Mitochondria maintain maturation and secretion of lipoprotein lipase in the endoplasmic reticulum

Materials

Cell culture media and substitutes, specific primers, DNA standards, Moloney murine leukaemia virus RT (reverse transcriptase) and first-strand buffer were obtained from Invitrogen Life Technologies. Foetal calf serum was from PAA Laboratories. Cell culture plasticware was purchased from Bertoni. The RNeasy Mini kit was from Qiagen. RQ1 RNase-free DNase I and TransFast™ were obtained from Promega. ³H-labelled and non-radioactive triolein, BHQ [2,5-di-(t-butyl)-1,4-hydroquinone], histamine and BSA fraction V were from Sigma–Aldrich. Fura 2/AM (fura 2 acetoxymethyl ester) was from Molecular Probes. RNAguard and random hexameric primers were purchased from Amersham Biosciences. CGP 37157 was purchased from Tocris Cookson. Heparin was from Boehringer Ingelheim. Hot Fire DNA polymerase I was from Solis BioDyne. Luminol reagent was obtained from Santa Cruz Biotechnology, and Ultima Gold scintillation cocktail was from PerkinElmer. X-ray films were from Siemens. Nitrocellulose, protein standards, dNTPs and all other chemicals were purchased from Roth.

Cell culture and adenovirus infection

CHO-K1 (Chinese-hamster ovary K1) cells were cultured on six-well dishes in Ham's F12 containing 10% foetal calf serum. Upon reaching confluence, cells were incubated with 1% (w/v) BSA in serum-free Dulbecco's minimum essential medium for 24 h before experiments. The human umbilical-vein-derived endothelial cell line EA.hy926 [22] at passages higher than 44 was grown on six-well dishes in culture medium containing 10% foetal calf serum and 1% HAT (5 mM hypoxanthine, 20 μM aminopterine and 0.8 mM thymidine). Subconfluent cells were infected with 100 MOI (multiplicity of infection) of adenovirus encoding human LPL [23] in culture medium containing 2% foetal calf serum which was replaced for regular medium after approx. 12 h.

RT-PCR

RT-PCR was performed as described previously [24]. Amplification was carried out with initial denaturation at 95 °C for 10 min followed by 40 cycles of denaturation at 95 °C and annealing at 55 °C for 30 s each and elongation at 72 °C for 60 s. The specific primers used for human LPL were 5′-GCATTGCAGGAAGTCTGACC-3′ and 5′-GATGTTCTCACTCTCGGCCA-3′ yielding a 632 bp product. As a control, a 537 bp fragment of GAPDH (glyceraldehyde-3-phosphate dehydrogenase) was amplified using the primers 5′-ACCACAGTCCATGCCATCAC-3′ and 5′-TCCACCACCCTGTTGCTGTA-3′. PCR products were separated on 2% TAE (Tris/acetate/EDTA)–agarose gels.

Western blot analysis

Western blots were carried out following our protocol published previously [25]. Briefly, 50 μg of total protein or 200 μl of vacuum-concentrated supernatant was mixed with sample buffer yielding a final concentration of 4% (v/v) glycerol, 1% (w/v) SDS, 0.03% (w/v) Bromophenol Blue and 0.6% (v/v) 2-mercaptoethanol in 12.6 mM Tris/HCl, pH 6.8, and was boiled at 95 °C for 10 min before separation on SDS/12.5% polyacrylamide gels. After electrotransfer of the proteins on to nitrocellulose membranes, non-specific binding sites were blocked with 5% (w/v) non-fat instant milk powder in PBS, and the membranes were probed at 4 °C overnight with rabbit polyclonal antibody against recombinant human LPL diluted 1:1000 in 5% (w/v) non-fat instant milk powder in PBS containing 137 mM NaCl, 2.7 mM KCl, 8 mM Na₂HPO₄·2H₂O, 1.5 mM KH₂PO₄, pH 7.4. Subsequently, membranes were washed with PBS containing 0.05% (v/v) Tween 20, incubated for 1 h with horse-radish-peroxidase-conjugated secondary antibody and detected by enhanced chemiluminescence after additional washing steps.

Determination of LPL enzyme activity

Following incubation with the respective substances in culture medium containing 0.5 μg/ml heparin, supernatants were collected on ice, and the cells were lysed by sonication in chilled 50 mM NH₃/NH₄Cl solution, pH 8.1. The protein concentration of cell lysates was determined using the established method of Lowry [26]. LPL activity was determined by the hydrolysis of ³H-labelled triacylglycerols as described previously [27]. [³H]Triolein (1 μCi) and 920 μg of non-radioactive triolein per sample were evaporated under a stream of nitrogen and mixed with 20 μl of 1 M Tris/HCl, pH 8.6, 20 μl of 1% (v/v) Triton X-100 and 80 μl of water by extensive sonication in ice-cooled water. Following addition of 40 μl of heat-inactivated (50 °C for 1 h) human serum containing apoCII as activator and 40 μl of 10% (w/v) BSA, 100 μl of cell lysate or supernatant was mixed with 200 μl of the resultant chilled substrate and were incubated for 30 min at 37 °C with continuous shaking. The reaction was stopped by adding 3.25 ml of an organic solvent mixture consisting of methanol/chloroform/n-heptane (1.41:1.25:1, by vol.). By addition of 1 ml of 0.1 M K₂CO₃/H₃BO₃, pH 10.5, non-esterified (‘free’) fatty acids became extractable from a two-phase system by vigorous shaking for 20 s and subsequent centrifugation at 2700 g for 20 min. A 1 ml sample of the upper phase was mixed with 8 ml of Ultima Gold scintillation cocktail and measured in a β-counter (Beckmann). LPL activity was expressed as amount of non-esterified fatty acids, hydrolysed per minute by lipolytic enzyme contained within or in the supernatant of 1 mg of total cellular protein as described previously [27]. This assay was not directly affected by BHQ, histamine or CGP 37187.

Determination of [Ca²⁺]_cyto (cytosolic free Ca²⁺ concentration)

[Ca²⁺]_cyto was measured as described previously [28]. Briefly, endothelial cells grown on glass coverslips or culture dishes of 3 cm diameter were loaded for 45 min at room temperature (22 °C) in the dark in loading buffer (2 mM CaCl₂, 135 mM NaCl, 1 mM MgCl₂, 5 mM KCl, 10 mM Hepes, 2.6 mM NaHCO₃, 0.44 mM KH₂PO₄, 10 mM D-glucose, 0.1% vitamins, 0.2% essential amino acids, 1% penicillin/streptomycin and 1% fungizone, pH 7.4) containing 2 μM fura 2/AM. Before experiments, cells were washed twice with loading buffer and were equilibrated for a further 15 min in experimental buffer (145 mM NaCl, 5 mM KCl, 2 mM CaCl₂, 1 mM MgCl₂ and 10 mM Hepes, pH 7.4) in the dark. Cells were illuminated alternately at excitation wavelengths of 340 and 380 nm (340HT15 and 380HT15; Omega Optical), and emission was monitored at 510 nm (510WB40; Omega Optical). [Ca²⁺]_cyto was expressed as (F₃₄₀/F₃₈₀)/F₀.

Measurement of [Ca²⁺]_ER

Intraluminal free Ca²⁺ concentration was monitored as described previously [3,29]. In brief, at a confluence of approx. 80%, endothelial cells grown on glass coverslips were transiently transfected with 2 μg of purified pcDNA3 encoding the improved versions of ER-_targeted D1_ER probe [30] using TransFast™ transfection reagent according to the manufacturer's protocol. At between 36 and 48 h after transfection, cells were used for experiment. D1_ER was excited at 440 nm (440AF21; Omega Optical) and emission was collected simultaneously at 535 and 480 nm using an optical beam splitter (Dual-View MicroImager™, Optical Insights, Visitron Systems). Changes of [Ca²⁺]_ER were expressed as (F₅₃₅/F₄₈₀)/F₀. Each data point represents the mean±S.E.M. for three to four cells over a measurement period of 3 min.

Statistics

Results are presented as means±S.E.M. Error bars are shown for every data point presented. If no bar is visible, the S.E.M. was smaller than the symbol size. ANOVA and Scheffe's post hoc F test were used for evaluation of the statistical significance. P<0.05 was considered to be significant.

Inhibition of SERCA (sarcoplasmic/endoplasmic reticulum Ca²⁺-ATPase) yields efficient Ca²⁺ depletion of the ER

As a consequence of an inhibition of SERCA with 15 μM BHQ, the ER was emptied of Ca²⁺ (Figure 1A) [3,15], despite the continuous presence of extracellular Ca²⁺. This effect was fully reversible upon removal of BHQ. Concomitantly, [Ca²⁺]_cyto increased and remained elevated as long as BHQ was present (Figure 1B).

SERCA inhibition attenuates activity and secretion, but not expression of endogenous LPL

To induce endogenous expression of LPL, CHO-K1 cells were propagated in serum-free medium for 24 h before experiments. Ca²⁺ depletion of the ER with 15 μM BHQ resulted in gradual decline of LPL activity in cell lysates (Figure 1C). Subsequent removal of BHQ yielded complete recovery of enzyme activity within 1 h (Figure 1C). LPL activity released into the supernatant of CHO-K1 cells rose constantly (Figure 1D). Application of 15 μM BHQ retained extracellular LPL activity at a low level, while, immediately after washout of BHQ, the secretion of active LPL increased with identical kinetics as in control (slope of 3.0±0.7 compared with 3.0±0.2 nmol·mg⁻¹·min⁻¹·h⁻¹ without BHQ pre-treatment) (Figure 1D). Using RT-PCR, the mRNA levels of LPL and GAPDH were found to be unaffected by 3 h of treatment with BHQ (Figure 1E).

These data suggest that endogenous transcription of LPL in wild-type CHO cells is not altered by depletion of [Ca²⁺]_ER, whereas its enzymatic activity and secretion are reversibly reduced.

Owing to the lack of a suitable antibody against LPL of hamster origin, possible alterations on the protein level could not be verified. Consequently, in order to correlate ER Ca²⁺ content further with protein transcription and expression, as well as maturation and secretion, human LPL was overexpressed in endothelial cells utilizing adenovirus [23].

Depletion of [Ca²⁺]_ER attenuates activity and secretion of adenovirus-encoded LPL

In endothelial cells, adenovirus-mediated expression of human LPL yielded a steady enzyme activity in cell lysates and continuous LPL secretion into the medium (Figure 2). In the presence of 15 μM BHQ, which achieved exhaustive ER Ca²⁺ depletion similar to that shown in CHO-K1 cells [3], LPL activity in cell lysates decreased and partially recovered upon removal of BHQ (Figure 2A).

In agreement with the results obtained on the constitutive LPL in CHO-K1 cells, activity of adenovirus-encoded human LPL increased gradually in the medium of untreated endothelial cells with a slope of 6.3±0.7 nmol·mg⁻¹·min⁻¹·h⁻¹ in the first 3 h. Addition of 15 μM BHQ arrested the enzyme activity at low level. In consequence of BHQ washout, in the first 3 h, LPL activity in the supernatant started to rise with a slope comparable with that initially seen in untreated cells (6.8±1.2 nmol ·mg⁻¹·min⁻¹·h⁻¹) (Figure 2B).

Adenovirus-mediated expression of LPL is unaffected by depletion of [Ca²⁺]_ER

As shown in Figure 3(A), mRNA levels of human LPL in relation to GAPDH remained unchanged in endothelial cells by up to 4 h of BHQ treatment and during the subsequent washout period. Moreover, the content of LPL protein within equal amounts of cell lysate was unaffected by BHQ (Figure 3B, top panel). However, the additional band of lower molecular mass that was detectable with the antibody increased slightly within the first 1 h after BHQ application and disappeared during the washout period. In the supernatant of adenovirus-infected endothelial cells, LPL protein content paralleled the raise in enzymatic activity found in the medium. Consecutively, BHQ abolished LPL secretion and LPL protein in the medium did not reach a detectable level up to 3 h after removal of the SERCA inhibitor (Figure 3B, middle panel). In comparison, the expression level of α-actin remained unaffected under each condition (Figure 3B, bottom panel).

These data indicate that strong depletion of the ER by BHQ diminished the activity and secretion of adenovirus-encoded human LPL in endothelial cells in a partially reversible manner, while mRNA levels and intracellular protein content remained unaffected by BHQ treatment.

Histamine causes moderate ER Ca²⁺ depletion despite strong elevation of cytoplasmic Ca²⁺

In order to analyse the impact of agonist-triggered ER Ca²⁺ depletion on LPL maturation and secretion, the effects of the IP₃-generating autacoid histamine on [Ca²⁺]_ER and the activity of adeno-virus-encoded human LPL were investigated in endothelial cells.

Stimulation of endothelial cells expressing LPL with 100 μM histamine in the presence of extracellular Ca²⁺ resulted in immediate and reversible reduction of [Ca²⁺]_ER (Figure 4A). This effect was comparable with that achieved with BHQ, but represented a transient phenomenon as ER Ca²⁺ content fully recovered within 1 h, even in the presence of the agonist. This histamine-initiated ER Ca²⁺ depletion was accompanied by a rapid initial cytosolic Ca²⁺ elevation and a subsequent elevated Ca²⁺ plateau until washout of the agonist (Figure 4B).

Agonist-induced ER Ca²⁺ reduction transiently attenuates activity and secretion of human LPL

Upon application of 100 μM histamine, intracellular LPL activity was reduced after 1 h and subsequently re-established slowly to reach complete recovery after 4 h, despite the presence of the agonist (Figure 4C). In the supernatant, histamine decreased LPL activity in the medium for 2 h, while further progression of LPL secretion was only slightly diminished (Figure 4D).

However, in the absence of extracellular Ca²⁺ (i.e. EGTA-containing solution), histamine (100 μM) yielded complete ER Ca²⁺ depletion (Figure 4A) that was accompanied by a transient cytosolic Ca²⁺ elevation (Figure 4B). Under these conditions, intracellular LPL activity upon histamine did not recover and decreased continuously over time (Figure 4C). Consistently, in the absence of extracellular Ca²⁺, secretion of LPL was prevented permanently by stimulation with histamine (Figure 4D). In agreement with our findings that point to a tight correlation between ER Ca²⁺ content and LPL folding, EGTA (i.e. without histamine stimulation), which has been shown previously to result in a slow ER depletion [32] and yielded approx. 50% ER depletion within 4 h (results not shown), gradually diminished intracellular activity and secretion of LPL over time (Figure 5).

Thus, in contrast with BHQ that yielded exhaustive and long-lasting ER Ca²⁺ depletion and, consequently, prevented post-translational processing of LPL, cell stimulation with an IP₃-generating agonist under physiological conditions (i.e. in the presence of extracellular Ca²⁺) only weakly and transiently (within 4 h) attenuated the maturation and secretion of LPL, most likely due to the moderate ER Ca²⁺ depletion under these conditions.

To investigate the contribution of mitochondrial Ca²⁺ flux to the maintenance of the Ca²⁺-dependent protein folding machinery of the ER during cell stimulation with the physiologically relevant agonist histamine, mitochondrial Ca²⁺ extrusion was prevented by CGP 37157, an inhibitor of the mitochondrial Na⁺/Ca²⁺ exchanger, that represents the main mechanism of Ca²⁺ extrusion from mitochondria in endothelial cells [3,33].

Inhibition of mitochondrial Ca²⁺ flux prevents LPL maturation and secretion

CGP 37157 has been demonstrated to inhibit mitochondrial Ca²⁺ flux and, thus, ER Ca²⁺ refilling upon stimulation in endothelial cells very efficiently [3]. As shown in Figure 6(A), in the presence of 20 μM CGP 37157, histamine-induced ER Ca²⁺ depletion was more pronounced than in the control, while [Ca²⁺]_cyto remained elevated (Figure 6B).

In the presence of CGP 37157, cytosolic LPL activity was largely decreased by stimulation with histamine and did not recover with time (Figure 6C). Accordingly, LPL secretion was prevented upon histamine stimulation in the presence of CGP 37157 (Figure 6D). In contrast, application of CGP 37157 without any further treatment did not affect ER Ca²⁺ content within 4 h (results not shown) and also had no effect on intracellular or secreted LPL activity (Figure 5).

These data suggest, that besides extracellular Ca²⁺, mitochondrial Ca²⁺ transfer essentially contributes to Ca²⁺-dependent maturation of LPL in the ER and its subsequent secretion during cell stimulation with an IP₃-generating agonist.