AbstractNano-junctions between the endoplasmic reticulum and the cytoplasmic surfaces of the plasma membrane and other organelles shape the spatiotemporal features of biological Ca2+ signals. Herein we propose that 2D Ca2+ exchange-diffusion on the negatively charged phospholipid surface lining the nano-junctions participates in guiding Ca2+ from its source (channel or carrier) to its target (transport protein or enzyme). Evidence provided by in vitro Ca2+ flux experiments using an artificial phospholipid membrane is presented in support of the above  proposed concept, and results from stochastic simulations of Ca2+ trajectories within nano-junctions are discussed to substantiate its possible requirements. Finally, we analyze recent literature on Ca2+ lipid interactions, which suggest that 2D interfacial Ca2+ diffusion may represent an important mechanism of signal transduction in other biological systems characterized by high phospholipid surface-to-aqueous volume ratios.IntroductionAll rapid functions in the body are selectively regulated by the universal biological messenger: ionic calcium (Ca2+). For this single ionic messenger to harmoniously control such a great range of biological mechanisms, it is crucial that its signals are delivered with pinpoint precision and millisecond  timing. The endoplasmic reticulum (ER) is the main organelle that orchestrates  this essential spatiotemporal precision of Ca2+ signalling via numerous different close contact sites or nano-junctions (NJ) with the plasma membrane (PM), mitochondria, lysosomes and other organelles \cite{23339179}. NJs have been defined as cytoplasmic sub-compartments where membranes of different organelles appose each other within the nano-metre scale. Typically the limiting membranes are separated by 20 nm or less and the specialized signalling function has been shown to fail at separation distances greater than 50 nm \cite{15541459,17418403,Pritchard2019}. The main function of the NJs in cells is to precisely localize the Ca2+ signals to specific Ca2+ sensors positioned within organellar membrane or PM domains, while bypassing the bulk cytoplasm. For example, refilling of the sarcoplasmic reticulum (SR) of vascular smooth muscle cells, in order to maintain asynchronous [Ca2+]cyt oscillations, is achieved by coupling Ca2+ entry via Ca2+-influx-mode NCX (rNCX) to SERCA at PM-SR NJs, thus bypassing the bulk cytoplasm \cite{Lee2002}. Employing stochastic particle simulator modelling software and using known NCX and SERCA turnover rates and surface densities, it was possible to generate a computational model of this cellular signalling process, which demonstrated that the rate of Ca2+ entry via rNCX/SERCA was indeed sufficient for replacing the Ca2+ released by periodic opening and closing of IP3Rs during the activation of asynchronous cytoplasmic Ca2+ waves that stimulate contraction \cite{17418403}. However, to generate plausible predictions by our computational model, it was necessary to implement a Ca2+ target size on SERCA of approximately 20 nm2, which is 2500 times larger than the area occupied by the dehydrated Ca2+ with a diameter of 1 Å. Assuming that short range local electrostatic forces of attraction between the fixed negatively charged binding sites on the SERCA macromolecule and the positively charged Ca2+ would increase the effective target size to an area several times larger than the size of the non-hydrated Ca2+, it would still be orders of magnitude smaller than 20 nm2. Therefore, in order to achieve effective functional transfer of Ca2+ from NCX on the PM to SERCA on the SR, it appears that an additional, as yet ignored, mechanism is operative to support the linkage between Ca2+ signalling elements (sources and sinks) within NJs, besides the proximity of the PM and SR in these regions. To this end, all available computational modelling data prompt us to conclude that, by itself, a 3D random walk of the Ca2+ between their sources and sinks on the two closely apposing membrane surfaces may be insufficient for efficient NJ Ca2+signalling. We thus propose that, in addition, 2D exchange diffusion of Ca2+ on the targeted surface of the NJ has the potential to increase the efficiency of Ca2+ reaching its target.BackgroundIt is well established that Ca2+ can move rapidly through negatively charged solid lattices, such as fluorapatite, by the process of exchange diffusion \cite{Jay_2011}. An analogous model, featuring negatively charged phospholipids (PLs), has been described some time ago \cite{TOBIAS1962,5723329,5803390}. This experimental model consists of a millipore filter impregnated with a mixture of phospholipids of animal origin, separating two aqueous phases and exhibits properties that are highly relevant to the topic of Ca2+ movements through narrow aqueous passages lined by PLs. Its salient feature is that it supports net transfer of Ca2+ through relatively long PL lined pores at a much faster rate than would be possible for free diffusion within the limited adjoining aqueous phase. The mechanism that was presented to explain Ca2+ transport through this solid ion exchange membrane involves the association of Ca2+ with a negatively charged phosphate or carboxyl group of the PL surface on the cis-side of the membrane, followed by transfer of Ca2+ within a 2-dimensional matrix of similar sites, constituted by the pore-lining PL layers, and a final step of dissociation of Ca2+ from the negatively charged PL head groups on the trans-side of the membrane. The rate of this PL-mediated transport of radioactive labeled Ca2+ through the membrane was decreased by removal of Ca2+ from the buffered solution on the trans-side of the membrane. Paradoxically, further removal of remaining Ca2+ from the trans-solution by the addition of the soluble, but non-permeant chelator EDTA, increased the rate of PL-mediated net Ca2+ transport across the PL lined millipore filter by more than one order of magnitude. The mechanism proposed to explain the latter observation is that dissociation of Ca2+ from PL head groups on the trans-side of the membrane is rate-limiting and addition of a freely diffusible, impermeant Ca2+ binding site on EDTA, to the trans-side of the flux chamber facilitates the dissociation of PL bound Ca2+. Once the Ca2+ transported from the cis-side of the flux chamber has been chelated by EDTA, which is dissolved in a large volume of buffered solution on the trans-side, it cannot rebind to the PL membrane, but is replaced by the next Ca2+arriving from the cis-side.  In this model system, electro-neutrality is preserved by the movement of Mg2+ and monovalent cations in the opposite direction. The mechanism envisioned for the transfer of Ca2+ from the PL membrane to EDTA involves an intermediary step of partial dissociation from the negative PL site and simultaneous association with a carboxyl group of EDTA. We propose herein, that a similar mechanism is involved in the transfer of Ca2+ bound to PL head-groups of the membranes lining the NJ to its biological target site on the Ca2+-receptor protein.Model DescriptionWhen we compare the artificial PL-mediated Ca2+ transport in the model described above with biological PL-lined nano-spaces some striking parallels become obvious. For example, rod outer segments of the bovine eye feature a 15-nm wide cytoplasmic phase between the intercalated discs, which stretch for several microns \cite{McLaughlin1981}. The lipid bilayers lining this narrow space contain 45% phosphatidylethanolamine, 36% phosphatidylcholine and 16% phosphatidylserine, calculated as percent of the total PL. At physiological pH, phosphatidylcholine and phosphatidylethanolamine are zwitterions and phosphatidylserine has one net negative charge \cite{McLaughlin1981}. In this example, it was calculated that of the Ca2+ released during stimulation, 90% to 99% was bound to the PL-head groups of membranes Iining the nano-spaces between the intercalated discs, which is similar to the high ratio of bound/free Ca2+ in the above model membrane. Although historically it has been accepted that such binding slows diffusion, due to a drastic decrease in freely diffusible Ca2+ in the aqueous phase, it is also possible that the PL-bound Ca2+ within the water-PL membrane interface of the nano-space continues its trajectory by the mechanism of 2D Ca2+ exchange diffusion. We therefore propose that the mechanism of 2D Ca2+/Mg2+, K+ exchange diffusion at the aqueous/phospholipid interfaces of NJs facilitates targeting of Ca2+ receptors located on organellar membranes and the inner PM.Returning to the example of NCX-mediated SR refilling in vascular smooth muscle, we propose that Ca2+ enters the NJ via the reverse mode of NCX located in the junctional domain of the PM. It will then perform a 3D random walk and when it hits the negatively charged PL membrane boundary of the junctional nano-space proceeds along the lipid-water interface by a series of steps of reversible binding to negatively charged oxygen molecules for some variable time before being released back into the aqueous phase to resume its 3D random walk. After a number of such cycles the Ca2+ is envisioned to hit the ER membrane surface in the proximity of its target protein, in this case SERCA, which will then be reached more effectively by 2D surface exchange diffusion than would be expected if it depended solely on 3D diffusion in the aqueous phase.