Independent Innexin Radiation Shaped Signaling in Ctenophores

Abstract

Innexins facilitate cell-cell communication by forming gap junctions or nonjunctional hemichannels, which play important roles in metabolic, chemical, ionic, and electrical coupling. The lack of knowledge regarding the evolution and role of these channels in ctenophores (comb jellies), the likely sister group to the rest of animals, represents a substantial gap in our understanding of the evolution of intercellular communication in animals. Here, we identify and phylogenetically characterize the complete set of innexins of four ctenophores: Mnemiopsis leidyi, Hormiphora californensis, Pleurobrachia bachei, and Beroe ovata. Our phylogenetic analyses suggest that ctenophore innexins diversified independently from those of other animals and were established early in the emergence of ctenophores. We identified a four-innexin genomic cluster, which was present in the last common ancestor of these four species and has been largely maintained in these lineages. Evidence from correlated spatial and temporal gene expression of the M. leidyi innexin cluster suggests that this cluster has been maintained due to constraints related to gene regulation. We describe the basic electrophysiological properties of putative ctenophore hemichannels from muscle cells using intracellular recording techniques, showing substantial overlap with the properties of bilaterian innexin channels. Together, our results suggest that the last common ancestor of animals had gap junctional channels also capable of forming functional innexin hemichannels, and that innexin genes have independently evolved in major lineages throughout Metazoa.

Keywords: ctenophore; gap junctions; innexin; innexon.

Fig. 1.

Ctenophores and innexins. (A) Three of the four ctenophore species in this study. From left to right: Mnemiopsis leidyi, Beroe ovata, and Pleurobrachia bachei (Hormiphora californensis, not pictured, is a tentaculate ctenophore with similar morphology to Pleurobrachia). (B) Based on phylogenomic evidence (supplementary Table 1, Supplementary Material online), Ctenophora is the sister group to the rest of animals. (C) Diagrammatic representation of potential subunit makeup of hemichannels and gap junctions (after Phelan and Starich [2001]). Innexin subunits oligomerize to form a hemichannel. Hemichannels in adjacent cells can dock to form gap junctions. Hemichannels are either homomeric (composed of a single type of innexin) or heteromeric. Gap junctions are homotypic if hemichannels are identical, heterotypic if hemichannels are distinct, and heteromeric if hemichannels differ in subunit composition. Mnemiopsis leidyi photo by Arianna Rodriguez. Other photos by Joseph Ryan.

Fig. 2.

Evolution of ctenophore innexins. (A) Maximum-likelihood tree of innexins from four ctenophore species Mnemiopsis leidyi (Ml), Pleurobrachia bachei (Pb), Beroe ovata (Bo), and Hormiphora californensis (Hc) as well as full sets of innexins from nonctenophores including Hydra vulgaris (Cnidaria), Nematostella vectensis (Cnidaria), Capitella teleta (Annelida), Lottia gigantea (Mollusca), Schistosoma mansoni (Platyhelminthes), Branchiostoma lanceolatum (Chordata), as well as Inx2 from D. melanogaster (Arthropoda). Solid circles at the nodes indicate bootstrap support greater than or equal to 90%. A version of this tree with all bootstrap values and without collapsed outgroup clades is available as supplementary figure 2, Supplementary Material online.

Fig. 3.

Temporal, cellular, and spatial expression of Innexins. (A) Innexin clusters in the genomes of Beroe ovata, Pleurobrachia bachei, Hormiphora californensis, and Mnemiopsis leidyi. Mnemiopsis leidyi has a four-gene Innexin cluster that includes INXA, INXB, INXC, and INXD. The genomes of B. ovata and P. bachei have a cluster that includes INXB, INXC, and INXD. Grey boxes represent three noninnexin genes between INXA and INXB in M. leidyi. All four of these innexins are on chromosome 10 in H. californensis, but there are 226 genes separating INXA and INXB. The clusters are not to scale. In all genomes, INXB, INXC, and INXD are within 20 kb of each other and in M. leidyi, the entire cluster including INXA is less than 80 kb. (B) Temporal gene expression in single M. leidyi embryos during the first 20 h of development (Levin et al. 2016) shows that INXB and INXD are both highly expressed and are tightly coordinated (top section). Comb plates are formed at 8 h postfertilization and at this point there is a spike in expression of several innexin genes that are expressed in comb-plate cell types and tissues (i.e., INXL, INXM, INXG.1, INXH, INXJ, and INXP). (C) Percentage of single cells expressing 0, 1, 2, or >3 innexins (supplementary Table 4, Supplementary Material online). (D) Columns represent metacells from single-cell data (Sebé-Pedrós et al. 2018). Shaded (pink) squares represent expression of the corresponding innexin in 50% or more of the cells that make up the specified metacell (full counts per metacell are in supplementary Table 5, Supplementary Material online). (E) Heatmap of coexpression of each innexin in individual cells. The number of individual cells that express each innexin is in parentheses (in the column header). The percentage was determined by taking the number of cells with coexpression divided by the lowest number when comparing the number of individual cells that express each innexin (full coexpression counts are in supplementary Table 6, Supplementary Material online). (F) Expression of M. leidyi INXB and INXD in three replicates of bulk tissue RNA-Seq from tentacle bulbs (TB), comb rows (CR), and aboral organ (AO). INXB and INXD are shown separately because they are very highly expressed relative to the other innexins. (G) Expression of M. leidyi innexins in three replicates of bulk tissue RNA-Seq from tentacle bulbs, comb rows, and aboral organ. *Metacells C33 and C34 are hypothesized to be neurons (Sachkova et al. 2021). **Metacells C52 are hypothesized to be colloblasts (Babonis et al. 2018).

Fig. 4.

Whole-mount in situ hybridization for four clustered Mnemiopsis leidyi innexin genes. (A) Cartoon depiction of the innexin genomic cluster. (B) Cartoon representation of spatial patterns of expression of INXA-D. Specific patterns within expression domains of INXB-D are not shown due to slight variations in patterns between individuals. Seemingly overlapping domains in cartoons may not indicate coexpression in the same cells. (C–Z) In situ expression of INXA-D. The label to the left of each row describes the view or tissue under focus. Columns correspond to positions of genes in the genomic cluster of A. (C, G, K) INXA is highly expressed in the lateral ridge of the tentacle bulb. (O) INXA is not expressed in the comb rows or underlying canals. (S, W) There are four distinct INXA domains of expression in the aboral organ. (D, H) INXB is widely expressed in the tentacle bulbs, pharynx, aboral organ, and meridional canals underlying the longitudinal comb rows. (L) INXB is expressed in the tentacle bulb, but not the tentacle. (P) There is punctate INXB expression in the meridional canals underlying the comb rows. (T, X) There are four distinct INXA domains of expression in the aboral organ. (E, I) INXC is widely expressed in the tentacle bulbs, comb rows, pharynx, and aboral organ. (M) INXC has a distinct expression domain in the tentacles. (Q) INXC is expressed in the comb rows and in the underlying meridional canals. (U, Y) There are four distinct INXA domains of expression in the aboral organ. (F, J) INXD is widely expressed in the tentacle bulbs, comb rows, pharynx, and aboral organ. (N) INXD is not expressed in the tentacles. (R) There is punctate INXD expression in the meridional canals underlying the comb rows. (V, Z) INXD is expressed in a ring in the aboral organ. Coloration in the tentacles of INXA, INXB, and INXD is likely background. No probe control is in supplementary figure 4, Supplementary Material online. All scale bars represent 100 μm.

Fig. 5.

Activity of putative innexons in Mnemiopsis leidyi muscle cells. Representative whole-cell currents recorded from isolated muscle cells in low (A, B) and high (C, D) intracellular calcium. (A, C) Voltage protocol diagram showing muscle cells were initially hyperpolarized by −50 mV and then 200 ms voltage steps were applied in 10 mV increments. Voltage-gated currents are depicted in cyan (also supplementary fig. S5, Supplementary Material online for details). For example, inward currents characterized by fast activation/inactivation kinetics (arrows) represent activity of voltage-gated sodium channels. Innexin channel activity (red rectangle) represents traces without active voltage-gated channels, which are shown in more detail in B and D. (B, D) Each panel displays portions of current traces obtained at different potentials (as indicated). Horizontal lines depict unitary current levels. Arrows depict possible short-lived subconductance states. Current values (y-axis) represent current values minus the basal current level. (E) Plot of the relationship between current (pA) and voltage (mV) based on the mean values (dark grey circles) of single-channel current amplitudes obtained at different voltages (as in D). Data were obtained from ten cells in total. Each light grey symbol represents the single-channel amplitude estimated for an individual cell. A linear approximation of this relationship corresponds to a slope conductance of ∼340 pS. Empty circles depict the predicted reversal potentials of ideal potassium (V_r, K⁺), chloride (V_r, Cl⁻), and monovalent cation (V_r, X⁺) selective channels in the given experimental conditions. Note, unitary current–voltage relationship suggests nonselective nature of the channel pore. Providing estimates of the relative permeability of the channel for inorganic and organic ions would require further detailed analysis. The blue symbols and lines (right y-axis) in D represent voltage dependence of channel open probability expressed as P_o = I/Ni (where I = integral current, N = number of channels detected in given conditions, i = single-channel amplitude). Voltage dependence of P_o was analyzed for four cells except for −30 mV where n = 1. Extracellular conditions in A and B are identical. Data presented were filtered at 1 kHz and reduced 10-fold.