Parallel adaptive feedback enhances reliability of the Ca2+ signaling system

Abstract

Despite large cell-to-cell variations in the concentrations of individual signaling proteins, cells transmit signals correctly. This phenomenon raises the question of what signaling systems do to prevent a predicted high failure rate. Here we combine quantitative modeling, RNA interference, and targeted selective reaction monitoring (SRM) mass spectrometry, and we show for the ubiquitous and fundamental calcium signaling system that cells monitor cytosolic and endoplasmic reticulum (ER) Ca(2+) levels and adjust in parallel the concentrations of the store-operated Ca(2+) influx mediator stromal interaction molecule (STIM), the plasma membrane Ca(2+) pump plasma membrane Ca-ATPase (PMCA), and the ER Ca(2+) pump sarco/ER Ca(2+)-ATPase (SERCA). Model calculations show that this combined parallel regulation in protein expression levels effectively stabilizes basal cytosolic and ER Ca(2+) levels and preserves receptor signaling. Our results demonstrate that, rather than directly controlling the relative level of signaling proteins in a forward regulation strategy, cells prevent transmission failure by sensing the state of the signaling pathway and using multiple parallel adaptive feedbacks.

Fig. 1.

Model simulations show that the Ca²⁺ signaling system fails to function when signaling components are increased or decreased twofold. (A) Schematic of the Ca²⁺ signaling model. (B) Output of the model at different levels of receptor activation. (C) When STIM, SERCA, and PMCA concentrations were varied from 50% to 200% of their setpoint concentrations (by regulating the synthesis rates), the transition points where receptor activation triggered Ca²⁺ oscillations or induced the plateau phase varied significantly from the control. In the control, all proteins were expressed at their normal, 1× concentrations. (D) Significant changes in basal ER and cytosolic Ca²⁺ levels that may affect cell health occurred when twofold changes were made to the synthesis rates of PMCA, SERCA, and STIM.

Fig. 2.

Selective reaction monitoring (SRM) mass spectrometry determination of copy numbers of important Ca²⁺ signaling components in Drosophila S2R+ cells. (A) Schematic of the targeted proteins and their regulatory connections. (B) SRM measurements were carried out by using a nano-HPLC and electrospray ionization source coupled to a triple-quadropole mass spectrometer. The transitions (precursor/fragment ion pairs) generated for each targeted peptide were read out as intensities by the detector. (C) A subset of transitions measured by SRM in a typical sample. (D) Fit of the peak of a heavy peptide to the light (endogenous) trace. Time-shifted decoy peaks were used in the fit to quantify background peptide activities. Four decoy shapes were generated by shifting the heavy peptide peak shape to the left and right, respectively. Decoy peaks significantly contributing to the fit indicated noise in the sample trace, and fragments were rejected. (E) Copy numbers of key Ca²⁺ signaling proteins per S2R+ cell measured in three biological replicates. The red and blue bars show quantitation of two different transitions for each peptide. Measurements of multiple independent peptides are shown for InsP3R, STIM, and SERCA.

Fig. 3.

Demonstration that multiple parallel adaptive feedback loops exist in the Ca²⁺ signaling system. The error bars in all plots show SE. (A and B) Effect of RNAi knockdown on cytosolic and ER Ca²⁺ levels (A) and protein levels (B). Knockdown of YFP or GL3, proteins that are not present in S2R+ cells, were used as controls. (C) The small changes in STIM and SERCA and PMCA protein concentration correlated with Ca²⁺ level changes, but data in B are not sufficient to determine whether the changes were regulated just by ER or just by cytosolic Ca²⁺. (D and E) Effect of treating cells for 24 h with low (2.5 mM EGTA plus 2.5 mM Mg²⁺) and high (10 mM Ca²⁺) external Ca²⁺ and thapsagargin addition (1 μM) on cytosolic and ER Ca²⁺ levels (D) and protein levels (E). (F) Summary of the three identified feedback loops.

Fig. 4.

Simulations show that including the identified parallel adaptive feedback loops into the model markedly enhances the reliability of the Ca²⁺ signaling system in response to changes in the synthesis rates of signaling components. (A) Marked stabilization of both basal cytosolic and ER Ca²⁺ levels (gray box). (B) Parallel adaptive feedback is well suited to preserve the setpoints where receptor stimuli first triggered Ca²⁺ oscillations and first triggered a plateau phase. Instead of the marked reduction in the range over which oscillations occurred in the model without adaptive feedback (from Fig. 1C), the signaling system with adaptive feedback (gray box) now shows minimal changes in the responsiveness to receptor stimulation.