Bridging the muscle genome to phenome across multiple biological scales

Abstract

Muscle is highly hierarchically organized, with functions shaped by genetically controlled expression of protein ensembles with different isoform profiles at the sarcomere scale. However, it remains unclear how isoform profiles shape whole-muscle performance. We compared two mouse hindlimb muscles, the slow, relatively parallel-fibered soleus and the faster, more pennate-fibered tibialis anterior (TA), across scales: from gene regulation, isoform expression and translation speed, to force-length-velocity-power for intact muscles. Expression of myosin heavy-chain (MHC) isoforms directly corresponded with contraction velocity. The fast-twitch TA with fast MHC isoforms had faster unloaded velocities (actin sliding velocity, Vactin; peak fiber velocity, Vmax) than the slow-twitch soleus. For the soleus, Vactin was biased towards Vactin for purely slow MHC I, despite this muscle's even fast and slow MHC isoform composition. Our multi-scale results clearly identified a consistent and significant dampening in fiber shortening velocities for both muscles, underscoring an indirect correlation between Vactin and fiber Vmax that may be influenced by differences in fiber architecture, along with internal loading due to both passive and active effects. These influences correlate with the increased peak force and power in the slightly more pennate TA, leading to a broader length range of near-optimal force production. Conversely, a greater force-velocity curvature in the near-parallel fibered soleus highlights the fine-tuning by molecular-scale influences including myosin heavy and light chain expression along with whole-muscle characteristics. Our results demonstrate that the individual gene, protein and whole-fiber characteristics do not directly reflect overall muscle performance but that intricate fine-tuning across scales shapes specialized muscle function.

Keywords: Mouse; Myosin isoforms; Shortening velocity; Soleus; Tibialis anterior; Transcriptomics.

Fig. 1.

Ergometry approach. (A) Soleus tested in vitro in a Ringer's solution bath. (B) Tibialis anterior (TA) tested in situ with preparation viability maintained by cardiovascular supply from the mouse. (C,D) Example data traces for a force–velocity trial and the measurements extracted (i, initial force, F_initial; ii, active force, F_act) to generate summary curves. (E,F) Measurement extraction for force–length (FL) (Figs 5 and 6) and force–velocity–power (FVP) (Fig. 7) relationships, respectively. F_final, final force; L_o, optimal length; L_slack, muscle slack length; P_o, peak isometric force; W_plateau, FL plateau width; β, passive muscle stiffness; V_max, peak contraction velocity; V_opt, velocity of peak power production; P_max, peak power; a/P_o, force–velocity curvature.

Fig. 2.

Transcriptomic signatures of the soleus and TA. (A) Topological distribution of the biological replicates of the soleus and TA create two distinct clusters across the first two principal components (PC) calculated for the differentially expressed genes. (B) Differentially expressed genes ranked by PC1 feature loading. The blue horizontal line indicates the cutoff value (0.07), which was utilized to extract the genes with the strongest contribution to PC1. (C) Heat map visualizing the level of relative up (red) or down (blue) regulation of selected genes in soleus and TA samples (Table S1). Columns represent biological replicates, rows represent genes. Color intensity in each sample is proportional to the expression deviation from the mean normalized count of the gene as the variance stabilizing transformation (VST).

Fig. 3.

Slow myosin expression is exclusively found in the soleus. (A) Expression levels of the four myosin isoforms in the TA and soleus shown in fragments per kilobase million (FPKM). The gene encoding the slow isoform Myh7 was expressed more in the soleus samples while genes encoding the fast isoforms Myh4 and Myh1 were primarily expressed in the TA samples (N=5 soleus, N=5 TA). (B) Myosin isoform distribution quantified by SDS-PAGE using densitometric analysis in ImageJ shows only fast isoforms (IIx and IIb) in the TA and a mixture of fast (IIa) and slow (I) isoforms in the soleus (N=3/muscle). (C) An unaltered representative image of the SDS-PAGE myosin gel (8% with 35% glycerol) with chicken pectoralis (ch.sk.), mouse cardiac, soleus and TA samples.

Fig. 4.

Mean actin sliding velocity from purified naive soleus and TA myosin using an in vitro motility assay. (A) ATP dependence (0–1 mmol l⁻¹) of actin sliding velocity (V_actin), fitted to the Michaelis–Menten equation, shows different mean V_actin and K_m for the soleus and TA. (B) Mean V_actin at 1 mmol l⁻¹ ATP shows a 3.3-fold difference between TA and soleus myosin. Data are the average of three technical repeats (respective myosin purified from 5 mice muscle pairs), with error bars representing s.d.

Fig. 5.

Comparison of active force-length properties for the soleus and TA. (A) Muscle mass-specific force plotted against strain (ε; muscle length normalized to tetanic L_o) for tetanic and twitch contractions. Solid symbols represent females; open symbols represent males. Note that L_o is left-shifted by ∼8% as activation increases (white arrow, see Results for details). (B,C). Plateau width for the ascending limb, measured at 90% P_o for (B) tetanic and (C) twitch contractions. (D–F) Box plots showing summary median-quartile and range data with individual measurements indicated by symbols for (D) muscle mass-specific tetanic force, (E) muscle mass-specific twitch force and (F) the ascending plateau width (W_plateau) at 90% P_o for the active tetanic (white boxes) and twitch (gray boxes) FL curves for the soleus (Sol) and TA. All summary variable comparisons between soleus and TA, except in F, are statistically significantly different (P<0.01; see Materials and Methods for GLM model details). Muscle sample sizes are shown next to curves.

Fig. 6.

Comparison of passive FL properties for soleus and TA. (A) Muscle mass-specific passive force plotted against strain (muscle length normalized to tetanic L_o). (B–D) Box plots showing summary median-quartile and range data with individual measurements indicated by symbols for (B) passive force at 10% beyond L_o, (C) muscle slack length and (D) slope of the passive FL relationship (β), measured from L_o to 110% L_o, representing the relative difference in stiffness of the two muscles. (E) Percent spliced-in index (PSI) plot for the PEVK regions of soleus and TA titin, showing the mean PSI values for six biological replicates of either TA or soleus with the shaded regions representing the s.d. The relatively lower values for the TA than the soleus (1=100% inclusion, 0=complete exclusion of an exon from the transcript pool) suggest expression of shorter titin isoform(s) in the TA than soleus. Exon numbering based on NCBI entry BN001114.1 and Ensemble titin transcript ENSMUST00000099981.9.

Fig. 7.

Comparison of FVP relationships for soleus and TA. (A) Force–velocity relationships, normalized to muscle mass and optimal length (L_o), respectively. Summary force–velocity curves (bold lines) for each muscle are from a custom implementation of the Hill equation in Igor Pro. (B–D) Box plots showing median-quartile and range data with individual measurements indicated by symbols, summarizing force–velocity data for (B) peak unloaded velocity, (C) force at optimal length (see dashed lines in A) and (D) force–velocity curvature (lower values indicate more force–velocity curvature, as defined by the Hill constant, a/P_o). (E) Power–velocity relationships, normalized to muscle mass and optimal length, respectively. (F) Muscle mass-specific peak power and (G) optimal velocity, where peak power occurs (dashed lines in E). Solid symbols represent females; open symbols represent males. All summary data comparisons between soleus and TA are statistically significantly different (P<0.01; see Materials and Methods for GLM model details).