Research article

GTPase of the immune-associated nucleotide-binding protein 5 (GIMAP5) regulates calcium influx in T-lymphocytes by promoting mitochondrial calcium accumulation

Xi Lin Chen, Daniel Serrano, Marian Mayhue, Hans-Joachim Wieden, Jana Stankova, Guylain Boulay, Subburaj Ilangumaran, Sheela Ramanathan


Mature T-lymphocytes undergo spontaneous apoptosis in the biobreeding diabetes-prone strain of rats due to the loss of the functional GIMAP5 (GTPase of the immune-associated nucleotide-binding protein 5) protein. The mechanisms underlying the pro-survival function of GIMAP5 in T-cells have not yet been elucidated. We have previously shown that GIMAP5 deficiency in T-cells impairs Ca2+ entry via plasma membrane channels following exposure to thapsigargin or stimulation of the T-cell antigen receptor. In the present study we report that this reduced Ca2+ influx in GIMAP5-deficient T-cells is associated with the inability of their mitochondria to sequester Ca2+ following capacitative entry, which is required for sustained Ca2+ influx via the plasma membrane channels. Consistent with a role for GIMAP5 in regulating mitochondrial Ca2+, overexpression of GIMAP5 in HEK (human embryonic kidney)-293 cells resulted in increased Ca2+ accumulation within the mitochondria. Disruption of microtubules, but not the actin cytoskeleton, abrogated mitochondrial Ca2+ sequestration in primary rat T-cells, whereas both microtubules and actin cytoskeleton were needed for the GIMAP5-mediated increase in mitochondrial Ca2+ in HEK-293 cells. Moreover, GIMAP5 showed partial colocalization with tubulin in HEK-293 cells. On the basis of these findings, we propose that the pro-survival function of GIMAP5 in T-lymphocytes may be linked to its requirement to facilitate microtubule-dependent mitochondrial buffering of Ca2+ following capacitative entry.

  • calcium flux
  • GTPase of the immune-associated nucleotide-binding protein 5 (GIMAP5)
  • mitochondrion
  • microtubule
  • T-lymphocyte


In the BB-DP (Biobreeding diabetes-prone) strain of rats, homozygous lyp mutation causes a 5–10-fold reduction in CD4+ T-lymphocyte numbers and virtual absence of CD8+ T-cells in secondary lymphoid organs [13]. Mature T-cells undergo spontaneous apoptosis in BB-DP rats soon after they emigrate from the thymus and enter the peripheral circulation. The half-life of recently emigrated mature T-cells is markedly reduced in BB-DP rats compared with non-lymphopenic diabetes-resistant (BB-DR) rats (3 days compared with >15 days) [48]. The recessive lyp allele arises from a frameshift mutation within the Gimap5 (GTPase of the immune-associated nucleotide-binding protein 5) gene, resulting in a truncated protein lacking 223 amino acids at the C-terminus that are replaced by 19 other residues [9,10]. GIMAP5 is a member of the highly conserved GIMAP family proteins that play important roles in immune responses and in protecting cells from apoptosis [11,12]. In mice, GIMAP5 was shown to interact with members of the Bcl-2 family of anti-apoptotic proteins [1316], whereas another study implicated GIMAP5 in the ER (endoplasmic reticulum) stress response [17]. In contrast, we have shown that endogenous GIMAP5 in rat T-cells resides in a cellular compartment distinct from the mitochondria and ER, and that neither endogenous rat GIMAP5 nor the overexpressed protein with intact C-terminus interacts with Bcl-2 [18]. However, the loss of GIMAP5 impairs mitochondrial membrane potential [18,19]. In agreement with the non-mitochondrial location of GIMAP5, a previous study showed an enrichment of human GIMAP5 in lysosomes [20]. Despite a decade of efforts by several groups, there is a lack of consensus on the subcellular location of GIMAP5. Likewise, molecular mechanisms underlying the pro-survival function of GIMAP5 remain largely unclear. Whereas the cell survival defect is confined to T-cells in BB-DP rats, Gimap5-knockout mice show defects in various haematopoietic cell types, including stem cells [15,16,21].

Even though BB-DP rat T-cells are short-lived, stimulation by mitogens such as concanavalin A or anti-CD3 mAb (monoclonal antibody) induces proliferation, suggesting that strong stimulation via the TCR (T-cell antigen receptor) rescues BB-DP rat T-cells from apoptosis [6,22,23]. We have shown that autoantigen-dependent rescue may underlie preferential expansion of autoreactive T-cells in BB-DP rats, leading to autoimmune diabetes [6]. In the absence of strong TCR stimulation, homoeostatic survival of naïve T-cells depends on two essential signals, one provided by IL-7 (interleukin-7) and the other by non-antigenic self-peptides that elicit a basal level of TCR stimulation [24]. Although the IL-7-mediated signals are well characterized, the TCR-dependent survival signals remain less clear, although it is known to require Lck, a non-receptor tyrosine kinase activated by TCR stimulation [25]. The events downstream of TCR signalling produce the second messenger Ins(3,4,5)P3, which binds to its receptor IP3R on the ER and triggers Ca2+ release from the ER store, resulting in a conformational change in the ER-localized STIM1 (stromal interaction molecule 1) protein [26,27]. This event relays a signal to open the CRAC (Ca2+ release-activated Ca2+ channel) on the plasma membrane, inducing capacitative Ca2+ entry from the extracellular milieu [28,29]. TCR stimulation by antigens induces sustained Ca2+ influx via CRACs leading to T-cell proliferation [30]. Non-stimulatory self-peptides, which provide survival signals without inducing cell proliferation, are also capable of eliciting a discernible Ca2+ response [3134]. Our earlier findings that TCR-induced Ca2+ flux is reduced in Gimap5lyp/lyp T-cells [35] suggested that their survival defect may be related to impairment of the basal homoeostatic Ca2+ response elicited by self peptides, whereas the decreased Ca2+ flux following strong TCR stimulation is still sufficient to induce proliferation [6,22,23].

Despite the uncertainties over the mechanisms underlying spontaneous apoptosis of Gimap5lyp/lyp T-cells, we and others have consistently observed that freshly isolated T-cells from BB-DP rats showed a loss of mitochondrial membrane potential that increased over time upon culture in vitro [18,19]. Several studies have clearly established a critical function for mitochondria in cellular Ca2+ homoeostasis [36,37]. Specifically, following sustained Ca2+ entry via the plasma membrane CRACs, the rising concentration of cytosolic calcium ([Ca2+]c) activates the Ca2+ uniporter on the mitochondrial membrane. This induces a slow membrane potential (Δψm)-driven uptake of Ca2+, which is released later via the Na+/Ca2+ exchanger [38]. This process ensures that Ca2+ entering via the CRAC does not cause a feedback inhibition of the CRAC activity [39,40]. Thus mitochondria act like a slow non-saturable non-linear buffer for intracellular Ca2+ as they sequester [Ca2+]c during periods of rapid Ca2+ entry and sustain the [Ca2+]c level by releasing it slowly, even after the cessation of Ca2+ influx [39]. Ca2+ uptake by mitochondria is facilitated by the movement of mitochondria along the microtubules to sites of Ca2+ entry [41].

In the present study, we investigated the mechanistic basis of the defective Ca2+ response in Gimap5lyp/lyp T-cells. Our results show that GIMAP5 is necessary for mitochondria to accumulate Ca2+ following capacitative entry, and that this process is dependent upon microtubules in T-cells.


Animals and GIMAP5-transfected cell lines

ACI.1u Gimap5lyp/lyp and ACI.1u Gimap5+/+ control rats [42] were maintained in specific pathogen-free conditions. All experiments involving animals were carried out in accordance with institutional ethics committee guidelines. Rat GIMAP5 with an N-terminal Myc or FLAG tag was subcloned from pCDNA3.1 [18] into the pIRES-Puro vector. HEK (human embryonic kidney)-293 cells were transfected using polyethylenimine (Polysciences) and the stable transfectants were selected using puromycin (2 μg/ml) and expression of Myc–GIMAP5 was confirmed by Western blotting using an anti-Myc antibody. mEGFP-a-tubulin-IRES-Puro2b (plasmid 21042) was obtained from Addgene.

Antibodies and reagents

Purified anti-(rat TCR) (R73) and anti-(rat CD28) antibodies were purchased from BD Biosciences. Polyclonal goat anti-mouse antibodies for cross-linking were purchased from Jackson ImmunoResearch Laboratories. Anti-Miro antibody was purchased from Abnova. TG (thapsigargin), CCCP (carbonyl cyanide m-chlorophenylhydrazone), BAPTA [1,2-bis-(o-aminophenoxy)ethane-N,N,N′,N′-tetra-acetic acid], nocodazole and latrunculin B were purchased from Calbiochem. Dynalbeads®-M450 tocylactivated beads, MitoTracker® Green, Texas Red® X-phalloidin, TMRE (tetramethylrhodamine ethyl ester), Fura-2 and Rhod-2 were from Molecular Probes. Antibodies against actin and tubulin, tissue culture medium, fetal bovine serum, poly-L-lysine and all other chemicals were from Sigma–Aldrich.

Cell isolation, measurement of Ca2+ flux and mitochondrial Ca2+

Purification of rat CD4+ T-lymphocytes from rat lymph nodes by negative magnetic selection and measurement of Ca2+ mobilization in individual cells have been described previously [35]. For measurement of mitochondrial Ca2+, control and Gimap5lyp/lyp CD4+ T-cells were equally loaded with MitoTracker® Green (100 nM) and Rhod-2 (5 μm) for 40 min at 37°C. Cells were washed in PBS, seeded at 5×106 cells/ml on to poly-L-lysine-coated Lab Tek II™ chamber slides (Nunc) and incubated for 30 min to allow complete de-esterification of intracellular AM (acetoxymethyl ester) esters before starting fluorescence measurements. Cells were examined with a scanning confocal microscope FV1000 (Olympus) coupled to an inverted microscope with a 60× oil-immersion objective (Olympus). Specimens were laser-excited at 488 nm (40 mW argon laser) and 543 nm (diode laser). All instrument settings (laser power, pixel excitation time, confocal slit opening, spectral bandwidth and photomultiplier sensitivity) were identical for control and Gimap5lyp/lyp cells. Serial optical sections of 512×512 pixels were taken at 45 s intervals through the scanning process. Images were acquired from three fields for each experimental condition. TG or anti-TCR antibody was added to the medium in the absence and presence of inhibitors after recording the resting level fluorescence. Accumulation of Ca2+ in mitochondria was determined on MitoTracker® Green/Rhod-2 merged fluorescence images for which the dots’ fluorograms were obtained by plotting pixel intensity values of each component on their respective axes. Quadrant markers based on threshold level signals were placed on the pixel data obtained before stimulation by defining background (C), red-only (D), green-only (A) and co-localization areas (B) (Figure 1A). The co-localization index was calculated as (B)/(B+D), and percentage co-localization as (B)/(B+D)×100 [43]. For the purpose of illustration, the images were contrast enhanced pseudocoloured according to their original fluorochromes, merged (FluoView software; Olympus), cropped and assembled (Adobe Photoshop software; Adobe Systems).

Figure 1 GIMAP5 deficiency impairs mitochondrial Ca2+ uptake in T-cells

CD4+ T-cells from lymphopenic Gimap5lyp/lyp and control rats, purified by negative magnetic selection, were loaded with Rhod-2 and MitoTracker® Green and incubated in medium containing 2 mM Ca2+. Upon binding to cytosolic Ca2+, Rhod-2 emits red fluorescence. After beginning data acquisition using a confocal microscope and establishing the baseline fluorescence signals, cells were stimulated with TG (1 mg/ml) or anti-TCR cross-linking as indicated. (A) Merged (red plus green) fluorescence images of a representative field at one time point after the addition of TG (upper panels), and the corresponding pixel intensity plots for green and red fluorescence (lower panels) are shown. (B) Representative images and the co-localization mask of single cells. (C) Data from one representative experiment showing the proportion of mitochondria that have taken up Ca2+ at different time points following stimulation with TG or anti-TCR mAb. Each data point represents more than 50 cells. Cells that were completely red prior to the addition of TG or anti-TCR were excluded from the analysis. (D) Percentage increase in mitochondrial Ca2+ concentration was calculated from the pooled data of co-localization (over total green fluorescence) from five independent experiments (means+S.E.M.). Arrows indicate the time of addition of reagents. ***P<0.001, **P<0.01 and *P<0.05.

Cytosolic Ca2+ was measured using Fura-2 as described previously [35]. To equalize the concentration of Ca2+ in the cytosol between different samples, cells were permeabilized with 0.1% saponin in cytosol-like buffer (20 mM Tris/HCl, pH 7.4, 110 mM KCl, 10 mM NaCl, 5 mM KH2PO4 and 2 mM MgCl2), supplemented with 20 units of creatine kinase, 20 mM phosphocreatine and 1 mM ATP, as described previously [44].

Analysis of conjugate formation, Western blotting and confocal microscopy

Mouse anti-(rat TCR) mAb (clone R73) and anti-(rat CD28) mAb (clone JJ319) were coated to Dynalbeads®-M450 tocylactivated polystyrene beads following the manufacturer's instructions. For the conjugate formation assay, the antibody-coated polystyrene beads (3×105) were washed and left to adhere to poly-L-lysine-coated glass coverslips for 20 min and washed. Purified CD4+ T-cells (5×105) from wild-type and Gimap5lyp/lyp rats were added to the monolayer of beads and incubated at 37°C for 30 min. After washing off the unbound cells with PBS, bound cells were fixed in 4% paraformaldehyde for 10 min, washed once with PBS and permeabilized in 0.1% of Triton X-100 in PBS for 10 min. After incubation with PBS containing 0.2% BSA for 30 min at room temperature (25°C), cells were labelled with anti-α-tubulin mAb overnight at 4°C. Alexa Fluor® 488-conjugated anti-mouse antibody and Texas Red® X-Phalloidin were added and incubated for 60 min. Nuclei were stained using DAPI (4′,6-diamidino-2-phenylindole) in PBS for 5 min and washed in PBS before mounting in Vectashield® (Vector Labs) medium. Cell–bead conjugates were identified by DIC (differential interference contrast) imaging, followed by acquisition of fluorescence images using an Olympus confocal microscope with a 60× oil-immersion objective. Among the cells that formed conjugates with antibody-coated beads, the ones that showed aggregation of tubulin were counted. In another set of experiments, 1×106 purified CD4+ T-cells were stimulated with anti-TCR/anti-CD28-coated polysytrene beads. At the indicated time points, the cells were lysed in sample buffer, and proteins were separated by SDS/PAGE (12% gel) followed by immunoblotting for tyrosine-phosphorylated proteins using the 4G10 antibody.

HEK-293 cells transiently transfected with Myc–GIMAP5 or FLAG–GIMAP5 expression constructs or the empty vector were labelled with anti-Myc and anti-α-tubulin antibodies. Cells were analysed by confocal microscopy as described previously [18]. For a merged image, a threshold flurogram was prepared and the pixel data from the area with a high degree of colocalization are shown in white pseudocolour.

Statistical analysis

Assay values obtained for Gimap5lyp/lyp and control T-cells, or HEK-293/vector and HEK-293/GIMAP5 cells were compared using the Student's t test or one-way ANOVA using the Prism software (Graph Pad). Pearson's coefficient was calculated using the equation given by Olympus microscopy ( [45].


Mitochondrial Ca2+ accumulation following capacitative Ca2+ entry is defective in Gimap5lyp/lyp rat T-cells

The observations that GIMAP5 deficiency does not completely abrogate TCR-induced Ca2+ response, but only reduces its magnitude [7,35], suggested that the impaired Ca2+ response of Gimap5lyp/lyp T-cells is unlikely to arise from defective CRAC activity. On the other hand, the loss of mitochondrial membrane potential in Gimap5lyp/lyp T-cells [18,19] suggested that their impaired Ca2+ response could arise from defective mitochondrial Ca2+ buffering, compromising the ability to sustain Ca2+ entry via the CRAC. To test this hypothesis, we loaded CD4+ T-cells purified from control and Gimap5lyp/lyp rats with the mitochondria-specific marker MitoTracker® Green and the Ca2+-indicator Rhod-2. We stimulated these cells with TG, which induces Ca2+ release from the ER, leading to opening of the CRAC and mitochondrial Ca2+ uptake. Rhod-2 preferentially accumulates in the mitochondria, but it also binds cytoplasmic Ca2+ [39]. Even though cytosolic Rhod-2 can be removed by electroporation [39], we could not use this method as it caused an unacceptable level of cell death in Gimap5lyp/lyp primary T-cells, rendering them unsuitable for further experiments. Therefore we determined mitochondrial Ca2+ accumulation by measuring co-localization of MitoTracker® Green and Rhod-2 (red) fluorescence (Figure 1 and Supplementary Movies S1 and S2 at, showing wild-type and Gimap5lyp/lyp cells respectively). In wild-type CD4+ T-cells incubated in medium containing 2 mM Ca2+, the addition of TG induced discernible Ca2+ accumulation within mitochondria at approximately 3 min, which continued with a faster kinetics for up to 7 min and then slowly reached a plateau (Figure 1). Throughout this period, mitochondrial Ca2+ accumulation in Gimap5lyp/lyp CD4+ T-cells was significantly lower in magnitude and did not reach the levels attained in control cells even up to 14 min. Gimap5lyp/lyp T-cells also showed impaired mitochondrial Ca2+ accumulation following TCR stimulation (Figures 1C and 1D). The defective mitochondrial Ca2+ accumulation in Gimap5lyp/lyp CD4+ T-cells was not due to cell death because they showed >98% viability at the time of purification and apoptosis was not evident until 9 h later [8], whereas the Ca2+ response experiments were completed within 3 h after cell purification. This contention is supported by the lower magnitude of mitochondrial Ca2+ accumulation in individual Gimap5lyp/lyp T-cells after TG or TCR stimulation (Figure 1B) and the similarity of its kinetics with control T-cells at the population level (Figures 1C and 1D). These results indicate that the loss of GIMAP5 in T-cells compromises Ca2+ sequestration within the mitochondria following a rise in intracellular Ca2+ level.

Forced expression of GIMAP5 in non-T-cells increases mitochondrial Ca2+ accumulation and reduces Ca2+ influx across the plasma membrane

Consistent with an earlier report [46], overexpression of GIMAP5 in human (Jurkat) and rat (C58) T-cell lines induced spontaneous cell death. Therefore we stably transfected GIMAP5 in HEK-293 cells in order to address the role of GIMAP5 on mitochondrial Ca2+ sequestration. We exposed control and GIMAP5-overexpressing HEK-293 cells to TG and measured mitochondrial Ca2+ accumulation. To simultaneously determine whether Ca2+ that accumulates within the mitochondria is derived from intracellular stores, or following Ca2+ entry from the extracellular milieu, we diminished the intracellular Ca2+ pool by equilibrating GIMAP5- and vector-transfected cells in a Ca2+-free medium. We assessed mitochondrial Ca2+ accumulation in these cells following exposure to TG, with or without the addition of Ca2+ to the culture medium. Prior to TG stimulation, GIMAP5-transfected cells maintained in a Ca2+-free or Ca2+-containing medium showed a comparable level of Rhod2 staining with minimal Ca2+ localization in the mitochondria (Figure 2A). Addition of TG caused an increase in mitochondrial Ca2+ content, which was comparable in vector- and GIMAP5-transfected cells maintained in Ca2+-free medium, but was markedly elevated in GIMAP5-expressing cells incubated in Ca2+-containing medium (Figure 2B). These findings indicate that GIMAP5 facilitates mitochondrial Ca2+ uptake following Ca2+ entry from the extracellular milieu.

Figure 2 Overexpression of Myc–GIMAP5 promotes mitochondrial Ca2+ accumulation

(A and B) Stable transfectants of HEK-293 cells expressing Myc–GIMAP5 or empty vector maintained in Ca2+-free medium for 2 h were loaded with Rhod-2 and MitoTracker® Green. Representative confocal images of Rhod-2 fluorescence, MitoTracker® Green and co-localization, before and after the addition of Ca2+, either alone or along with TG, are shown. (C) The percentage increase in mitochondrial Ca2+ in cells exposed to Ca2+ and TG, calculated from three independent experiments as described in Figure 1, is shown as means+S.E.M. **P<0.01 and *P<0.05. The arrow indicates the time of the addition of reagents. (D) Stable transfectants of HEK-293 cells overexpressing Myc–GIMAP5 or the control vector were stained with fura 2/AM, and deposited on to poly-L-lysine-coated slides. The bathing medium was changed to Ca2+-free buffer and stimulated with TG to measure Ca2+ release from the ER. Extracellular Ca2+ was added as indicated to measure Ca2+ entry across the plasma membrane. Fura 2/AM fluorescence, representing free cytosolic Ca2+, was recorded by fluorescent digital video microscopy using the Metaflor® imaging system (Olympus). The mean cytosolic Ca2+ concentration in 100–150 cells from one experiment, representing three independent experiments, is shown.

Whereas mitochondrial Ca2+ buffering is necessary to delay feedback inhibition of CRACs and to sustain Ca2+ influx across the plasma membrane, saturation of the mitochondrial Ca2+ store results in the early attenuation of the Ca2+ influx [3840]. To investigate the effect of increased mitochondrial Ca2+ levels on Ca2+ influx across the plasma membrane in GIMAP5-overexpressing HEK-293 cells, we maintained them in Ca2+-free medium, added TG to elicit Ca2+ release from intracellular stores, then added Ca2+ to the culture medium to induce Ca2+ influx and measured the total intracellular Ca2+ levels (Figure 2C). In cells maintained in Ca2+-free medium, the TG-induced Ca2+ release from intracellular stores was comparable between GIMAP5- and vector-transfected control cells. However, subsequent to Ca2+ addition, the magnitude of Ca2+ influx was significantly reduced in GIMAP5-transfected cells compared with control cells. These observations suggest that the increased mitochondrial Ca2+ content in GIMAP5-overexpressing HEK-293 cells would compromise mitochondrial Ca2+ buffering capacity, resulting in reduced Ca2+ influx across the plasma membrane.

GIMAP5 does not directly influence mitochondrial Ca2+ accumulation

In the above experiments, it is possible that the presence or absence of GIMAP5 directly influenced the ability of mitochondria to accumulate Ca2+, rather than due to the reduced availability of cytosolic Ca2+. To determine whether GIMAP5 directly influenced the capacity of mitochondria to accumulate Ca2+, we permeabilized primary T-lymphocytes and HEK-293 cells overexpressing GIMAP5 with saponin to equilibrate the cytosolic Ca2+ and assessed Rhod-2 fluorescence directly. Permeabilization results in the leaching of residual cytosolic Rhod-2 [39]. As shown in Figure 3(A), equalization of cytosolic Ca2+ resulted in comparable levels of Rhod-2 staining in T-cells from Gimap5lyp/lyp and control rats. Similarly, the intensity of Rhod-2 fluorescence was comparable between vector and GIMAP5-expressing HEK-293 cells (Figure 3B). Addition of CCCP, which uncouples the electron transport chain in mitochondria, decreased the accumulation of Ca2+ in the mitochondria in wild-type and GIMAP5-deficient T-lymphocytes to a comparable extent (Figure 3C). Ca2+ release from the mitochondria has been shown to induce Ca2+ flux from the extracellular milieu [47]. We observed that the CCCP-induced increase in cytosolic Ca2+ concentration was elevated in GIMAP5-expressing HEK-293 cells (Figure 3E), further supporting a role for GIMPA5 in facilitating mitochondrial Ca2+ accumulation. To ascertain that the subcellular source of mitochondrial Ca2+ is indeed the cytosol, we treated Gimap5lyp/lyp and control CD4+ T-cells with BAPTA/AM [BAPTA tetrakis(AM)], a cell-permeable Ca2+ chelator [48], and then exposed the cells to TG. As shown in Figure 3(D), BAPTA/AM completely prevented Ca2+ accumulation within the mitochondria in control as well as in Gimap5lyp/lyp T-cells, indicating that the cytosol is the main source of mitochondrial Ca2+ in T-cells following capacitative Ca2+ entry.

Figure 3 Impaired accumulation of [Ca2+]m is not an intrinsic mitochondrial defect in Gimap5lyp/lyp T-cells

(A) CD4+ T-cells from Gimap5lyp/lyp and control rats or (B) GIMAP5-transfected HEK-293 cells were loaded with Rhod-2 and MitoTracker® Green and incubated in medium containing 2 mM Ca2+. At 5 min before the start of data acqusition, saponin was added to a final concentration of 0.1% to equilibrate the cytosolic Ca2+ concentration with that of the extracellular milieu. Samples were analysed as described in Figure 2. As permeabilization with saponin results in the loss of free Rhod-2 from the cytoplasm, red fluorescence represents the mitochondrial Ca2+. (C and D) Purified CD4+ T-cells from Gimap5lyp/lyp and control rats were loaded with Rhod-2 and MitoTracker® Green and then treated with CCCP (C) or BAPTA/AM (D), both at 100 μM final concentration for 30 min in Ca2+-containing medium. The cells were then stimulated with TG at t0 and Ca2+ uptake by mitochondria was evaluated as described in Figure 2. Representative data from three independent experiments, each representing more than 50 cells per time point, are shown. Note that the scale is different between wild-type and Gimap5lyp/lyp T-cells. (E) Stable transfectants of HEK-293 cells expressing GIMAP5 or empty vector were loaded with fura 2/AM and transferred to medium containing Ca2+ and CCCP throughout the length of the experiment. Fura 2/AM fluorescence representing free [Ca2+]c was recorded as described in Figure 2.

GIMAP5-dependent mitochondrial Ca2+ accumulation requires tubulin

Ca2+ uptake by mitochondria is dependent on their active displacement, which in turn relies on cytoskeletal structures consisting of actin filaments and microtubules [40,49]. In T-lymphocytes stimulated with immobilized anti-TCR antibodies, mitochondrial movement towards Ca2+-rich subplasmalemmal microdomains, formed at the site of initiation of TCR signalling, requires actin polymerization [50,51]. On the other hand, sustained Ca2+ influx following stimulation of Jurkat T-cells with soluble anti-TCR antibody or TG requires microtubule-dependent movement of mitochondria to the sub-plasmalemmal area [50,51]. To address whether this cytoskeleton-dependent movement of mitochondria, subsequent to the induction of Ca2+ influx across the plasma membrane, is influenced by GIMAP5, we used nocodazole to inhibit the microtubule-mediated movement, and latrunculin B to inhibit actin polymerization. We pretreated Rhod-2- and MitoTracker® Green-loaded Gimap5lyp/lyp and wild-type T-cells with nocodazole or latrunculin B before the addition of TG. Consistent with the observations of Quintana et al. [50], latrunculin B did not inhibit TG-induced mitochondrial Ca2+ accumulation either in the control or in Gimap5lyp/lyp T-cells (Figure 4A). On the other hand, nocodazole treatment significantly reduced Rhod-2 fluorescence in wild-type T-cells, but it did not cause any further reduction in mitochondrial Ca2+ uptake in Gimap5lyp/lyp T-cells (Figure 4B). These results suggest that GIMAP5 is involved in microtubule-dependent mitochondrial Ca2+ sequestration in primary T-cells.

Figure 4 Disruption of microtubules impairs mitochondrial Ca2+ uptake in wild-type but not in Gimap5lyp/lyp T-cells

Purified CD4+ T-cells from Gimap5lyp/lyp and control rats, maintained in Ca2+-free medium for 2 h and loaded with MitoTracker® Green and Rhod-2, were pre-incubated with vehicle, latrunculin B (15 μg/ml) for 15 min (A) or nocodazole (2 μM) for 35 min (B), before the addition of Ca2+ and TG. Localization of Rhod-2 fluorescence within mitochondria was evaluated as described in Figure 1. Arrows indicate the time of the addition of reagents. The data shown are representative of ten independent experiments. Pooled data from these experiments, expressed as percentage inhibition of mitochondrial Ca2+ accumulation by latrunculin B or nocodazole, is shown in the lower panels. Results are means+S.E.M.

Next, we examined the effect of nocodazole and latrunculin B on mitochondrial Ca2+ uptake in GIMAP5- or vector-transfected HEK-293 cells following exposure to TG. In contrast with primary T-cells, both nocodazole and latrunculin B abolished mitochondrial Ca2+ uptake in GIMAP5-transfected cells (Figure 5). On the other hand, disruption of either microtubules or actin polymerization had a minimal effect on mitochondrial Ca2+ accumulation in vector-transfected cells (Figure 5). These results suggest that GIMAP5 may participate in both actin- and tubulin-dependent mitochondrial Ca2+ accumulation in non-T-cells, or it might arise from overexpression.

Figure 5 GIMAP5-mediated mitochondrial Ca2+ accumulation in non-T-cells is dependent on both microtubules and actin cytoskeleton

Stable transfectants of HEK-293 cells expressing Myc–GIMAP5, maintained in Ca2+-free medium and loaded with Rhod-2 and MitoTracker® Green, were pre-incubated with vehicle, nocodazole (2 μM) for 35 min (A) or latrunculin B (15 μg/ml) for 15 min (B), before the addition of Ca2+ and TG. Localization of Rhod-2 fluorescence in mitochondria in a representative experiment is shown in the upper panels. The percentage inhibition of mitochondrial Ca2+ accumulation by nocodazole or latrunculin B, calculated from three independent experiments (means+S.E.M.), is shown in lower panels. Arrows indicate the time of addition of reagents. ***P<0.001, **P<0.01 and *P<0.05.

TCR-induced cytoskeletal reorganization and proximal TCR signalling are not affected by GIMAP5 deficiency

Rearrangement of the actin and microtubule cytoskeleton play important roles in T-cell activation by facilitating TCR clustering at the IS (immune synapse) formed between the T-cell and the antigen-presenting cell, the establishment of T-cell polarity and directional secretion of cytokines [5256]. These cytoskeletal structures are also required to achieve maximal Ca2+ flux during T-cell activation [30,57]. However, GIMAP5 deficiency does not compromise T-cell proliferation induced by alloantigens [6,23], suggesting that GIMAP5 does not participate in TCR-stimulated cytoskeletal reorganization at the IS or in proximal TCR signalling events. To address this question, we assessed the reorganization of actin and microtubules in GIMAP5-deficient CD4+ T-cells stimulated via the TCR. As shown in Figure 6(A), actin and tubulin translocated to the site of contact with anti-TCR/anti-CD28-coated beads in control and Gimap5lyp/lyp T-cells with comparable efficiency (Figure 6A; 43% compared with 50%). As the interaction of T-cells with anti-CD3/CD28-antibody coated beads mimics T-cell stimulation by strongly agonistic antigenic peptide by professional antigen-presenting cells, the above results suggest that GIMAP5 is dispensable for T-cell activation by strong TCR agonists. This was further supported by a similar pattern and magnitude of protein tyrosine phosphorylation induced by TCR cross-linking between control and Gimap5lyp/lyp T-cells (Figure 6B). In these experiments, we were unable to identify the key components of the TCR signalling cascade, such as Lck, Zap70, LAT, etc., because commercially available reagents do not recognize rat proteins. Nonetheless, these results indicate that GIMAP5 deficiency affects mitochondrial Ca2+ sequestration without affecting the proximal TCR signalling events induced by strong TCR stimulation.

Figure 6 Cytoskeletal reorganization at the immunological synapse is not affected in Gimap5lyp/lyp T-cells

(A) Purified CD4+ T-cells stimulated for 30 min with anti-TCR (R73) and anti-CD28 antibody-coated polysytrene beads that were previously adhered to poly-L-lysine-coated slides, were fixed and stained for actin and α-tubulin as described in the Experimental section. The number of cells that showed cytoskeletal reorganization at the contact site between the bead and the cell over the total number of cells that was in close proximity with the beads is shown. (B) Purified CD4+ T-cells stimulated for the indicated time with anti-TCR (R73) and anti-CD28 coated polysytrene beads were lysed and immunoblotted for tyrosine-phosphorylated proteins with the 4G10 antibody. Molecular mass is shown on the left-hand side in kDa.

Overexpressed Myc–GIMAP5 co-localizes with proteins involved in the movement of mitochondria

We have shown previously that neither the endogenous nor the overexpressed GIMAP5 is enriched in the subcellular fraction containing mitochondria [18]. However, the results presented above indicate that GIMAP5 facilitates mitochondrial Ca2+ uptake, which is dependent on cytoskeletal elements. Several reports have shown variable involvement of actin filaments and microtubules in mitochondrial movement, and that this variability depends on the cell type and the nature of the stimuli [49,5860]. Mitochondrial cargo is moved on the cytoskeletal track by the adaptor proteins Miro and Milton that link mitochondria to the motor proteins such as kinesin and dynein [41,59]. Therefore we examined whether the GIMAP5-dependent modulation of mitochondrial Ca2+ accumulation is related to the potential interaction of GIMAP5 with the cytoskeletal elements and the mitochondrial adaptor proteins. As shown in Figure 7, overexpressed GIMAP5 co-localized with tubulin, as represented by the white pixels and the Pearson's coefficient (Figure 7A). These results suggest that the interaction between GIMAP5 and cytoskeletal elements probably underlies the GIMAP5-dependent modulation of mitochondrial Ca2+ sequestration. Furthermore, overexpressed GIMAP5 protein co-localized with endogenous Miro (Figure 7B), thereby implicating a role for GIMAP5 in the movement of mitochondria.

Figure 7 Overexpressed GIMAP5 co-localizes with proteins involved in mitochondrial movement

HEK-293 or HeLa cells grown on coverslips and transfected with N-terminal Myc- or FLAG-tagged GIMAP5 were stained for (A) FLAG–GIMAP5 (red, Alexa Fluor® 633) and EGFP (enhanced green fluorescent protein)–Tubulin (green) or (B) Myc–GIMAP5 (green) and Miro1 (red). The nuclei are stained blue with DAPI. White regions represent the overlap between GIMAP5 and the mentioned proteins. Representative results from three independent experiments are shown. The Pearson's coefficient for co-localization was calculated from all three experiments.


We have previously shown that rat T-cells lacking GIMAP5 display a decreased Ca2+ influx across the plasma membrane subsequent to the emptying of the ER Ca2+ stores by TG or following TCR stimulation [35]. In the present study, we have traced this defective Ca2+ response in GIMAP5-deficient T-cells to the inability of their mitochondria to buffer Ca2+ entering from the extracellular milieu [39], that can be either at the level of uptake, storage or release of Ca2+ from mitochondria. The observation that mitogenic activation of T-cells from Gimap5lyp/lyp rats results in robust proliferation additionally suggests that the mitochondria themselves may not be defective, and that the reduced accumulation of Ca2+ may be the result of factors extrinsic to mitochondria. Even though Gimap5lyp/lyp T-cells showed an increase in the frequency of cells that have lost the mitochondrial membrane potential after 8 h in culture (Supplementary Figure S1 at [18,19], there was no significant increase in the generation of reactive oxygen species in the mitochondria or in the cytosol (Supplementary Figure S2 at

In the absence of antigen stimulation, long-term survival of naïve T-cells in the quiescent state is dependent on two essential non-redundant signals, one delivered via the IL-7 receptor and the other via the TCR that constantly engage the MHC molecules presenting peptides derived from self antigens [6166]. At the intracellular level, IL-7 receptor signalling sustains cellular metabolism and up-regulates the pro-survival protein Bcl-2 [24]. The precise nature of the TCR-induced survival signal mediated via Lck remains poorly understood [24,25,67], and hence it is unclear to what extent the signals emanating from IL-7 receptor and TCR converge on common signalling pathways, or provide complementary signals that act in synergy to promote cell survival [24,68]. Normal development of B-cells and T-cells in GIMAP5-deficient rats indicate that these animals are not deficient for IL-7, and that IL-7 signalling is intact [7,69]. Gimap5lyp/lyp rat T-cells proliferate robustly to cross-linked anti-TCR antibody in vitro and nominal antigens in vivo [6,7,23], indicating that these T-cells express functional TCRs and respond to antigens presented by MHC molecules. Accordingly, we observed that proximal TCR signalling was comparable between Gimap5lyp/lyp and control rat T-cells. Therefore we propose that even though the magnitude of Ca2+ flux induced by strong TCR stimulation is reduced in Gimap5lyp/lyp rat T-cells, it is still sufficient to induce cell proliferation. On the other hand, the threshold level of Ca2+ response elicited by weak TCR stimulation following engagement of self-peptide MHC complexes may be so reduced in Gimap5lyp/lyp rat T-cells that it is insufficient to provide cell survival signals in vivo. Even though this prediction is supported by defective homoeostasis of naïve T-cells in mice lacking the Ca(v)1.4 calcium channel [70], formal proof would require T-lymphocyte-specific ablation of the Gimap5 gene in mice, also bearing transgenic TCRs of differing affinities towards cognate peptides.

T-cell proliferation in response to optimal TCR stimulation induced by cognate antigenic peptides, or supra-optimal TCR stimulation induced by TCR cross-linking by mAbs, is associated with the formation of a stable immunological synapse, leading to sustained TCR signalling and a strong Ca2+ response [71]. In contrast, the physiological survival signal elicited in naïve T-cells by MHC–self-peptide complexes may not result in the formation of a stable immunological synapse necessary to elicit sustained TCR signalling. Nonetheless the latter interaction stimulates a small, but clearly discernible, Ca2+ response [3234,72]. Whether this basal level of Ca2+ response is crucial to naïve T-cell survival, and whether reaching that threshold level of Ca2+ response is dependent on mitochondrial Ca2+ uptake, are important questions that remain to be answered. Even though direct experimental evidence for these events would be difficult to obtain, our findings on GIMAP5-deficient rat T-cells shown in the present study suggest that it is a probable scenario. As a corollary, overexpression of GIMAP5 in T-lymphocytes would result in Ca2+ overload in the mitochondria, resulting in cell death, which could explain our inability to establish T-cell lines overexpressing GIMAP5, as also reported by other investigators [46]. On the other hand, lack of any discernible effect of GIMAP5 overexpression on the survival and proliferation of HEK-293 cells may arise from their tolerance to apoptosis induced by mitochondrial Ca2+ overload. Overall, our data suggest that GIMAP5-dependent regulation of mitochondrial Ca2+ homoeostasis may be a critical determinant of naïve T-cell survival and homoeostasis.

The GIMAP5-dependent mitochondrial Ca2+ sequestration relies on cytoskeletal machinery that presumably facilitates the movement of mitochondria towards Ca2+-rich microdomains that arise transiently as sub-plasmalemmal hotspots following capacitative Ca2+ entry [39]. Whereas the disruption of microtubules, but not actin polymerization, prevented GIMAP5-dependent mitochondrial Ca2+ accumulation in primary T-cells (Figure 4) and Jurkat T-cells [50], both were equally effective in HEK-293 cells overexpressing Myc–GIMAP5 (Figure 5). The varying contributions of distinct cytoskeletal structures may reflect the type and intensity of the stimulus as well as the cell-type-specific regulation. Further studies are needed to elucidate the details of the interaction of GIMAP5 with the cytoskeletal elements on one hand, and with the mitochondrial adaptor protein Miro on the other (Figure 7). T-cell stimulation following strong TCR activation by exogenous antigen presented on antigen-presenting cells, or by immobilized anti-TCR antibodies on polystyrene beads, facilitate sustained Ca2+ entry at the immunological synapse, which is accompanied by actin-dependent enrichment of mitochondria beneath the CRACs [51]. On the other hand, soluble anti-TCR antibodies may provide qualitatively different signals to the TCR, such that they rely on microtubule-dependent mitochondrial movement to the sub-plasmalemmal area [50,51]. In the light of these reports, we propose that weak TCR stimuli induced by self-peptides would rely on GIMAP5 to facilitate mitochondrial movement along the microtubules in order to sequester Ca2+. Generation of GIMAP5-deficient TCR transgenic mice would help to resolve this issue.

On the basis of our findings shown in the present study, we propose that GIMAP5 plays an important role in integrating TCR-induced cell survival signals in naïve T-cells. We propose that GIMAP5-mediated mitochondrial Ca2+ sequestration may regulate the activity of plasma membrane Ca2+ channels under conditions where the IS is not formed, in order to achieve a threshold level of intracellular Ca2+ that is essential to sustain cell survival.


Xi Lin Chen, Daniel Serrano and Marian Mayhue carried out the experiments. Hans-Joachim Wieden and Jana Stankova helped with the cloning of GIMAP5. Guylain Boulay contributed to the expertise on videomicroscopy experiments. All of the authors contributed to the planning of the experiments. Xi Lin Chen, Subburaj Ilangumaran and Sheela Ramanathan wrote the paper.


This work was funded by the Canadian Institute for Health Research [grant number MOP-86530 (to S.R.)]. X.L.C. is a recipient of a FRSQ graduate student fellowship. The Centre de Recherche Clinique Etienne-Le Bel is an FRSQ-funded research centre.


We thank Dr Leonid Volkov for technical assistance with confocal microscopy. We thank Dr Markus Hoth for critical reading of the paper prior to submission.

Abbreviations: AM, acetoxymethyl ester; BAPTA, 1,2-bis-(o-aminophenoxy)ethane-N,N,N′,N′-tetra-acetic acid; BAPTA/AM, BAPTA tetrakis(AM); BB-DP, Biobreeding diabetes-prone; CCCP, carbonyl cyanide m-chlorophenylhydrazone; CRAC, Ca2+-release-activated Ca2+ channel; DAPI, 4′,6-diamidino-2-phenylindole; ER, endoplasmic reticulum; GIMAP5, GTPase of the immune-associated nucleotide-binding protein 5; HEK, human embryonic kidney; IL-7, interleukin-7; IS, immune synapse; mAb, monoclonal antibody; TCR, T-cell antigen receptor; TG, thapsigargin


View Abstract