1 SUPPLEMENTARY METHODS AND MATERIALS

2 Antibodies

3 The following antibodies were used: rabbit anti-PNPLA3 (Sigma-Aldrich), mouse anti-

4 PNPLA2 (Sigma-Aldrich), mouse anti-LPL (Sigma-Aldrich), mouse anti-LIPC (Sigma-

5 Aldrich), rabbit anti-LIPE (Sigma-Aldrich), mouse anti-V5 (Invitrogen), rabbit anti-AKT

6 (Cell Signaling), rabbit anti-phospho (Ser473) AKT (Cell Signaling) and rabbit anti-

7 calnexin (Sigma-Aldrich).

9 Real-time PCR assay

10 Real time qPCR was used to assess PNPLA3 mRNA expression in 48 human tissues using

11 TissueScan Human Normal cDNA Array (Origene). To assess mRNA expression in

12 cultured cells, total RNA was isolated from primary human stellate cells, primary human

13 hepatocytes, HEPG2 and CACO-2 cells using the RNeasy Mini Kit (Qiagen). First-strand

14 cDNAs were synthesized from 2 μg RNA using a reverse-transcription kit (Applied

15 Biosystems). The mRNA expression levels were determined in cDNA samples using the

16 delta-delta CT method and normalized to β-actin. TaqMan probes and master mix (Life

17 Technologies) were used in a total volume of 20 µl per reaction. Real-time PCR assay

18 was performed in a 7900HT Fast Real-Time PCR System (Life Technologies).

19

20 Immunoblot analysis

21 After washing twice in PBS, cells were harvested and incubated on ice in M-PER®

22 Mammalian Protein Extraction Reagent (Pierce, Thermo Fisher Scientific, Rockford, IL)

23 containing complete protease inhibitor cocktail (Roche). Immunoblot analysis was

24 performed according to standard procedures. Bands were visualized by Chemidoc XRS

25 System and Image Lab Software (Biorad).

1 Lipid droplet analysis

2 The total area of Oil Red O-stained lipid droplets was determined as described

3 previously (1). Pictures were obtained using Axio KS 400 Imaging System and

4 AxioVision 4.8 Software (Zeiss) at 100X magnification.

7 Transfection with siRNA and vectors

8 ON-TARGET plus PNPLA3 siRNA was obtained from Thermo Scientific. Transfection was

9 performed according to the manufacturer’s instructions. Scrambled siRNA was used as

10 control.

11 Wild type and 148M mutant PNPLA3 cDNA were cloned in aV5-6His tagged pcDNA3.1

12 vector (Life Technologies) and transiently transfected into cells using Turbofect

13 liposomes (Fermentas). Transfection efficiency was assessed using an anti-V5 antibody.

14 An empty vector was used as a negative control.

15

16 HEPG2/LX-2 co-culture

17 HEPG2 and LX-2 cells were plated in the upper and lower chamber, respectively, of a 24

18 well trans-well plate with 0.4 μm pore membrane. LX-2 in the lower chamber were

19 transfected with wild type or mutant PNPLA3 and 36 hours after transfection HEPG2 in

20 the upper chamber were incubated with 1μCi/ml of [3H]-glycerol plus 300 μM palmitic

21 acid and 0,3% albumin. After 8 hours, cells were harvested and radiolabeled

22 triglycerides were isolated by thin layer chromatography and measured by scintillation

23 counting.

24

25
1 PNPLA3 effect on triglyceride metabolism analysis in LX-2 cells

2 LX-2 cells were incubated with 1 μCi/ml of [ 14C]-palmitic acid plus 300 μM palmitic acid

3 and 0,3% albumin transfected with wild type or mutant PNPLA3 and next incubated

4 with 1 μCi/ml of [14C]-palmitic acid plus 300 μM palmitic acid and 0,3% albumin. After

5 48 hours, cells and media were harvested and intracellular triglycerides and released

6 palmitc acid were measured by scintillation counting after thin layer chromatography

7 lipid fractionation.

9 Retinyl-palmitate hydrolysis assessment McA-RH 7777 cells

10 McA-RH 7777 stably overexpressing the wild type or the mutant PNPLA3 were

11 harvested and broken in non-denaturing lysis buffer (phosphate buffer saline ph 7.4

12 plus 1% triton) and cell lysates were incubated 15 minutes at 37 degrees with retinyl-

13 [14C]-palmitate. Released [14C]-palmitic acid was next measured by scintillation counting.

14 An empty vector was used a s a control.

15

16

17 1 Nicoletti, A., Kaveri, S., Caligiuri, G., Bariéty, J. and Hansson, G.K. (1998)

18 Immunoglobulin treatment reduces atherosclerosis in apo E knockout mice. J Clin Invest, 102,

19 910-918.

20

21

22

23

24

25
1

3 SUPPLEMENTARY FIGURE LEGENDS

4 Supplementary figure 1. PNPLA3 expression pattern in human tissues. Human

5 PNPLA3 mRNA expression was measured using cDNA from 48 different human tissues

6 by quantitative real-time PCR. The tissue with the highest CT value was assigned the

7 value of 1.

9 Supplementary figure 2. PNPLA2, LIPE, LIPC and LPL proteins show low or no

10 expression in primary human hepatic stellate cells. Immunoblot showing PNPLA3,

11 PNPLA2, LIPE, LIPC and LPL protein amount in primary hepatic stellate cells and human

12 hepatocytes. Blots were obtained from a single PAGE run.

13

14 Supplementary figure 3. PNPLA3 is upregulated by insulin in primary hepatic

15 stellate cells. Immunoblot shows PNPLA3 protein amount in primary hepatic stellate

16 cells before and after incubation in medium containing insulin (1 nM) for 24 hours.

17 Calnexin was used as loading control. Insulin signaling activation was confirmed by cell

18 lysate immunoblot with anti-phospho-AKT (Ser473) and anti-AKT antibodies at time 0

19 (-) and after 10 min (+) of insulin treatment.

20

21 Supplementary figure 4. Insulin-mediated PNPLA3 upregulation reduces lipid

22 droplet accumulation in LX-2 cells. (A) Lipid droplet content visualized by Oil Red O

23 (ORO) staining in LX-2 cells incubated with or without retinol (10 µM) and palmitic acid

24 (300 µM) for 48 hours, transfected with PNPLA3 siRNA or control siRNA for 12 hours

25 and incubated with or without insulin (1 nM) for a further 48 hours. (B) The ORO-
1 stained area was quantified by BioPix. Data are mean ± s.d. of 3 experiments. *P < 0.05.

2 (C) Immunoblot showing PNPLA3 and PNPLA2 expression in LX-2 cells under

3 conditions described in A. Calnexin was used as a loading control. Scale bars, 10 µm.

5 Supplementary figure 5. Overexpression of wild type and mutant PNPLA3 does not

6 affect intracellular triglyceride content in LX-2 cells. (A) Intracellular [14C]-

7 tripalmitin after incubation with radiolabeled palmitic acid and subsequent

8 overespression of wild type and mutant PNPLA3. Empty vector (EV) was used as

9 negative control. (B) Immunoblot showing transfection efficiency.

10

11 Supplementary figure 6. Overexpression of wild type but not mutant PNPLA3

12 reduces lipid droplet content in LX-2 cells. (A) Lipid droplets visualized by ORO-

13 staining in LX-2 overexpressing V5-tagged 148I or 148M PNPLA3 and incubated with

14 retinol-palmitic acid for 48 hours. Empty vector (EV) was used as negative control. (B)

15 ORO-stained area quantified by BioPix. (C) Immunoblot showing transfection efficiency.

16

17 Supplementary figure 7. Overexpression of wild type and mutant PNPLA3 in LX-2

18 cells does not affect intracellular lipid content of HEPG2. (A) Intracellular 3H-

19 tripalmitin in HEPG2 incubated with 3H-glycerol and palmitic acid after co-culture with

20 LX-2 overexpressing the wild type or mutant PNPLA3. An empty vector (EV) was used as

21 a control. (B) Immunoblot showing transfection efficiency in LX-2 cells.

22

23 Supplementary figure 8 Stable overexpression of wild type and mutant PNPLA3

24 does not affect retiny-palmitate esterase activity in McA-RH 7777 cells. (A) [14C]-

25 palmitate production after incubation of retinyl-[14C]-palmitate with non-denatured cell

1 lysates from McA-RH 7777 stably overexpressing the wild type or the mutant PNPLA3

2 for 15 min. An empty vector (EV) was used as a control. (B) Immunoblot showing stable

3 transfection efficiency in McA-RH 7777 cells.

6 Supplementary table 1

PNPLA3 I148M genotypes

Pdo
Cohort characteristic II IM MM Padd Prec
m

N 58 65 23 - - -

0.65 0.59
Men (%) 76 82 74 0.599
2 8

0.50 0.53
Age (years) 47±13 51±12 47±12 0.148
0 1

0.08 0.14
2
BMI (Kg/m ) 26.7±4 27.0±3 25.6±3 0.156
4 6

0.88 0.69
Glucose (mg/dl) 95±20 101±34 97±16 0.624
6 9

45.6±1 0.04 0.00

RBP4 (μg/ml) § 50.9±11 52.6±14 0.303
7 0 9

0.60 0.32
NASH (%) 42 44 56 0.600
4 5

0.13 0.89
Diabetes or IFG (%) 26 43 36 0.055
5 9
1

2 Data were analyzed using linear regression analysis after adjusting for confounding factors

3 (age, gender, BMI, NASH, Diabetes or IFG). Values have been log-transformed before

4 entering the model if not normally distributed.

5 § LSD Post hoc test: P value= 0.782 II compared to IM; P value=0.023 II compared to MM

6 P value=0.012 IM compared to MM

7 Abbreviations: PNPLA3, patatin-like phospholipase domain-containing protein 3; Padd, P

8 value under an additive model; Prec, P value under a recessive model; Pdom, P value under a

9 dominant model; II, homozygous for the PNPLA3 148I allele; IM, heterozygous; MM,

10 homozygous for the PNPLA3 148M allele; N, number; BMI, body mass index; RBP4, retinol

11 binding protein 4; NASH, non-alcoholic steato-hepatitis; IFG, impaired fasting glucose; LSD,

12 Fisher's least significant difference.

13

14

15

16

17