of fouling and also the nature of fouling that occurred. Factors considered were whether the fouling surface coverage was continuous or discontinuous (or patchy), whether the fouling was a simple monolayer or a matrix and the fouling particle types and sizes. These observations were made for zooids from different colonies (Colony, C), from zooids in two different positions within colonies (growing edge and ten zooids back from the growing edge) (Age, A) and from regions within zooids (proximal and distal) (Location, L).
For quantitative analysis of fouling coverage each image was resized to 20 × 18 cm and a regular grid of 336 dots (spaced 1 cm apart) was overlaid. Percentage cover was calculated by dividing the number of dots which directly overlaid surface fouling by the total number of dots (336). Some images contained regions that were difficult to analyse in this way, for example when there were large cracks in the bryozoan surface. Where dots overlaid such regions they were excluded from the calculation.
Statistical analyses
Levene's test for equal variance was used to determine the suitability of the data for parametric analysis. A balanced ANOVA of the percentage surface cover was performed using Minitab 14 (Minitab Ltd., 2006) on the raw data without data transformation, using the percentage surface cover as the variable and ‘Colony’, ‘Age’, ‘Zooid’ and ‘Location’ as the factors. Only the valid interaction ‘Colony x Age’, and the three main factors ‘Colony’ (C), ‘Age’ (A) and ‘Location’ (L) were tested.
RESULTS
Qualitative analysis
COLONY
Figure 2 illustrates the difference in fouling cover between colonies 2 and 3, particularly in the older zooids (B & D). The type of cover shown in Figure 4A & C is an agglutinated biofilm; according to Gerdes et al. (Reference Gerdes, Kadagies, Kaselowsky, Lauer and Scholz2005) this is the most common type of microbial fouling. Colonies 1 and 2 have a monolayer in comparison with colonies 3, 4 and 5 which have a complex, multilayered biofilm. Fouling coverage is discontinuous in colonies 1, 2 and 5, whereas fouling coverage is continuous in colonies 3 and 4. The particle size and type in colonies 1 and 2 is quite different from colonies 3, 4 and 5; the mixture of particles on the surface of colonies 1 and 2 is less complex and the particles appear to be smaller.
Fig. 2. Typical micrographs taken from colonies 3 (A & B) and 2 (C & D). Images A & C are taken from new zooids at the growing edge of the colony; these surfaces are substantially fouled. B & D are images of zooids from the centre of the colony; these surfaces show partial fouling. Images of colony 3 are representative of colonies 3, 4 and 5. Colony 2 images are representative of both colonies 1 and 2. The scale bar in each image is 10 µm.
AGE OF ZOOIDS
Figure 2 shows the difference between the fouling observed from images of zooids selected from the growing edge of colonies (A & C) compared to zooids selected from the older part of colonies (B & D). The images of the older parts of the colonies clearly show a decrease in surface coverage in comparison with newer zooids from the colony edge. Whereas the older regions consist of a single thin layer of fouling, newer regions have multilayered fouling. The multilayered fouling of newer zooids is a matrix of debris, diatoms, bacterial cells and other fouling particles. In contrast, the fouling on older zooids consists of occasional identifiable micro-organisms such as a diatom.
LOCATION WITHIN ZOOID
There is no appreciable qualitative difference in fouling coverage between distal and proximal regions within the zooids. The surface coverage appears to be the same and the nature of the coverage is also similar.
Quantitative analysis
Levene's test showed that variances between samples were equal (H0: variances of percentage surface cover between zooids taken from different colonies were equal; P = 0.1, N = 300, d.f. = 299).
INTERACTION COLONY X AGE
ANOVA showed that the interaction ‘Colony x Age’ is significant (H0: no interaction between factors ‘Colony’ and ‘Age’ in terms of proportion of surface fouled; P = 0.001, N = 300, d.f. = 4). The graph shown in Figure 3 illustrates that the difference in surface coverage between old and new zooids varies between colonies. The smallest difference is observed in colony 4, and the largest difference is observed in colony 2. Although there is some evidence of interaction, fouling cover in new zooids is always greater than fouling cover in older zooids from the same colony. Therefore the factors age and colony are also considered separately, particularly as the F values of these factors are at such a high level of significance (P = 0.001).
Fig. 3. Bar graph illustrating colony differences split by age of zooids. Dark shading indicates older zooids, light grey shading indicates new zooid growth. Error bars show 1 standard deviation.
COLONY
Table 1 shows that this factor is significant at P = 1.06 × 10−5, N = 300, d.f. = 4 (H0: there is no difference in the proportion of zooid surface fouled between colonies), therefore at least one colony is significantly more or less fouled than one or more other colonies.
Table 1. Estimated surface fouling cover split by colony, age of zooids and position within the zooids. All 3 factors shown here are significant at or below the level of P = 0.001 in ANOVA tests.
Results of Tukey's test (Table 2) showed that colony 1 is significantly less fouled than colonies 3, 4 and 5. The level of fouling observed on colony 2 is not significantly different to that of colony 1 or colonies 3 and 5, but colony 2 is less fouled than colony 4. Fouling coverage on colonies 3, 4 and 5 are not significantly different.
Table 2. Results of Tukey's test. Underlines indicate series where no significant difference in fouling is evident.
AGE
ANOVA shows that there is a significant difference in fouling between older zooids and the areas of new growth (H0: no difference in proportion of surface fouled between old and new regions of bryozoan colonies; P = 4.46 × 10−12, N = 300, d.f. = 1). Results are given in Table 1, showing that in all colonies, the older parts were less fouled than the newer parts.
LOCATION WITHIN ZOOID
Table 1 illustrates that the average level of surface fouling in the distal areas of the zooids is approximately 87%. This is 2% lower than the average level of fouling found in the proximal areas. Although this difference is small, it is significant at P = 0.001, N = 300, d.f. = 50 (H0: no difference in proportion of surface fouled between proximal and distal end of individual zooids) in an ANOVA.
COMPARISON OF FOULING STUDIES
Table 3 summarizes the published accounts of the extent of fouling in a range of bryozoan species with various colony morphologies and from different geographical areas. The present study provides the first systematic quantitative analysis of microbial fouling on a bryozoan surface.
Table 3. Previous studies of bryozoan fouling.
Y, yes; N, no; 1 10 pseudoreplicates examined; 2 results recorded from a single zooid; 312 replicates from each colony, one colony examined per week for 4 weeks.
DISCUSSION
A previous, largely qualitative, study on fouling of F. foliacea by Pukall et al. (Reference Pukall, Kramer, Rohde and Stackebrandt2001) (Table 3) had indicated that there may be differential fouling of bryozoan zooids. We aimed to undertake a comprehensive analysis of microbial fouling distribution of biotic surfaces to gain systematic quantitative evidence focusing on the model bryozoan P. fascialis.
Observations of fouling on a F. foliacea zooid (Pukall et al., Reference Pukall, Kramer, Rohde and Stackebrandt2001) suggested that the proximal region was more heavily fouled than the distal region. In our extended analysis of a total of 50 zooids taken from 5 colonies of P. fascialis, the increased fouling in proximal regions of zooids compared to distal regions was found to be statistically significant. It has been speculated that high levels of fouling encrusting the operculum might interfere with the feeding activity of the zooid (Pukall et al., Reference Pukall, Kramer, Rohde and Stackebrandt2001) and therefore inhibition of surface fouling through production of antimicrobial compounds in this sensitive region may be advantageous to the bryozoan.
Walls et al. (Reference Walls, Blackman and Ritz1991) demonstrated that metabolites from the bryozoan Amathia wilsoni are differentially distributed and Shellenberger & Ross (Reference Shellenberger and Ross1998) found higher levels of antibacterial metabolites in bryozoan species which were less fouled than other species. Metabolites produced within Bryozoa may be responsible for anti-predation activity (Montanari et al., Reference Montanari, Fenical, Lindquist, Lee and Clardy1996); some natural compounds extracted from Bryozoa have been shown to exhibit various activities including antimicrobial, nematocidal and larvicidal activity (Sharp et al., Reference Sharp, Winson and Porter2007).
In B. pacifica and T. occidentalis (Table 3), lower levels of fouling had previously been observed at the colony tips (Shellenberger & Ross, Reference Shellenberger and Ross1998). We investigated the distribution of microbial fouling on P. fascialis zooids of different ages as we had observed general differences in colour in new growing tips of the colony compared to older, more basal parts of the colony. It is possible that a difference in fouling at this level could arise because new zooids have been subjected to a shorter exposure period than older zooids. In our study, statistical analysis indicated that the greater fouling of younger than older zooids was significant. The complexity and depth of fouling cover was also increased. It has been previously documented that some bryozoan species shed and replenish the exterior cuticle (Winston & Hakansson, Reference Winston and Hakansson1989). However, if P. fascialis zooids were able to shed their outer cuticle, similar levels of fouling might be expected on old and new surfaces. Although we have observed this in SEM photographs of the encrusting bryozoan species, Umbonula littoralis Hastings, no evidence for loss of the cuticle was seen in P. fascialis.
At the level of the bryozoan colony we found a significant difference in the proportion of zooid surface fouled between colonies. This indicated that at least one colony was significantly more or less fouled than one or more of the other 4 colonies. The colonies from which fragments were collected were of different sizes and therefore of different ages and this may go some way to explaining the differences observed. Some variation may also be expected to occur depending on which side of the bryozoan frond the zooids are sampled (although this factor was not analysed in our study). In this study we have focused specifically on obtaining quantitative information about the distribution of microbial fouling on a biotic surface. However, although outside the scope of the present study, it would also be of interest to characterize the bacterial species present using 16S rDNA analysis.
Studies of similar scope to this one could be performed with a range of other Bryozoa with different colony morphologies. This would help to determine whether the species, colony type, three dimensional structure and habitat contribute to the distribution of microbial fouling on colony surfaces. We can speculate that deterrence of biofilm formation on the surface of bryozoans could be caused by differential spatial production of antibacterial metabolites, or compounds which inhibit bacterial quorum sensing control, produced by the bryozoan or microbial endosymbionts. Differential metabolite production has been shown to occur in A. wilsoni, although this has not been conclusively linked to differential fouling within colonies (Walls et al., Reference Walls, Blackman and Ritz1991). Mediterranean populations of P. fascialis are known to produce 3 secondary metabolites, pentaporins A, B and C, although no antimicrobial or antifouling activity is recorded from these compounds (Eisenbarth et al., Reference Eisenbarth, Gehling, Harder and Steffan2002). An interesting question arising from the study is what mechanisms are responsible for the observed localized fouling on P. fascialis? In a study of the bryozoan F. foliacea by Peters et al. (Peters et al., Reference Peters, Konig, Wright, Pukall, Stackebrandt, Eberl and Riedel2003) using quorum sensing based biosensors inducible by N-acyl homoserine lactones (Winson et al., Reference Winson, Swift, Hill, Sims, Griesmayr, Bycroft, Williams and Stewart1998; Steidle et al., Reference Steidle, Sigl, Schuhegger, Ihring, Schmid, Gantner, Stoffels, Riedel, Givskov, Hartmann, Langebartels and Eberl2001), extracts of the bryozoan were shown to interfere with the quorum sensing process. These types of interference mechanisms are currently of considerable interest because it is likely that through understanding the processes employed by Bryozoa to reduce fouling on their surfaces we may find a more diverse range of natural compounds for the control of commercial marine fouling on abiotic surfaces.
ACKNOWLEDGEMENTS
We acknowledge the assistance of Dr David Causton for advice on statistical methods. We also thank Dr Peter Hayward for editorial expertise and comments on this manuscript.
