Tautogolabrus adspersus (Walbaum, 1792)

Boulding, Christopher J. D. Bender Elizabeth G., Moir, Camden D. & Hajibabaei, Mehrdad, 2025, Morphological and metabarcoding dietary analysis of the cunner wrasse (Tautogolabrus adspersus) revealed significant regional variation, with large overlap between its common prey species and biofouling animals living on salmonid sea cages, Journal of Fish Biology 107 (1), pp. 143-160 : 152-154

4.1 | Dietary composition of T. adspersus across Atlantic Canada

This study represents the first in-depth investigation of T. adspersus dietary ecology using dDNA metabarcoding technology across a geographic range spanning 1086 km. This study also compared results from traditional morphological dietary analysis to metabarcoding using the same set of gut samples. Of the two methods, only metabarcoding routinely identified soft-bodied taxa and often enabled resolution to the species level. We detected statistically significant dietary variation between regional, locational and temporal groupings of individuals. As hypothesized, presence and/or absence of observed prey taxa within the two groups of Newfoundland samples was the most distinct from the Nova Scotian samples, suggesting unique prey taxa were located and consumed within both study sites in Newfoundland. These results suggest that geographical separation and ecological variation in prey taxa diversity contribute to differences in observed diets between the mainland and the island of Newfoundland.

Significantly consumed prey items, in terms of composition by weight (%WC), were mussels, bryozoans, ascidians, gastropods, unidentified material and barnacles (Figure 2). However, the high relative weights of the bivalves and gastropods represented mostly indigestible shell fragments, which artificially inflated their contribution to the nutrition of this wrasse. Thus, the metric %WC, used for only the morphological dataset, was biased towards hard-bodied prey taxa because of their longer gut residence times relative to soft-bodied taxa. Many of the common prey taxa detected here have been previously documented in strictly morphological studies, reinforcing the consensus that T. adspersus is a demersal feeder preferentially consuming readily available, slow-moving or sessile organisms ( Olla et al., 1975; Patricio Ojeda & Dearborn, 1991; Whoriskey, 1983).

By contrast, dDNA metabarcoding identified soft-bodied prey taxa such as polychaetes and sea anemones that had not been detected in the morphological analysis (Figure 3). High values of percentage FOO (%FOO) from the metabarcoding results for genera like Mytilus and Ciona were in agreement with %FOO from the morphological results for bivalves and ascidians, the respective classes for these genera (Figure 3).

Two metrics exclusive to the dDNA metabarcoding (%ESV and RRA) resulted in different rankings of prey taxa from those seen when ranking had been by FOO. Classes including Insecta, Arachnida and Malacostraca made up a high proportion of the total ESVs that were detected in the entire dataset (Figure 4a). Classes including Malacostraca, Insecta and Ascidiacea made up a high proportion of the RRA (Figure 4b). These discrepancies may result from either primer bias (i.e., the overamplification of specific taxa that are particular targets of any one chosen amplicon) or from database bias within the BOLD database itself; BOLD has an excellent coverage of many species (i.e., Canadian insects) which are therefore more frequently assigned to the genus or species level ( Porter et al. 2019), than are Canadian marine taxa which are comparatively poorly covered ( Schultz & Hebert, 2024).

Some of the prey species found in this study were frequently detected in a recent study of natural diet variation in wild populations of two of the five commonly used cleaner wrasses within European salmon aquaculture ( Bourlat et al., 2021). Metabarcoding of gut contents of wild-caught goldsinny ( C. rupestris ) and corkwing ( S. melops ) wrasse collected from three sites across a 12-km marine protected area in southwestern Sweden found large proportions of ASVs (amplicon sequence variants, analogous to the ESVs in the current study) were identified as arthropods in the diet of both species (69% for C. rupestris and 97% for S. melops ). A large percentages of these arthropods belonged to the class Malacostraca (51% for C. rupestris and 90% for S. melops ; Bourlat et al., 2021). Malacostracan ESVs were not as common in the current study on T. adspersus yet represented the third most frequently detected group at the class level, composing 9.81% of all detected ESVs (Table S11). A solely morphological study on European cleaner wrasse diets found that common prey items among ballan wrasse ( L. bergylta ) dietary samples included various hard-shelled species like decapods and bivalves ( Deady & Fives, 1995). Commonalities in the diet of T. adspersus and the diet of the European wrasses used as cleaner fish options highlight a shared ecological niche and dietary breadth, suggesting, based on the success of cleaner wrasses in Europe to combat this issue, the potential for success in the event that T. adspersus is employed in a similar manner in the Northwestern Atlantic.

In the current study, the salmon louse ( L. salmonis ) was uniquely detected in three gut samples (%FOO>1.5%) from three different sites in the metabarcoding dataset specifically, which would not have been detected had only a morphological approach to dietary analysis been employed (Table S4). dDNA metabarcoding detected the salmon louse ( L. salmonis ) inside the ethanol-preserved stomachs of lumpfish ( C. lumpus ) living within sea cages containing Atlantic salmon ( S. salar ) that were frequently heavily infected with lice, but careful morphological analyses of the same sets of stomachs found no evidence of lice ( Roy & Boulding, 2024). The low %FOO of L. salmonis for the morphological analyses within this C. lumpus study is likely a result of crustaceans such as L. salmonis possessing a thinner cuticle, which can be digested at a faster rate than other crustaceans with thicker cuticles such as krill ( Roy & Boulding, 2024). L. salmonis placed in C. lumpus stomachs with tubing had a mean digestion time of 15 h at 9 C ( Staven et al., 2021), and the digestion rate would likely be even quicker inside a stomach-less (agastric) species of wrasse ( Le et al., 2019), such as T. adspersus . This demonstrates that, for taxa with soft body parts, or taxa that possess thinner or frail hard parts that quickly breakdown during digestion, the use of metabarcoding is invaluable as a detection technique. This is especially true when the detection of specific frail taxa (such as problematic species such as the salmon louse within sea cage environments) is of particular interest.

Many of the prey in this study are widely considered pest biofouling species, supporting the integration of T. adspersus as a potential cleaner fish in marine sea cages. The genus Ciona was one of the most frequently identified prey items, especially in more southern samples, and is of particular interest for aquaculture managers in more northern waters given its recent invasion of the area ( Deibel et al., 2014; Ma et al., 2017; McKenzie et al., 2015; Therriault & Herborg, 2008). The large contribution of ascidian sequences (Figure 4b), as well as the frequent detection of Ciona sequences in different samples (Figures 3b and S6), implies a significant contribution of this group to overall dietary composition of T. adspersus , especially for the individuals from southern regions (Figure 4c,d). This suggests potential for T. adspersus to act as a direct method for prevention of the colonization of this invasive species on sea cages.