NAD+ plays essential roles in cellular energy homoeostasis and redox state, functioning as a cofactor along the glycolysis and citric acid cycle pathways. Recent discoveries indicated that, through the NAD+-consuming enzymes, this molecule may also be involved in many other cellular and biological outcomes such as chromatin remodelling, gene transcription, genomic integrity, cell division, calcium signalling, circadian clock and pluripotency. Poly(ADP-ribose) polymerase 1 (PARP1) is such an enzyme and dysfunctional PARP1 has been linked with the onset and development of various human diseases, including cancer, aging, traumatic brain injury, atherosclerosis, diabetes and inflammation. In the present study, we showed that overexpressed acyl-CoA-binding domain containing 3 (ACBD3), a Golgi-bound protein, significantly reduced cellular NAD+ content via enhancing PARP1's polymerase activity and enhancing auto-modification of the enzyme in a DNA damage-independent manner. We identified that extracellular signal-regulated kinase (ERK)1/2 as well as de novo fatty acid biosynthesis pathways are involved in ACBD3-mediated activation of PARP1. Importantly, oxidative stress-induced PARP1 activation is greatly attenuated by knocking down the ACBD3 gene. Taken together, these findings suggest that ACBD3 has prominent impacts on cellular NAD+ metabolism via regulating PARP1 activation-dependent auto-modification and thus cell metabolism and function.
Acyl-CoA-binding domain containing 3 (ACBD3) is known as a Golgi-resident protein that can be released into cytosol during cell mitosis because of disassembly of the Golgi apparatus and regulates Numb-controlled asymmetric division and neurogenesis of neural progenitor cells [1,2]. Early studies on this protein found that ACBD3 physically interacts with protein kinase A regulatory subunit γ and plays essential roles in the biosynthesis of steroid hormones in the mitochondria . ACBD3 is also functionally connected with structural homoeostasis of Golgi, intracellular membrane trafficking and viral RNA replication in the host cells [4–6].
Our previous studies identified a neuronal nitric oxide synthase (nNOS)/Dexras1/ACBD3/divalent metal transporter 1 (DMT1) signalling pathway through which excitatory amino acids initiate overwhelming iron influx following N-methyl-D-asparate (NMDA) receptor activation and contribute to excitoneurotoxicity [7,8]. Because of the N-terminal ACBD, ACBD3 has been presumably considered as a member of lipid-binding protein family and may modulate lipid transport, metabolism, storage and/or signalling like other family members [9–11]. One of our previous studies demonstrated that ACBD3 inhibits the intracellular processing pathway of the sterol-regulatory element-binding protein 1A (SREBP1A), a master regulator of cellular lipid metabolism, attenuates maturation of this transcriptional factor and eventually down-regulates the de novo biosynthesis of fatty acids . Thus, these new findings reveal that ACBD3 plays regulatory roles in cellular metabolism via impacting on the lipogenic pathway.
Poly(ADP-ribosyl)ation (PARylation) is an intensively-studied protein post-translational modification (PTM), which is catalysed by poly(ADP-ribose) polymerases (PARPs) in the cells [13–15]. PARylation is regulated by the enzymes via transferring and polymerizing ADP-ribose units from NAD+ to form linear or branched polyanionic chains that are covalently linked with acceptor proteins. Even though there are as many as 17 proteins classified as PARP family members due to sequence homology, the founding member nuclear PARP1 is claimed to be responsible for synthesis of more than 85% of PAR in cells. One of the major substrate proteins of PARP1 is itself and this auto-modification has been suggested to cause a negative feedback on its catalytic capacity. Like other formats of PTM, PARylation may affect the function of acceptor proteins via changing their enzymatic activities directly or binding affinities with their partners (proteins or nts).
Earlier findings from Krishnakumar et al.  demonstrated that PARP1 binds with not only the enhancers but also the promoters of most actively transcribed genes therefore having a direct impact on the transcription of those genes. It has been suggested that PARP1 executes these regulatory effects on gene transcription via functioning as a transcriptional co-regulator, promoter-specific exchange factor, insulator or DNA methylation regulator [14,17–19]. In addition, the capability to remodel chromatin structure (loose or compact) by interacting with nucleosomes, chromatin histones and chromatin-associated proteins also contributes to its influences on gene transcription [14,20,21].
Another well-defined cellular function of PARP1 is DNA damage detection and repair and it has been implicated in several different DNA repair pathways such as base excision repair (BER), single-strand break (SSB) and double-strand break (DSB) repair [13,14,22]. Once DNA damage signals are detected by this sensor protein, it will relocate to the damage sites, recruit/modify proteins (XRCC1, X-ray repair cross-complementing protein 1; DNA-PK, DNA-dependent protein kinase; ATM, ataxia telangiectasia mutated etc.) and assemble the repairing machinery to correct the sequences, maintain the genomic integrity and rescue cells from genetic mutations and even cell death. However, if the DNA-damaging stimulus is too strong and lasts for a long time, over-activated PARP1 may result in cell death through the necrosis pathway or the caspase-independent apoptosis pathway, due to NAD+/ATP depletion and translocation of mitochondrial pro-apoptotic protein called apoptosis-induction factor to the nucleus respectively [13,15].
In the recent years, a large body of studies has been performed to investigate the potential roles of PARP1 in regulating cellular metabolism because the hyperactivated enzyme may deplete nuclear and cytoplasmic NAD+, whose function is fundamental for cellular ATP biosynthesis along both the cytoplasmic glycolysis pathway and the mitochondrial oxidative phosphorylation pathway . Moreover, it is well established that the NAD+-dependent type III deacetylase Sirtuin 1 (SIRT1) plays important roles in cell metabolic homoeostasis . As a consequence, it is very likely that PARP1-mediated NAD+ reduction may affect energy metabolism through down-regulating SIRT1 activity. In fact, interesting findings from Bai et al.  strongly support this hypothesis: both NAD+ content and SIRT1 activity were enhanced in brown adipose tissue and muscle from PARP1−/− C57BL/6 mice. PARP1 was also found to be involved in the white adipocyte differentiation and lipid metabolism by influencing PPARγ-dependent gene expression and mitochondrial biogenesis . Furthermore, several lines of evidence indicate that PARP1-regulated PAR metabolism may have significant effects on mitochondrial gene expression and respiratory function [25,27,28]. Taken together, these findings are undoubtedly in line with the regulatory roles played by PARP1 in cellular energy homoeostasis.
Considering the fact that both ACBD3 and PARP1 are implicated in cellular metabolism, we were greatly interested in investigating whether there is functional cross-talk between these two proteins and how ACBD3 regulates energy homoeostasis in a PARP1-related manner. In the present studies, we found that overexpressed ACBD3 provoked extracellular signal-regulated kinase (ERK)1/2- and SREBP1-involved PARP1 activation, contributing to the significantly decreased cellular NAD+ content. Importantly, suppression of ACBD3 expression markedly reduced oxidative stress-mediated PARP1 activation.
Reagents and antibodies
Dulbecco's modified Eagle's medium (DMEM, 1×) was purchased from Mediatech. NUPAGE precast SDS/PAGE gel, FBS, penicillin–streptomycin (PS) mixture and Alexa Fluor 594 goat anti-mouse IgG Neon were acquired from Invitrogen. Polyfect transfection reagent was purchased from Qiagen. Pure-yield plasmid midiprep system and dNTP were obtained from Promega. Quick ligation kit and restriction enzymes were from New England Biolabs; SuperSignal West Pico chemiluminescence reagent, microscope cover glasses and slides were purchased from Thermo Scientific. PARP inhibitor ABT-888 and Olaparib were obtained from Selleckchem. The following reagents were purchased from the indicated companies: universal PARP colorimetric assay kit w/histone-coated strip wells from Trevigen; phosphatase inhibitor cocktail 2 and 3, hydrogen peroxide from Sigma–Aldrich; fatty acid synthase (FASN) inhibitor C75 from Cayman Chemical; DNA damage assay kit from Active Motif; SIRT1 fluorometric assay kit from Enzo Life Sciences; highly reactive oxygen species (ROS) detection kit from Cell Technology; Pfu Ultra High Fidelity DNA polymerase from Stratagene; FractionPREP cell fractionation kit from BioVision; anti-Myc mouse monoclonal antibody from Roche; anti-ACBD3 mouse monoclonal antibody from Novus Biologicals; primary antibodies against PARP1, glyceraldehyde-3-phosphate dehydrogenase (GAPDH), ERK1/2 and phospho-ERK1/2 and ERK1/2 inhibitor U0126 from Cell Signaling Technology; horseradish peroxidase (HRP)-conjugated goat anti-mouse and donkey anti-rabbit IgGs from Jackson ImmunoResearch; glutathione 4B sepharose beads and HRP-conjugated anti-GST antibody from GE Healthcare; anti-H2AX and anti-SREBP2 antibodies from Abcam; anti-PAR antibody from BD Biosciences; VECTASHIELD mounting medium with DAPI and normal goat serum from Vector Laboratories.
Human embryonic kidney (HEK)293T , HeLa (A.T.C.C.) and NIH3T3 (A.T.C.C.) cells were cultured in complete growth medium containing DMEM supplemented with 10% FBS and 100 units/ml PS at 37°C with 5% CO2 atmosphere in a humidified incubator.
Generation of constructs
Human SREBP1A (GenBank ID: U00968) cDNAs encoding the two nuclear mature forms (1–490 and 1–460) were generated from construct pSREBP-1A (A.T.C.C.) and cloned into pCMV–Myc (Clontech) between SalI and NotI sites. The plasmids encoding Myc- and GST-tagged rat full-length ACBD3 were constructed as described before . cDNAs of ΔC deletion and ΔN mutants of ACBD3 were made using the full-length as the template and then introduced into SalI/NotI sites of pCMV–Myc vector. Plasmid pEYFP–Golgi was obtained from Clontech.
Cells were transfected with tested plasmids using PolyFect (Qiagen) reagent according to the manufacturer's protocol and then harvested at the indicated time point after transfection. The cell pellet was solubilized in buffer A (100 mM Tris, pH 7.4, 150 mM NaCl, 1% Triton X-100, 15% glycerol, phosphatase inhibitor cocktail 2 and 3, 1 mM PMSF, 25 μg/ml antipain, 50 μg/ml leupeptin, 50 μg/ml aprotinin, 25 μg/ml chymostatin and 25 μg/ml pepstatin). Total protein of 10 μg for each sample was loaded for SDS/PAGE followed by immunoblotting with appropriate antibodies. Densitometric analysis was performed using ImageJ software. All the chemicals were obtained from Sigma–Aldrich.
GST–ACBD3-encoding vector was co-transfected with empty vector or plasmid expressing Myc–Dexras1 into HEK293T cells. Forty-eight hours later, cells were harvested and lysed in buffer A. Total protein of 400 μg was incubated with 40 μl of glutathione 4B sepharose beads overnight at 4°C. Then, beads were washed three times with the lysis buffer and bound proteins were eluted with 2× SDS-loading sample buffer by boiling for 10 min. Samples were separated by SDS/PAGE and analysed by immunoblotting. Total cell lysate (10 μg of protein) was loaded as input.
Immunofluorescence staining and confocal microscopy
HEK293T cells were plated on to microscope cover glass pre-treated with poly-D-lysine (0.1 mg/ml) and left overnight at room temperature. After recovery for 24 h, cells were transfected with pEYFP–Golgi and vectors encoding Myc-tagged full-length, ΔC or ΔN ACBD3 using PolyFect (Qiagen). Forty hours after transfection, immunocytochemistry staining and image capture and analysis were performed as described before .
Cells were transfected with pCMV—Myc–ACBD3 or empty vector using PolyFect on day 1. On day 3, cells were harvested followed by cell fractionation to extract cellular membrane and nuclear fractions using FractionPREP kit according to the manufacturer's protocol. Aliquots of the nuclear extracts (9 μg of protein) and membranes (9 μg of protein) were subjected to SDS/PAGE/immunoblotting analysis.
Knockdown of endogenous ACBD3
Two shRNA-expressing vectors for knocking down endogenous human ACBD3 were generated as described before . Cells were transfected with the mixture of the two plasmids (1:1) or the empty control using Lipofectamine 2000 (Invitrogen) according to the manufacturer's protocol. Cells were harvested 72 h after transfection and then lysed for immunoblotting to determine protein expression.
Measurement of NAD+ content
Intracellular NAD+ levels were quantified using a cycling assay-based method . Briefly, cells transfected with empty or pCMV—Myc–ACBD3 vector were harvested 48 h later and snap-frozen in liquid nitrogen. To the frozen cell pellets were then added pre-chilled 0.6 M perchloric acid or 0.25 M KOH in 50% ethanol for preparing extracts containing total NAD (NAD+ plus NADH) and NADH respectively. The alkaline supernatants were incubated at 55°C for 5 min to decompose NAD+. Both acidic and alkaline supernatants were added to the cycling reaction mixture, containing phosphate buffer (0.1 M, pH 8.0), BSA (0.1 mg/ml), nicotinamide (10 mM), ethanol (2%, v/v), flavin mononucleotide (10 μM), resazurin (20 μM), alcohol dehydrogenase (0.1 mg/ml) and diaphorase (10 μg/ml). Fluorescent product resazurin (λex: 530 nm, λem: 590 nm) was quantified using Synergy 4 Multi-mode Microplate Reader (BioTek Instruments). NADH was employed to generate a standard curve and to calculate the total NAD and NADH. The intracellular NAD+ content was obtained by subtracting NADH from total NAD. The protein amount in each cell pellet was used for normalization.
Polymerase activity of PARP was analysed with a colorimetric PARP assay kit according to the manufacturer's protocol. In brief, cells were lysed in buffer containing 1× PARP buffer, 0.4 M NaCl, 1% (v/v) Triton X-100 and protease inhibitor cocktail. Then the lysates were mixed with biotinylated NAD+-containing 1× PARP cocktail and were allowed to react for 60 min at room temperature. The PARylated histone products were eventually determined after adding HRP-conjugated streptavidin and colorimetric substrate (absorbance at 450 nm). The protein amount in the lysate was used for normalization.
All quantitative data are presented as mean±S.E.M., if they were derived from at least three experiments. For comparison of multiple groups, the data were analysed by one-way ANOVA followed by a Tukey post hoc test. Student's t test was employed for directly testing the difference between two sets of independent samples. If the P-value less than 0.05, the difference was defined as significant. All statistical analyses were performed using GraphPad Prism (GraphPad Software).
ACBD3 reduces cellular NAD+ level via stimulating PARP1 activity and increasing PAR formation
NAD+ metabolism could be strongly influenced by activation status of the polymerase. Therefore, in order to know how ACBD3 functionally interacts with PARP1, we first measured the NAD+ content in the cells overexpressing ACBD3. Figure 1(A) shows that NAD+ cellular level was significantly decreased (~33%) by overexpression of Myc-tagged ACBD3. Because PAR production is one of the major NAD+-consuming pathways in the cells and auto-PARylation is one major cellular event following PARP1 activation [13,14,30,31], we next tested whether overexpression of ACBD3 in cells might enhance auto-PARylation of PARP1. As we expected, overexpression of either Myc-tagged or GST-tagged ACBD3 protein significantly enhanced (~300%) PARylated PARP1 in the HEK293T cells, with a very marginal change in the total PARP1 expression level (Figures 1B and 1C). In addition, this ACBD3-induced PAR formation was both dose- and time-dependent (Figures 1D and 1E). Furthermore, those ACBD3-mediated effects on NAD+ and PAR formation are not cell-specific as they were observed in HeLa cells and NIH 3T3 cells (Supplementary Figure S1).
ACBD3 reduces cellular NAD+ level via stimulating PARP1 activity and increasing PAR formation in HEK293T cells
Subsequently, we grew cells in the medium containing Olaparib (0.1 μM), a specific PARP enzyme inhibitor and found that NAD+ content reduced by ACBD3 overexpression was recovered completely by 1-h of Olaparib treatment and ACBD3 overexpression-induced PAR formation was significantly blocked as early as in 30 min (Figures 1F and 1G). This quick inhibitory effect on PAR formation by Olaparib appears to be matched with the reported fast dynamics of PAR metabolism [32,33]. In order to directly evaluate ACBD3's impact on the enzymatic activity of the polymerase, we performed a PARP assay using lysate from cells with or without overexpressed ACBD3 and found that ACBD3 indeed initiated enhancement of the enzyme activity, which was significantly blocked by another PARP1 inhibitor, ABT-888 (0.1 μM, 2 h; Figure 1H). In addition, we constructed vectors expressing shRNAs, specifically targeting human ACBD3. Seventy-two hours after transfection, endogenous ACBD3 gene was successfully knocked down and protein expression was significantly reduced (Figures 1I and 1J). Concomitantly, the level of PARylated PARP1 was found to be decreased by approximately 80% in the cells overexpressing ACBD3 shRNAs. Consequently, all the findings suggest that ACBD3 may function to activate PARP1 and enhance auto-PARylation by consuming cellular NAD+.
ACBD3-mediated ERK1/2 activation contributes to PARP1 activation
ERKs have been demonstrated to directly bind with and regulate PARP1 activity . Thus, we examined whether the effect of ACBD3 on PARP1 is mediated by ERK1/2 in the cells. We found that overexpression of ACBD3 increased phosphorylation of ERK1/2 in a dose-dependent manner (Figure 2A; Supplementary Figures S1A and S1B). Moreover, the amount of the phosphorylated mitogen-activated protein kinase (MAPK) was tightly associated with the expression level of ACBD3 at all the time points following the transfection (Figure 2B; Supplementary Figures S1C and S1D). Observation from the cell fractionation experiments suggested that extranuclear ERK1/2 translocated into cell nuclei after being phosphorylated, where they might interact physically with and activate nuclear PARP1 (Figure 2C). Subsequently, we examined whether ACBD3-stimulated PARP1 activation was sensitive to the ERK1/2 inhibitor U0126 and confirmed that ACBD3-stimulated PARylation of PARP1 was significantly reduced by the MAPK inhibitor, U0126 (10 μM) by approximately 50% (Figures 2D and 2E). Collectively, these results indicate that ACBD3-induced ERK1/2 activation and translocation are involved in the ACBD3-mediated PARP1 activation and auto-modification.
ACBD3-mediated ERK1/2 activation contributes to PARP1 activation in HEK293T cells
ACBD3-inhibited de novo synthesis of fatty acids is implicated in PARP1 activation
Our recent observations revealed that ACBD3 plays an inhibitory role in the de novo biosynthesis of fatty acids via blocking the intracellular maturation pathway of transcriptional factor SREBP1A, a master regulator of lipid metabolism . In order to assess whether lipid metabolism contributes to a modulation of ACBD3-mediated regulation of PARP1 activity, cells were incubated with C75, a specific inhibitor targeting FASN. Figure 3(A) shows that FASN inhibition by C75 (5 μg/ml) for 4–6 h was effective to enhance auto-PARylation of PARP1 as well as ERK1/2 activity. Next, we tested whether re-instatement of cellular fatty acid contents could mitigate ACBD3-induced PARP1 stimulation. For recovering cellular fatty acids in cells overexpressing ACBD3, In order to restore cellular fatty acids in cells overexpressed with ACBD3, N-terminal fragments of SREBP1A which bypass the requirement of protease-mediated activation of a full length SREBP1, can be overexpressed directly in the nucleus as an active transcription factor. We found that overexpression of nuclear active form of SREBP1A significantly attenuated ACBD3's stimulatory impact on PARP1 activity (Figure 3C). These results strongly suggest that the ACBD3-inhibited lipogenic pathway contributes to the stimulated auto-PARylation of PARP1.
ACBD3-inhibited biosynthesis of fatty acids is implicated in PARP1 activation in HEK293T cells
ACBD3 overexpression does not cause DNA damage, oxidative stress or ER stress
We also conducted a series of experiments to further define whether there are ERK1/2- and SREBP1-independent molecular mechanisms underlying ACBD3's stimulatory impact on PARP1. Since damaged DNA strands (SSB and DSB) are well-established inducers of PARP1 activation [15,22] we firstly examined one well-recognized DNA damage marker, phosphorylated (Ser139) histone protein H2AX (γ-H2AX) [22,35]. We confirmed that ACBD3 overexpression did not significantly up-regulate the DNA damage marker, which was apparently responsive to the DNA damage inducer etoposide (Supplementary Figures S3A and S3B). Results from experiments quantifying cellular ROS was in agreement with those from DNA damage experiments and ACBD3 did not initiate significant production of ROS in the cells (Supplementary Figure S3C). As described earlier, ACBD3 may augment cellular iron uptake and initiate cellular oxidative stress that has been linked with neuroexcitotoxicity [7,8]. In order to clarify the possible role of an iron-involved pathway, cells overexpressing ACBD3 were treated with iron chelator deferoxamine (DFO) or salicylaldehyde isonicotinoyl hydrazone (SIH) which is highly membrane permeable. We found that iron chelation did not block ACBD3-stimulated PARylation of PARP1 (Supplementary Figure S3D). In addition, observations in the co-immunoprecipitation experiment showed that ACBD3 did not physically bind with PARP1 (Supplementary Figure S4), implying that the ACBD3 indirectly exerts its regulatory function on PARP1 through linking intermediates.
It is also known that treatment with brefeldin A (BFA), an endoplasmic reticulum (ER) stress inducer, can disassemble Golgi structure and promote protein PARylation [36,37]. Considering that ACBD3 is also known as a Golgi-resident protein, we investigated whether ACBD3 stimulated auto-modification of PARP1 through disturbing Golgi structure or inducing ER stress. As shown in Supplementary Figure S5(A), cellular localization of Golgi apparatus in cells treated with BFA became more dispersive, indicating the disintegrated Golgi structure. On the other hand, full-length ACBD3 overexpression had no influence on the distribution pattern of Golgi in the cells. Thus, it was clear that ACBD3 did not cause Golgi disassembly. We also examined whether overexpression of ACBD3 induces ER stress. Slowed migration pattern of phosphorylated PRKR-like endoplasmic reticulum kinase (PERK) on SDS/PAGE gel was employed as the marker of ER stress . According to Supplementary Figure S5(B), treatment with two well-recognized ER stress inducers, BFA and tunicamycin , did slow down the migration of PERK protein and increase PAR formation. On the other hand, overexpressed ACBD3 had no effect on PERK migration pattern, but up-regulated PARylated PARP1 prominently. All these results exclude the possibility that ACBD3 modulates PARP1 activation status through disassembling the Golgi structure and enhancing ER stress.
N-terminal sequence is necessary for ACBD3-induced PARP1 activation
We have previously shown that the ACBD-containing N-terminal sequence (1–172) was important for the ACBD3-regulated lipogenic process . Importantly, the results in Figure 3 showed that the ACBD3-mediated disruption of lipid pathway is associated with the PARP1 activation. Accordingly, we wondered whether the ACBD was also required for the ACBD3's function affecting the NAD+ metabolic pathway. We generated serial deletion mutants of ACBD3 and found that the C-terminal Golgi dynamic (GOLD) domain (residues 384–527), which is responsible for Golgi association was not necessary for this newly-defined role of ACBD3. Besides the GOLD domain, experiments using mutants with further C-terminal deletion implied that all the residues from 230 to 527 appeared to be not required (Figures 4A–4C). On the other hand, it was shown that the ACB domain-containing N-terminal residues (1–172) were indispensable for ACBD3-mediated PARP1 stimulation as deletion of this region severely attenuated the auto-PARylation of PARP1. Similar to what we reported before, our imaging data showed that the C-terminal, but not the N-terminal sequence, was essential for ACBD3's physical association with Golgi structure [40,41]. The results indicate that Golgi residence is not necessary for ACBD3-mediated activation of PARP1. These studies further corroborate that ACBD3-mediated regulation of PARP1 activity is associated with the cellular lipogenic pathway as our previous study demonstrated that N-terminal fragment is necessary for cellular lipid regulation .
N-terminal sequence is necessary for ACBD3-induced PARP1 activation in HEK293T cells
ACBD3 knockdown attenuates H2O2-induced PARP1 activation
Cellular oxidative stress stimulated by ROS may trigger DNA damage and concomitant PARP1 activation consequently leading to an initiation of the cell death pathway if the damage exceeds the cellular repair capacity [42–44]. PARP inhibition is supposed to be an effective way to mitigate ROS-induced cell death and has been extensively investigated in many studies . As shown in Figures 1(I) and 1(J), shRNA-mediated down-regulation ACBD3 was capable of reducing PARP1's basal activity. Thus, we determined whether ACBD3 knockdown could alleviate PARylation of PARP1 mediated by cellular oxidative stress. Cells were treated with hydrogen peroxide (H2O2) to increase cellular oxidative stress, which was shown to enhance PAR formation in a dose-dependent manner (Figure 5). However, reduced ACBD3 expression following gene knockdown apparently decreased the level of PARylated PARP1, implying that a blockade of ACBD3–PARP1 signalling pathway may be beneficial for oxidative stress-involved pathological conditions.
ACBD3 knockdown significantly attenuates H2O2-induced PARP1 activation in HEK293T cells
In the present study, we have identified lipid-binding protein ACBD3 as an upstream regulator to stimulate PARP1 activity, induce auto-PARylation and subsequently down-regulate cellular NAD+ content. Furthermore, we found that enhanced ERK1/2 activity and inhibited SREBP1-controlled fatty acid biosynthesis play crucial roles in this ACBD3-regulated PARP1 activation. These data significantly broaden our views on the biological roles played by ACBD3 and may provide novel insights into the mechanisms underlying how lipid-binding proteins affect metabolic homoeostasis via PARP1-involved signalling pathway.
Our findings demonstrate that ACBD3 overexpression may give rise to phosphorylation and activation of serine/threonine-specific ERK1/2, the two classic members of the MAPK family that are involved in various cellular functions such as differentiation and migration . Phosphorylated/activated ERK1/2 have been implicated in PARP1 activation and auto-PARylation in both DNA-independent and DNA-dependent manners [34,47,48]. Our observations revealed that ACBD3 did not enhance cellular ROS nor induce DNA damage, which strongly suggests that modulation of PARP1 by MAPK is the DNA-independent way in cells overexpressing with ACBD3. The findings from cell fractionation assays (Figure 2C; Yong Chen and Sangwon F. Kim, unpublished data) implied that phosphorylated extranuclear ERK1/2 translocated into the cell nucleus where it might bind and activate PARP1 directly. The PARylated PARP1 has been suggested to function as a nuclear anchoring protein for ERK1/2 and promote ERK1/2-mediated phosphorylation of transcriptional factor E26 transformation-specific domain containing protein (Elk1) . However, the signalling molecules upstream of ERK1/2 remain to be investigated. There are several candidate pathways, including Raf–mitogen-activated protein kinase kinase (MEK)1/2, protein kinase A, protein kinase C and histone acetyltransferase P300/CBP, all of which have been linked with direct ERK1/2 activation [49–51].
The PARP1-involved signalling pathway is responsive to various extracellular and intracellular stimuli such as hormones (retinoid, oestrogen, platelet-derived growth factor (PDGF) etc.), stresses (heat shock, inflammation, bacterial infection, genotoxic reagent, oxidative stressor etc.) and even membrane depolarization [14,52–54]. PARP1 activation and protein PARylation have been associated with many cellular functions, including chromatin remodelling, gene transcription, DNA repair, cell survival and division, pluripotency and energy metabolism [14,55,56]. Normal expression and activity of PARP1 are essential for maintaining genomic integrity, memory formation and consolidation, longevity, circadian clock and so forth [14,57–59]. Dysregulated PARP1-related signals play causative roles in various human pathological conditions, for example cancer, inflammation, aging, atherosclerosis, obesity, diabetes and traumatic brain injury [14,55,60–62]. One of our recent experiments showed that expression of ACBD3 is prominently enhanced in the breast tissues from two mouse breast cancer models (Yong Chen and Sangwon F. Kim, unpublished data). More studies are ongoing for obtaining expression profiles of other ACBD3-related signalling molecules. In addition, ACBD3 was found to be up-regulated in the livers of mice with a high-fat diet (Yong Chen and Sangwon F. Kim, unpublished data), under which PARP1 activity was also increased . Hence, those novel findings warrant more future investigations on how ACBD3 is regulated under these pathophysiological conditions and whether manipulation of the ACBD3–PARP1 signalling axis has therapeutic potential in the clinic.
To the best of our knowledge, the current study is the first report showing a functional connection between ACBD3 and PARP1. Our findings provide evidence that ACBD3 is a crucial regulator of PARP1 activity, protein PARylation and cellular NAD+ metabolism through ERK1/2- and SREBP1-dependent pathways. More studies are required to define the exact molecular mechanisms underlying ACBD3's regulatory function with respect to the PARP1 protein and to probe how interventions modulating ACBD3–PARP1 signalling cascades can be targeted for PARP1-associated diseases such as cancer, obesity, diabetes and traumatic injury of the nervous system.
Yong Chen and Sangwon Kim were involved in the concept, design and interpretation of data. Sookhee Bang constructed ACBD3 mutants and Soohyun Park and Hanyuan Shi performed PARP1 enzyme activity assays and immunoblotting for PAR formation. The paper was written by Yong Chen and Sangwon Kim with contributions from all authors.
We are grateful for Catherine Steenstra for laboratory support.
This work was supported by the National Institute of Health [grant numbers HD026979, MH079614 and DK084336 (to S.F.K.)].
acyl-CoA-binding domain containing 3
Dulbecco’s modified Eagle’s medium
extracellular signal-regulated kinase
fatty acid synthase
human embryonic kidney
mitogen-activated protein kinase
reactive oxygen species
sterol regulatory element-binding protein 1A