Uncoupling proteins 2 and 3 are fundamental for mitochondrial Ca2+ uniport

Abstract

Mitochondrial Ca(2+) uptake is crucial for the regulation of the rate of oxidative phosphorylation, the modulation of spatio-temporal cytosolic Ca(2+) signals and apoptosis. Although the phenomenon of mitochondrial Ca(2+) sequestration, its characteristics and physiological consequences have been convincingly reported, the actual protein(s) involved in this process are unknown. Here, we show that the uncoupling proteins 2 and 3 (UCP2 and UCP3) are essential for mitochondrial Ca(2+) uptake. Using overexpression, knockdown (small interfering RNA) and mutagenesis experiments, we demonstrate that UCP2 and UCP3 are elementary for mitochondrial Ca(2+) sequestration in response to cell stimulation under physiological conditions - observations supported by isolated liver mitochondria of Ucp2(-/-) mice lacking ruthenium red-sensitive Ca(2+) uptake. Our results reveal a novel molecular function for UCP2 and UCP3, and may provide the molecular mechanism for their reported effects. Moreover, the identification of proteins fundemental for mitochondrial Ca(2+) uptake expands our knowledge of the physiological role for mitochondrial Ca(2+) sequestration.

Figure 1. Expression of UCP orthologues and functional tests on Ca²⁺ carrier function of UCP2 or UCP3 in human endothelial cells.

(A) RT–PCR of UCPs in EA.hy926 (EA) cells and human white adipose tissue (WAT). Methods and primer sequences are available shown in the Supplementary Information, Methods. (B) Targeting of fusion proteins of UCP2 and UCP3. a and d show the localization of the fusion constructs of UCP2 and UCP3; b and e the localization of mito-tracker orange; and c and f the colocalization in the respective cell. Mitochondria structure was visualized in endothelial cells transfected with mitochondria-targeted DsRed (g) or DsRed and UCP2 alone (h). Further information is shown in the Supplementary Information, Fig. S1b. For comparison with mitochondrial fragmentation, DsRed transfected cells were treated with 2 μM FCCP for 5 min (i). The scale bars represent 10 μm in a-f and 2 μm in g–i. (C) UCP2 and UCP3 overexpression increased mitochondrial Ca²⁺ uptake. Mitochondrial Ca²⁺ elevation in response to 100 μM histamine was measured with mitochondria-targeted ratiometric-pericam in intact endothelial cells overexpressing the sensor, either alone (control; n = 32) or together with UCP2 (n = 23) or UCP3 (n = 18). The asterisk indicates P <0.05 versus control. (D) UCP2 and UCP3 overexpression had no effect on histamine-induced changes in the mainly pH-sensitive wavelength of pericam (F₄₈₅; data shown correspond to that shown in C). (E) UCP2 and UCP3 increased capacity, but not sensitivity, of mitochondrial Ca²⁺ uptake. Concentration response curves of histamine on mitochondrial Ca²⁺ elevation were obtained in single endothelial cells overexpressing mitochondria-targeted ratiometric pericam either alone (control, n = 6–12) or together with UCP2 (n = 6–12) or UCP3 (n = 6–12). The asterisk indicates P <0.05 versus control. The respective data on cytosolic Ca²⁺ concentration are shown in the Supplementary Information, Fig. S1k. (F) UCP2 and UCP3 increased mitochondrial Ca²⁺ uptake independently of mitochondrial pH, mitochondrial Na⁺–Ca²⁺ exchanger or mitochondrial membrane potential. In single endothelial cells overexpressing UCP2 (n = 11) or UCP3 (n = 14) mitochondria were depolarized by 2 μM FCCP and 2 μM oligomycin and the mitochondrial Na⁺–Ca²⁺ exchanger was inhibited by CGP 37157, as described previously. The asterisk indicates P <0.05 versus control (n = 14). The error bars represent s.e.m.

Figure 2. UCP2 or UCP3 knockdown by siRNAs resulted in diminution and/or elimination of mitochondrial Ca²⁺ uptake.

(a) siRNAs against UCP2 or UCP3 (see Supplementary Information, Fig. S2a) diminished expression of the respective protein, indicated by western blot analysis. The bars indicate the optical density of the respective band in lower panels, which are normalized according the corresponding β-actin band. Representative blots from three independent experiments are shown. (b) Silencing UCP2 and/or UCP3 expression attenuated and/or eliminated mitochondrial Ca²⁺ uptake. Mitochondrial Ca²⁺ sequestration in response to histamine was measured in single mitochondria-targeted ratiometric pericam-transfected endothelial cells that were cotransfected with control siRNA (see Supplementary Information, Methods), or with siRNAs against UCP2 and/or UCP3. Average curves are shown on the left (control, n = 16; UCP2 siRNA, n = 17; UCP3 siRNA, n = 17; and UCP2/3 siRNA, n = 22). The distribution of individual independent experiments is shown on the right. The bars indicate the mean value. The single asterisk indicates P <0.05 versus control and the double asterisk indicates P <0.05 versus cells transfected with UCP2 siRNA. The corresponding data on cytosolic free Ca²⁺ are provided in the Supplementary Information, Fig. S2b. (c) In single HeLa cells, which express UCP3 but not UCP2 (see Supplementary Information, Fig. S1a), only the UCP3 siRNA (n = 14) was suitable to prevent mitochondrial Ca²⁺ sequestration. The asterisk indicates P <0.05 versus control (n = 12) and UCP2 siRNA (n = 9). (d) In single HeLa cells transfected with UCP3 siRNA (n = 10), overexpression of UCP2 rescued mitochondrial Ca²⁺ uptake, as monitored with cotransfected mitochondria-targeted ratiometric-pericam. The asterisk indicates P <0.05 versus control (n = 11) and UCP2 + UCP3 siRNA (n = 10). (e) Overexpression of UCP2 rescued mitochondrial Ca²⁺ uptake in suspended digitonin-permeabilized HeLa cells that were transfected with UCP3 siRNA (n = 6). Calcium Green-5N (0.1 μM) was used to measure the reduction of extra-mitochondrial Ca²⁺ by mitochondria in digitonin-permeabilized and suspended HeLa cells, which were transiently transfected with either the vector alone (control), UCP3 siRNA or UCP2 + UCP3 siRNA. As indicated (arrows), 5 μM digititonin (Dig.), 90 nmol Ca²⁺ and 2 μM FCCP were added in the presence of 1 μM thapsigargin. The asterisk indicates P <0.05 versus control (n = 6) and UCP2 + UCP3 siRNA (n = 6). The error bars represent s.e.m.

Figure 3. In isolated liver mitochondria from Ucp2^−/− mice, no ruthenium red (RR)-sensitive Ca²⁺ uniporter exists.

(a) Ca²⁺ uptake of suspended isolated liver mitochondria from wild-type (WT) animals was measured by monitoring the reduction of extra-mitochondrial Ca²⁺ on application of 140 nmol Ca²⁺ mg⁻¹ mitochondrial protein with Calcium Green-5N (0.1 μM). Experiments were performed in the absence (n = 7) and presence (WT + RR, n = 4; single asterisk indicates P <0.05 versus WT) of ruthenium red, an established inhibitor of mitochondrial Ca²⁺ uniporter (100 nM). (b) In suspended liver mitochondria isolated from Ucp2^−/− mice, Ca²⁺ uptake (measured as described in a), was reduced (Ucp2^−/−; n = 7). The remaining Ca²⁺ uptake was insensitive to ruthenium red (100 nM; Ucp2^−/− + RR; n = 4). n.s. versus Ucp2^−/−. (c) Kinetics of mitochondrial Ca²⁺ uptake within 30 s after addition of Ca²⁺ calculated from the data presented in a and b. The asterisk indicates P <0.05 versus WT. (d) Fluorescence image of fura-2 loaded single mitochondria freshly isolated from mouse liver. The scale bar represents 10 μm. (e) In isolated single liver mitochondria from wild-type animals, mitochondrial Ca²⁺ sequestration was measured in the absence (WT, n = 44) and presence (WT + RR, n = 22) of ruthenium red (1 μM) by mitochondrial fura-2. The asterisk indicates P <0.05 versus WT. (f) Ca²⁺ uptake in isolated single mitochondria from Ucp2^−/− mice measured by mitochondrial fura-2 was reduced (Ucp2^−/−, n = 34) compared with that obtained in the wild-type littermates. The remaining Ca²⁺ uptake into the liver mitochondria of Ucp2^−/− mice was not ruthenium red-sensitive (Ucp2^−/− + RR, n = 21). n.s. versus Ucp2^−/−. (g) Ruthenium red-sensitive mitochondrial Ca²⁺ uptake in single mitochondria freshly isolated from wild-type and Ucp2^−/− (Ucp2^−/−, asterisk indicates P <0.05 versus WT) mice calculated from data in e and f. The error bars represent s.e.m (a–f) or indicate 95% confidencence intervals at the times shown (g).

Figure 4. Mutagenesis of UCP2 and UCP3 resulted in loss of Ca²⁺ function of the proteins and provided dominant-negative UCP homologues.

(a) Sequence alignment and identification of specific domains common for UCP2 and UCP3, but not UCP1. Additional information is shown in the Supplementary Information. (b) Using the ROBETTA full-chain protein structure prediction server (http://robetta.bakerlab.org), the protein structure of UCP2 and UCP3 was predicted to form a channel-like structure. As indicated, the mutation was introduced in a helix that is located on the matrix side of the inner mitochondrial membrane and results in shortening of the helix loop and loss of polarity. (c) Functional consequences on mitochondrial Ca²⁺ uptake in single endothelial cells transiently expressing mitochondria-targeted ratiometric-pericam overexpressing mutated UCP2 (n = 25) or mutated UCP3 (n = 15). The asterisk indicates P <0.05 versus control (n = 17). (d) Mutated UCP2 failed to rescue mitochondrial Ca²⁺ uptake in single HeLa cells transfected with UCP3 siRNA. HeLa cells were transfected with either a non-functional control siRNA (n = 14), UCP3 siRNA (n = 9) or UCP3 siRNA together with mutated UCP2 (UCP2^mut + UCP3 siRNA, n = 9). Asterisk indicates P <0.05 versus control siRNA.