Palmitate-induced lipotoxicity alters acetylation of multiple proteins in clonal β cells and human pancreatic islets

Type 2 diabetes is characterized by progressive β cell dysfunction, with lipotoxicity playing a possible pathogenetic role. Palmitate is often used to examine the direct effects of lipotoxicity and it may cause mitochondrial alterations by activating protein acetylation. However, it is unknown whether palmitate influences protein acetylation in β cells. We investigated lysine acetylation in mitochondrial proteins from INS-1E β cells (INS-1E) and in proteins from human pancreatic islets (HPI) after 24 h palmitate exposure. First, we confirmed that palmitate damages β cells and demonstrated that chemical inhibition of deacetylation also impairs INS-1E function and survival. Then, by 2-D gel electrophoresis, Western Blot and Liquid Chromatography-Mass Spectrometry we evaluated the effects of palmitate on protein acetylation. In mitochondrial preparations from palmitate-treated INS-1E, 32 acetylated spots were detected, with 13 proteins resulting over-acetylated. In HPI, 136 acetylated proteins were found, of which 11 were over-acetylated upon culture with palmitate. Interestingly, three proteins, glutamate dehydrogenase, mitochondrial superoxide dismutase, and SREBP-1, were over-acetylated in both INS-1E and HPI. Therefore, prolonged exposure to palmitate induces changes in β cell protein lysine acetylation and this modification could play a role in causing β cell damage. Dysregulated acetylation may be a target to counteract palmitate-induced β cell lipotoxicity.


Function and survival of INS-1E β cells.
First we assessed the effects of prolonged palmitate exposure on glucose-stimulated insulin secretion from INS-1E β cells. As shown in Supplementary Fig. 1, insulin release in response to 2.5 and 16.7 mM glucose was not apparently influenced by culture in palmitate-containing medium for 6 and 14 h. However, after 24 h exposure to the fatty acid, INS-1E β cells showed increased basal secretion and impaired ability to proportionately augment the release of insulin at higher glucose concentration; in addition, compared to control samples, palmitate-treated samples showed reduced cell survival (−32.0 ± 9.4%, p < 0.05).
These results confirm what reported in previous studies on β cell deleterious effects of palmitate 27,28 . Then, in another set of experiments, INS-1E β cells were exposed for 24 h to a combination of deacetylase inhibitors (trichostatin A and nicotinamide). As illustrated in Supplementary Fig. 1, cells exposed to the deacetylase inhibitors showed a slightly higher release of insulin at basal (2.5 mM) glucose concentration, and the secretion of the hormone at stimulating glucose level (16.7 mM glucose) was significantly lower compared to that of control cells. Furthermore, in the presence of deacetylase inhibition, cell survival was 33.1 ± 10.8% lower than that of cells cultured in control medium (p < 0.05). When INS-1E β cells were incubated for 24 h with both palmitate and deacetylase inhibitors, we observed a further deterioration of insulin release with respect to palmitate alone, particularly at basal conditions (release at 2.5 mM glucose: 3.98 ± 2.74%; p < 0.05; release at 16.7 mM glucose: 3.44 ± 1.33%, ns), which was accompanied by a modest, additional decrease of β cell survival (− 9.1 ± 1.2%, ns).
These results confirm that 24 h exposure to palmitate alters insulin secretion and survival of INS-1E β cells, and demonstrate that inhibition of deacetylation further impairs the function and survival of palmitate-exposed cells.

Protein acetylation in mitochondria from INS-1E β cells. Mitochondrial enriched preparations
were obtained using differential centrifugations, according to the protocols implemented within the Italian human mitochondrial proteome project (MtHPP) network 29 and the quality of enrichment was checked as previously described 19 . A mitochondrial protein fraction of approximately 150 µg (starting from 20 million cells) was obtained from each of 7 independent experiments of control and palmitate-treated INS-1E β cells. These mitochondrial preparations were pooled together and analyzed in triplicate. We first verified global acetylation levels by WB analysis. Proteins were separated by 1-D gel electrophoresis and immunoblotted using a specific anti-acetylated lysine antibody. As shown in Fig. 1 panel a, several immunoreactive bands, ranging from 7 to 167 kDa, were detected and an apparent increase of acetylation was observed after palmitate treatment ( Fig. 1  panel b).
Then, identification of specific acetylated proteins was performed. To this aim, aliquots of mitochondrial fractions were analyzed by coupling 2-DE and WB (Fig. 1 panel c and d). Immediately after WB, membranes were stained with Ruthenium II tris (bathophenanthroline disulfonate) tetrasodium salt (RuBP) and subsequently incubated with the specific anti-acetylated lysine antibody. This procedure allowed us to exactly match the image of the stained membrane containing all transferred proteins (Fig. 1, panel c) with the image of the same membrane obtained by detecting the immunoreactive spots ( Fig. 1, panel d). Each spot intensity was normalized by the total protein intensity as visualized in the image of membrane stained with RuBP. Overall, we detected 32 acetylated spots, of which 17 were over-acetylated in palmitate-treated samples. These differentially acetylated spots were analyzed by tandem mass spectrometry to obtain their identification and collapsed to 13 proteins ( Table 1). The molecular weight (MW), isoelectric point (pI), and statistical data of identified proteins are also shown in Table 1. For some proteins (voltage-dependent anion-selective channel protein 2, 60 kDa heat shock protein, succinate dehydrogenase flavoprotein subunit, ATP synthase subunit δ, pyruvate dehydrogenase E1 component subunit α, protein disulfide-isomerase A3) the corresponding acetylated spot was absent in the control samples. This finding does not necessarily indicate the absence or under expression of the protein but only of its acetylated form. Four of these proteins (78 kDa glucose-regulated protein, protein disulfide-isomerase A3, sterol regulatory element-binding protein 1 (SRBP1) and protein disulfide-isomerase A1) belong to the ER, which is partially physically and functionally associated with mitochondria [30][31][32][33] .
Scientific RePoRts | 7: 13445 | DOI:10.1038/s41598-017-13908-w Mitochondrial fractions from untreated and palmitate-treated cells were also digested into peptides and lysine-acetylated peptides were immune-captured using the specific anti-acetyl lysine antibody. The lists of proteins inferred from the identification of the acetylated peptides are supplied as Supplementary Tables S1  and S2. Consistently, the number of detected acetylation sites was higher in the mitochondrial fraction of palmitate-treated cells compared to control cells.
Protein acetylation in human pancreatic islets. We subsequently analyzed pancreatic islets from 5 non-diabetic human donors. For each experiments, islets were incubated for 24 h with or without palmitate. Approximately 350 µg of protein was extracted from each preparation. Acetylated proteins in human pancreatic In order to check the global acetylation level, one-dimensional electrophoresis was performed. Aliquots of mitochondrial samples corresponding to 10 µg of proteins were mixed with Laemmli solution, heated at 100 °C for 5 min, loaded onto 12% polyacrylamide gels and separated by electrophoresis before WB. After protein transfer, membranes were firstly stained with RuBP and subsequently incubated with the specific anti-acetylated lysine antibody. The RuBP-stained images were used for normalization. (a) Immunoreactive bands of the mitochondrial enriched fraction from INS-1E β cells detected using an anti acetylated-lysine antibody and the same membrane stained with RuBP. CTRL: control samples: PALM: palmitate-exposed samples. MW: molecular weight standards. (b) Histograms of the normalized optical density calculated for antigen-specific bands. The optical density was measured for each band of the whole lane. This density was then normalized by the density of total proteins showed by RuBP staining. Data are presented as mean ± standard deviation of three technical replicates (M ± SD) ***p ≤ 0.001. (c) A representative nitrocellulose with 2-DE protein map of mitochondrial enriched fraction of INS-1E β cells. Proteins were detected by RuBP staining. Proteins were firstly separated according to pI on Immobiline Dry-Strips (18 cm, linear gradient pH 3-10) and then according to molecular weight on 12.5% polyacrylamide gels. Then, proteins were transferred onto nitrocellulose membranes. Immediately after WB, membranes were stained with RuBP. (d) The same membrane after detection of the immunoreactive spots using the anti-acetylated lysine antibody. Spots of interest are encircled.

Discussion
To the best of our knowledge, this is the first study showing that: 1) palmitate exposure induces acetylation of mitochondrial proteins in insulin secreting INS-1E β cells and 2) some of these proteins are also over-acetylated in primary human pancreatic islets. Additional findings of the present report are the confirming evidence that palmitate-treated cells show impaired insulin release capability as well as reduced viability and the novel data indicating that treatment with deacetylase inhibitors also impairs β cell function and survival.
The possibility that palmitate might induce protein acetylation was already suggested in our previous study 19 where 2-D gel analysis showed the presence of two close spots identified as DHE3 among those found differentially expressed in palmitate treated INS-1E β cells 19 . The present work now demonstrates that 24 h exposure to 0.5 mM palmitate causes over-acetylation of several proteins in INS-1E mitochondrial preparations.
As recently reviewed 34 protein acetylation in mitochondria typically leads to loss of function, which, in turn, causes alterations of mitochondria integrity. Interestingly, three of the proteins that we identified in INS-1E β cell preparations (DHE3, SODM and the ER protein SRBP1) were also found over-acetylated in palmitate-exposed human pancreatic islets.
All the proteins that we observed to show increased acetylation in the present study are associated with important β cell functions. The mechanisms whereby glucose stimulates insulin secretion require the metabolism of glucose to pyruvate. In pancreatic β cells pyruvate metabolism also depends on the activity of the pyruvate dehydrogenase complex (PDH) and pyruvate carboxylase (PC) 35 . We found that palmitate exposure induced PDH and PC hyper-acetylation. PDH, the rate-limiting enzyme of glucose oxidation, catalyzes the irreversible decarboxylation of pyruvate into acetyl-CoA and its enzymatic activity is under acetylation control 36,37 . SIRT3 increases PDH activity while aberrant PDH hyperacetylation might result in metabolic reprogramming 38 . PC catalyzes the ATP-dependent carboxylation of pyruvate to form oxaloacetate and its activity is decreased in diabetic pancreatic islets compared to non-diabetic samples 39 . It has been previously reported that long-term exposure of rodent β cells to palmitate reduces PC expression and activity 40 . Since we found PC acetylation after cell exposure to palmitate this PTM could be responsible for the described decrease of enzyme activity.
It has been proposed that palmitate hampers mitochondrial function by causing a decrease in cellular ATP production. ATP is needed for the elevation of cytosolic calcium which is required for insulin exocytosis 41 . In our study, acetylated ATP synthase levels were greater in mitochondria from palmitate-treated INS-1E β cells than control cells. These data are consistent with findings showing that acetylation decreases ATP synthase catalytic activity. Indeed, ATP synthase contains multiple SIRT3-dependent reversible acetyl-lysines, which are altered in several conditions of metabolic stress, including exercise, calories restriction, fasting, and high-fat diet 42 . In particular, acetylation is accompanied by a decrease in the activity of this protein 43,44 .
Chronic exposure to FFAs is known to induce β cells apoptosis 8,9,45 . Voltage-dependent anion channels (VDAC) are considered key players in mitochondria-mediated apoptosis, but the molecular mechanisms by which VDAC influences cell death execution are not totally understood 46 . VDAC was shown to be co-translationally or post-translationally modified by phosphorylation or acetylation 21,46,47 but the relevance of these PTMs in the regulation of VDAC function remains unclear. Interestingly, it was reported that lipid accumulation triggers PTM of VDAC, causing early mitochondrial dysfunction and apoptosis 46,48 . Herein, we detected over-acetylation of VDAC1 by immunoprecipitation (IP) and VDAC2 by WB in mitochondrial preparations from palmitate-treated INS-1E β cells. The molecular consequence of this modification deserves further studies for better understanding the pathophysiological relevance of the various VDAC PTMs especially considering that they can behave as a lipid sensor 46,48 .
A major finding of the present study is that three proteins (DHE3, SODM and SRBP1) were over-acetylated in both INS-1E preparations and human pancreatic islets after exposure to palmitate. DHE3 catalyzes the oxidative deamination of endogenous glutamate, which is present at high concentrations in pancreatic β cells contributing to insulin secretion [49][50][51] . Post-translational DHE3 acetylation regulates the enzyme activity with progressive loss of activity as the degree of acetylation increases 52,53 . On the other hand, SIRT3 is able to deacetylate and therefore activate DHE3 52,54 . Therefore the impaired DHE3 activity can contribute to the detrimental effect of fatty acids on insulin secretion.
Pancreatic β cells are vulnerable to oxidative stress due to the relatively low expression of antioxidant enzymes 55,56 . Since palmitate can increase the intracellular production of ROS 57 , it can also cause β cell damage by enhancing oxidative stress. In this context, the role of mitochondrial SODM is crucial. This enzyme plays a pivotal role against oxidative stress since mitochondria dissipate over 90% of intracellular oxygen producing a large flux of ROS 58 . The genetic association of sod2 with T2D has been already reported 59,60 and the interest in the post-translational regulation of SODM is growing. As a result of palmitate treatment, we observed increased SODM acetylation both in INS-1E β cells mitochondria and human pancreatic islets. Acetylation inhibits SODM catalytic activity while, at the same time, palmitate increases NADH/NAD + ratio impairing SIRT3 activity, which deacetylates and hence activates SODM 58 . Therefore we speculate that the β cell oxidative stress may be also due to the enhanced acetylation by palmitate of key antioxidant enzymes.
In our study we also found over-acetylated proteins of the ER which is strictly associated with the mitochondria and can be present in mitochondrial enriched cellular preparations [30][31][32][33] . Among these, we identified the SRBP1, which is a member of the sterol regulatory element binding-proteins (SRBPs), transcriptional factors that control cholesterol and lipid metabolism 61 . SRBP1 mRNA expression is induced in palmitate-treated INS-1E β cells 40 . In our study, we detected increased acetylation levels of SRBP1 in palmitate-treated INS-1E β cells and human pancreatic islets. Moreover acetylation was directly detectable during identification of the corresponding spot by MS (data not shown) in INS-1E β cells and human pancreatic islets. A previous work showed that the activity of SRBPs is regulated by PTMs including acetylation 62 . Of interest, these enzymes are important targets of SIRT1 63 . In particular, acetylation levels were found to be increased in response to feeding and decreased by fasting and therefore it was proposed that deacetylation by SIRT1 inhibits SRBP 63,64 . Indeed, SIRT1 decreases the stability of SRBP through deacetylation, which in turn affects the level of occupancy at the lipogenic target gene promoters 63 . On the other hand, acetylation enhances SRBP activity by increasing its stability and recruitment to its lipogenic target promoters, which favours lipogenesis and contribute to fatty liver disease (hepatosteatosis), a condition associated with obesity and insulin resistance 63 . Therefore, since acetylation of SRBPs is considered a  Table 2. Each pair of connected points represents one experiment using a sample of human pancreatic islets obtained from a single subject. The p-values were determined by paired t test. (b) Enlarged images of the acetylated spots detected in human pancreatic islets. CTRL: control samples; PALM: palmitate-exposed samples. crucial event for the pathogenesis of metabolic disorders 64-66 its deacetylation through SIRT1 targeting could be useful in the treatment of T2D.
In conclusion, this study demonstrates that prolonged exposure to palmitate induces changes of mitochondrial protein lysine acetylation, which could play a role in β cell dysfunction through several possible mechanisms some of which are summarized in Fig. 4. Therefore, the balance between acetylation and deacetylation may be critical in pancreatic β cell adaptation to lipotoxicity. Whereas additional studies are needed to unveil the fine molecular mechanisms that regulate lysine acetylation/deacetylation pathways the present results provide novel data that could pave the way for future targeted therapeutic strategies modulating acetyl transferase and deacetylase activity.

Cell culture. INS-1E β cells. INS-1E
β cells (kindly provided by Prof. C. Wollheim Centre Médical Universitaire, Geneva, Switzerland) were handled as previously described 19 . For the purposes of the present study, batches of cells were seeded in 24 well plates at density of 2 × 10 5 cells/well and cultured up to 24 h in RPMI medium containing 0.5 mM palmitate, 1% fatty acid free-bovine serum albumin (BSA) and 1% FBS. Palmitate was dissolved in ethanol 90% (v/v). The control was represented by cells exposed to the ethanol 90% (v/v).The procedures for the preparation of the palmitate solution have been previously described in detail 67 .
In additional experiments, INS-1E β cells were exposed for 24 h to control or 0.5 mM palmitate containing medium, with or without the addition of deacetylase inhibitors (combined 500 nM trichostatin A and 10 mM nicotinamide) 68 . For insulin release studies, the cells were first exposed to RPMI medium without glucose for 2 h. Then, after a 30 min washing period with a Krebs' solution containing 2.5 mM glucose, cells were exposed to 2.5 or 16.7 mM glucose for 30 min. At the end, the supernatant was collected and stored at −20 °C until insulin measurement. Total insulin content was extracted by an ethanolic acid solution. Insulin was quantified by the High Range Rat Insulin ELISA kit following the manufacturer's protocol (Mercodia AB, Uppsala, Sweden). INS-1E β  The increased acetylation of DHE3 causes a progressive loss of activity impairing insulin secretion. SODM plays a pivotal role against oxidative stress and acetylation inhibits its catalytic activity contributing to oxidative stress. The hyper-acetylation of SRBP1 enhances its activity by increasing its stability and its recruitment to lipogenic promoters, which could promote lipogenesis.
cell viability was monitored using the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide-formazan (MTT) assay. Briefly, a solution containing 5 mg/mL of tetrazolium bromide salt (Sigma-Aldrich) dissolved in phosphate buffer saline (PBS) was added to cells exposed to different incubation conditions to obtain a final concentration of 5 μg/mL. Then, plates were kept in an incubator at 37 °C and 5% CO 2 for 3 h. At the end, supernatants were aspirated and the formed salt dissolved in DMSO. After 20 minutes of shaking, plates were read at 570 nm with baseline at 650 nm. Viability was calculated as percentage of absorbance of treated versus untreated cells.
Human pancreatic islets. Islets were isolated by enzymatic digestion and gradient purification 9,69 from the pancreas of 5 multiorgan donors who gave informed consent (age: 79 ± 6 years; 2 M/3 F; BMI: 25.8 ± 2.5 kg/m 2 ), with the approval of the local Ethics Committee (Comitato Etico di Area Vasta Nord Ovest, Azienda Ospedaliera Universitaria Pisana, Pisa) since the glands were not suitable for clinical purposes. The average β cell proportion in islets prepared in our laboratory over the past 10 years is 50 ± 17% (mean ± standard deviation of 163 preparations). In 2 of the 5 islet preparations used in the present study, the percentage of β cells was evaluated: insulin immunofluorescence staining after islet dispersion 70 identified 40 and 48% β cells. Following isolation, islets were cultured in M199 medium containing 5.5 mM glucose, supplemented with 10% (v/v) serum, 100 U/mL penicillin, 100 μg/mL streptomycin, 50 μg/mL gentamicin, and 750 ng/mL amphotericin B (Sigma-Aldrich, St. Louis, MO). Approximately 1,000 islets were then incubated with or without 0.5 mM palmitate (Sigma-Aldrich) for 24 h, as described above and previously detailed 6,17,69 . All experiments were performed in accordance with relevant guidelines and regulations.
Mitochondria enrichment. Mitochondrial enriched fractions were obtained from INS-1E β cells using differential centrifugations, according to the protocols implemented within the MtHPP network 29 and the quality of enrichment was checked as previously described 19 . Briefly, cells (about 20 millions) were collected from the flasks, suspended in the isolation buffer (IB) (20 mM TRIS pH 7.4, 250 mM sucrose, 2 mM EDTA, 50 mM NaF, 2 mM Na 3 VO 4 , 1 μL/10 6 cells protease inhibitors, 1 µM trichostatin A, 10 mM nicotinamide), transferred into a pre-cooled potter and homogenized. Consecutive centrifugation cycles were used to isolate mitochondria from unbroken cells, nuclei and cytosol. Finally, a volume of IB, depending on the size of the pellet, was added to resuspend the mitochondrial fraction. Protein amounts were determined by the Bio-Rad DC-protein assay (Bio-Rad; Hercules, CA, USA) using BSA as calibration standard. Mass spectrometry analysis of our mitochondrial preparations showed protein yields of 45 ± 7% and 20 ± 5% respect to total proteins, for mitochondrial and ER proteins, respectively. These yields were in accordance to previous results 19,29 . The integrity of mitochondria was assayed by measuring the activity of cytochrome oxidase (Cytochrome Oxidase Activity Colorimetric Assay Kit; BioVision, CA, USA). WB analysis was performed to confirm the enrichment of the mitochondria using a cocktail of 4 different antibodies each targeting a specific cellular marker (1:250 dilution; Abcam, UK) as previously described 19 . Mitochondria preparations which were isolated as described above from seven independent experiments consisting of palmitate-treated and -untreated cells (control) were pooled to perform proteomics and WB analysis.
Proteins extraction from human pancreatic islets. Isolated islets were collected and washed twice with PBS (37 °C). Cells were suspended in the rehydration solution (7 M urea, 2 M thiourea, 4% CHAPS, 60 mM dithiothreitol (DTT), 0.002% bromophenol blue) added with 50 mM NaF, 2 mM Na 3 VO 4 , 1 μL/10 6 cells protease inhibitors, 1 µM trichostatin A, 10 mM nicotinamide. After stirring and sonication (4 seconds, 5 times) cells were allowed to rehydrate for 1 h at room temperature (RT) with occasional stirring. Thereafter, the solution was centrifuged at 17,000 g for 5 min at RT. Protein concentration of the resulting supernatant was determined using the Bio-Rad RC/DC-protein assay (Bio-Rad). BSA was used as a standard. The global protein content was analyzed with focus on mitochondrial proteins. . The reaction was quenched by adding 10% formic acid (FA). The resulting peptides were purified using reversed-phase purification cartridges (Oasis ® ; Waters, MA, USA). Purified peptides were eluted using a solution with 80% ACN and 0.1% FA. The elutes were dried in SpeedVac Concentrator (Thermo Scientific; Bremen, Germany). Then, the dried elutes were redissolved in IP buffer (50 mM Tris, pH 8.0, 1 mM EDTA, 100 mM NaCl) and incubated with beads (Protein A/G PLUS-Agarose; Santa Cruz Biotechnology Inc., Dallas, TX, USA) and anti-acetylated lysine antibody (1:12.5 w/w; mouse monoclonal; Cell Signaling Technology Inc.) for 18 hrs at 4 °C on a rotation wheel. The immunoprecipitated peptides were washed 3 times with IP buffer, eluted using 1% TFA, 40% acetonitrile in H 2 O and dried in SpeedVac Concentrator. After re-solubilisation in 0.1% FA, the enriched acetylated peptides were analyzed by LC-MS as already reported 19 . Raw data were processed with PEAKS 7.5 to provide protein IDs and PTMs mapping.

Electrophoresis and Western Blot. Sodium Dodecyl Sulphate-Polyacrylamide Gel
Data analysis. Images from WB of 1-D gel electrophoresis were analyzed using the Image Quant-TL (GE Health Care). The antigen-specific bands and the total proteins after RuBP staining were quantified. The optical density was measured for all the immunoreactive bands of the whole lane. In order to normalize the volume of each band, the ratio of optical density of the antigen-specific bands with those of total proteins, obtained from RuBP staining, was calculated. The results were expressed as a ratio of optical density. The significance of the differences of global acetylation was calculated by unpaired Student's t-test making a comparison between the two classes. The analysis of 2-DE images was performed using the Same Spot (v4.1, TotalLab; Newcastle Upon Tyne, UK) software. The volume of each spot obtained after the detection of immunocomplexes was normalized by the total protein content obtained with RuBP staining of membranes. A comparison between cells treated with and without palmitate was performed. The significance of the differences of normalized volume for each acetylated spot was calculated by paired Student's t-test. Therefore the protein spots of interest were selected and preparative gels for the identification were prepared. Data Availability. All data generated or analyzed during this study are included in this published article (and its Supplementary Information files). The data not shown are available from the corresponding author.