GRAM marks the spot for STIM. Commentary on "GRAM domain proteins specialize functionally distinct ER-PM contact sites in human cells"

Article

Reference

GRAM marks the spot for STIM. Commentary on "GRAM domain proteins specialize functionally distinct ER-PM contact sites in human

cells"

NUNES-HASLER, Paula, DEMAUREX, Nicolas

Abstract

GRAM domain proteins were reported as novel ER-PM tethers defining specific membrane contact sites (MCS) subdomains. GRAMD2a pre-marks the sites occupied by STIM1 at MCS and its ablation impairs STIM1 translocation, but not store-operated Ca2+ entry. We discuss these apparently counterintuitive findings in the context of STIM/ORAI signaling at MCS.

NUNES-HASLER, Paula, DEMAUREX, Nicolas. GRAM marks the spot for STIM. Commentary on "GRAM domain proteins specialize functionally distinct ER-PM contact sites in human cells".

Cell Calcium , 2018, vol. 73, p. 70-71

PMID : 29684786

DOI : 10.1016/j.ceca.2018.04.002

Available at:

http://archive-ouverte.unige.ch/unige:109621

Disclaimer: layout of this document may differ from the published version.

1 / 1

Contents lists available atScienceDirect

Cell Calcium

journal homepage:www.elsevier.com/locate/ceca

Commentary

GRAM marks the spot for STIM. Commentary on “ GRAM domain proteins specialize functionally distinct ER-PM contact sites in human cells ”

Paula Nunes, Nicolas Demaurex

Department of Cell Physiology and Metabolism, Rue Michel-Servet, 1, University of Geneva, Switzerland

A B S T R A C T

GRAM domain proteins were reported as novel ER-PM tethers defining specific membrane contact sites (MCS) subdomains. GRAMD2a pre-marks the sites occupied by STIM1 at MCS and its ablation impairs STIM1 trans- location, but not store-operated Ca²⁺entry. We discuss these apparently counterintuitivefindings in the context of STIM/ORAI signaling at MCS.

Store-operated Ca²⁺ entry (SOCE) mediated by STIM and ORAI proteins couple the Ca²⁺depletion of the endoplasmic reticulum (ER) to the activation of Ca²⁺permeable plasma membrane (PM) channels [1,2]. As the free Ca²⁺concentration within the ER lumen decreases, STIM proteins accumulate in cortical ER sheets via electrostatic inter- actions between membrane phosphoinositides and STIM polybasic cy- tosolic tails, which can then trap and gate ORAI channels at ER-PM membrane contact sites (MCS) [3]. The resulting Ca²⁺fluxes fuel local and global Ca²⁺signals that control the developmental program of immune and muscle cells[4–6], and patients bearing loss of function mutations in STIM1 and ORAI genes suffer from severe im- munodeficiency and muscle weakness [7].

Recombinant STIM1 can gate ORAI1 channels in yeast vesicles, but a host of accessory proteins is required for productive STIM/ORAI in- teractions at MCS inside cells [8]. CRACR2A, junctate and UNC93B1 promote STIM1 clustering, SARAF, POST and Surf4 regulate STIM/

ORAI coupling, septins maintain permissive lipid microdomains, and STIMATE/TMEM110 regulates MCS stability. Extended synaptotagmins (E-Syts) stabilize MCS by tethering the ER to membranes enriched in phosphoinositides in a Ca²⁺-dependent manner, but are dispensable for SOCE [8–10].

In the March issue ofElife, Besprozvannaya et al identify mamma- lian homologs of a yeast ER-PM tether sharing a pleckstrin homology (PH)-like GRAM (forGlucosyltransferases, RAb-like GTPase activators and Myotubularins) domain known to bind phosphoinositides [11]. The two isoforms tested, GRAMD1a and GRAMD2a, co-localized with ER and PM markers in peripheral foci typical of MCS on thefluorescence microscope. Each homologue marked a distinct site, with GRAMD2a but not GRAMD1a co-localizing extensively with E-Syts and STIM1.

Accumulation of GRAMD2a in foci required the PI(4)P and PI(4,5)P2-

binding GRAM domain and was disrupted by PI(4,5)P2 depletion but not by E-Syt ablation. STIM1, but not a STIM1 mutant lacking the K- polybasic tail, accumulated in GRAMD2a foci upon Ca²⁺depletion and subsequently migrated to distinct yet likely contiguous cortical ER structures. The authors provide data suggesting that GRAMD2a ablation may impair STIM1 translocation to the PM, yet this did not impact SOCE or the distribution of E-Syt2/3 isoforms. However, enhanced PM recruitment of E-Syt1 in naïve, store-replete cells was observed. GRAM domain proteins therefore mark the spots initially targeted by STIM1 and may facilitate its PM translocation, but they are dispensable for SOCE and their functional role remains unknown.

Because STIM1 must translocate to MCS to gate ORAI channels, the reduced PM recruitment reported by the authors would be expected to impair SOCE, yet this was not observed. Instead, when SOCE was ac- tivated with thapsigargin (Tg) to release Ca²⁺from stores, none of the Ca²⁺ handling parameters, i.e. resting [Ca²⁺]cyt levels, constitutive Ca²⁺entry, amount of Ca²⁺mobilized from stores, and SOCE ampli- tude and kinetics, were affected by the ablation of GRAMD2a. Thus, reducing STIM1 translocation by twofold does not translate into visible Ca²⁺handling defects. Although counterintuitive, this discrepancy is, in fact, consistent with earlier studies. First, a preserved SOCE was observed in cells lacking all three E-Syts isoforms, despite a 50% re- duction in cortical MCS measured by electron microscopy (EM) [10], indicating that there is an excess of contact sites that can be populated by endogenous STIM1. Whether GRAMD2a ablation reduces cortical ER abundance would need to be verified by EM, but a deficit of less than 50% is not expected to translate into a SOCE defect. Second, STIM1 PM recruitment was measured by light microscopy approaches that cannot distinguish between pre-cortical ER and cortical MCS. Artificial struc- tures are generated by STIM1 overexpression [12], and the reduced

https://doi.org/10.1016/j.ceca.2018.04.002 Received 29 March 2018

⁎Corresponding author.

E-mail address:Nicolas.Demaurex@unige.ch(N. Demaurex).

Cell Calcium 73 (2018) 70–71

Available online 18 April 2018

0143-4160/ © 2018 Elsevier Ltd. All rights reserved.

T

mCherry-STIM1 signal could reflect reduced STIM1 accumulation in pre-cortical ER sites unable to engage ORAI channels. A third point is that overexpression of STIM1 increases cortical MCS by 3 fold [12], but does not significantly increases SOCE without co-expression of ORAI. In this setting, GRAMD2a ablation might reduce the accumulation of su- perfluous STIM proteins accumulating around saturated endogenous ORAI channels. Lastly, GRAMD2a ablation affected only the size, but not the number of mCherry-STIM1 fluorescent puncta. The size of STIM1 fluorescent clusters does not correlate with the amplitude of SOCE as shown for the long STIM1L isoform, which mediates robust SOCE without enlarging cortical MCS [13]. Based on these priorfind- ings, one would predict that a 50% reduction in mCherry-STIM1 fluorescent puncta measured by spinning disk microscopy should not translate into a SOCE defect, consistent with the phenotype of GRAMD2a-null cells. The defective clustering might also reflect partial store depletion due to the short Tg exposure in this particular case.

One interesting aspect of this study is the non-overlapping pattern of GRAMD1a and GRAMD2a, as this highlights the fact that cortical MCS are heterogeneous and likely functionally distinct. Perhaps even more intriguing, however, is the observation that upon store depletion, the STIM1 and GRAMD2a signals initially co-localize but then diverge. This implies that functionally discrete subdomains exist even within the same MCS. Indeed previous studies have suggested that STIM1 initially migrates into a PI(4,5)P2-poor domain but then, via its K domain, partitions into a PI(4,5)P2-rich domain formed and maintained by E- Syt1 and septin4, where it interacts with SARAF to mediate slow Ca²⁺- dependent inactivation (SCDI) of ORAI1 [14]. From the sequence of events, one might expect that GRAMD2a demarcates the PI(4,5)P2-poor domain, yet this is counterintuitive since GRAMD2a preferentially binds PI(4,5)P2. Another intriguing and incongruent observation is that although the STIM1-ΔK mutant lacking the polybasic domain clearly translocated to membrane-proximal puncta upon store depletion, these did not co-localize with GRAMD2a. One possibility is that it might generate or populate distinct MCS, and another that it could localize directly to the GRAMD2a-contiguous MCS subdomain to which full- length STIM1 partitions subsequently. However the latter is again counterintuitive, since the STIM1 polybasic domain is thought to bind to PI(4,5)P2. One possible explanation could be that GRAMD2a has a very high affinity for PI(4,5)P2 and prevents the accessibility of re- porter proteins, causing the initial STIM1 subdomain to appear as being PI(4,5)P2-poor. The partitioning of STIM1 to discrete subdomains was clear from the data of Maléth and colleagues, however, that these subdomains are PI(4,5)P2-rich or poor was only inferred, and not di- rectly visualized. Whether other lipids or proteins contribute to the segregation of discrete MCS subdomains thus remains to be verified.

If GRAMD2a marks the spot for STIM1 but does not affect SOCE, then a critical unanswered question is what is the physiological func- tion of GRAMD2a? The only other clue so far is the unexplained pre- recruitment of E-Syt1 to MCS in GRAMD2a knockout cells, that could not be linked to global changes in PM PI(4,5)P2, cholesterol or ca- veolin. Clearly more work will be needed to decipher whether GRAMD2a has an impact on more subtle SOCE parameters such as SCDI, whether differences in expression, recruitment or even sub- domain localization of other SOCE players may be affected, or whether

other lipid species are involved. Regardless, the discovery that GRAMD2afirst merges with and then segregates from STIM1 during store depletion, suggests it may serve as a novel marker for MCS sub- domains and will surely serve as a useful new tool for probing changes in MCS composition and function during SOCE.

Competing interests

The authors declare no competingfinancial interests.

Authors contributions

ND and PNH outlined and wrote the review. Both authors accept the final version of the manuscript as submitted.

Funding

ND’s research is funded by grants 31003A-149566 and CRSII3_

160782 of the Swiss national science foundation (SNF). PNH is sup- ported by a Novartis Foundation for Medical-Biological Research (17B078).

References

[1] P.G. Hogan, A. Rao, Store-operated calcium entry: mechanisms and modulation, Biochem. Biophys. Res. Commun. 460 (2015) 40–49.

[2] M. Prakriya, R.S. Lewis, Store-operated calcium channels, Physiol. Rev. 95 (2015) 1383–1436.

[3] W.W. Shen, M. Frieden, N. Demaurex, Remodelling of the endoplasmic reticulum during store-operated calcium entry, Biol. Cell 103 (2011) 365–380.

[4] N. Demaurex, P. Nunes, The role of STIM and ORAI proteins in phagocytic immune cells, Am. J. Physiol. Cell Physiol. 310 (2016) C496–508.

[5] S. Srikanth, J.S. Woo, Z. Sun, Y. Gwack, Immunological disorders: regulation of Ca (2+) signaling in T lymphocytes, Adv. Exp. Med. Biol. 993 (2017) 397–424.

[6] Z. Pan, M. Brotto, J. Ma, Store-operated Ca2+ entry in muscle physiology and diseases, BMB Rep. 47 (2014) 69–79.

[7] R.S. Lacruz, S. Feske, Diseases caused by mutations in ORAI1 and STIM1, Ann. N. Y.

Acad. Sci. 1356 (2015) 45–79.

[8] J.J. Lopez, L. Albarran, L.J. Gomez, T. Smani, G.M. Salido, J.A. Rosado, Molecular modulators of store-operated calcium entry, Biochim. Biophys. Acta 1863 (2016) 2037–2043.

[9] S. Maschalidi, P. Nunes-Hasler, C.R. Nascimento, I. Sallent, V. Lannoy, M. Garfa- Traore, N. Cagnard, F.E. Sepulveda, P. Vargas, A.M. Lennon-Dumenil, P. van Endert, T. Capiod, N. Demaurex, G. Darrasse-Jeze, B. Manoury, UNC93B1 interacts with the calcium sensor STIM1 for efficient antigen cross-presentation in dendritic cells, Nat. Commun. 8 (2017) 1640.

[10] F. Giordano, Y. Saheki, O. Idevall-Hagren, S.F. Colombo, M. Pirruccello, I. Milosevic, E.O. Gracheva, S.N. Bagriantsev, N. Borgese, P. De Camilli, PI(4,5)P(2)- dependent and Ca(2+)-regulated ER-PM interactions mediated by the extended synaptotagmins, Cell 153 (2013) 1494–1509.

[11] M. Besprozvannaya, E. Dickson, H. Li, K.S. Ginburg, D.M. Bers, J. Auwerx, J. Nunnari, GRAM domain proteins specialize functionally distinct ER-PM contact sites in human cells, Elife (2018) 7.

[12] L. Orci, M. Ravazzola, M. Le Coadic, W.W. Shen, N. Demaurex, P. Cosson, From the cover: STIM1-induced precortical and cortical subdomains of the endoplasmic re- ticulum, Proc. Natl. Acad. Sci. U. S. A. 106 (2009) 19358–19362.

[13] S. Sauc, M. Bulla, P. Nunes, L. Orci, A. Marchetti, F. Antigny, L. Bernheim, P. Cosson, M. Frieden, N. Demaurex, STIM1L traps and gates Orai1 channels without remodeling the cortical ER, J. Cell. Sci. 128 (2015) 1568–1579.

[14] J. Maleth, S. Choi, S. Muallem, M. Ahuja, Translocation between PI(4,5)P2-poor and PI(4,5)P2-rich microdomains during store depletion determines STIM1 con- formation and Orai1 gating, Nat. Commun. 5 (2014) 5843.

P. Nunes, N. Demaurex Cell Calcium 73 (2018) 70–71

71