Keystone predation and molecules of keystone significance

Abstract

Keystone species structure ecological communities and are major determinants of biodiversity. A synthesis of research on keystone species is nonetheless missing a critical component - the sensory mechanisms for behavioral interactions that determine population- and community-wide attributes. Here, we establish the chemosensory basis for keystone predation by sea stars (Pisaster ochraceus) on mussels. This consumer-resource interaction is prototypic of top-down driven trophic cascades. Each mussel species (Mytilus californianus and M. galloprovincialis) secretes a glycoprotein orthologue (29.6 and 28.1 kDa, respectively) that acts, singularly, to evoke the sea star predatory response. The orthologues (named "KEYSTONEin") are localized in the epidermis, extrapallial fluid, and organic shell coating (periostracum) of live, intact mussels. Thus, KEYSTONEin contacts chemosensory receptors on tube feet as sea stars crawl over rocky surfaces in search of prey. The complete nucleotide sequences reveal that KEYSTONEin shares 87% (M. californianus) or 98% (M. galloprovincialis) homology with a calcium-binding protein in the shell matrix of a closely related congener, M. edulis. All three molecules cluster tightly within the Complement Component 1 Domain Containing (C1qDC) protein family; each exhibits a large globular domain, low complexity region(s), coiled coil, and at least four of five histidine-aspartic acid tandem motifs. Collective results support the hypothesis that KEYSTONEin evolved ancestrally in immunological, and later, in biomineralization roles. More recently, the substance has become exploited by sea stars as a contact cue for prey recognition. As the first identified compound to evoke keystone predation, KEYSTONEin provides valuable sensory information, promotes biodiversity, and shapes community structure and function. Without this molecule, there would be no predation by sea stars on mussels.

Keywords: Mytilus californianus; Mytilus galloprovincialis; Pisaster ochraceus; chemical cue; community ecology; foundation species; glycoprotein; intertidal ecology; keystone predation; mussel; sea star; sensory ecology.

Fig. 1

Series 1 results showing feeding preferences by sea stars for faux prey. Separate experiments were performed, testing either extrapallial fluid (EF), extrapallial fluid bioactive precipitate (EFBP, 142 μg protein per g gel), purified KEYSTONEin (67 μg protein per g gel), or EFBP minus KEYSTONEin (75 μg protein g gel) against bovine serum albumin (BSA, control for organic enrichment; 142 μg protein per g gel) or 0.45 μm-filtered seawater (FSW). N = number of replicate trials in a given experiment; subscript ‘F’ is field, ‘L’ is lab. An asterisk (*) denotes a significant difference between a given treatment and seawater (control) (G-test for homogeneity with Williams’ and Bonferoni’s corrections: G² ≥ 5.74, d.f. = 1, P ≤ 0.016). Results of lab and field experiments (where performed) did not differ significantly across treatments (G² ≤ 0.30, df = 2, P ≥ 0.90). Trials were conducted using faux prey prepared with materials harvested either from (A–D) Mytilus californianus, or from (E–H) M. galloprovincialis.

Fig. 2

Series 2 results showing sea star feeding preferences for EF, EFBP, KEYSTONEin, or seawater faux prey relative to live (intact) mussels. All procedures, protein concentrations, abbreviations and sample sizes are as explained in the Fig. 1 caption. An asterisk (*) denotes a significant difference between a given treatment and seawater (control) (G-test for homogeneity with Williams’ and Bonferoni’s corrections: G² ≤ 5.51, d.f. = 1, P ≤ 0.019). Results of lab and field experiments (where performed) did not differ significantly across treatments (G² ≤ 0.20, df = 3, P ≥ 0.90).

Fig. 3

Structural architectures and domain organizations of Mytilus californianus KEYSTONEin, M. galloprovincialis KEYSTONEin, and M. edulis heavy metal binding protein (as identified in Yin et al., 2005). The signal peptide and conserved motifs (low complexity regions, coiled coils, and globular domains) are denoted by color-specific rectangles. The amino acid position of the N-glycosylation site is indicated by a closed circle, whereas vertical lines (pink) with numbers denote cysteine positions. The estimated location of a disulfide bridge is marked with a horizontal line (pink) connecting cysteine residues. Abbreviation: ‘C1q’ is the first subcomponent of the C1 complex of the classical pathway in invertebrate immune systems.

Fig. 4

Feeding by sea stars in response to faux prey with purified KEYSTONEin and shells from Mytilus californianus (‘M. calif’) or M. galloprovincialis (‘M. gallo’). All procedures, protein concentrations, abbreviations and sample sizes are as explained in the Fig. 1 caption. (A) Faux prey matched shell and KEYSTONEin cue for each mussel species. (B) Faux prey paired KEYSTONEin cue from one species with shell of the other species. Significant agreement between results from experiments in (A) and (B) indicate that sea stars did not discriminate between KEYSTONEin molecules, irrespective of shell type (G-test for homogeneity with Williams’ and Bonferoni’s corrections: G² ≥ 0.13, d.f. = 1, P ≥ 0.718). An asterisk (*) denotes a significant difference between a given treatment and seawater (control) (G-test for homogeneity with Williams’ and Bonferoni’s corrections: G² ≥ 7.20, d.f. = 1, P ≤ 0.0072).

Fig. 5

SDS-PAGE and immunohistochemical (IHC) surveys of KEYSTONEin expression in select mussel tissues. (A) SDS-PAGE profiles of the extrapallial fluid bioactive precipitates and purified KEYSTONEin molecules from Mytilus californianus (‘M. calif’) and M. galloprovincialis (‘M. gallo’). (B, D) Immunolabeling of KEYSTONEin in tissue sections of M. californianus. Red/brown color indicates the presence of a monoclonal antibody raised against a KEYSTONEin epitope, whereas blue represents hematoxylin counter-stain of cell nuclei. IHC reactivity is limited to epidermal cells in association with the mantle epithelium (Me), columnar epithelial cells (Cec) and foot epithelial layer (El), as well as with an organic shell coating (periostracum, P) secreted by the epidermis. No reactivity, however, occurs between the IHC label and muscle fiber bundles (Mfb), or other tissues/products of non-epidermal origin. Abbreviations: Pg, periostracal groove; Mc, mucus-secreting cell. (C, E) The absence, here, of red/brown color indicates no cross-reactivity between non-immune mouse serum (control) and mussel proteins. Abbreviation: Pgl, Periostracal gland. (F) IHC reactivity is found in the organomineral material, newly-deposited along the shell margin (Sm) of a live mussel. (G) In contrast, no IHC reactivity occurs along the shell margin of a dead (empty) mussel shell and in the absence of new, active organo-mineral deposition. Results are representative of five replicate surveys per tissue, and each image was captured from a transverse section and thin slice (7 μm-thick). Scale bars = 100 μm.