Muscle cell-type diversification is driven by bHLH transcription factor expansion and extensive effector gene duplications

Abstract

Animals are typically composed of hundreds of different cell types, yet mechanisms underlying the emergence of new cell types remain unclear. Here we address the origin and diversification of muscle cells in the non-bilaterian, diploblastic sea anemone Nematostella vectensis. We discern two fast and two slow-contracting muscle cell populations, which differ by extensive sets of paralogous structural protein genes. We find that the regulatory gene set of the slow cnidarian muscles is remarkably similar to the bilaterian cardiac muscle, while the two fast muscles differ substantially from each other in terms of transcription factor profiles, though driving the same set of structural protein genes and having similar physiological characteristics. We show that anthozoan-specific paralogs of Paraxis/Twist/Hand-related bHLH transcription factors are involved in the formation of fast and slow muscles. Our data suggest that the subsequent recruitment of an entire effector gene set from the inner cell layer into the neural ectoderm contributes to the evolution of a novel muscle cell type. Thus, we conclude that extensive transcription factor gene duplications and co-option of effector modules act as an evolutionary mechanism underlying cell type diversification during metazoan evolution.

Fig. 1. Distinct muscle cell types in the sea anemone Nematostella vectensis are identified from single-cell sequencing data.

a Muscle cell relationships in vertebrates (Bilateria) and the sea anemone (Cnidaria). Two ancestral cell types corresponding to a fast (purple circles) and slow (green circles) contracting phenotype derive from the mesoderm in vertebrates. Muscles can arise from both cell layers in the diploblastic sea anemone, but their contractile properties are undescribed (gray). Both the intra- and interspecies evolutionary relationships of these muscle types are unclear. Ec: ectoderm, Me: mesoderm, En: endoderm, Em: endomesoderm b Schematic view of Nematostella muscle systems: tentacle and mesentery anatomy are illustrated schematically in cross section with the positions of the muscles indicated; blue: tentacle retractor (TR); light green: circular muscle (CM); dark green: parietal muscle (PM); red: mesentery retractor (MR); ocher: the intermuscular membrane (ImM) c Heatmap of differentially expressed genes of the mesentery-derived bulk transcriptomes. Average gene expression from differentially expressed genes with less than a twofold change of expression between library source (mCherry negative non-muscle, two replicates, vs. mCherry+ mesentery retractor muscle cells, two replicates) is imaged. The indicated muscle-related genes are upregulated in the muscle-cell libraries. d Heat map of differentially expressed genes across eight cell populations identified by single-cell RNAseq. The same muscle-related genes are detected in the retractor (1) cluster. Cell clusters carried forward as the muscle subset are indicated in the box. d′ Dimensional reduction cell plot (UMAP) of the full dataset showing expression of the muscle marker myosin heavy chain (myhc-st) in the retractor (1) as well as the gastrodermis (2). e UMAP cell plot of the muscle subset annotated according to cluster identity. Four differentiated muscle cell clusters are identifiable and color-coded as in part (b). Differentiated muscle cells represent approximately one-quarter of the data subset (pie chart). e′ Expression profiles of markers indicative of the retractor muscles (melc4), the bodywall muscles (myph-like8), and the intermuscular membrane (melc3), with rainbow expression profile: gray: no expression; blue: low; red: high.

Fig. 2. Structural protein gene expression reveals two functional classes of muscle cells.

a Dot plot showing expression profiles for the 10 most significant differentially expressed genes for each muscle cluster. Cluster identity is indicated by the same coloration shown in (Fig. 1b, e). The relative expression profile of the same gene set within the bulk dataset of the non-muscle cells (non-M BULK) and the muscle cells (MR BULK) are illustrated as square blocks. Common gene sets that unite the retractor muscle (TR and MR) and the bodywall musculature (PR and CM) are boxed in purple and green, respectively. CM: circular muscle; MR: mesentery retractor; PM: parietal muscle; tentacle retractor: TR. b Validation of paralogous muscle gene sets by in situ hybridization. Schematic (top) and in situ hybridization for genes specific to the retractor muscles (purple boxes) or bodywall muscles (green boxes) are shown. One labeled mesentery is shown for each gene, and scRNA expression profiles are indicated in (c). bwm: bodywall muscle (PM & CM); mes: mesenteries; tb: tentacle buds. The oral pole is to the left in all whole mount images. Scale bars are 50 µm. c Expression profile of paralogous genes illustrating differential use across clusters, set-up as in (a). Fast (purple boxes) and slow (green boxes) paralogs are highlighted. Genes shown in (b) are indicated with an asterisk (*). d Measured contraction speeds for tentacle and body column retraction (blue and red) versus peristaltic bodywall contractions (green). Data are presented as a boxplot of log10 values of measured contraction speeds, illustrating the median, first, and third quartiles, with whiskers indicating the 95% percentile of the data. “*” denotes p-value < 0.001 in paired two-sample Student’s t-test (6e-27 and 3e-30), ns = nonsignificant (p = 0.078); n = 17 (body column retraction), 18 (tentacle retraction), and 51 (peristaltic contractions) independent measurement observations. See Supplementary Data 2—t-test for the full output of the statistical test. e Schematic overview of reconstructed gene family relationships is shown for select muscle proteins. Diversification of these protein families occurred independently in vertebrates and cnidarians. See Supplementary Fig. 6 for full-resolution trees.

Fig. 3. Regulatory profiles of muscle cells indicate bHLH complex as key to cell type individuation.

a Dot plot showing relative expression profiles of selected regulatory genes across the entire dataset. Ectodermal derivatives are indicated in blue. Gene family relationships are indicated according to the color scheme shown in the legend. Gene sets are organized according to expression profile (Muscle Candidates, Tentacle RM, Mesentery RM, Bodywall Slow M). The relative positions of the muscle cells are shown schematically below the dot plot. b Validation of selected profiles by in situ hybridization. Scale bars are 20 µm.

Fig. 4. Gene perturbation selectively eliminates muscle.

Wild-type animals are shown in the upper row panels and mutant animals in the lower row panels. a e-protein−/− mutant: Phalloidin staining in control (upper) and mutant (lower). b nem64−/− mutant: Control animals (upper) show extended tentacles with intact phalloidin-positive retractor muscles, whereas mutant animals (lower) exhibit droopy nonresponsive tentacles and the absence of phalloidin-positive muscle within the tentacle retractors. c Structural marker genes (myhc-st, melc4) are lost within the tentacles of nem64−/− mutant animals (bottom). Outline of animal indicated with a dashed line, arrows: tentacles. d Single-cell sequencing of nem64−/− animal heads lose the retractor muscle cluster present in the wild type (TR & MR: top), but the retractor muscle signal is present in a few cells of the gastrodermis (MR). e Mapping mutant cells onto the wild-type UMAP confirms the presence of MR cells but the loss of TR cells. f The distribution of cell types is otherwise similar between wild type and mutant heads. Scale bars are 100 µm (embryos) and 20 µm (confocal images).

Fig. 5. Summary of anthozoan muscle transcriptome and model of muscle cell evolution.

a The left side shows key regulatory genes and their association with the four muscle types shown in the center panel. The dendrogram on the left indicates the developmental relationships between the cell types. Individual genes within the cell type boxed in the center panel are proposed here as cell-type-specific regulatory factors, whereas those positioned on the dendrogram are shared between descendants of these branches. On the right side, the described effector modules, or aponemes, are indicated. Note that the composition of this list is identical, but the individual paralogs are specific to each effector module. b Hierarchical clustering with expressed DNA-binding proteins vs. structural proteins shows noncongruence between regulatory gene profile and effector gene profile among the four muscle cell types in Nematostella. c Hypothesized events underlying the evolution of myocytes in the sea anemone. The fast-contracting effector module is hypothesized to have evolved within the endo(meso)dermal muscle populations and was then co-opted by the ectodermal epithelium after the radiation of PaTH-related bHLH proteins (Paraxis, Twist and Hand (see Supplementary Fig S1.5), red) and recruitment of one bHLH TF paralog, nem64, to the tentacle ectoderm within the sea anemone lineage.