Exercise-dependent formation of new junctions that promote STIM1-Orai1 assembly in skeletal muscle

Abstract

Store-operated Ca²⁺ entry (SOCE), a ubiquitous mechanism that allows recovery of Ca²⁺ ions from the extracellular space, has been proposed to limit fatigue during repetitive skeletal muscle activity. However, the subcellular location for SOCE in muscle fibers has not been unequivocally identified. Here we show that exercise drives a significant remodeling of the sarcotubular system to form previously unidentified junctions between the sarcoplasmic reticulum (SR) and transverse-tubules (TTs). We also demonstrate that these new SR-TT junctions contain the molecular machinery that mediate SOCE: stromal interaction molecule-1 (STIM1), which functions as the SR Ca²⁺ sensor, and Orai1, the Ca²⁺-permeable channel in the TT. In addition, EDL muscles isolated from exercised mice exhibit an increased capability of maintaining contractile force during repetitive stimulation in the presence of 2.5 mM extracellular Ca²⁺, compared to muscles from control mice. This functional difference is significantly reduced by either replacement of extracellular Ca²⁺ with Mg²⁺ or the addition of SOCE inhibitors (BTP-2 and 2-APB). We propose that the new SR-TT junctions formed during exercise, and that contain STIM1 and Orai1, function as Ca ²⁺ Entry Units (CEUs), structures that provide a pathway to rapidly recover Ca²⁺ ions from the extracellular space during repetitive muscle activity.

Figure 1

Following exercise, membranes at the I band rearrange into stacks of flat-cisternae. (A–D) Longitudinal (A and C) and cross (B and D) EM sections in proximity of the I bands. Empty arrows in (C and D) point to newly formed stacks of flat-cisternae; arrows in insets point to strands between SR vesicles (B) and flat-cisternae (D). (E and F) Incidence of stacks. Sample size: control, 3 mice; exercised, 3 mice. (G) Junctional gap size in membrane stacks (empty arrows) and in triads (black arrows) measured as shown in the cartoon. Samples size: control, 3 mice/76 measurements/21 stacks analyzed; exercised, 3 mice/119 measurements/34 stacks analyzed. (H) Average size of SR vesicles at the I band (some larger ones are marked by asterisks in panel D). Sample size: control, 3 mice/561 measurements; exercised, 3 mice/309 measurements. Data are shown as mean ± SEM; *p < 0.01. Numbers in bars (n) indicate the number of fibers analyzed (E and F) and the number of SR vesicles measured (H). Scale bars: (A and C) (and insets in B and D), 0.1 μm; (B) and (D), 0.2 μm.

Figure 2

Following exercise, TT extensions at the I band are more frequent and become part of stacks of flat-cisternae. (A–D) Representative EM images (A and C, longitudinal; B and D, cross) showing TTs (stained black with ferrocyanide-precipitate) extending from the triad into the I band to make contact with SR vesicles (A and B) and stacks of flat-cisternae (C and D). (E) TT and SR (green and yellow, respectively) extending from a triad to form a lateral stack of flat-cisternae. (F) Quantitative analysis of the TT network extension at the I band measured in cross sections (see also Fig. S3C–F). Sample size: control, 3 mice; exercised, 3 mice. Data are shown as mean ± SEM; *p < 0.01. Numbers in bars (n) indicate the number of fibers analyzed. Scale bar: 0.1 μm.

Figure 3

Co-localization between STIM1 and Orai1, low in control samples, increases significantly following exercise. Representative immunofluorescence images obtained from EDL fibers double-labeled for RyR1-STIM1 (A and D), RyR1-Orai1 (B and E), and STIM1-Orai1 (C and F). Each panel also contains a fluorescence intensity profile along 4 sarcomeres (see dashed line in A) and the Pearson’s correlation coefficient value, a measure of covariance of pixel intensities, given as the mean ± SEM; *p < 0.01, compared to fibers from control mice. Numbers in figure (n) indicate the number of images analyzed. Scale bar: 2.5 µm (insets: 1 µm).

Figure 4

Immunogold labeling shows i) preferential I band localization of STIM1 and ii) increased presence of Orai1 at the I band following exercise. Left) Representative immunogold EM images showing the localization of RyR1 (top), STIM1 (center), and Orai1 (bottom) under control and exercised conditions. Right) Histograms of distances of immunogold particles from the Z line. The cyan line marks the position of the TT at the triad. Brackets with numeric values indicate the percentage of gold particles that fall within the triad-area vs. those that fall within the I-band area. Grey arrow points to the increased presence of Orai1 at the I-band following exercise. Sample size: control, 2 mice/4–8 fibers analyzed; exercised, 3 mice/3–8 fibers analyzed. *p < 0.05. Scale bar: 0.2 µm.

Figure 5

EDL muscles from exercised mice exhibit an enhanced resistant to fatigue that depends of Ca²⁺ entry. (A) Time course of average relative force decay during 30 consecutive frequency stimulus trains (1 s–60 Hz every 5 s), normalized to the 1^st stimulus train, in EDL muscles from control and exercised WT mice using a KH solution containing 2.5 mM Ca²⁺. (B) Bar plot showing the average fold increase in force produced by EDL muscles from exercised mice recorded during the 15^th stimulus train (arrow in A) normalized to that of control muscles. (C and E) Time courses of average relative force decay during repetitive stimulation of EDL muscles from control and exercised mice in the presence of 2.5 mM Ca²⁺, in Ca²⁺-free KH solution, or presence of either 10 µM BTP-2 or 100 µM 2-APB. (D and F) Bar plots summarizing average force reduction during the 15^th stimulus train in the presence of 2.5 mM Ca²⁺, in Ca²⁺-free KH solution, or presence of either 10 µM BTP-2 or 100 µM 2-APB. Data are shown as mean ± SEM; *p < 0.05. Number of experiments (n) reflect the number of EDL muscles analyzed for each condition.

Figure 6

Model for exercise-dependent formation of Calcium Entry Units (CEUs). Main events leading to CEU formation: (A-to-B) fusion of free-SR into larger vesicles and elongation of TTs toward the I band; (B-to-C) flattening of SR membranes into parallel cisternae and final docking to elongated TTs. (D and E) Enlargements of dashed boxes in A: under control conditions, some membranes within the I band may (E) or may not (D) be of TT origin (and contain Orai1). (F) Enlargement of dashed box in C with modeling of the Ca²⁺entry pathway that would allow flow-back of ions to triads during exercise/fatigue. Empty arrows in A point to SR-TT contacts in proximity of triads, as imaged by EM in Fig. S8.