|Home | About | Journals | Submit | Contact Us | Français|
Uncoupling proteins 2 and 3 (UCP2/3) are essential for mitochondrial Ca2+ uptake but both proteins exhibit distinct activities in regard to the source and mode of Ca2+ mobilization. In the present work, structural determinants of their contribution to mitochondrial Ca2+ uptake were explored. Previous findings indicate the importance of the intermembrane loop 2 (IML2) for the contribution of UCP2/3. Thus, the IML2 of UCP2/3 was substituted by that of UCP1. These chimeras had no activity in mitochondrial uptake of intracellularly released Ca2+, while they mimicked the wild-type proteins by potentiating mitochondrial sequestration of entering Ca2+. Alignment of the IML2 sequences revealed that UCP1, UCP2 and UCP3 share a basic amino acid in positions 163, 164 and 167, while only UCP2 and UCP3 contain a second basic residue in positions 168 and 171, respectively. Accordingly, mutants of UCP3 in positions 167 and 171/172 were made. In permeabilized cells, these mutants exhibited distinct Ca2+ sensitivities in regard to mitochondrial Ca2+ sequestration. In intact cells, these mutants established different activities in mitochondrial uptake of either intracellularly released (UCP3R171,E172) or entering (UCP3R167) Ca2+. Our data demonstrate that distinct sites in the IML2 of UCP3 effect mitochondrial uptake of high and low Ca2+ signals.
The elementary importance of mitochondrial Ca2+ uptake for the regulation of the organelle's physiological functions  and dysfunctions [2,3] is convincingly reported. Mitochondria accumulate Ca2+ from high Ca2+ microdomains that are generated particularly in the vicinity of Ca2+ release channels of the endoplasmic reticulum (ER) [4–6] and Ca2+ entry channels of the plasma membrane (PM) upon acute cell stimulation [7,8]. Hence, mitochondria sequester Ca2+ from moderate submicromolar cytosolic Ca2+ elevations in intact cells by an obvious shift in their Ca2+ sensitivity of the Ca2+ sequestration pathway [9–11]. This versatile competency of mitochondria to decode various cellular Ca2+ signals is accomplished by Ca2+ channels and Ca2+ exchangers at the inner mitochondrial membrane (IMM) that govern different modes of mitochondrial Ca2+ uptake . In addition the transfer of Ca2+ across the IMM appears to be modulated by posttranslational modifications, such as phosphorylation [12,13], small diffusible messengers such as nucleotides  or polyamines , the proton concentration within the matrix of mitochondria  and Ca2+ itself [17,18], which all are thought to affect the activity of mitochondrial Ca2+ channels in intact cells. Such a sophisticated regulation of mitochondrial Ca2+ channels by multiple processes is thought to be essential in order to balance mitochondrial Ca2+ accumulation, which not only stimulates respiration and ATP synthesis  but also can evoke cell death . However, the exact molecular mechanisms as well as their putative interdependencies, which actually control mitochondrial Ca2+ signaling in intact cells remain unclear over a large extent.
Mitochondrial Ca2+ channels are functionally well characterized and convincing electrophysiological recordings of single Ca2+ channel activities on mitoplasts (isolated swollen mitochondria lacking the outer mitochondrial membrane) have been presented [21,22]. Notably, such studies not only proved unambiguously the existence of a functional mitochondrial Ca2+ channel at the IMM but also unveiled the presence of distinct mitochondrial Ca2+ channels with different properties , challenging the concept of a unique mitochondrial Ca2+ uniporter (MCU) . In contrast to these remarkable achievements regarding the functional characterization of mitochondrial Ca2+ channels, mitochondrial Ca2+ channel proteins have not been conclusively identified so far. Nevertheless, very recently an EF-hand containing protein referred to as mitochondrial calcium uptake 1 (MICU1) has been identified to essentially contribute to the mitochondrial Ca2+ uptake machinery . Although MICU1 has been shown to be required for high capacity mitochondrial Ca2+ uptake, based on the predicted structural features of MICU1 this protein seems to rather function as a regulator of (a) mitochondrial Ca2+ channel(s) than to act itself as a Ca2+ channel of the IMM .
In recent reports we  and others  described the fundamental importance of UCP2/3 for mitochondrial Ca2+ accumulation in intact cells , indicating that these proteins either exhibit conductive subunits of a mitochondrial Ca2+-selective ion channel or function as essential modulators of one of the major mitochondrial Ca2+ uptake routes. However, these findings were challenged  and it became apparent that the disclosure of a UCP2/3-dependent mitochondrial Ca2+ uniport critically depends on the experimental procedures and models chosen . However, we have recently demonstrated that constitutively expressed UCP2/3 preferentially contribute to mitochondrial Ca2+ uptake from the intraorganelle gap between the mitochondria and the ER and account for mitochondrial Ca2+ uptake of intracellularly released Ca2+, while these proteins did not contribute to mitochondrial Ca2+ uptake of entering Ca2+. However, once UCP2/3 have been overexpressed, mitochondrial Ca2+ uptake was significantly enhanced independently from the source and mode of supplied Ca2+ (i.e. intracellularly released Ca2+ vs. Ca2+ entering the cell via the so called store-operated Ca2+ entry, SOCE). Thus, the contribution of UCP2/3 to mitochondrial Ca2+ sequestration appears to be determined by the source of supplied Ca2+ as well as their expression levels.
In the present work we used site-directed mutagenesis of UCP3 (and UCP2) to elucidate the structural determinants of such a distinct contribution of the novel UCPs to mitochondrial Ca2+ uptake depending on the source, mode and hence strength of supplied Ca2+.
Cell culture chemicals were purchased at Invitrogen (Vienna, Austria) and fetal calf serum and media supplements were from PAA laboratories (Pasching, Austria). Dulbecco's modified eagle's medium (DMEM), 2,5-di-tert-butylhydroquinone (BHQ), histamine, oligomycin, and digitonin were obtained from Sigma–Aldrich (Vienna, Austria). 7-Chloro-5-(2-chlorophenyl)-1,5-dihydro-4,1-benzothiazepin-2(3H) (CGP 37157), and carbonyl cyanide-p-trifluoromethoxyphenylhydrazone (FCCP) were from Tocris Bioscience (Bristol, UK). Ru360 was ordered from EMD Chemicals Inc. (Gibbstown, NJ, USA) Fura-2/AM and 5,5′,6,6′-tetrachloro-1,1′,3,3′-tetraethylbenzimidazolylcarbocyanine iodide (JC-1) was obtained from Molecular Probes Europe (Leiden, Netherlands). All other chemicals were from Roth (Karlsruhe, Germany).
The human umbilical vein endothelial cell line EA.hy926  (passage 45–85) stably expressing RPmt was used in this study. Cells were grown in DMEM containing 10% FCS, 1% HAT (5 mM hypoxanthin, 20 μM aminopterin, 0.8 mM thymidine), 50 units/ml penicillin, 50 μg/ml streptomycin, and were maintained at 37 °C in 5% CO2 atmosphere. For experiments, cells were plated on 30 mm glass cover slips 2–4 days before use. After reaching 70–80% of confluence, cells were transiently transfected with the different plasmids using the Transfast® reagent according to the protocol supplied by the manufacturer. For measurements of mitochondrial Ca2+ concentration, the protein/mutant of interest was co-transfected with a nuclear-targeted GFP (NLS-GFP)  in a ratio of 3:1 to identify overexpressing cells. In analogy the protein/mutant of interest was co-transfected with mtDsRed  and/or D1ER  for structural analysis of the organelles. All constructs were confirmed by restriction digestion and sequencing. Experiments were performed between 42 and 48 h after transfection.
Full length human UCPs, UCP1 (GenBank accession no. NM_021833.4), UCP2 (GenBank accession no. NM_003355.2) and UCP3 (GenBank accession no. NM_003356.3) were subcloned from previously described constructs  into the CMV-controlled multiple cloning site of the mammalian expression vector pcDNA3.1 (Invitrogen, Carlsbad, CA) by restriction digests using NotI-XhoI (UCP1) or KpnI-XhoI (UCP2 and UCP3) respectively as recently described .
Applying a three-step PCR approach, the UCP2/3 specific intermembrane loop 2 (IML2) of UCP2 and UCP3 was replaced by that of UCP1 to provide the UCP2UCP1, UCP3UCP1 chimeras. Moreover, the IML2 of UCP1 was replaced by that of UCP2 (UCP1UCP2). All chimeras were cloned into pcDNA3.1 using either KpnI-XhoI (UCP2UCP1 and UCP3UCP1) or EcoRI-XhoI (UCP1UCP2).
For UCP3 point mutations suitable primers (5′-CTATGGACGCCTACGGAACCATCGCCAGG-3′ and 5′-CCTGGCGATGGTTCCGTAGGCGTCCATAG-3′ for UCP3R167G or 5′-CAGAACCATCGCCGGGGGGGAAGGAGTCAGG-3′ and 5′-CCTGACTCCTTCCCCCCCGGCGATGGTTCTG-3′ for UCP3RE171/172GG, respectively) exchanged the corresponding codons from AGA to GGA (UCP3R167G) or from AGG GAG to GGG GGG (UCP3RE171/172GG), respectively. Mutants were constructed by using the QuikChange XL site-directed mutagenesis kit (Stratagene, La Jolla, CA).
Cell loading with Fura-2/AM and at rest prior to experiments, cells were kept in a Hepes-buffered solution containing (in mM): 135 NaCl, 5 KCl, 2 CaCl2, 1 MgCl2, 10 Hepes acid, 2.6 NaHCO3, 0.44 KH2PO4, 10 glucose, 0.1% vitamins and 0.2% essential amino acids, 1% penicillin/streptomycin, pH adjusted to 7.4 with NaOH. For experiments in the nominal absence of extracellular Ca2+ the Ca2+ free buffer composed of (in mM): 138 NaCl, 5 KCl, 1 MgCl2, 1 EGTA, 10 glucose and 10 Hepes acid, pH adjusted to 7.4 with NaOH. Prior switching to EGTA-containing solution, and for experiments in the presence of extracellular Ca2+, cells were perfused with a solution containing (in mM): 138 NaCl, 5 KCl, 1 MgCl2, 2 CaCl2, 10 d-glucose and 10 Hepes acid, pH adjusted to 7.4 with NaOH. If not otherwise indicated, this solution was normally used in the Ca2+ readdition experiments with intact cells. For mild cell permeabilization cells were perfused with 3 μM digitonin for 3 min in high KCl-buffer composed of (in mM): 110 KCl, 0.5 KH2PO4, 1 MgCl2, 20 Hepes, 0.03 EGTA, 5 succinate, 10 d-glucose, pH adjusted to 7.4 with KOH. To trigger mitochondrial Ca2+ uptake cells were partially permeabilized and the actual intracellular Ca2+ concentration ([Ca2+]a) was set to 174 ± 18 nM (n = 17) (referred as “low Ca2+”), 319 ± 21 nM (n = 20) (referred as “middle Ca2+”) and 921 ± 119 nM (n = 17) (referred as “high Ca2+”) as calculated using Fura-2 and the following equation: [Ca2+]a = 350 nM·(FCa − Fmin)/(Fmax − FCa).
Free mitochondrial Ca2+ concentration was measured with mitochondria-targeted ratiometric-pericam  using a fluorescence microscope (Zeiss Axiovert 100/AxioObserver, Zeiss, Vienna, Austria) as previously described [35,36]. Due to the sensor sensitivity to changes in pH at 480 nm excitation data presented are normalized to 1 − (F430/F0) as described previously . Cytosolic free Ca2+ was recorded with Fura-2/am  using fluorescence microscope as previously described . Cytosolic free Ca2+ is expressed as ratio of the fluorescence at 340 and 380 nm excitation (F340/F380). The free Ca2+ concentration within the ER lumen was monitored using D1ER  as previously described [35,39]. D1ER was excited at 440 ± 21 nm (440AF21, Omega Optical) and emission was collected simultaneously at 535 and 480 nm with one given camera using an optical beam splitter (535 and 480 nm, Dual-View MicroImagerTM, Optical Insights, Visitron Systems).
High-resolution imaging of cells expressing mtDsRed and/or D1ER was performed using a Nipkow-disk-based array confocal laser scanning microscope (ACLSM) as described previously [24,36]. The ACLSM consisted of a Zeiss Axiovert 200M (Zeiss Microsystems, Jena, Germany) with a 100× objective (α Plan-Fluar 100×/1.45 oil objective, Zeiss Microsystems, Jena, Germany), equipped with VoxCell Scan® (VisiTech, Sunderland, UK), and an air cooled argon ion laser system (series 543, CVI Melles Griot, CA, USA). The laser line 488 nm was used to excite fluo4 and D1ER, whereas alternatively wavelength 514 nm was used to excite mtDsRed. Emitted light was collected at 535 nm (535AF26; Omega Optical, Brattleboro, VT, USA) for fluo4 and D1ER or 570 nm (Omega optical) for mtDsRed using a high resolution CCD camera (Photometrics CoolSNAPfx-HQ, Roper Scientific, Tucson, AZ, USA). Acquisition and analysis were performed with Metamorph 6.2r6 (Universal Imaging, Visitron Systems, Puchheim, Germany). The deconvolution of z-scans with the iterative quick maximum likelihood estimation algorithm (QMLE) was performed with Huygens 2.4.1p3, software (Hilversum, Netherlands). 3D-rendering of ER and mitochondria were done with the Imaris 3.3 software (Bitplane AG, Zürich, Switzerland) [24,39].
Analysis of variance (ANOVA) and Dunnett's Multiple Comparison Test were used for the analysis. P < 0.05 was determined to be significant.
In our recent work the UCP2/3-specific intermembrane loop 2 (IML2), in which the highest heterogeneity with UCP1 exists (Fig. 1A), was found to be critical for the contribution of UCP2/3 to mitochondrial Ca2+ uptake . This assumption was built on mutations in IML2 with a replacement of amino acids 160–169 (UCP2) and 163–172 (UCP3), respectively, with equivalent l-glycines. Interestingly, UCP1, which has been proven to establish a proton leak in intact cells under certain conditions [40,41], was found to be inactive in terms of mitochondrial Ca2+ uptake . Because of the striking heterogeneity of UCP2/3 vs. UCP1 in the IML2, we first investigated whether the substitution of IML2 of UCP2 and UCP3 by that of UCP1 (Fig. 1B) affects the Ca2+ function of the two novel uncoupling proteins. Both chimeras, i.e. UCP2UCP1 and UCP3UCP1, had no effect on mitochondrial sequestration of intracellularly released Ca2+ while these UCP mutants mimicked the effects of the wild-type isomers regarding a strong enhancement of mitochondrial uptake of entering Ca2+ (Fig. 1C). These data differed from the effects of an overexpression of the wild-type proteins that yielded significantly enhanced mitochondrial Ca2+ uptake from intracellularly released and entering Ca2+ (Fig. 1D), thus, pointing to the respective IML2 to be essential for the contribution of UCP2/3 to mitochondrial Ca2+ uptake of intracellularly released Ca2+. Notably, like their wild-type isoforms , UCP2UCP1 and UCP3UCP1 had no effect on cytosolic Ca2+ signaling (data not shown).
Since the substitution of the entire IML2 by l-glycine in UCP2/3 was found to be inactive in terms of mitochondrial Ca2+ uptake independently from its source , we speculated that the IML2 of the UCP1 contains a distinct site that is suitable to accomplish the contribution to mitochondrial uptake of entering Ca2+ but not of intracellularly released Ca2+ if it replaces the IML2 of UCP2/3.
In contrast, expression of the UCP1UCP2 chimera in which the IML2 in UCP1 was replaced by that of UCP2 failed to reveal any changes in mitochondrial Ca2+ uptake, indicating that the exchange of just the IML2 from UCP2 to UCP1 is insufficient to establish some Ca2+ transport function (Fig. 1E).
Alignment of the respective IML2 sequences (Fig. 1A) revealed that UCP1, UCP2 and UCP3 strikingly share a basic amino acid residue (l-arginine or l-lysine) in positions 162, 164 and 167, respectively. Hence, only UCP2 and UCP3 contain a second basic residue (i.e. l-arginine) at positions 168 and 171, respectively, while UCP1 has an l-threonine at the respective position (166). In order to investigate the importance of the two basic domains in the IML2 for the discriminating function of UCPs regarding the two different sources (i.e. intracellular release and entry), the l-arginine residue in position(s) 167 (UCP3R167G) or 171/172 (UCP3RE171/172GG) in the IML2 of UCP3 was substituted by l-glycine(s) (Fig. 2A).
As in our previous studies the mutation in IML2 of UCP3 with a replacement of amino acids 163–172 by l-glycines did not affect the mitochondrial targeting  it can be assumed safely that the mutants UCP3R167G and UCP3RE171/172GG are also well targeted towards the IMM. In line with these findings, UCP3R167G and UCP3RE171/172GG mutants did neither affect basal ER Ca2+ content (Fig. 2B), agonist-triggered ER Ca2+ depletion, subsequent ER Ca2+ refilling (Fig. 2C), nor the architecture of the ER (Fig. 2D). Both UCP3 mutants did not affect the overall structure and shape of mitochondria (Fig. 2E and F) or the average surface area, volume and number of mitochondria per cell (Fig. 2G). Moreover, a more detailed analysis on potential effects of these mutants on the architectural organization of mitochondria revealed that the expression of the UCP3R167G or UCP3RE171/172GG mutant had no effect on the cellular distribution of these organelles that was verified by analyzing the relative position of mitochondria to the cell center (Fig. 2H and I).
Like their wild-type isoforms and the UCP2UCP1/UCP3UCP1 chimeras, the UCP3R167G and UCP3RE171/172GG mutants had no effect on cytosolic Ca2+ signals that were triggered by ER Ca2+ mobilization and store operated Ca2+ entry (Fig. 3A and B, respectively). However, in agreement with our experiments with the UCP2UCP1/UCP3UCP1 chimeras described above, the UCP3RE171/172GG mutant had no effect on mitochondrial Ca2+ uptake of intracellularly released Ca2+ but mimicked wild-type UCP3 in the strong elevation of mitochondrial sequestration of entering Ca2+ (Fig. 3C). These data demonstrate a crucial importance of a basic residue in position 171 of the UCP3 exclusively for the protein's contribution to mitochondrial Ca2+ uptake of intracellularly released Ca2+. Strikingly, the UCP3R167G mutant mimicked its wild-type isoform in regard to its elevating effect on mitochondrial Ca2+ accumulation of intracellularly released Ca2+ but failed to affect mitochondrial Ca2+ uptake of entering Ca2+ (Fig. 3D), pointing to a crucial importance of the basic residue in position 167 of UCP3 exclusively for mitochondrial Ca2+ uptake of entering Ca2+.
Like that of the wild-type UCP3, mitochondrial Ca2+ uptake by both UCP3 mutants was sensitive to mitochondrial depolarization, while under these conditions their principal differences in responsiveness remained. In particular, despite the overall mitochondrial Ca2+ uptake was largely reduced under depolarizing conditions, the UCP3R167G mutant still behaved like an overexpressed wild-type UCP3 and yielded almost 3-fold elevation of mitochondrial Ca2+ accumulation, while the UCP3RE171/172GG mutant had no effect on mitochondrial uptake of intracellularly released Ca2+ (Fig. 4A and B). In contrast, neither the wild-type UCP3 nor its mutants yielded effects on mitochondrial sequestration of entering Ca2+ under conditions of mitochondrial depolarization (Fig. 4C and D), possibly due to the strongly reduced Ca2+ entry under these conditions (data not shown).
Accordingly, two distinct residues were described herein that are essential for the contribution of UCP3 to mitochondrial sequestration of Ca2+ provided from distinct sources, the ER and the extracellular area. These findings are in line with the current concept of specific coupling between the ER and the mitochondria [4,42–45] and lead us to hypothesize that limited, optimized Ca2+ transfer sites between the ER and the mitochondria exist . Two recent studies convincingly demonstrate that Ca2+ hot spots on the surface of mitochondria are established by intracellularly released Ca2+ but not store-operated Ca2+ entry [46,47]. In agreement with this work, the promptitude of mitochondrial sequestration of intracellularly released Ca2+ (i.e. the time to reach maximum rate, ) appears to be very similar despite independently from the actual amount of Ca2+ release (Fig. 5A), thus, indicating a rather direct functional coupling between the ER and the mitochondria. In contrast, the promptitude () of mitochondrial Ca2+ uptake of entering Ca2+ critically depended on the actually strength of Ca2+ entry (Fig. 5A and B), thus, indicating that the functional plasma-membrane-to-mitochondria Ca2+ coupling critically depends on the strengths and dynamics of Ca2+ entry. Moreover, the analysis of mitochondrial contact sites with the ER and the plasma membrane revealed a large proportion of mitochondria to be in close proximity with the ER while the proximity of the mitochondria with the plasma membrane was very limited in the cell type used for this study (Fig. 5B).
Altogether, there is evidence that the actual Ca2+ concentration at the mitochondrial surface differs from the source of Ca2+ (i.e. intracellular Ca2+ release vs. Ca2+ entry) mobilization. Consequently, one might speculate that the two UCP3 mutations that exclusively mimicked the wild-type UCP3 in sequestration of either intracellularly released (UCP3R167G) or entering Ca2+ (UCP3RE171/172GG) might exhibit different Ca2+ sensitivities. In order to test possible differences in the Ca2+ sensitivity of UCP3R167G and UCP3RE171/172GG compared with that of the wild-type UCP3, experiments on mitochondrial Ca2+ uptake in digitonin-permeabilized cells that (over)expressed the respective UCP3 isoforms were performed. The Ca2+ sensitivities of the UCP3 isoforms were tested by the addition of a low, middle and high Ca2+ concentration to permeabilized cells. Thereby the actual Ca2+ concentrations that were reached within digitonin-permeabilized cells were 174 ± 18 nM at low Ca2+ concentration, 319 ± 21 nM at middle Ca2+ concentration and 921 ± 119 nM at high Ca2+ concentration. Overexpression of wild-type UCP3 yielded similar increase in mitochondrial Ca2+ sequestration upon all Ca2+ concentrations applied (Fig. 6A and B). The expression of the UCP3RE171/172GG mutant mimicked the effect of an overexpression of wild-type UCP3 only upon the addition of low Ca2+ while its effect was reduced upon the addition of middle and completely lost upon the addition of high Ca2+ (Fig. 6A and B). In contrast, expression of the UCP3R167G mutant had no effect on the mitochondrial uptake upon low concentration of Ca2+, while it was more effective than the wild-type UCP3 upon the addition of middle and high Ca2+ (Fig. 6A and B).
Experiments under these conditions with the ruthenium amine derivate Ru360, a specific inhibitor of mitochondrial Ca2+ uniport , revealed similar sensitivities of the UCP3 mutants to Ru360 on the respective Ca2+ signals than that found for the wild-type UCP3 (Fig. 6C and D).
Overall these findings point to distinct Ca2+ sensitivities that are determined by two probably interacting sites within the IML2 of UCP3. Moreover these results reveal a sophisticated molecular switch that is able to adjust the UCP3-dependent mitochondrial Ca2+ uniport to versatile Ca2+ signals (Fig. 7).
While there is increasing evidence that support the concept of UCP2/3 as being part of the/a mitochondrial Ca2+ uniporter [24–26,33,49], the putative involvement of UCP2 and UCP3 to mitochondrial Ca2+ uptake is still a matter of controversy [1,50–52]. Moreover, the co-existence of more than one function- and molecular-distinct mitochondrial Ca2+ uptake route has been recently demonstrated [22,33,53,54], thus, challenging the concept of one protein/phenomenon to be responsible for mitochondrial Ca2+ sequestration. Accordingly, the aim of this work was to provide the first mechanistic insights into the UCP2/3-dependent mitochondrial Ca2+ uptake machinery in order to proceed with the dissection of the different mitochondrial Ca2+ sequestration mechanisms. Overall our present data suggest that UCP3 works as a complex molecular switch with two levels of a Ca2+ sensitivity for its contribution to mitochondrial Ca2+ uptake (Fig. 7) of which the essential structural determinants are embedded in the IML2.
In our previous work, UCP1 was found to be inactive in regard to mitochondrial Ca2+ uptake . Moreover, a supplementation of the characteristic IML2 sequence in UCP2 and UCP3 by l-glycines (UCP2/36G) prevented any Ca2+ function of these proteins, thus indicating that the IML2 is crucial for the contribution of UCP2/3 to mitochondrial Ca2+ uptake. Our present findings that the chimera in which the IML2 of UCP1 was supplemented by that of UCP2 (UCP1UCP2) could not establish any mitochondrial Ca2+ uptake (Table 1) indicate that for establishing the contribution of UCPs to mitochondrial Ca2+ uptake other domains of the UCP2/3 proteins besides the IML2 are essential.
Nevertheless, the distinct role of the IML2 for the Ca2+ function of UCP2/3 was further demonstrated by using the converse chimeras in which the IML2 of UCP2 and UCP3 was replaced by that of UCP1 (UCP2UCP1 and UCP3UCP1). In particular, our findings that both chimeras had no effect in terms of mitochondrial sequestration of intracellularly released Ca2+ while they mimicked the effect of wild-type UCP2/3 on mitochondrial uptake of entering Ca2+ significantly differ from the previous results with UCP2/36G (Table 1). These differences indicate that the sequence of the IML2 of UCP1 in either UCP2 (UCP2UCP1) or UCP3 (UCP3UCP1) exclusively supports the uptake of entering Ca2+. Since none of the chimeras affected the respective cytosolic Ca2+ signaling, these data point to specific changes in the activity of mitochondrial Ca2+ uptake upon IML2 variations rather than changes in the Ca2+ supply to the surface of the mitochondria.
Using sequence-based annotation analysis, the comparison of the IML2 of UCP1 vs. that of UCP2/3 revealed that all proteins share a basic residue at the N-terminal side of the IML2. This is l-arginine in position 162 of UCP1 and 167 of UCP3 as well as l-lysine in position 164 of UCP2. In view of our previous findings with the inactive UCP2/36G mutants  and the properties of UCP2UCP1 and UCP3UCP1 to exclusively sequester entering but not intracellularly released Ca2+, we speculated that the N-terminal basic amino acid residue in the IML2 at position 164 and 167 of the IML2 of UCP2 and UCP3, is essential for the contribution of these proteins to the uptake of entering Ca2+. This hypothesis was further proven by our experiments with the UCP3 mutant, in which the l-arginine at position 167 was supplemented by an l-glycine (UCP3R167G). Strikingly, expression of UCP3R167G yielded elevated mitochondrial uptake of intracellularly released Ca2+ while this mutant was obviously inactive to contribute to the mitochondrial Ca2+ uptake of entering Ca2+. Overall these data indicate that the N-terminal basic amino acid residue of the IML2 in UCP2/3 is crucial for the contribution of these proteins to mitochondrial uptake of entering Ca2+.
Since neither the UCP2/36G nor the UCP2UCP1 and UCP3UCP1 exhibited any activity in regard to mitochondrial Ca2+ uptake of intracellularly released Ca2+, a crucial role of amino acid residues at the C-terminal site of IML2 for this particular function of UCP2 and UCP3 was hypothesized. In line with this assumption, the C-terminal region of IML2 in UCP1 significantly differs from that of UCP2/3. In particular, the protonated A-R-E-E sequence of positions 168–170 of UCP2 and 171–173 of UCP3, respectively, significantly differs from the polar A-T-T-E domain of UCP1. Accordingly, we speculated this sequence to be crucial for the contribution of UCP2/3 to mitochondrial Ca2+ uptake of intracellularly released Ca2+. Our findings that the UCP3 mutant in which the l-arginine and l-glutamate in positions 171 and 172, respectively, were substituted with l-glycines (UCP3RE171/172GG) exclusively exhibited elevated mitochondrial uptake of entering but not intracellularly released Ca2+ strongly support our conclusion that the C-terminal protonated residue of the IML2 in UCP2/3 is crucial for the contribution of these proteins to mitochondrial uptake of intracellularly released Ca2+.
Importantly, our findings that neither UCP3RE171/172GG nor UCP3R167G affected histamine-induced cytosolic Ca2+ signaling or ER Ca2+ depletion, ER architecture, or mitochondrial structure and cellular distribution indicate that different properties of the two UCP3 mutants for mitochondrial Ca2+ accumulation upon stimulation are due to their effect on the competence of mitochondria to sequester either intracellularly released or entering Ca2+. This assumption is further supported by the sensitivity of both mutants to mitochondrial depolarization and Ru360 that was not distinguishable from that of wild-type UCP3. Accordingly, these data describe two specific motifs in the IML2 of UCP2 and UCP3 to be essential for the Ca2+ function of these proteins (Table 1). Interestingly, these particular sequences in the IML2 of UCP2 and UCP3 are specific for these proteins and are not shared by any other member of the UCP family and, thus, do not belong to the so-called UCP signatures in the first, second and fourth α-helix [55,56]. The lack of such specific sequences might explain the negative findings in a genome-wide Drosophila RNA interference screen  as the Drosophila UCPs only share app. 30% homology with UCP2 and UCP3 and do not contain the particular motifs in their predicted IML2 (Table 2).
While these findings highlight the essential existence of two sites in the IML2 for the contribution of UCP2/3 to mitochondrial uptake from Ca2+ from the two major sources (i.e. intracellular Ca2+ release and store-operated Ca2+ entry) in non-excitable cells, the actual molecular cause of the differences in the activity of the two UCP3 mutants in terms of the source of supplied Ca2+ remains illusive. However, our present finding on the different Ca2+ sensitivities of the UCP3 mutants and the comparison with the wild-type UCP3, provide important information that can explain the observed differences in the contribution of UCP3 mutants to mitochondrial uptake of Ca2+ supplied varyingly strong from different sources.
A specific coupling between the ER and the mitochondria [4,42–45,57] and the existence of Ca2+ hot spots on the surface of mitochondria by intracellularly released but not entering Ca2+ have been demonstrated . Our data on the different kinetics of mitochondrial sequestration of Ca2+ either supplied by intracellular Ca2+ release or store-operated Ca2+ entry are in agreement with these reports and further point to differences in the kinetics and local concentration of supplied Ca2+ by intracellular stores (i.e. fast kinetics and high Ca2+ concentration) vs. the store-operated Ca2+ entry pathway (i.e. slow kinetics with moderate Ca2+ elevation) at the mitochondrial surface. Such diversity in the superficial Ca2+ concentration at the mitochondrial surface requires either multiple carriers and/or different uptake modes in order to achieve the decoding of versatile cellular Ca2+ signals by mitochondria.
Though multiple carriers for mitochondrial Ca2+ uptake have been described [22,33,53], reviewed in , there is strong evidence that the given mitochondrial Ca2+ uniport phenomenon might work in different modes depending on distinct Ca2+ sensitivities [9–11]. In this study the wild-type UCP3 was found to cover a large range of Ca2+ signals to be sequestered into the mitochondria, thus indicating that one given protein is capable to contribute to mitochondrial Ca2+ uptake at low and high Ca2+ concentrations. Moreover, two Ca2+ sensitivities of the UCP3 became obvious when testing the UCP3 mutants, thus supporting the concept of various modes of mitochondrial Ca2+ uptake described by others [10,11,58]. In particular, the experiments with digitonin-permeabilized cells revealed that the UCP3R167G mutant exhibits almost normal function at high Ca2+ concentration but lacks Ca2+ uptake activity under low Ca2+ conditions. These characteristics perfectly match the results in intact cells in which UCP3R167G mimicked the wild-type protein in regard to mitochondrial uptake of intracellularly released Ca2+, while it appeared to be inactive to sequester entering Ca2+. On the other hand, the UCP3RE171/172GG mutant, which mimicked the wild-type UCP3 exclusively for mitochondrial uptake of entering Ca2+ in intact cells, exhibited its activity in digitonin-permeabilized cells only at low Ca2+ concentrations while it appeared to be inactive at high Ca2+ concentrations. These data suggest that the two residues of IML2 are crucial for two distinct Ca2+ sensitivities of UCP3 in its contribution to mitochondrial Ca2+ uptake. It appears that UCP3 works as a sophisticated molecular switch that is conducive to mitochondrial Ca2+ uptake in a high and low Ca2+ sensitive manner and allows the protein to meet the demands for mitochondrial Ca2+ uptake from different sources (Fig. 7).
In order to explain our findings with the two UCP3 mutants, one might speculate that the individual domains independently account for distinct Ca2+ sensitivities of the UCP3, thus, a mutation of one domain exclusively affects only the respective Ca2+ sensitivity. Accordingly, the mutation in the 171/172 amino acids results in a mutant with remaining high Ca2+ sensitivity for achieving uptake of entering Ca2+, while the contribution of this UCP3 mutant to mitochondrial uptake under high Ca2+ conditions (i.e. intracellular Ca2+ release) is prevented. On the other hand, a mutation at position 167 of UCP3 yields a mutant that appears to lack the high Ca2+ sensitivity while it achieves mitochondrial Ca2+ uptake under high Ca2+ conditions.
In agreement with our present findings that the UCP2/3-dependent mitochondrial Ca2+ uptake exhibits two Ca2+ sensitivities, the modulation of the activity of the mitochondrial Ca2+ uniport may be a putative signature of an altered Ca2+ sensitivity. Possibly this is accomplished by kinases such as the serine/threonine kinases p38 MAPK [13,59,60], a novel PKC isoform  or a ER or ER-related protein located in mitochondria . Although it is too speculative to consider a context between these reports and the present findings at the present stage, the report on the existence of respective phosphorylation sites in UCP1  and the existence of a threonine in the IML2 of UCP2 (position 165) and UCP3 (position 168) (Fig. 2) but no other human (and Drosophila) UCP family members (Table 2) allow to predict such link and animates for further investigation.
Another very interesting possibility on the underlying mechanisms of these two distinct sites that predict the Ca2+ sensitivity of the mitochondrial Ca2+ uniport, comes with the discovery that the MICU1-encoded mitochondrial EF hand protein is required for the Ca2+ uptake in this organelle in HeLa cells . Unlike UCP2/3, MICU1 is not predicted to form a channel-like structure but rather to serve as a mitochondrial Ca2+ sensor as member of a set of proteins that establishes (one particular type of) mitochondrial Ca2+ uptake . While an interaction between UCP2/3 and MICU1 has not been studied so far, one can speculate that the mutations of the two distinct sites of IML2 of UCP2/3 might affect the protein interactions with the Ca2+ sensor MICU1, resulting in an altered Ca2+ sensitivity of the Ca2+ uptake complex.
This work provides further mechanistic information on the contribution of UCP2/3 to mitochondrial Ca2+ sequestration. In particular two distinct sites in the IML2 of these proteins have been identified as to be responsible for two distinct Ca2+ sensitivities of the UCP2/3-consisting mitochondrial uptake machinery of either intracellularly released or entering Ca2+. These findings add to recent reports on the versatility of Ca2+ signals that are decoded by the mitochondrial Ca2+ uptake machinery and are in agreement with reports on different Ca2+ sensitivities of the mitochondrial Ca2+ uniporter. The identification of these two molecular sites for Ca2+ sensitivity of the contribution of UCP2/3 to mitochondrial Ca2+ uptake represents a crucial step in the understanding of the molecular mechanisms of UCP2/3-dependent Ca2+ uptake in mitochondria.
All authors state that they do not have any kind of conflict of interest regarding this paper.
We thank A. Miyawaki, Riken, Japan, for mitochondria-targeted ratiometric pericam, C.J.S. Edgell, NC, USA, for the EA.hy926 cells, and N. Demaurex, University of Geneva Medical Center, Geneva, Switzerland for the NLS-GFP, and T. Pozzan, University of Padua, Italy for sending us mtDsRed. The excellent technical assistance of Anna Schreilechner, BSc, is highly appreciated by the authors. This work was supported by the Austrian Science Funds (P20181-B5; P21857-B18; F3010-B05). The Institute of Molecular Biology and Biochemistry was supported by the infrastructure program of the Austrian Ministry of Education, Science and Culture.