Introduction

Bok (Bcl-2-related ovarian killer) is a member of the Bcl-2 protein family that governs the intrinsic apoptosis pathway [1,2,3]. These proteins contain conserved Bcl-2 homology (BH) domains (BH1–4) and the family is made up of anti-apoptotic proteins (e.g., Bcl-2 and Mcl-1), pro-apoptotic proteins (e.g., Bak and Bax), and BH3-only proteins (e.g., Bid and Bad). A complex network of interactions between these family members controls apoptosis and cell viability [1,2,3]. Bok has been generally grouped with Bak and Bax on the basis of high sequence homology and initial in vitro studies showing that overexpressed Bok induced apoptosis in cultured cells [4,5,6]. Endogenous Bok is predominately localized to the endoplasmic reticulum (ER) [7], whereas many Bcl-2 family members predominately localize to the mitochondrial membrane [1, 8]. Although Bok is ubiquitously expressed, Bok-knockout (KO) mice are phenotypically normal [4].

Currently, the role that Bok plays in apoptosis is unclear, with various studies showing both pro- and anti-apoptotic characteristics. On the one hand, overexpressed exogenous Bok can promote apoptosis [7, 9,10,11], purified Bok can permeabilize the outer membrane of isolated mitochondria and artificial liposomes [9, 11, 12], and Bok-KO mouse embryonic fibroblasts (MEFs) exhibit reduced susceptibility to ER stress-induced apoptosis [13, 14]. On the other hand, Bok-KO MEFs have also been shown to exhibit increased susceptibility to ER stress-induced apoptosis [7, 15], and Bok is neuroprotective during Ca2+-induced neuronal injury, by suppressing poly ADP-ribose polymerase (PARP)-dependent cell death [16]. Additionally, evidence has been presented that Bok plays no role in apoptosis, since Bok-KO neurons, hematopoietic cells, and non-neuronal cells have all been shown to respond normally to several apoptotic stimuli [4, 7, 16].

Uniquely for the Bcl-2 protein family, Bok binds very strongly to inositol 1,4,5-trisphosphate (IP3) receptors (IP3Rs) and, indeed, in unperturbed cells all Bok appears to be IP3R bound [17, 18]. IP3Rs are ~300 kDa proteins that form tetrameric, IP3- and Ca2+-gated Ca2+ channels in the ER membranes of mammalian cells and play an essential role in intracellular Ca2+ signaling [19, 20]. There are three highly homologous IP3R forms (IP3R1–3) that can form both homo- and heterotetramers [19, 20]. Binding of IP3 and Ca2+ to IP3Rs induces a yet-to-be-defined conformational change in the tetrameric channel that allows Ca2+ to flow from stores within the ER lumen into the cytosol to increase cytosolic Ca2+ concentration [19, 20]. In addition, the Ca2+ released by IP3Rs can be taken up by adjacent mitochondria where it plays a role in regulating ATP production [21, 22]. IP3Rs also play a role in intrinsic apoptosis, and various effects of Bcl-2 family members on IP3R activity and ER Ca2+ content have been reported [22,23,24]. Bok binds strongly to only IP3R1 and IP3R2, with the binding mediated by the BH4 domain of Bok and a small region within the IP3R coupling domain at approximately amino acid ~1900 [17, 18]. The interaction of Bok with IP3Rs protects them from proteolysis either by chymotrypsin in vitro or by caspase-3 in vivo, apparently by steric hindrance of protease access [17, 18]. In addition, Bok is dramatically stabilized by binding to IP3Rs and unbound Bok is ubiquitinated and degraded by the proteasome [17].

While Bcl-2 family members are well known for their ability to influence apoptosis, they also regulate mitochondrial dynamics (e.g., fusion and fission), morphology, and function [25, 26]. Mitochondrial fusion and fission allow for growth, the exchange of mitochondrial content, the elimination of damaged mitochondria, and the balance between fusion and fission controls overall mitochondrial morphology [25, 27]. Bcl-2 family members, such as Bax, Bak, Bcl-xl, and Mcl-1, have been shown to regulate mitochondrial dynamics through interactions with either mitofusin 2 (Mfn2) (a dynamin-like GTPase that regulates fusion of the outer mitochondrial membrane) or dynamin-related protein 1 (Drp-1) (a dynamin-like GTPase that regulates mitochondrial fission) [28,29,30,31,32,33,34]. Thus, the network of interactions between Bcl-2 family members also likely govern mitochondrial morphology [25, 26, 30].

Here we report that despite the strong interaction between Bok and IP3Rs, Bok deletion does not alter IP3R-mediated Ca2+ mobilization or Ca2+ influx into the mitochondria. Rather, Bok deletion results in mitochondrial fragmentation, due to a decrease in mitochondrial fusion rate, and an enhancement of mitochondrial spare respiratory capacity and membrane potential. Lastly, Bok deletion does not alter responsiveness to several apoptotic stimuli. These findings provides new insight into the cellular role of Bok.

Materials and methods

Materials

MEFs, αT3, and HEK 293T cells were cultured as described [35, 36]. DT40 cells were maintained at 39 °C and 5% CO2 in RPMI-1640 supplemented with 1% chicken serum, 10% fetal bovine serum, 10 μM β-mercaptoethanol, 100 U/mL penicillin, and 100 µg/mL streptomycin. Antibodies raised in rabbits were: anti-IP3R1 and anti-IP3R2 [35], anti-erlin2 [37], anti-IP3R1–3 [38], anti-Mcl-1 #D35A5, anti-Bcl-xL #54H6, anti-Bcl-2 #50E3, anti-caspase-3 #9662, anti-PARP #9542, anti-Mfn2 #9482, anti-VDAC1 #4661, and anti-pDrp1-616 #3455 (Cell Signaling Technology), anti-Bak #06-536 (Millipore), anti-Bax #N-20 and anti-Tom20 #FL145 (Santa Cruz Biotechnology Inc.), anti-Bok, raised against amino acids 19–32 of mouse Bok [4, 7], and anti-Miro1 #HPA010687 (Atlas Antibodies). Mouse monoclonal antibodies were: anti-FLAG epitope clone M2 (Sigma), anti-IP3R3, raised against amino acids 22–230 of human IP3R3 (BD Transduction Labs), anti-Mfn1 #D-10 (Santa Cruz Biotechnology Inc.), anti-Drp-1 #14647 (Cell Signaling Technology), anti-electron transport chain complexes #110413 (Abcam), and anti-p97 (Research Diagnostics Inc.). Horseradish peroxidase-conjugated secondary antibodies, gonadotropin-releasing hormone (GnRH), oleoyl-l-α-lysophosphatidic acid (LPA), protease inhibitors, Triton X-100, 3-[(3-cholamidopropyl)-dimethylammonio]-1-propanesulfonate (CHAPS), brefeldin A (BFA), tunicamycin (TN), and thapsigargin (TG) were purchased from Sigma. Dithiothreitol, Precision PlusTM Protein Standards, and sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) reagents were from Bio-Rad. Protein A-Sepharose CL-4B was from GE Healthcare. MG132 was from Biomol. Staurosporine (SST) was from Enzo Life Sciences. Linear, MW ~25,000 polyethylenimine was from Polysciences Inc. MitoTrackerTM Red and Deep Red, tetramethylrhodamine ethyl ester perchlorate (TMRE), and Lipofectamine LTX were from Thermo Fisher.

Cell lysis, immunoprecipitation, SDS-PAGE, crude mitochondrial fractionation, and mass spectrometry

To prepare cell lysates for SDS-PAGE or for immunoprecipitation (IP), cells were harvested with ice-cold lysis buffer (150 mM NaCl, 50 mM Tris-HCl, 1 mM EDTA, 1% CHAPS, or 1% Triton X-100, 10 µM pepstatin, 0.2 mM phenylmethylsulfonyl fluoride, 1 mM dithiothreitol, and 0.2 µM soybean trypsin inhibitor, pH 8.0). CHAPS was used for all experiments involving IP. Lysates were incubated on ice for 30 min and clarified by centrifugation at 16,000 × g for 10 min at 4 °C. To IP-specific proteins, clarified lysates were incubated with antisera and protein A-Sepharose CL-4B for ~16 h at 4 °C, and IPs were washed thoroughly with lysis buffer. To obtain crude mitochondrial fractions, cells were harvested with ice-cold mitochondrial isolation buffer (MIB) (20 mM HEPES, 220 mM mannitol, 70 mM sucrose, 1 mM EGTA, 2 mg/mL bovine serum albumin (BSA), and 0.5 mM phenylmethylsulfonyl fluoride, pH 7.6), incubated on ice for 5 min, and homogenized with 40 strokes of a drill-fitted pestle. Homogenates were centrifuged at 800 × g for 5 min at 4 °C, supernatants were centrifuged at 10,000 × g for 20 min at 4 °C, pellets were resuspended in MIB without BSA, were centrifuged at 10,000 × g for 10 min at 4 °C, and finally pellets were resuspended in MIB without BSA. All samples were resuspended in gel-loading buffer, incubated at 37 °C for 30 min, subjected to SDS-PAGE, and proteins were transferred to nitrocellulose for probing as described [17]. Immunoreactivity was detected using Pierce ECL reagents and a Genegnome imager (Syngene Bio Imaging) or a Chemidoc (Bio-Rad). For mass spectrometry, Coomassie blue-stained IP3R1 and Bok bands were cut from SDS-PAGE gels, were destained, combined, trypsinized, and analyzed by liquid chromatography with tandem mass spectrometry, using a LTQ Orbitrap XLTM Hybrid Ion Trap-Orbitrap mass spectrometer (Thermo Fisher).

Generation and analysis of Bok-KO MEF, αT3, and DT40 cell lines

The CRISPR/Cas9 system [39, 40] was used to target exons within the Bok gene. Oligonucleotides that contained the target sequences (MEFs and αT3, exon 2: ACGCGCGGCTTTTGCGCGC, or exon 4: CCCCGCGGCCACGGAATAC; DT40, exon 2: GAGCTTGTGTCCCAAGCCA, or exon 4: GCCATGGTCCACACCATCG) were annealed and ligated into AflII-linearized guide RNA (gRNA) vector (AddGene). MEFs, αT3, and DT40 cells were NEON transfected with Bok-gRNA constructs and vectors encoding hCas9 (Addgene) and enhanced green fluorescent protein (EGFP) (Clontech). EGFP-expressing cells were selected 48 h later by fluorescence-activated cell sorting and plating at 1 cell/well in a 96-well plate. Colonies were expanded and screened in immunoblots for Bok immunoreactivity, which yielded Bok-KO cell lines. Multiple independent cell lines from each exon target were used for all experiments.

Generation and analysis of reconstituted MEF cell lines

Stable expression of exogenous Bok was obtained by NEON transfection of Bok-KO MEF cell lines with 3F-WTBok or 3F-L34GBok complementary DNAs [17], followed by selection in 0.6 mg/mL G418 for 72 h, and plating at 1 cell/well in a 96-well plate. Colonies were expanded and screened in immunoblots for FLAG immunoreactivity, which yielded Bok-KO+WTBOK and Bok-KO+L34GBOK cell lines that were maintained in 0.3 mg/mL G418.

Bok depletion

Bok depletion was achieved using a Mission short hairpin RNA (shRNA) pLKO.1 plasmid (Sigma-Aldrich) expressing shRNA targeting mouse Bok (5′-CCGGAGCTTCTTGCTTAGGCTATTTCTCGAGAAATAGCCTAAGCAAGAAGCTTTTTTG-3′). This plasmid and a non-targeting shRNA control (SHC002) were transiently co-transfected with pRSV-REV (Addgene plasmid #12253), pMDL/pRRE (Addgene plasmid #12251), and pCMV-VSVG (Addgene plasmid #8454) into HEK 293T cells using polyethylenimine as described [17]. Media were changed 24 h later, virus-containing supernatants were collected 48 and 72 h later, combined, filtered through a 0.45 µm filter (Corning), and stored at −80 °C until used to transduce MEFs in the presence of 9 µg/mL polybrene, followed by selection with 2 µg/mL puromycin for 72 h.

Cytoplasmic Ca2+ measurements

Cytosolic-free Ca2+ concentration ([Ca2+]c) in cell suspensions was measured as described [18], except that cells were loaded with 1 µM Fura 2-AM for 1 h at 37 °C and were washed three times with a Krebs/HEPES buffer. [Ca2+]c was calculated using the Grynkiewicz equation [41]. For single-cell [Ca2+]c measurements, cells were plated on glass coverslips, incubated for 15 min at 37 °C with 1 μM Fura 2-AM, 0.01% (vol/vol) pluronic acid in a standard extracellular solution (135 mM NaCl, 6 mM KCl, 1 mM CaCl2, 1 mM MgCl2, 10 mM HEPES-NaOH, 5.5 mM d-glucose, and 0.2% BSA, pH 7.4), then placed into a temperature-controlled microperifusion chamber mounted on an inverted fluorescence spinning disk confocal microscope (Olympus IX81), and continuously superfused with standard extracellular solution (2 mL/min) at 37 °C. Cells were visualized with a ×40 oil-immersion objective. Fura 2 excitation wavelengths were 340 and 380 nm with emission measured at 510 nm. Data are expressed as the fold-change in 340/380 nm ratio normalized to resting, unstimulated values.

Mitochondrial Ca2+ measurements

To measure mitochondrial-free Ca2+ concentration ([Ca2+]m), MEFs were transfected with a vector encoding 2mtD3cpv, a fluorescence resonance energy transfer (FRET)-based indicator localized to the mitochondrial matrix [42], using the NEON Transfection System (Invitrogen; 10 µg total DNA per 100 µL cell suspension at 3 × 107 cells/mL, 1 pulse, 20 ms, 1500 V), and 48 h later cells were plated on glass coverslips. Single-cell [Ca2+]m was measured as described for single-cell [Ca2+]c measurements, but 2mtD3cpv was excited at 440 nm and emission was measured at 485 nm (FRET donor) and 535 nm (FRET acceptor). Data are expressed as the fold-change in 535/485 nm ratio normalized to resting, unstimulated values.

Single IP3R1 channel measurements

Preparation of DT40 cell nuclei, on-nuclear patch clamp, and data analysis was performed as described [43]. Data shown are from exon 4-targeted DT40-3KO + rIP3R1 cells, but essentially identical data were obtained from exon 2-targeted DT40-3KO + rIP3R1 cells.

Live cell fluorescence microscopy of mitochondrial morphology

MEFs were NEON transfected with a plasmid encoding DsRed2-Mito (Clontech), and 48 h later, cells were set onto 1.5 glass cover dishes (Cellvis). Cells were imaged with a Nikon structured illumination microscope (N-SIM, Ti microscope) equipped with a ×100 SR Apo TIRF oil-immersion objective, an electron multiplying charged-coupled device camera (iXon Em-CCD, Andor), using 561 nm excitation/605 nm emission. Z-stacks of entire cells were taken, followed by reconstruction of the three-dimensional z-stacks and deconvolution using NIS-Elements software (Nikon). For additional confocal analysis, MEFs were either NEON transfected with a plasmid encoding DsRed2-Mito or stained with MitoTracker Red as per the manufacturer’s instructions, and cells were imaged with a Nikon C2 microscope (Ti microscope) equipped with a ×60 Plan Apo VC oil-immersion objective lens, using 543.5 nm excitation/585 nm emission. To measure mitochondrial particle velocity, DsRed2-Mito-expressing MEFs were continuously imaged for 5 min and velocity of mitochondrial particles that could be tracked for at least 30 s was measured using the Tracking module in NIS-Elements software. All microscopy was done at 37 °C with 5% CO2. All images were subjected to one iteration of blind deconvolution and mitochondrial particle quantification was done using NIS-Elements software.

Mitochondrial fusion assay

MEFs were transiently transfected using Lipofectamine LTX as per the manufacturer’s instructions to express photoactivatable GFP targeted to the mitochondrial matrix (mito-PAGFP) [30]. After 48 h, MEFs were stained with MitoTracker Deep Red as per the manufacturer’s instructions, and cells were imaged with a Zeiss 880 Airy microscope equipped with a Plan-Apochromat ×63/1.40 oil objective. A small region of interest was photoactivated with a 405 nm laser and z-stacks of the entire cells were acquired at 0 and 30 min post-photoactivation. GFP was imaged using 488 nm excitation/516 nm emission and MitoTracker Deep Red was imaged using 633 nm excitation/654 nm emission. A custom MATLAB script was used to determine the fraction of mitochondria positive for GFP. Briefly, images were background subtracted, PAGFP-positive cells were manually selected, and mitochondria were defined on the basis of size and an automated intensity threshold for each z-stack. Mitochondria were deemed GFP-positive if GFP intensity was above a GFP intensity threshold for each cell, defined as the mean plus two times the SD of pixel intensities before photoactivation. The fraction of mitochondria positive for GFP was then calculated for 0 and 30 min after photoactivation to determine the fraction of mitochondrial particles above the GFP intensity threshold.

Measurement of oxygen consumption rate

MEFs were seeded in XFe96 plates (Agilent) at 40,000 cells/well, and 24 h later, oxygen consumption rate (OCR) was measured using the Seahorse XFe96 Extracellular Flux Analyzer and the XF Cell Mito Stress Test Kit (Agilent), as per the manufacturer’s instructions. Cells were washed three times with unbuffered XF media supplemented with 10 mM d-glucose, 2 mM l-glutamine, and 1 mM pyruvate, pH 7.4, were incubated at 37 °C in a CO2-free environment for 1 h, and triplicate measurements of baseline followed by triplicate measurements after the addition of each drug (1 μM oligomycin, 1 μM carbonyl cyanide-p-trifluoromethoxy-phenyl-hydrazone (FCCP), and 0.5 μM rotenone/antimycin A, respectively) were collected. Afterwards, a cell normalization assay was performed (cyQUANT, Thermo Fisher), as per the manufacturer’s instructions and measured with a Synergy H1 microplate reader and Gen 5 software (BioTek) using 485 nm excitation/528 nm emission. OCR measurements were normalized to the cell densities.

Measurement of mitochondrial membrane potential

MEFs were seeded at 1.0 × 106 cells/9.6 cm2 well, incubated with 500 nM TMRE for 20 min at 37 °C and 5% CO2, rinsed and harvested with 10 mM HEPES, 154 mM NaCl, 1 mM EDTA, pH 7.4, centrifuged at 3000 × g for 1 min, resuspended with phosphate-buffered saline and 2% fetal bovine serum, and analyzed by flow cytometry using a LSR II Cell Analyzer (Becton Dickinson). For all samples, 10,000 cell events were recorded and data were analyzed using the Flow Jo software to determine the geometric mean for TMRE fluorescence.

Data analysis

All experiments were repeated two or more times (N = the number of independent experiments) and representative images of gels and traces are shown. Quantitated data are expressed as mean ± SEM. Statistical analysis was performed using a Student’s t test and p values of <0.05 were considered statistically significant.

Results

Bok deletion does not alter IP3R-dependent Ca2+ signaling

We have shown previously that Bok binds strongly to IP3Rs and that essentially all cellular Bok is IP3R bound [17, 18]. To assess stoichiometry, IP3R1-Bok complex was immunopurified from αT3 cells, as described [18], and was subjected to mass spectral analysis. The number of unique Bok and IP3R1 peptide spectrum matches obtained was 36 and 329, respectively (an index of relative abundance), and divided by the Bok and IP3R1 molecular weights (23.4 and 313 kDa, respectively) yielded a Bok to IP3R1 ratio of 1.5:1. Thus, it is probable that most, if not all of the four, IP3R1 polypeptides in a tetrameric channel are bound to Bok. To determine if this IP3R-bound Bok regulates IP3R activity, we deleted Bok in MEFs via CRISPR/Cas9-mediated gene editing (Fig. 1a) and measured IP3R-mediated Ca2+ mobilization (Fig. 1b, c). Bok deletion was specific, as the levels of other pertinent proteins (IP3Rs and various Bcl-2 family members) were unchanged (Fig. 1a). Examination of [Ca2+]c showed that LPA, which stimulates IP3 formation in fibroblasts [44], induced essentially identical responses in wild-type (WT) and Bok-KO MEFs, in both cell suspension and single-cell analyses (Fig. 1b and Fig. 1c, respectively). Thus, Bok does not appear to regulate IP3R-mediated Ca2+ mobilization in intact cells.

Fig. 1
figure 1

Bok (Bcl-2-related ovarian killer) deletion does not alter calcium mobilization via inositol 1,4,5-trisphosphate receptors (IP3Rs) or calcium influx into mitochondria. a Levels of Bok and other pertinent proteins in lysates from control and Bok-knockout (KO) mouse embryonic fibroblasts (MEFs) obtained by targeting exon 2 (lane 2) or exon 4 (lane 4), with p97 serving as a loading control. b MEFs in suspension were loaded with Fura 2-AM, exposed to 1 µM oleoyl-l-α-lysophosphatidic acid (LPA) (box), and cytosolic-free Ca2+ concentration ([Ca2+]c) was calculated from the 340/380 nm fluorescence ratio. EGTA (3 mM) was added 1 min prior to LPA to reduce extracellular [Ca2+] to ~100 nM [51] and eliminate Ca2+ entry. Traces shown are mean ± SEM; N = 6. c MEFs on coverslips were loaded with Fura 2-AM, exposed to 1 µM LPA (box) and the 340/380 nm fluorescence ratio was measured in single cells. Data are expressed as the fold-change in ratio normalized to baseline values. Traces shown are mean ± SEM; n = 10 control cells, n = 7 Bok-KO cells. d MEFs transfected with 2mtD3cpv on coverslips were exposed to 1 µM LPA (box) and the 535/485 nm fluorescence ratio was measured in single cells. Data are expressed as the fold-change in ratio normalized to baseline values. Traces shown are mean ± SEM; n = 104 control cells, n = 134 Bok-KO cells. e Levels of Bok and other pertinent proteins in lysates from αT3 control and Bok-KO cells obtained by targeting exon 2 (lane 2) or exon 4 (lane 4), with p97 serving as a loading control. f αT3 cells in suspension were loaded with Fura 2-AM, exposed to 30 nM gonadotropin-releasing hormone (GnRH) (box), and [Ca2+]c was calculated from the 340/380 nm fluorescence ratio. Traces shown are mean ± SEM; N = 5. g Lysates and anti-IP3R1 immunoprecipitations (IPs) from various DT40 cell lines. h Levels of Bok and other pertinent proteins in lysates from DT40-3KO+rIP3R1 control and Bok-KO cells obtained by targeting exon 4, with p97 serving as a loading control. i Single-channel recordings of rIP3R1 activity upon stimulation by 1 µM IP3, 5 mM ATP, and 200 nM Ca2+ in nuclear patches from DT40-3KO + rIP3R1 control and exon 4 Bok-KO cells; N = 3. j Histogram of the open probability (P0) for rIP3R1 in control and exon 4 Bok-KO cells; mean ± SEM, N = 3

Full size image

Due to the close proximity and interaction sites between the ER and mitochondria, IP3R-mediated Ca2+ mobilization from the ER results in Ca2+ transfer to the mitochondrial matrix [45,46,47]. To determine if Bok deletion altered mitochondrial Ca2+ levels, [Ca2+]m was measured using 2mtD3cpv, a FRET-based indicator [42]. However, LPA caused an identical [Ca2+]m increase in control and Bok-KO MEFs (Fig. 1d), indicating that Bok does not regulate Ca2+ transfer to the mitochondrial matrix.

We have shown previously that Bok binds preferentially to IP3R1 and IP3R2, but not IP3R3 [18], and since MEFs express all IP3R types (Fig. 1a), it is possible that unregulated IP3R3 could be masking the effect of Bok on IP3R activity. Thus, to directly assess IP3R1 activity we examined αT3 cells, the cell type in which we originally discovered that endogenous Bok and IP3R1 interact strongly [18], and in which IP3R1 constitutes ~99% of total IP3R content [37]. Bok was specifically deleted in αT3 cells via CRISPR/Cas9-mediated gene editing (Fig. 1e) and [Ca2+]c was measured (Fig. 1f). Examination of [Ca2+]c showed that GnRH, which stimulates IP3 formation in αT3 cells [48], induced essentially identical responses in WT and Bok-KO αT3 cells (Fig. 1f). Thus, Bok does not appear to regulate IP3R1-mediated Ca2+ mobilization.

Since feedback regulation of the various proteins that regulate [Ca2+]c [20] could potentially account for global increases in [Ca2+]c being unaltered in the Bok-KO cell lines, we also directly measured IP3R1 channel activity using the on-nuclear patch-clamp technique applied to IP3R1-3KO DT40 chicken B cells reconstituted with rat IP3R1 (DT40-3KO + rIP3R1 cells) [43] (Fig. 1g, lane 3). Bok binds strongly to both endogenous and exogenous IP3R1 (Fig. 1g, lanes 1 and 3, respectively), but the absence of IP3Rs in DT40-3KO cells caused a dramatic decline in Bok levels (Fig. 1g, lane 2), similar to that seen previously in IP3R1-KO αT3 cells [17]. Bok was specifically deleted from DT40-3KO + rIP3R1 cells via CRISPR/Cas9-mediated gene editing (Fig. 1h), and single-channel recordings of rIP3R1 produced similar traces in control and Bok-KO cells (Fig. 1i). Further, quantification of the IP3R1 channel activity showed that Bok deletion did not alter channel open probability (P0) (Fig. 1j), burst lengths, or interburst intervals (data not shown). Overall, the data in Fig. 1 show that despite Bok having a strong interaction with IP3Rs, it does not appear to regulate their channel activity.

Bok deletion and depletion results in mitochondrial fragmentation

While measuring [Ca2+]m with 2mtD3cpv (Fig. 1d), we noticed that the mitochondria in Bok-KO MEFs were morphologically different from control MEFs (data not shown). To better define this difference, mitochondria were labeled with DsRed2-Mito, a fluorescent protein that labels mitochondria specifically, via transient transfection and high-resolution images were taken using SIM. While mitochondria in control MEFs displayed the typical elongated morphology [29], mitochondria in Bok-KO MEFs were more fragmented (Fig. 2a). Quantification of images showed that Bok KO resulted in significantly more mitochondrial particles per cell, a decrease in particle volume, and an increase in particle sphericity (Fig. 2b–d, respectively). To determine if acute depletion of Bok has a similar effect to long-term CRISPR/Cas9-mediated Bok KO, shRNA was used to acutely “knockdown” (KD) Bok expression in MEFs (Fig. 2e). Confocal images of mitochondria stained with MitoTracker showed that control MEFs displayed the typical elongated mitochondrial morphology, but mitochondria in Bok-KD MEFs were more fragmented, similar to Bok-KO cells (Fig. 2f). Quantification of images showed that Bok KD resulted in significantly more mitochondrial particles per cell, a decrease in particle area, and an increase in particle roundness (Fig. 2g–i, respectively). Thus, both long-term deletion and acute depletion of Bok causes mitochondrial fragmentation.

Fig. 2
figure 2

Bok (Bcl-2-related ovarian killer) deletion and acute depletion causes mitochondrial fragmentation. a Representative confocal images of two independently derived control and Bok-KO mouse embryonic fibroblast (MEF) cell lines (either exon 2 or exon 4 targeted), transiently expressing DsRed2-Mito, and captured using a structured illumination microscope (SIM) (scale bar = 5 µm). The white boxed regions are further magnified in the expanded images (right side). b–d Quantified measurements of mitochondrial particle number/cell, particle volume, and particle sphericity in control and Bok-KO MEFs. Histograms are mean ± SEM, with p values comparing control to Bok-KO cells via unpaired Student’s t test, and n designating cell number or mitochondrial particle number. e Levels of Bok in lysates from control and Bok-KD MEFs, with p97 serving as a loading control. f Representative confocal images of control and Bok-KD MEFs loaded with MitoTracker Red (scale bar = 10 µm). The white boxed regions are further magnified in the expanded images (right side). g–i Quantified measurements of mitochondrial particle number/cell, particle area, and particle roundness in control and Bok-KD MEFs. Histograms are mean ± SEM, with p values comparing control to Bok-KD cells via unpaired Student’s t test, and n designating cell number or mitochondrial particle number

Full size image

Exogenous Bok normalizes mitochondrial morphology

To confirm that the mitochondrial fragmentation seen upon Bok deletion or depletion (Fig. 2) is specifically due to loss of Bok and not possibly caused by off-target effects of genome editing, exogenous Bok (Bok tagged at the N terminus with the triple FLAG epitope; 3F-WTBok) was stably expressed in Bok-KO MEFs at levels similar to endogenous Bok (Fig. 3a). Note that exogenous 3F-WTBok, like endogenous Bok, expresses as multiple species due to alternative translation initiation [17]. Similar to endogenous Bok, exogenous 3F-WTBok interacts strongly with IP3R1, as indicated by co-IP (Fig. 3b, lanes 1 and 3, respectively). To assess mitochondrial morphology, mitochondria were labeled with DsRed2-Mito via transient transfection. Bok-KO+WTBok MEFs displayed the typical elongated mitochondrial morphology seen in control MEFs, rather than the fragmented morphology seen in Bok-KO MEFs (Fig. 3c), and quantification of images showed that reconstitution of Bok expression with 3F-WTBok restored mitochondrial morphology parameters to the levels observed in control cells (Fig. 3d–f). Thus, the mitochondrial fragmentation seen upon Bok deletion or depletion is specifically due to the loss of Bok.

Fig. 3
figure 3

Exogenous Bok (Bcl-2-related ovarian killer) reverses mitochondrial fragmentation in Bok-KO mouse embryonic fibroblasts (MEFs). a Levels of Bok and other pertinent proteins in lysates from control and Bok-KO MEFs and Bok-KO cell lines stably expressing 3F-WTBok (Bok KO+WTBok, lane 3) or 3F-L34GBok (Bok KO+L34GBok, lane 4), with p97 serving as a loading control. b Anti-IP3R1 immunoprecipitations (IPs) and input cell lysates were probed in immunoblots for the proteins indicated. c Representative confocal images of control, Bok-KO, Bok-KO+WTBok, Bok-KO+L34GBok MEF cell lines transiently expressing DsRed2-Mito (scale bar = 10 µm). Further magnified regions are shown in expanded white boxes (right side). d–f Quantified measurements of mitochondrial particle number/cell, particle area, and particle roundness in control, Bok-KO, Bok-KO+WTBok, and Bok-KO+L34GBok MEFs. Histograms are mean ± SEM, with p values comparing control to Bok-KO, Bok-KO+WTBok, and Bok-KO+L34GBok via unpaired Student’s t test, and n designating the cell number or mitochondrial particle number. P values >0.05 were considered not significant (NS)

Full size image

To determine whether the maintenance of normal mitochondrial morphology by Bok requires the Bok-IP3R1 interaction, 3F-L34GBok, an IP3R1 binding-deficient Bok mutant [17], was also stably expressed in Bok-KO MEFs (Fig. 3a, lane 4). Unlike 3F-WTBok, 3F-L34GBok does not co-IP with IP3R1 (Fig. 3b, lane 4). Interestingly, 3F-L34GBok also normalized mitochondrial morphology parameters to the levels observed in control cells (Fig. 3c–f), and actually enhanced mitochondrial elongation, as indicated by a decrease in mitochondrial particle roundness (Fig. 3f). Thus, maintenance of normal mitochondrial morphology does not require that Bok be IP3R bound, but rather merely requires that Bok be present in the ER membrane.

Bok deletion reduces mitochondrial fusion rate and motility

To begin to examine why Bok deletion or depletion causes mitochondrial fragmentation, mitochondrial fusion rate was measured by assessing mito-PAGFP mobility [30], since other Bcl-2 protein family members (i.e., Bak and Bax) have been reported to regulate this process [29]. After photoactivation, GFP signal spread rapidly throughout the mitochondrial network in control MEFs (Fig. 4a), indicative of efficient mitochondrial fusion [30]. In contrast, little to no spread was seen in Bok-KO MEFs (Fig. 4a), indicative of deficient mitochondrial fusion. Quantification showed that the increase in the fraction of GFP-positive mitochondrial particles from 0 min to 30 min was dramatically lowered in the Bok-KO MEFs (Fig. 3b). Thus, Bok KO results in a decrease in mitochondrial fusion rate.

Fig. 4
figure 4

Bok (Bcl-2-related ovarian killer) deletion decreases mitochondrial fusion rate and velocity without altering the levels of fusion/fission and Miro proteins. a Representative confocal images of control and Bok-knockout (KO) mouse embryonic fibroblasts (MEFs) transiently expressing mito-PAGFP and stained with MitoTracker Deep Red, followed by photoactivation of one region of interest per cell. Cells were imaged immediately after photoactivation and 30 min later (scale bar = 10 µm). b Quantification of the increased fraction of green fluorescent protein (GFP)-positive mitochondria from 0 to 30 min. Histograms are mean ± SEM, with p values comparing control to Bok-KO cells via unpaired Student’s t test, and n designating cell number. c Representative confocal image of Bok-KO MEFs transiently expressing DsRed2-Mito, pre-treatment and post-treatment with 50 µM Mdivi-1 for 1 h (scale bar = 10 µm). d–f Quantified measurements of mitochondrial particle number/cell, particle area, and particle roundness in Bok-KO MEFs pre- and post-treatment. Histograms are mean ± SEM, with p values comparing Bok-KO MEFs pre-treatment to post-treatment via unpaired Student’s t test, and n designating cell number or mitochondrial particle number. g Quantified measurements of mitochondrial particle velocity in control and Bok-KO MEFs transiently expressing DsRed2-Mito. Histograms are mean ± SEM, with p values comparing control to Bok-KO MEFs via unpaired Student’s t test, and n designating the number of mitochondrial particles measured in 12 cells. h Levels of fission and fusion mediators and other pertinent proteins in lysates from control and Bok-KO MEFs, with erlin2 serving as a loading control. i Levels of fission and fusion mediators in crude mitochondrial fractions. j Levels of Miro1 and other pertinent proteins in lysates from control and Bok-KO MEFs, with erlin2 serving as a loading control

Full size image

In order to determine if Bok deletion completely blocks or just inhibits mitochondrial fusion, mitochondria were labeled with DsRed2-Mito via transient transfection and treated with Mdivi-1 (a mitochondrial fission inhibitor) [49]. Bok-KO MEFs treated with Mdivi-1 converted the fragmented mitochondria to a more elongated phenotype, typical of that seen in control cells (Fig. 4c–f) and indicating that some fusion must still be occurring. To determine if the decrease in mitochondrial fusion rate could be due to a change in mitochondrial motility, mitochondrial velocity was measured in control and Bok-KO MEFs. Interestingly, Bok deletion significantly reduced mitochondrial velocity (Fig. 4g). Thus, fusion is not blocked, but is significantly reduced in Bok-KO MEFs, perhaps because of a decrease in mitochondrial motility.

The decrease in mitochondrial fusion rate and mitochondrial fragmentation seen in Bok-KO MEFs could be due to changes in the levels of proteins that regulate mitochondrial fusion or fission. However, this was not the case as the levels of Mfn1 and Mfn2 (fusion mediators) and Drp-1 (fission mediator) in cell lysates were unaltered by Bok KO (Fig. 4h). Phosphorylation of Drp-1 at amino acid 616 induces mitochondrial fission activity [27], but pDrp-616 levels in cell lysates were also unaltered by Bok KO (Fig. 4h). Finally, Drp-1 translocation from the cytosol to mitochondria regulates fission [27], but Drp-1 levels in crude mitochondrial fractions were unaltered by Bok KO (Fig. 4i). Likewise, the level of Miro1, the mitochondrial GTPase that sustains mitochondrial motility [25, 27], was unaltered by Bok KO (Fig. 4j). Thus, the decreases in mitochondrial fusion rate and velocity seen in Bok-KO MEFs are not due to changes in the levels of proteins that regulate mitochondrial dynamics.

Bok deletion does not inhibit mitochondrial bioenergetics or cell growth

Since Bok KO results in mitochondrial fragmentation, we sought to determine if this affected mitochondrial bioenergetics. Mitochondrial respiration was analyzed in control and Bok-KO MEFs (Fig. 5a), before and after the addition of various electron transport chain (ETC) complex inhibitors; oligomycin (complex V inhibitor) to determine the amount of ATP production linked to respiration, FCCP (ionophore that collapses the inner membrane gradient) to cause the ETC to function at its maximal rate, and a combination of rotenone and antimycin A (complex I and III inhibitor, respectively) to shut down the ETC and determine non-mitochondrial respiration. Control and Bok-KO MEFs displayed a similar bioenergetic profile (Fig. 5a), but, interestingly, Bok-KO MEFs had significantly higher spare respiratory capacity (difference between maximal and basal OCR) (Fig. 5b), despite the levels of ETC protein complexes being unaltered by Bok KO (Fig. 5c). Surprisingly, Bok-KO MEFs also exhibited a significantly higher mitochondrial membrane potential, as indicated by enhanced TMRE fluorescence (Fig. 5d, e). Thus, the mitochondrial fragmentation seen in Bok-KO MEFs does not inhibit mitochondrial bioenergetics, and in fact, correlates with enhanced spare respiratory capacity and membrane potential. Consistent with this lack of inhibition of bioenergetics, cellular proliferation was not impaired (Fig. 5f).

Fig. 5
figure 5

Bok (Bcl-2-related ovarian killer) deletion increases mitochondrial spare respiratory capacity and membrane potential, but has no effect on cell growth. a Oxygen consumption rate (OCR) in control and Bok-knockout (KO) mouse embryonic fibroblasts (MEFs) treated with oligomycin, carbonyl cyanide-p-trifluoromethoxy-phenyl-hydrazone (FCCP), and a combination of rotenone and antimycin A (traces are mean ± SEM, N = 4). b Spare respiratory capacity in control and Bok-KO MEFs. Histograms are mean ± SEM, N = 4, with p values comparing control to Bok KO via unpaired Student’s t test. c Levels of electron transport chain (ETC) complexes in lysates from control and Bok-KO MEFs, with p97 serving as a loading control. d Representative histogram of control and Bok-KO MEFs incubated with 500 nM tetramethylrhodamine ethyl ester perchlorate (TMRE) and analyzed by flow cytometry. Unstained control cells are also shown for comparison. e Mean TMRE fluorescence in control and Bok-KO MEFs. Histograms are mean ± SEM, N = 4, with p values comparing control to Bok KO via unpaired Student’s t test. f Cell growth curves from control and Bok-KO MEFs (mean ± SEM; N = 5)

Full size image

Bok deletion does not affect apoptosis

Since Bok is a Bcl-2 family member, we sought to determine if endogenous Bok contributes to apoptotic signaling by exposing control and Bok-KO MEFs to various apoptotic stimuli; SST (a non-selective protein kinase inhibitor), MG132 (a proteasome inhibitor), TG (a sarco/ER Ca2+ ATPase pump inhibitor), BFA (a Golgi/ER transport inhibitor), and TN (a N-linked glycosylation inhibitor). However, Fig. 6 (lanes 3–12) shows that regardless of the stimulus, control and Bok-KO MEFs displayed very similar levels of apoptotic activity as assessed by PARP and caspase-3 cleavage (cC3), and quantification of cC3 showed no significant changes (p > 0.05, as determined by a unpaired Student’s t test) (Fig. 6). Thus, despite Bok being a Bcl-2 protein family member, endogenous Bok does not seem to play a significant role in stimulus-induced apoptosis.

Fig. 6
figure 6

Bok (Bcl-2-related ovarian killer) deletion does not alter apoptotic responses. Control and Bok-knockout (KO) mouse embryonic fibroblasts (MEFs) (odd and even numbered lanes, respectively) were incubated as indicated with dimethyl sulfoxide (DMSO), 1 µM staurosporine (SST), 10 µM MG132, 1 µM thapsigargin (TG), 0.5 µg/mL brefeldin A (BFA), or 0.5 µg/mL tunicamycin (TN), and cell lysates were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and probed in immunoblots for the proteins indicated, with erlin2 serving as a loading control. The histogram shows quantitated cleaved caspase-3 (cC3) immunoreactivity from multiple independent control and Bok-KO MEF cell lines (N = 5)

Full size image

Discussion

In summary, we find that (i) despite Bok having a strong interaction with IP3Rs, it does not regulate IP3R channel activity, (ii) both long-term deletion and acute depletion of Bok results in mitochondrial fragmentation, (iii) maintenance of normal mitochondrial morphology is dependent on the presence of Bok, but not the Bok-IP3R interaction, (iv) the mitochondrial fragmentation caused by Bok deletion is due to a decrease in mitochondrial fusion rate, perhaps because of a decrease in motility, (v) Bok deletion enhances mitochondrial bioenergetics, and, lastly, (vi) despite Bok being a Bcl-2 family member, it does not play a significant role in stimulus-induced apoptosis. Overall, these data reveal a novel cellular role for Bok in maintaining mitochondrial morphology by sustaining mitochondrial fusion.

We reported previously that control MEFs and Bok-KO MEFs derived from Bok-KO mice had identical IP3R-mediated Ca2+ responses [18]. However, a caveat was that IP3R expression was altered by Bok KO, with IP3R1 levels increased and IP3R2 and IP3R3 levels decreased, possibly due to adaptation to the long-term germline loss of Bok [18]. In contrast, the Bok-KO MEFs generated in the present study using CRISPR/Cas9-mediated gene editing did not have altered IP3R expression (Fig. 1a). Despite the thorough investigation of these MEFs and other Bok-KO cells shown in Fig. 1, Bok was not seen to regulate IP3R-induced Ca2+ mobilization in intact cells or IP3R channel activity. This is surprising for several reasons: (i) Bok binds strongly to IP3Rs, (ii) essentially all cellular Bok is IP3R bound [17, 18], (iii) each of the four polypeptides in an IP3R tetramer appear to be Bok bound, and (iv) Bok binds much more efficiently to IP3Rs than other Bcl-2 protein family members (e.g., Bcl-2 and Bcl-xl) [18] that have been reported to regulate IP3Rs [24]. Further, Bok KO did not alter [Ca2+]m, indicating that Ca2+ transfer to mitochondria is not regulated by Bok and suggesting that ER–mitochondria contact sites are normal in Bok-KO cells. It remains a possibility, however, that Bok could regulate IP3R-independent Ca2+ homeostasis at the ER (e.g., Ca2+ uptake or passive leak) [50], or that effects of Bok on IP3R activity could be revealed under different conditions.

Rather than affecting IP3R activity, Bok deletion was found to cause mitochondrial fragmentation by inhibiting mitochondrial fusion. Interestingly, mitochondrial fragmentation is also seen when other Bcl-2 protein family members are deleted (e.g., Bak, Bax, Mcl-1, Bcl-xl, and Bcl-2), apparently due to their interactions with fusion or fission mediators [28,29,30,31,32,33,34]. For example, Bak and Bax have been shown to maintain normal mitochondrial morphology by regulating mitochondrial fusion rate via interactions with Mfn2 [28,29,30,31]. However, using co-IP we were unable to detect interactions between Bok and Mfn1/2 or Drp-1 (data not shown). This does not rule out the possibility that such interactions occur, since co-IP identifies only strong protein–protein interactions, and Bok localization to the ER may limit it from interacting strongly with fusion or fission mediators, which predominately localize to the mitochondria [27]. Alternatively, Bok could sustain mitochondrial fusion indirectly by interacting with a Bcl-2 protein family member that in turn has a direct interaction with a fusion or fission mediator. For example, Bok has been reported to interact with Mcl-1 by a yeast two-hybrid system [5] and Mcl-1 interacts with Drp-1 to regulate mitochondrial dynamics [33]. This idea is consistent with the notion that Bcl-2 family members form a network that acts cooperatively to regulate mitochondrial dynamics [25, 26, 30]. Finally, since we found that Bok deletion decreases mitochondrial particle velocity (Fig. 4g), it is possible that this accounts for the reduced fusion rate in Bok-KO MEFs; a decrease in mitochondrial motility will diminish the likelihood of two mitochondria being in close proximity, which is necessary for their fusion [25, 27]. At present, it is unclear why motility is reduced, as the level of Miro1 was unaltered by Bok KO (Fig. 4j).

Interestingly, both exogenous WTBok and L34GBok, an IP3R1 binding-deficient Bok mutant, were able to normalize mitochondrial morphology (Fig. 3). Thus, maintenance of a normal mitochondrial morphology is dependent merely on the presence of Bok, rather than the Bok-IP3R interaction. Intriguingly, L34GBok caused a more elongated mitochondrial morphology than WTBok, indicating that “free” Bok (that not bound to IP3Rs) may be a more potent promoter of mitochondrial fusion than tethered Bok, and perhaps that the Bok-IP3R interaction restrains Bok activity in order for it to maintain normal mitochondrial dynamics. At the same time, it is important to note that endogenous Bok is profoundly stabilized by binding to IP3Rs [18]. Therefore, the interaction between endogenous IP3Rs and Bok is critical for maintaining Bok at levels that can sustain normal mitochondrial fusion.

Mitochondrial fragmentation induced by KO of Bcl-2 family members can inhibit mitochondrial bioenergetics in some cases, for example, Mcl-1 KO in MEFs [32], but not always, for example, Bax/Bak double KO in MEFs [29]. Interestingly, the mitochondrial fragmentation induced by Bok KO did not inhibit mitochondrial bioenergetics, but, surprisingly, enhanced mitochondrial spare respiratory capacity (Fig. 5b) and membrane potential (Fig. 5d, e). These changes are not due to an increase in the levels of ETC protein complexes and may simply be due to the increase in mitochondrial surface area that will inevitably result from mitochondrial fragmentation. Such an increase in surface area could enhance ETC activity by deconstraining diffusion. Clearly, additional work is needed to resolve these questions.

Since Bok is a Bcl-2 protein family member, many studies have focused on the possible role that it could play in apoptosis. However, this has not led to a clear consensus view, with studies showing both pro- and anti-apoptotic characteristics [4, 7, 9, 11,12,13,14,15,16]. Indeed, a concern with the numerous studies showing that overexpressed exogenous Bok is pro-apoptotic [7, 9, 11, 12] is that overexpressed Bok will likely saturate the available IP3R binding sites and generate “free Bok” that can cause non-physiological deleterious effects and perturbations to the Bcl-2 family network. The Bok-KO approach to defining the role of Bok has also led to conflicting findings; for example, Bok-KO MEFs derived from germline Bok deletion by two independent groups have shown Bok to have either an anti- or pro-apoptotic role in response to ER stress agents [7, 13,14,15]. In contrast, the CRISPR/Cas9-derived Bok-KO MEFs used in the present study reveal that Bok has no role in apoptosis induced by various stimuli, including ER stress agents (Fig. 6). These differences may be due to adaptation to the long-term loss of Bok in the MEFs derived from KO mice, which do not occur in the Bok-KO MEFs derived from CRISPR/Cas9-mediated gene editing. In summary, our data support the idea that Bok plays little or no role in stimulus-induced apoptosis.

Overall, our findings reveal that an important role of Bok is to maintain mitochondrial morphology by sustaining mitochondrial fusion rate. Further research is needed to determine the mechanism by which Bok regulates mitochondrial fusion, motility, and bioenergetics and the proteins with which Bok interacts, beyond IP3Rs. Another major question that remains unsolved is precisely how the Bcl-2 family network [1,2,3] collectively regulates mitochondrial dynamics.