The round goby genome provides insights into mechanisms that may facilitate biological invasions

Abstract

Background: The invasive benthic round goby (Neogobius melanostomus) is the most successful temperate invasive fish and has spread in aquatic ecosystems on both sides of the Atlantic. Invasive species constitute powerful in situ experimental systems to study fast adaptation and directional selection on short ecological timescales and present promising case studies to understand factors involved the impressive ability of some species to colonize novel environments. We seize the unique opportunity presented by the round goby invasion to study genomic substrates potentially involved in colonization success.

Results: We report a highly contiguous long-read-based genome and analyze gene families that we hypothesize to relate to the ability of these fish to deal with novel environments. The analyses provide novel insights from the large evolutionary scale to the small species-specific scale. We describe expansions in specific cytochrome P450 enzymes, a remarkably diverse innate immune system, an ancient duplication in red light vision accompanied by red skin fluorescence, evolutionary patterns of epigenetic regulators, and the presence of osmoregulatory genes that may have contributed to the round goby's capacity to invade cold and salty waters. A recurring theme across all analyzed gene families is gene expansions.

Conclusions: The expanded innate immune system of round goby may potentially contribute to its ability to colonize novel areas. Since other gene families also feature copy number expansions in the round goby, and since other Gobiidae also feature fascinating environmental adaptations and are excellent colonizers, further long-read genome approaches across the goby family may reveal whether gene copy number expansions are more generally related to the ability to conquer new habitats in Gobiidae or in fish.

Keywords: Adaptation; Detoxification; Epigenetics; Evolution; Fish; Gene duplication; Genomics; Innate immunity; Invasive species; Neogobius melanostomus; Olfaction; Osmoregulation; PacBio; Vision.

Fig. 1.

The round goby, an invasive benthic fish. a Wild-caught round goby in aquaria. Individuals usually feature a light gray or gray-brown mottled coloration, with a characteristic black dot on the first dorsal fin. Adults measure between 8 and 20 cm. During the reproductive season, males may become territorial and develop a black body color (first panel). b Phylogenetic position of the round goby among fishes. The round goby is a neoteleost and member of the Percomorpha. c Current geographical distribution of round goby. The round goby has spread from its native region (green) in the Ponto-Caspian area in Eurasia to many European rivers and lakes, the Baltic Sea, the Great Lakes and their tributaries (orange). d The growing relevance of the species as research model is reflected by increasing publication numbers. Web of Science records on round goby (orange) have risen since its first detection in the Great Lakes in contrast to records on a non-invasive fish with similar ecology (European bullhead, Cottus gobio; gray)

Fig. 2.

Phylogenetic tree of vertebrate opsin gene sequences. Maximum likelihood phylogenetic tree based on the cone and rod visual opsins and using VA opsins and pinopsins as outgroup. The round goby genome contains two LWS gene copies, which seem to be the results of an ancient gene duplication event, and two more recently duplicated RH2 gene copies. Round goby is indicated in orange. Red opsin branches are indicated in red; green opsin branches are indicated in green. Non-teleost species and the outgroup (VA opsins and pinopsins) are indicated in gray. Gray boxes highlight Gobiidae. See Additional file 2: Figure S1 for a tree including expanded SWS1, SWS2, and RH1 branches and Additional file 3: Figure S2 for an exon-based tree

Fig. 3.

Red skin fluorescence in the round goby. Round gobies exhibit red fluorescence above the eyes when exposed to green light.

Fig. 4.

Phylogenetic tree of percomorph olfactory receptor protein sequences. a Phylogenetic relationship among five analyzed percomorph species, i.e., three gobiids (Neogobius melanostomus, Boleophthalmus pectinirostris, Periophthalmodon magnuspinatus), one cichlid (Oreochromis niloticus), and one stickleback (Gasterosteus aculeatus). b Maximum likelihood phylogenetic tree constructed with adrenergic receptors as outgroup. Sequences were identified de novo except for nile tilapia (Oreochromis niloticus; blue). Branches magnified in c and d are highlighted with gray boxes. An expanded view of the tree is available as Additional file 4: Figure S3. c Branch of the 7tm4 family featuring large independent expansions in all species analyzed. d Branch of the 7tm1 family featuring several expansions in Gobiidae (red, orange) that are not paralleled in other percomorph species (blue)

Fig. 5.

Phylogenetic tree of vertebrate CYP protein sequences. Maximum likelihood phylogenetic tree of round goby (Neogobius melanostomus), zebrafish (Danio rerio), human (Homo sapiens), chicken (Gallus gallus), frog (Xenopus laevis), mouse (Mus musculus), and rat (Rattus norvegicus), with 100 bootstraps, rooted with the CYP51 family. Detoxification genes CYP1–3 do not feature expansions, while a family with largely unknown function, CYP8, is expanded to six members (see gray boxes). Non-fish vertebrates are indicated in gray. Gene fragments too short for tree building but attributable to a certain family are indicated by orange half circles next to the root of the respective family

Fig. 6.

Phylogenetic tree of fish aquaporin proteins. Maximum likelihood tree with 100 bootstraps of round goby (Neogobius melanostomus, orange) in relation to cyprinid zebrafish (Danio rerio) and percomorph threespine stickleback (Gasterosteus aculeatus), nile tilapia (Oreochromis niloticus), and blue-spotted mudskipper (Boleophthalmus pectinirostris). The main classes of aquaporins are labeled with human genes names

Fig. 7.

Phylogenetic tree of human and fish sodium/potassium/chloride co-transporter proteins (NKCC). Maximum likelihood tree with 100 bootstraps of round goby (Neogobius melanostomus, orange), zebrafish (Danio rerio), threespine stickleback (Gasterosteus aculeatus), nile tilapia (Oreochromis niloticus), blue-spotted mudskipper (Boleophthalmus pectinirostris), and as non-fish representative human (Homo sapiens, gray). Gobiidae feature more NKCC2 genes (gray box). Potassium/chloride co-transporters (KCC) indicated in gray type and with gray lines on top are used as outgroup

Fig. 8.

The inflammasome pathway. Pathogen-associated patterns are recognized by pattern recognition receptors such as Toll-like receptors at the cell surface (TLRs), or NACHT domain and Leucine-rich Repeat containing receptor (NLRs) in the cytoplasm. This interaction triggers the transcription of cytokine precursors via NFkB, and the activation and assembly of inflammasome components (NLRs, Pro-Caspase-1, and ASC). Inflammasome-activated Caspase 1 then initiates the maturation of cytokines and an acute phase inflammatory response (CRP, APCS proteins), and/or pyroptosis through gasdermin. Several components of the pathway are expanded in the round goby (gene numbers in round goby, or novel groups for NLRs, are indicated in orange)

Fig. 9.

Phylogenetic tree of teleost Toll-like receptor protein sequences. A maximum likelihood phylogenetic tree run with the JTT substitution model and 500 bootstrap replicates on the transmembrane, linker, and TIR domain of all TLRs found in a selected set of teleosts in the Ensembl database, the Atlantic cod genome version 2, and all manually investigated Gobiiformes. A TLR sequence from the lancelet Branchiostoma belcheri was used as an outgroup and the root was placed upon its corresponding branch. Green triangles, Atlantic cod. Orange circles, round goby. Gray box, TLR22 and TLR23

Fig. 10.

Phylogenetic tree of the NACHT domain and Leucine-rich Repeat containing receptor (NLR) nucleotide-binding domain sequences in round goby, blue-spotted mudskipper (Boleophthalmus pectinirostris), zebrafish (Danio rerio), (Ictalurus punctatus), miiuy croaker (Miichthys miiuy), and human (Homo sapiens). Maximum Likelihood phylogenetic tree with 500 bootstraps rooted at the split between NB-ARC (found in APAF) and NACHT domains (present in all the other NLRs). NB-ARC domains from APAF1 orthologs were used as an outgroup. Bootstrap values are shown for nodes that determine an entire cluster. The tree resolves all three major classes of vertebrate NLRs (NLR-A, NLR-B, NLR-C). NLR-A genes were conserved in all analyzed species; no NLR-B genes were found in the gobies. Six groups of NLR-C genes were identified, four of which are exclusive to zebrafish (Danio rerio) (groups 1–4) and two contain only sequences from gobies (groups 5 and 6, gray boxes and bold print). Lineage-specific expansions are displayed with colored endpoints. Within the goby-specific groups, lineage-specific expansions can be seen for both round goby (orange) and blue-spotted mudskipper (Boleophthalmus pectinirostris) (brown). The placement of sparse miiuy croaker genes in group 3 and round goby genes in NLR-A clusters is not well supported and presumably an artifact. The characteristic Walker A motifs are shown next to each subgroup, with group 5 featuring 2 different motifs

Fig. 11.

a Phylogenetic tree of gnathostome ASC protein sequences. Maximum Likelihood phylogenetic tree with 500 bootstraps rooted at the split between tetrapods and ray-finned fish. Tetrapods were used as outgroup. Round goby is indicated in orange. Gobiidae are highlighted with a gray box. The goby sequences form a clear separate cluster, with a large expansion apparent in the round goby. b Phylogenetic tree of gnathostome Caspase 1 protein sequences The Caspase 1 tree comprises all protein sequences annotated as CASP1 in the investigated Gobiidae genomes aligned together with reference sequences from Ensembl and GenBank. The root was placed on the branch containing the mammalian sequences

Fig. 12.

Phylogenetic tree of vertebrate EZH proteins. Midpoint-rooted Bayesian phylogenetic tree. The Australian ghost shark (potential outgroup) is positioned within the poorly supported EZH2 branch. When rooting with Australian ghost shark, teleost EZH2 genes cluster with EZH1 (data not shown). Round goby is indicated in orange

Fig. 13.

Phylogenetic tree of vertebrate DNMT3 proteins. Midpoint-rooted Bayesian phylogenetic tree. The Australian ghost shark (potential outgroup) is positioned among DNMT3A genes. Round goby is indicated in orange. Zebrafish (Danio rerio), the only other fish with well-annotated DNMT3 genes, is indicated in green