Adaptive effect of sericin on hepatic mitochondrial conformation through its regulation of apoptosis, autophagy and energy maintenance: a proteomics approach

We recently demonstrated that in addition to its protective effect on pancreatic and adrenal biosynthesis, antioxidant properties of sericin decrease blood cholesterol levels and improve the liver mitochondrial architecture. However, little is known about the detailed mechanisms underlying these effects. Using proteomics and electron microscopy, we identified mitochondrial proteins that play important roles in the preservation of the mitochondrial ultrastructure and cholesterol-lowering properties of sericin. Our results showed that sericin maintains the mitochondrial architecture during conditions of high blood cholesterol by regulating apoptotic (NADH-ubiquinone oxidoreductase 75 kDa subunit) and autophagic (mitochondrial elongation factor Tu and prohibitin-2) proteins as well as energy maintenance proteins [haloacid dehalogenase-like hydrolase domain-containing protein 3, succinate dehydrogenase (ubiquinone) flavoprotein subunit, ATP synthase-α subunit precursor, enoyl-CoA hydratase domain-containing protein 3 and electron transfer flavoprotein subunit-α]. Sericin also exerts anti-oxidative properties via aconitate hydratase and Chain A, crystal structure of rat carnitine palmitoyltrasferase 2 proteins. Together, these activities may reduce hepatocytic triglyceride deposition, thereby decreasing steatosis, as demonstrated by the modulatory effects on ornithine aminotransferase, mitochondrial aspartate aminotransferase, acyl-CoA synthase, hydroxyacyl-CoA dehydrogenase and D-beta-hydroxybutyrate dehydrogenase. Sericin activity further balanced nitrogenous waste detoxification, characterised by carbamoyl-phosphate synthase (ammonia), aldehyde dehydrogenase and uricase, or folate biosynthesis via sarcosine dehydrogenase and dimethyl glycine dehydrogenase. These results suggest that sericin maintains the hepatic mitochondrial architecture through apoptotic, autophagic, energy maintenance and anti-oxidative mitochondrial proteins for alleviating hepatic steatosis and promoting liver function under conditions of hypercholesterolaemia.

Immunoelectron microscopy and Western blot results. Immunoelectron microscopy and Western blot were performed for validating the results of the proteomics analysis. Up-and downregulated mitochondrial proteins, NDUFS1, TUFM and HDHD3, were chosen for verifying their expression levels at each stage of liver mitochondrial degeneration fine morphology: intact (Fig. 3a,f,l and q), swelling ( Fig. 3b,g,m and r), spheroid (Fig. 3c,h,n and r), clumping (Fig. 3d,i,o and t) and necrosis or ghost (Fig. 3e,j,p and u) mitochondria, as described previously 1 . Swelling mitochondria were represented by singular or multiple distensions in the intercellular matrix in association with cristae loss and intact double membranes. Mitochondrial spheroidicity was categorised by a high level of swelling with severe distension of the intercellular matrix, formation of multiple cysts and loss of double membranes. Clumping mitochondria were indicated by a complete loss of cristae and double membranes accompanied by electron-dense material clumping of the degenerative or fragment mitochondria. Lastly, ghost mitochondria were the final step of mitochondrial alterations in which the remaining mitochondrial fragments presented as a fine granular substance of electron-dense material with or without its boundary.
In addition, to relate the function of these selected proteins to their roles in the pathogenesis of hepatic steatosis and dysmorphic mitochondrial under conditions of hypercholesterolaemia, the expression of apoptotic protein (caspase-3), mitochondrial fission protein (DRP1), mitochondrial fusion protein (OPA1) and autopagic protein SCIenTIFIC RepoRts | (2018) 8:14943 | DOI:10.1038/s41598-018-33372-4 (LC3) was examined. The results indicated that the expression of NDUFS (Fig. 3a-k) and TUFM (Fig. 4a-k) in the hepatic mitochondria from sericin-treated rats was significantly lower than that the hepatic mitochondria from non-treated rats, particularly in ghost cells. Conversely, the expression of HDHD3 in sericin-treated rats was significantly upregulated in most stages of mitochondrial alterations compared with that in non-treated rats ( Fig. 5a-k). In addition, Western blot showed a high TUFM protein level in non-treated rat compared with that in sericin-treated rat (Fig. 4w). Therefore, the protein expression level determined by gold labelling was in agreement with the results from our proteomics analysis, thereby reflecting consistency in our proteomics results.
Caspase-3 ( Fig. 3l-v) and LC3 (Fig. 4l-v) gold labelling was decreased in ghost mitochondria from sericin-treated rats compared with that in ghost mitochondria from non-treated rats. Considering fission and fusion phenomena, dynamic processes involved in mitochondria homeostasis and DRP1 and OPA1 gold labelling were evaluated in the present study. The results revealed that compared with non-treated rats, almost all mitochondrial stages in sericin-treated rats had lower expression of DRP1 ( Fig. 6a-k). In contrast to the expression of OPA1, the total expression in sericin-treated rats was significantly increased compared with that in non-treated rats, particularly in intact and ghost mitochondria ( Fig. 7a-k). The present study also revealed correlations contributing to the role of sericin in mitochondrial alterations under high blood cholesterol levels using Spearman's rho correlation test. Low-energy mitochondria had an increased apoptotic and fission protein labelling levels, as characterised by a negative correlation between HDHD3 and caspase-3 (r = −0.456, p = 0.029) and between HDHD3 and DRP1 (r = −0.554, p = 0.006) (Fig. 8a-b). A positive correlation was also found between DRP1 and high-cholesterol-fed rats was compared and represented by percentage area fraction or number of lipid/ cell, respectively (c,f). Compared with sericin-treated rats, micro-and macrovesicular steatosis were highly presented in the periportal area of non-treated rats (H&E staining). Electron microscopy demonstrated that lipid accumulation (*) in hepatocytes in non-treated rats (d) was higher than that in sericin-treated rats (e). The bar graph indicates blood cholesterol and HDL levels in sericin-treated and non-treated rats with hypercholesterolaemic; data is represented by mean ± SEM; *p < 0.05, **p < 0.01, and ***p < 0.001 (f). Comparison of electron micrograph of the mitochondria architecture between sericin-treated (g) and nontreated (h) rats. caspase-3 (r = 0.423, p = 0.044) (Fig. 8d). Finally, although NDUFS1 showed a positive correlation with caspase-3 (r = 0.558, p = 0.006), a negative correlation was observed between NDUFS1 and OPA1 (r = −0.546, p = 0.007) (Fig. 8c,e).

Discussion
In this study, we used proteomics for characterising and comparing liver mitochondrial proteins from sericin-treated and non-treated rats with hypercholesterolaemia. The highly significant proteins were categorised into five groups depending on their functions and association with the pathogenesis of steatosis: (i) apoptosis and autophagy, (ii) energy maintenance, (iii) anti-oxidative stress, (iv) liver injury (e.g. fat deposition) and (v) liver function improvement (e.g. nitrogen waste detoxicfication and folate-binding protein). These findings are consistent with those of our recent study, which reported that lipid dysregulation leads to mitochondrial alteration and dysfunction and high apoptosis levels in association with defective energy synthesis 1 .
Apoptosis contributes to different stages of mitochondrial biosynthesis, fission (apoptotic induction process) or fusion (maintenance of mitochondrial function) in relation to dysmorphic mitochondria: (i) swelling, (ii) spheroidicity, (iii) collapse or clumping and (iv) necrosis 1 . During apoptosis, disturbed mitochondrial function is mediated by mitochondrial NDUFS1, a critical caspase substrate in the mitochondria 12 . Cleavage of NDUFS1 by caspase activation due to permeabilisation of the outer mitochondrial membrane and cytochrome release is mainly associated with apoptosis. High NDUFS1 levels were found in the liver mitochondria from non-treated rats. Our proteomics results also showed that an increase in the autophagy regulation proteins TUFM and PHB2 was observed in non-treated rats. In agreement with the results of gold labelling, TUFM was higher in the ghost mitochondria of non-treated rats than in that of sericin-treated rats. High expression of TUFM induces autophagy 13 . PHB2 is an inner mitochondrial membrane mitophagy receptor that is imperative for targeting mitochondria for autophagic degradation 14 . This may be associated with the high prevalence of dysmorphic mitochondria in non-treated rats due to the central role autophagy plays in mitochondrial homeostasis to remove unwanted or damaged mitochondria 15 . Unfortunately, instead of upregulated LC3 in the mitochondria of non-treated rats, there was no difference in LC3 gold labelling between sericin-treated and non-treated rats.
The main function of mitochondria is being the powerhouse of the cell. This study provided evidence of high expression of some mitochondrial energy maintenance proteins in sericin-treated rats, such as HDHD3, SDHA, ATPSA, ECHDC3 and ETFA [16][17][18][19][20] . In addition, it is well known that impairment of fatty acid oxidation or energy production in the mitochondria is significantly associated with the pathogenesis of non-alcoholic steatohepatitis (NASH) 21 . This indicates that sericin alleviates the severity of hepatic steatosis by preserving the generation of mitochondrial energy and improves the mitochondrial architecture by modulating DRP1 and caspase-3 activity.
The present study demonstrated that sericin induces an antioxidative environment characterised by upregulation of several mitochondrial proteins, such as mitochondrial ACO2 and CPT-II. ACO2 is a catalytic enzyme in the tricarboxylic acid cycle 22 , which plays a role in oxidative stress tolerance, leading to cell death prevention 23 , whereas CPT-II controls the long-chain fatty acid beta-oxidation level, thereby contributing to appropriate oxidative stress and inflammatory inductions 24,25 . In agreement with our previous report indicating anti-oxidative properties of sericin in the heart and liver mitochondria, as shown by oxidative stress markers (e.g. superoxide dismutase, MDA and NRF-2) 1,2 , the present proteomics study confirmed that sericin preserves liver mitochondria from oxidative stress by maintaining ACO2 and CPT-II functions.
Moreover, abnormal mitochondria cause an accumulation of OAT, leading to cytosolic protein degradation 26 . In agreement with liver damage-related protein, GOT2, a transamination catalyst for kynurenic acid formation, is responsible for metabolic exchange between the mitochondria and cytosol 27 . Elevated GOT2 levels may be associated with some pathologies in the liver, particularly ischaemic necrosis 28 and liver injuries 29 . The present study indicated that in high-cholesterol-fed rats, a decrease in hepatic injury from the hypercholesterolaemic stage might be attributable to the absence of GOT2.
Steatosis, a hepatocytic fatty change, is an imbalanced fat metabolism in the liver characterised by over-accumulation of triglyceride fat, reflecting impairment in synthesis or elimination of lipids. The present study demonstrated that sericin modulates mitochondrial ACSM1, which plays a major role in fatty acid biosynthesis 30 . Overexpression of ACSM1 results in the deposition of large amounts of triglycerides in hepatocytes 31 and an increase in free fatty acid transportation into the cell 32 . The presence of ACSM1 in non-treated rats indicated high deposition of triglycerides in hepatocytes and led to higher severity of steatosis than that in sericin-treated rats. Moreover, mitochondrial HADH, a mitochondrial protein that enhances lipid deposition 33 , was significantly expressed in non-treated rats. In addition, BDH1 was significantly higher in non-treated rats than in sericin-treated rats. BDH1 plays a role in the regulation of ketone body biosynthesis, particularly in adipose tissue development 34 . High BDH1 levels may induce lipid deposition, as demonstrated in non-treated rats.
Mitochondrial CPS, a catalytic enzyme in the urea cycle, is an important modulator of ammonia detoxification 35,36 that is activated by N-acetyl-L-glutamate, a first step activator of the urea cycle in mitochondria 37 . High production of CPS is responsible for high ammonia levels and leads to increased urea synthesis. Disruption of the urea cycle can cause steatosis and focal hepatic necrosis 38 . Increases or decreases in the levels of enzymes essential for the mitochondrial urea cycle reflect mitochondrial impairment in NASH and steatosis 39 . Our results suggest that sericin balances urea synthesis in high-cholesterol-fed rats as high CPS levels were observed in non-treated rats in association with the high severity of hepatic steatosis.
Mitochondrial AD plays a vital role in the detoxification of the ethanol metabolite, acetaldehyde 40 . High AD levels seem to be beneficial to ethanol intoxication. However, a significant difference in AD activity is not frequently observed between alcoholic and non-alcoholic liver, but its activity is highly altered following liver damage 41 . Non-treated rats showed high AD levels compared with sericin-treated rats, which possibly contributed to the high severity of hepatic steatosis in relation to the presence of some toxic metabolites.
Mitochondrial UOX, a homotetrameric enzyme, converts uric acid to allantoin, which is a more soluble, simple form for excretion 42 . Endogenous uric acid is mainly produced in the liver 43 . High blood cholesterol level is closely related to uric acid levels 44 under oxidative stress conditions 45 . Non-treated rats had higher uricase levels than sericin-treated rats. This indicated that non-treated rats had high uric acid levels that may lead to hepatic steatosis and an enhanced oxidative stress environment.
There is strong evidence to show that folate depletion is usually found in non-alcoholic fatty liver or other conditions of liver damage because of perturbation of hepatic metabolism 46,47 . Mitochondrial SARDH and DMGDH are the major folate-binding proteins that contribute to sarcosine metabolism 48 . Increased SARDH causes folate deficiency 49 . Therefore, folate depletion may have caused the presence of SARDH and DMGDH in non-treated rats but not in the sericin-treated group.
In summary, the mitochondrial ultrastructure in the liver was improved by sericin activity under hypercholesterolaemic conditions. Dysmorphic mitochondria were decreased in sericin-treated rats because of sericin's modulatory effects on apoptotic (NDUFS1) and autophagic (TUFM and PHB2) mitochondrial proteins; this is in agreement with the caspase-3, OPA1 and DRP1 levels in the hepatic mitochondria. Sericin not only maintains energy production through enhancement of HDHD3, SDHA, ATPSA, ECHDC3 and ETFA function but also demonstrates anti-oxidative activity via ACO2 and CPT-II proteins in the hepatocytes. Our proteomics study also revealed that lipid deposition and hepatocyte injury in the liver were decreased in association with the expression of OAT, GOT2, ACSM1, HADH and BDH1 proteins. These mechanisms ameliorated hepatic steatosis and improved other liver functions, characterised by expression of CPS, AD, UOX, SARDH and DMGDH. Our findings may contribute to the understanding of mechanisms by which sericin recovers hypercholesterolaemia, based on liver mitochondria conformation and hepatic steatosis, and highlights sericin as a leading candidate and as a blood cholesterol-lowering food additive.

Materials and Methods
Ethics statement. This  Sericin extraction. Silkworm (Bombyx mori) cocoons were obtained from Chul Thai Silk Co., Ltd.
(Petchaboon Province, Thailand). Sericin was extracted by heat induction to achieve appropriate purity amino acids, using methods described previously 50 . After autoclaving the fresh cocoon shells in distilled water at 120 °C for 60 min, the supernatant was filtered, frozen and lyophilised.  Animal experiment protocol. Induction of cholesterolaemia and sericin treatment. Eight-week-old female Wistar rats (200-300 g) were obtained from the National Laboratory Animal Center, Mahidol University (NLAC-MU). To induce hypercholesterolaemia, they were fed 6% cholesterol-coated diets (Perfect Companion Ltd., Thailand) ad libitum for 6 weeks prior to the experiment 1 . Next, rats were separated into two groups according to the treatment, i.e. sericin-treated and non-treated rats (six rats per group), and were administered 1000 mg/kg  of sericin or distilled water, respectively, by oral gavage for 4 weeks. All rats were humanely sacrificed by an overdose of isoflurane inhalation.
Specimen collection. Blood specimens were collected by cardiac puncture and centrifuged at 1500 × g for 15 min. Serum (six samples/group) was separated; blood lipid (cholesterol and HDL) profiles were determined by the Quality Control Division, NLAC-MU. To exclude technical artefacts from mitochondrial architecture studies, the specimen must be maintained under ice-cold condition throughout the experiment. Liver specimens were immediately collected and stored in ice-cold homogenate buffer (0.32 M sucrose, 1 mM EDTA, and 10 mM Tris-HCl; pH 7.4) when performing a standard protocol for mitochondrial extraction, as described previously 1,2 . Liver specimens were also collected and separated for histopathology and electron microscopy analysis and preserved in 10% neutral buffer formalin for 48 h and in 2.5% glutaraldehyde in 0.1 M sucrose phosphate buffer (SPB), pH 7.4, respectively.
Mitochondrial extraction. After specimen collection, mitochondrial extraction was performed within 1-2 h to prevent cellular damage. Liver specimens from each group were pooled (six livers/group), weighed (15.24 g of liver from sericin-treated rats and 13.75 g of liver from non-treated rats), chopped and washed in homogenate buffer. Liver specimens were homogenised in a glass Potter-Elvehjem tissue grinder with an appropriate volume of the homogenate buffer (4 mL of homogenate buffer/1 g of liver specimen). During this step, several up and down strokes were performed using a motor-driven Teflon pestle at 600 rpm. Next, homogenates were centrifuged at 1000 × g at 4 °C for 5 min. The supernatants were collected and centrifuged at 15,000 × g at 4 °C for 2 min. The mitochondrial pellets were collected and washed several times in homogenate buffer. The pellets were resuspended in ice-cold final equilibrated buffer [250 mM sucrose, 5 mM KH 2 PO 4 , 10 mM Tris-HCl, and 2 mg/mL bovine serum albumin (BSA); pH 7.2], and 200 μL of resuspended pellet was fixed in 2.5% glutaraldehyde in 0.1 M SPB for electron microscopy analysis. The mitochondrial protein content was measured by protein assay (Bio-Rad ® ) using a spectrophotometer (NanoDrop-1000, Thermo Scientific).
Proteomic studies. Two-dimensional gel electrophoresis. Each mitochondrial protein was loaded onto Immobiline ™ Drystrip immobilised pH gradient (IPG) strips (13 cm, pH 3-10 NL, GE healthcare ® ) using the passive rehydration method for 12 h. Isoelectric focussing was performed using Ettan IPGphor (GE healthcare ® ) at 20 °C, and the gel strips were equilibrated with the equilibration buffer twice for 15 min each, as described previously 51 . The second dimension run was performed as follows. Each equilibrated gel strip was loaded onto a 12.5% sodium dodecyl sulphate polyacrylamide gel, sealed with the agarose sealing solution and supplied with current, using a vertical slab gel electrophoresis unit (SE 600 Chroma Hoefer ® ), of 15 mA/gel for 30 min and Image analysis. Gels were scanned using a Typhoon Trio Variable Mode Imager (GE Healthcare ® ) under a 532-nm green laser and 555-nm long pass emission filter. Image Master 2D software (GE healthcare ® ) was used for detecting, quantifying and comparing the spots depending up on their OD values. Spots were considered to be significant at p < 0.05 using analysis of variance (ANOVA). Twenty-five significant spots were chosen for protein identification.
In-gel tryptic digestion. Selected spots (Fig. 1)  Morphological studies. Histopathology analysis. Histopathology was used for measuring the severity of hepatic steatosis from hypercholesterolaemia. Fixed liver specimens (six livers/group) were used for standard tissue processing and stained with haematoxylin and eosin. The severity of hepatic steatosis between sericin-treated and non-treated rats was examined under a light microscope, and the percentage of fat deposition/high-power field was quantified using image J programme (NIH, USA), as described previously 1,2,51-53 . Briefly, colour images (10 images/liver) were acquired using light microscope (BX41, Olympus ® , Japan) and digital camera (DP20, Olympus ® , Japan) at 400× magnification. The images were then transformed to a grey scale. The fat deposit area was localised and measured into a percentage area fraction using threshold adjustment. Conventional electron microscopy. Electron microscopy was used for examining the ultrastructure of hepatic steatosis. Secondary fixation of the six liver specimens from each group was performed using 1% osmium tetroxide, dehydrated in graded ethanol, infiltrated in a series of LR white resin (EMS ® , USA), embedded in pure LR white (EMS ® , USA), polymerised at 60 °C for 48 h, cut into 100-nm-thick sections and stained with lead citrate and uranyl acetate. The liver ultrastructure was examined under transmission electron microscope (model HT7700, Hitashi, Japan). Fat droplets in hepatocytes were counted (50 cells per group) and compared between sericin-treated and non-treated rats.
Immunogold labelling technique. Immunogold labelling was performed for confirming and validating the proteomics study results on pathogenesis of hypercholesterolaemia-induced hepatic steatosis and dysmorphic mitochondria. In this study, several markers (MyBioSource, USA) were used as primary antibodies: apoptotic markers, mouse polyclonal NDUFS1 and caspase-3; autophagic markers, mouse polyclonal TUFM and rabbit polyclonal anti-LC3; fusion and fission markers, rabbit polyclonal OPA1 and DRP1 and energy marker, rabbit polyclonal anti-HDHD3. The mitochondrial pellet from the pooled liver extract in each group was secondary fixed, and tissue processing was performed as described previously. Sections were blocked using 50 mM glycine in phosphate-buffered saline (PBS) followed by 5% BSA (EMS, USA) in PBS for 30 min each. The sections were incubated with 1:50 diluted primary antibodies for 1 h prior to applying goat anti-rabbit or mouse IgG conjugated with ultra-small (3-5 nm) or 10-nm gold particles (EMS ® , USA). Between each step, sections were washed using 0.1% BSA in PBS several times. To improve the contrast of the gold particle labelling, a silver enhancement kit (Aurion R-Gent SE-EM kit, EMS, USA) was used after rigorously washing the sections with distilled water. Finally, the sections were stained with lead citrate and uranyl acetate prior to transmission electron microscope examination. The number of labelled gold particles was counted for each ultra-characteristic stage of liver mitochondrial degeneration: intact, swelling, spheroidicity, collapse or clumping and necrosis. At least 100 mitochondria (20 mitochondria/stage) were evaluated per group.

Western blot analysis.
To confirm the proteomic results, Western blot was used for determining the level of one selected marker, TUFM. Fifty micrograms of mitochondrial protein from treated and non-treated rats was loaded onto 10% sodium dodecyl sulphate polyacrylamide gel electrophoresis and transferred to a transfer membrane. The membrane was blocked with 10% skin milk, incubated in mouse polyclonal anti-mitochondrial TUFM overnight and incubated in HRP-conjugated secondary antibody for 1 h. The signal was visualised using a DAB chromogen. Statistics. Statistical analysis was performed using GraphPad Prism ® version 5. Differences between quantitative parameters were compared using Student's t-test and presented by mean ± SEM. Bivariate correlation was examined using Spearman's test. The level of statistical significance was set at p < 0.05.

Data Availability
Datasets generated and/or analysed during the current study are available from the corresponding author upon request.