A mechanism of lysosomal calcium entry

Lysosomal calcium (Ca2+) release is critical to cell signaling and is mediated by well-known lysosomal Ca2+ channels. Yet, how lysosomes refill their Ca2+ remains hitherto undescribed. Here, from an RNA interference screen in Caenorhabditis elegans, we identify an evolutionarily conserved gene, lci-1, that facilitates lysosomal Ca2+ entry in C. elegans and mammalian cells. We found that its human homolog TMEM165, previously designated as a Ca2+/H+ exchanger, imports Ca2+ pH dependently into lysosomes. Using two-ion mapping and electrophysiology, we show that TMEM165, hereafter referred to as human LCI, acts as a proton-activated, lysosomal Ca2+ importer. Defects in lysosomal Ca2+ channels cause several neurodegenerative diseases, and knowledge of lysosomal Ca2+ importers may provide previously unidentified avenues to explore the physiology of Ca2+ channels.


Localization of WT LCI
Our data show that WT human LCI is largely localized to the Golgi in HeLa cells, with minimal colocalization with TMR-dextran.However, previous experiments have found human LCI in lysosome fractions (24).In addition, we have detected overexpressed and endogenous human LCI on membranes of lysosomes (Fig. S4, S18A,B).Even a small fraction of human LCI present on the lysosomes would be highly active, given the higher pH gradient there than across the Golgi membrane.Thus, the low lysosome localization does not preclude human LCI from having a physiologically relevant role on the lysosome membrane.

Lysosomal calcium measurements
The O/R ratio of ~50% of lysosomes of lci-1 +/-worms was below the O/Rmin of CalipHluor2.0,indicating a Ca 2+ concentration <100 nM that is not quantifiable by our probe (Fig. S6D).
Conversely, the O/R ratio of ~50% of lysosomes of lci-1 +/-worms expressing WT human LCI was above the O/Rmax of CalipHluor2.0,indicating a Ca 2+ concentration >1 mM that is not quantifiable by our probe (Fig. S6D).Thus, our reported effect of human LCI on lysosomal Ca 2+ levels in worms is actually an underestimation.
Similarly, the O/R ratio of over 60% of lysosomes of TMEM165 KO HeLa cells was below the O/Rmin of CalipHluor mLy (Fig. S13C).Only about 30% of lysosomes of WT HeLa cells had an O/R below the O/Rmin (Fig. S13C).Thus, the reported effect of human LCI on lysosomal Ca 2+ levels in cells is also an underestimation.
The lysosomal Ca 2+ measurements of worms expressing mutants of human LCI are complicated by the heterozygous knockout background.Specifically, we see a surprisingly high level of lysosomal Ca 2+ in lci-1 +/-worms expressing the G304R, E108A, and E248A mutants of human LCI.Yet, in all other assays, these mutants impair lysosomal Ca 2+ import.Thus, it is likely that human LCI acts as a dimer, and that the remaining copy of endogenous worm lci-1 can dimerize with the mutant human LCI, and form a partially functional transporter.Given the smaller size of human LCI compared to other Ca 2+ transporters and exchangers, we hypothesize that it acts as a dimer.

Homologous regions of human LCI
The regions of human LCI that show homology to vcx1 offer clues to how human LCI may transport Ca 2+ with high capacity, dependent on the lysosomal pH gradient.In vcx1, the proton motive force across the vacuole drives a conformational change where active site glutamate residues face the cytosol and maintain a negative charge (29).Under conditions of high cytosolic Ca 2+ , as seen in our experiments in Fig. 3C in yeast and Fig. 4 and S12A in mammalian cells, Ca 2+ ions are coordinated by the cytosolic acidic helix to bring them near the active site.
Coordination by the active site displaces water molecules to move helix M2b (designated in yellow in Fig. 3A,B) towards the active site.This movement closes the cytosolic vestibule lined by M7b (designed in purple in Fig. 3A,B) and opens a vacuolar cleft, such that the acidic pH of the vacuole lowers the Ca 2+ affinity of active site glutamate residues and leads to release of Ca 2+ .This cyclical pumping occurs because of flexible helices around the active site and more rigid piston-like helices further away from the pore.Given that human LCI possesses two Ca 2+ -binding sequences near regions homologous to the flexible internal helices of vcx1 and an acidic cytosolic helix in proximity, it follows that human LCI functions similarly in response to high cytosolic Ca 2+ .However, the fact that human LCI is much smaller than vcx1 and most other exchangers may indicate that it functions differently, for example as a dimer or as a pH-activated transporter instead of an exchanger.

Yeast color change
Strains of S. cereivisiae that have mutations in certain steps of the adenine biosynthetic pathway (such as the ade2-1 mutation in K665 ( 60)) accumulate an adenine-intermediate-derived red pigment inside vacuoles (61).The intermediate phosphoribosylaminoimidazole (AIR) is transported glutathione-dependently into vacuoles where it is polymerized and modified to form the characteristic red pigment.The structure of this pigment has yet to be fully established.
Importantly, development of red pigmentation in ade2 mutants requires normal vacuolar function (62).This concept has even been used to screen for chemicals that disrupt vacuolar function by loss of red pigmentation (63).Interestingly, we see that K665 colonies grown on SD-Leu plates do not exhibit red pigmentation, but that human LCI-transformed K665 colonies appear red.This red pigmentation is lost when plated on plates with high Ca 2+ .This implies that human LCI rescues vacuolar dysfunction in K665 under normal osmotic conditions, but that high Ca 2+ causes vacuolar dysfunction even as human LCI rescues lethality.While we cannot rule out the effect of human LCI on other aspects of the adenine-derived pigment biosynthetic pathway, its lysosomal roles in humans and nematodes established elsewhere in this manuscript support the hypothesis that it rescues vacuolar dysfunction here.lci-1

Table S1.
Genes screened for cup-5 +/-worm survival rescue, with brood size difference and significance calculated with respect to empty vector (EV).Table S2.

Gene
Sequences of DNA oligos used in this study.D1 and D2 were used for labeling of coelomocyte lysosomes for lysosome size assay and mammalian cell lysosomes for endocytic tracking.C1, C2, and C3 were used to prepare CalipHluor 2.0.OG-C1, C2, and C3 were used to prepared Table S3.
Internal symmetry of human LCI.

Figure S4 :
Figure S4: Lysosomal localization of human LCI.(a) Representative fluorescence images of COS-

Figure S7 .
Figure S7.Conservation within the UPF0016 family and homology-based model of human LCI.

Figure S8 .
Figure S8.Human LCI rescues phenotypes of the K665 strain.(a) The rescue assay in S. cerevisiae

Figure S15 .
Figure S15.Sensitivity of human LCI to the pH gradient across the lysosomal membrane.(a)

Figure S16 .
Figure S16.Electrophysiological characterization of human LCI on the plasma membrane.(a)

Fig
Fig. 5B.(d) Zoomed-in version of Fig. 5D to show reversal potentials under the indicated

Table S5 .
Templates used for homology-based modelling of human LCI.

Table S6 .
Reversal potentials expected for exchanger with pipette buffer of pH 7.5 and 1µM Ca and bath buffer of 100µM Ca and indicated pH.