A “mix-and-match” approach to designing Ca²⁺ microdomains at membrane-contact sites

Abstract

Ca²⁺ microdomains are critical for regulating cellular activity and often form at membrane contact sites. Such sites between lysosomes and the ER potentially provide a platform for signaling by the Ca²⁺ mobilizing messenger NAADP. However, at present we know little of how Ca²⁺ release events are coordinated at these experimentally intractable junctions. We therefore developed a computational model of lysosome-ER microdomains, which suggested that small leaks of Ca²⁺ from the lysosome couple to Ca²⁺-sensitive Ins(1,4,5)P₃ receptors on the ER to generate global, microdomain-dependent Ca²⁺ signals. Here we discuss how the “mix-and-match” arrangement of different Ca²⁺ signaling proteins on the “source” and “target” membranes might generate functionally heterogeneous Ca²⁺ microdomains.

Keywords: :

• Ca²⁺
• Membrane Contact Sites
• Microdomains
• NAADP
• 2-Pore Channel
• Lysosomes
• computational modeling
• SERCA
• Endoplasmic Reticulum

This article refers to:

Membrane contact sites (MCS) are specialized structures, whereby cellular membranes are closely apposed (10–50 nm).¹ Such an arrangement facilitates information flow between different cellular compartments.¹ Ca²⁺ is a universal signaling currency that often makes use of MCS to direct cellular activity.² The restricted space within MCS allows formation of local Ca²⁺ microdomains where the Ca²⁺ concentration ([Ca²⁺]) is thought to reach high, micromolar levels (~10 μM) levels.² These microdomains facilitate activation of downstream Ca²⁺-dependent processes. A high [Ca²⁺] within the dyadic junction of cardiac cells, for example, couples excitation of the sarcolemma to cellular contraction, driven by Ca²⁺ release from the sarcoplasmic reticulum (SR).³ Similarly, high [Ca²⁺] microdomains at ER-mitochondria MCS drive excitation-metabolism coupling.⁴ Recently, MCS between lysosomes and the endoplasmic reticulum (ER) have been identified.⁵ These sites provide an anatomical basis for functional coupling of lysosomal Ca²⁺ stores with the ER through the Ca²⁺ mobilizing messenger, NAADP.⁶ NAADP is thought to act via 2-pore channels to evoke Ca²⁺ signals from acidic organelles that are then amplified by the ER to drive complex Ca²⁺ behavior.^7,8 However, the function of lysosome-ER MCS and putative Ca²⁺ microdomains contained therein is currently unknown.

In our recent paper,⁹ we developed a model of Ca²⁺ dynamics within these putative lysosome-ER microdomains and the continuous bulk cytosol to simulate cellular Ca²⁺ responses. We modeled leaks of Ca²⁺ from the “source” lysosome membrane, such as those produced in response to NAADP. These leaks activated Ca²⁺-sensitive inositol trisphosphate receptors (Ins(1,4,5)P₃Rs) on the “target” ER membrane and generated global Ca²⁺ responses that were microdomain-dependent. Interestingly, we also found that varying either the distribution or density of the lysosomal Ca²⁺ leak altered the mode of coupling. For example, selectively increasing the magnitude of the lysosomal Ca²⁺ leak into the microdomain, akin to channel clustering, broadened the concentration range over which NAADP produced responses. A uniform increase in Ca²⁺ leak density into both the microdomain and non-microdomain compartments, mimicking channel overexpression, also had a broadening effect. However, in these simulations, microdomains acted to modulate the frequency of global Ca²⁺ oscillations rather than initiate them. Lysosome-ER Ca²⁺ microdomains can either drive or shape complex Ca²⁺ signals, depending on their composition.⁹

Unlike other microdomains,^4,10 the lysosome-ER microdomains simulated within our model did not achieve high [Ca²⁺]. Although kinetically distinct, Ca²⁺ signals in the microdomain were only marginally greater in amplitude than within the bulk cytosol.⁹ However, this result is not as counterintuitive as it may first seem. As Ca²⁺-release from Ins(1,4,5)P₃Rs on the target membrane can be activated by relatively low [Ca²⁺],¹¹ small leaks of Ca²⁺ from lysosomes can raise microdomain [Ca²⁺] sufficiently to initiate Ca²⁺-induced Ca²⁺-release via the Ins(1,4,5)P₃R. This arrangement differs from other Ca²⁺ microdomains. For example, ryanodine receptors (RyRs) on the target membrane in the cardiac dyadic cleft have a lower affinity for Ca²⁺-induced activation than Ins(1,4,5)P₃Rs (~0.8–1 μM, compared with ~0.1–0.3 μM).¹² They therefore require higher [Ca²⁺] for activation to generate a high [Ca²⁺] microdomain. Moreover, the mitochondrial Ca²⁺ uptake machinery on the target membrane in ER-mitochondria microdomains has an even lower affinity for Ca²⁺ (estimated between 10–100 μM).^13,14 Thus, microdomain [Ca²⁺] may “match” Ca²⁺ affinity of target proteins.

The unexpectedly low microdomain [Ca²⁺] in our model was due to the presence of sarco/endoplasmic reticulum Ca²⁺ ATPase (SERCA) pumps within the microdomains, which efficiently take up the released Ca²⁺.³ SERCA pumps were required to balance a basal Ca²⁺ leak through the Ins(1,4,5)P₃R at a resting level of Ins(1,4,5)P₃. This arrangement differs from other studies utilizing RyR models¹⁰ because, unlike Ins(1,4,5)P₃Rs, RyRs can be gated directly by Ca²⁺. As such, these models contain no basal RyR leak and therefore do not require balancing Ca²⁺ uptake mechanisms within the microdomain. This allows for the accumulation of high [Ca²⁺] upon activation of the RyRs. Our hypothesis, that peak [Ca²⁺] within microdomains can be determined by the level of microdomain SERCA, may also extend to ER-mitochondria microdomains. Here, the ER acts as the source membrane whereby Ins(1,4,5)P₃Rs create high microdomain [Ca²⁺] to facilitate low affinity mitochondrial Ca²⁺ uptake. It is therefore of note that Ins(1,4,5)P₃Rs are enriched within ER-mitochondria MCS, while SERCA is not.¹⁵ The potential absence of SERCA within these microdomains might permit spontaneous activity within these junctions, potentially accounting for the proposed maintenance of mitochondrial bioenergetics by basally active Ins(1,4,5)P₃Rs.¹⁶ Ca²⁺ signaling proteins may be “mixed” together to generate functionally heterogeneous microdomains.

Computational models of Ca²⁺ dynamics often have a modular design, whereby individual models for each Ca²⁺ transport process are assembled together to generate an appropriate, relevant system. This rule is consistent at different levels of computational complexity.^9,10,17 We suggest that viewing “real” microdomains in this modular fashion can aid our understanding of their architecture in live cells. As discussed, the variable expression of individual Ca²⁺ channels and pumps within MCS, on either the source or target membrane, can profoundly alter the properties of microdomains. This “mix-and-match” approach may account for the functionally diverse behaviors that microdomains coordinate. (Fig. 1)

Figure 1. Heterogeneity of Ca²⁺ microdomains at membrane contact sites. MCS between the source (top) and target (bottom) membranes allow functional Ca²⁺ microdomains to form between them. Ca²⁺ influx through voltage gated Ca²⁺ channels (Ca_v) in the PM-SR MCS of the dyadic cleft (left) forms a high [Ca²⁺] microdomain (dark circle) to initiate Ca²⁺ release from low-affinity ryanodine receptors (RyR). Ca²⁺ release through inositol trisphosphate receptors (Ins(1,4,5)P₃R) in ER-mitochondria MCS (center) also forms a high [Ca²⁺] microdomain to facilitate mitochondrial Ca²⁺ uptake by the low-affinity mitochondrial uniporter (MCU). Ca²⁺ release through 2-pore Channels (TPC) in lysosome-ER MCS (right) forms a low [Ca²⁺] microdomain (light circle) due to the presence of SERCA (S) but which is nevertheless able to activate high-affinity Ins(1,4,5)P₃Rs.

Abbreviations:
Membrane contact sites (MCS)	=	endoplasmic reticulum (ER), sarcoplasmic reticulum (SR), Ca2+ concentration ([Ca2+]) Nicotinic acid adenine dinucleotide phosphate (NAADP), inositol trisphosphate (IP3), inositol trisphosphate receptor (IP3R), sarco/endoplasmic reticulum Ca2+ ATPase (SERCA), ryanodine receptor (RyR)

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Acknowledgments

This was work was supported by a Biotechnology and Biological Sciences Research Council studentship BB/J014567/1 (to C.J.P.); an IMPACT studentship from University College London (to B.S.K.); and the Marsden Fund of the Royal Society of New Zealand. We thank Gyorgy Szabadkai and Michael Duchen for helpful discussion.