Number and volume of peroxisomes are increased and peroxisomal ?-oxidation enzymes such as acyl-CoA oxidase (AOX) are induced when responsive animals are exposed to peroxisome proliferators. This phenomenon is known as peroxisome proliferation and it has been observed in several fish species after treatment with fibrate hypolipidemic drugs and a diverse array of environmental organic pollutants. In zebrafish Danio rerio, exposure to the organochlorine pesticide methoxychlor, the phthalate dibutylphthalate, the alkylphenol 4-tert-octylphenol and estrogens causes an increase in liver peroxisomal volume, surface and numerical densities together with a significant induction of AOX activity. Because of this significant response to pollutant exposure, peroxisome proliferation has been proposed in recent years as a novel early warning biomarker that may be useful in marine pollution monitoring.

Peroxisome proliferation is mediated by peroxisome proliferator-activated receptors (PPARs), which belong to the nuclear receptor superfamily. These receptors are ligand-activated transcription factors. Hitherto, three PPAR subtypes have been identified in several mammals and some amphibians: PPAR, PPAR? and PPAR?. In addition, PPARs have been found in several fish species such as salmon (Salmo salar), turbot (Scophthalmus maximus) and plaice (Pleuronectes platessa). In zebrafish (Danio rerio), 100 bp of the DNA-binding domain of each of the three subtypes, and two different isotypes of PPAR? have been cloned. In spite of our increased knowledge on PPAR genes in fish, tissue and cell distribution patterns of PPAR proteins and their specific functions are not well understood.

In a previous study, we have detected expression of the three subtypes of PPARs in various tissues of adult zebrafish by immunohistochemistry using commercially available antibodies against PPAR, PPAR? and PPAR?. PPAR was predominantly expressed in tissues that catabolize large amounts of fatty acids, such as liver parenchymal cells, proximal tubules of kidney, enterocytes and pancreas. PPAR? showed a widespread distribution and was expressed in liver, proximal and distal tubules and glomeruli of the kidney, pancreas, enterocytes and smooth muscle of the intestine, skin epithelium, lymphocytes and male and female gonads. PPAR? expression was weak and was only detected in pancreatic cells, intestine and gonads. Although zebrafish is a very convenient model species for laboratory studies, this tropical freshwater species cannot be used as sentinel for marine pollution monitoring studies. Therefore, these investigations were extended in the present study to other fish species such as the gray mullet Mugil cephalus, which is very abundant in estuaries and coastal areas in Europe.

Recently, liver peroxisomes of gray mullet have been characterized ultrastructurally, immunocytochemically and morphometrically. Furthermore, mullet liver peroxisomes exhibit variations depending on physiological and environmental conditions, thus constituting an appropriate model to study regulation of PPAR expression. The aim of the present study was to investigate immunohistochemically expression and distribution patterns of PPAR, PPAR? and PPAR? in the various cell types of gray mullet liver.

Three-month-old BALB/c mice were used as positive controls and were obtained from the central animal facility of the University of Heidelberg (for immunohistochemistry) and of the University of the Basque Country (for Western blotting).

Adult mullets of 27–38.5 cm length (3–8 g liver weight) were sampled from the Plentzia estuary (43°24?N, 2°56?W). For all studies, fish were anesthetized by immersion in a saturated solution of 3-aminobenzoic acid ethyl ester (Sigma, St. Louis, MO, USA) before being sacrificed.

Rabbit anti-mouse PPAR, PPAR? and PPAR? were raised by the Genosys Division of Sigma (St. Louis, MO, USA) against the 1–18, 1–14 and 1–16 amino acids, respectively, of the A domains of the respective proteins.

Livers from gray mullet and mouse were dissected and immediately homogenized in RIPA buffer (PBS buffer, 1% NP-40, 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulfate) using a Potter S homogenizer (Braun, Melsungen, Germany) in an ice-water cooled bath. The samples were ultracentrifuged in a Centrikon T-1065 with a TFT 65.13 rotor (Kontron Instruments, Munich, Germany) at 100,000g for 1 h at 4°C. The resulting pellet was removed and the supernatant was frozen at ?80°C. Total protein content was determined using a DC protein assay kit (BioRad, Hercules, CA, USA) which is based on the Lowry assay using ?-globulin as standard.

SDS-PAGE was performed using minigels consisting of a 4% stacking gel and 12% running gel in a Mini-Protean II electrophoresis cell (BioRad). Samples were prepared with a final protein concentration of 5 ?g/?l and 50 ?g of protein as well as molecular weight standards were loaded on the gels. After electrophoresis, separated proteins were transferred to nitrocellulose membranes (Amersham Pharmacia Biotech, Amersham, Buckinghamshire, UK). Membranes were washed in phosphate-buffered saline (PBS) containing 0.05% Tween 20 (PBST) and were blocked for 1 h at room temperature with 5% milk powder dissolved in PBST. Incubations with the diluted specific antibodies (1:200 in PBST containing 0.1% milk powder) were performed overnight at 4°C. Membranes were washed for 4×10 min in PBST and incubated with an HRP-coupled anti-rabbit secondary antibody (Sigma; diluted 1:10,000 in PBST containing 0.1% milk powder) for 1 h at room temperature. After 2×5 min and 3×10 min washes with PBST, HRP activity was visualized by chemiluminescence using an ECL kit (Amersham Biosciences).

Mullet liver was immersion fixed in Carnoy fixative for 1 h. The tissue was dehydrated through a graded series of ethanol, and embedded in paraffin. Sections (1.5 ?m thick) were cut using a 2000R microtome (Leica Microsystems, Wetzlar, Germany) and picked up onto SuperFrost Plus glass slides (Menzel-Glaser, Braunschweig, Germany). Mouse liver samples were fixed in paraformaldehyde and processed similarly for paraffin embedding.

Sections were deparaffinized in xylol and rehydrated through a graded series of ethanol ending in water. Sections were stained with or without antigen retrieval. For antigen retrieval, sections were treated using a microwave or with trypsin. Microwaving was performed three times for 5 min in a conventional household microwave (740 W; Sanyo, Chatsworth, CA, USA), using citrate buffer (0.01 M, pH 6.0), followed by 20 min cooling without changing the buffer. Digestion was performed with 0.001% (w/v) trypsin for 10 min or 0.01% (w/v) trypsin for 5 min (lyophilized porcine pancreas trypsin; Sigma) at 37°C. Only microwave pretreatment was applied to mouse liver. After antigen retrieval, sections were washed 2×5 min in TBS (0.05 M Tris–HCl, 0.15 M NaCl) and incubated in 3% hydrogen peroxide dissolved in methanol for blocking endogenous peroxidase. Sections were washed 3×10 min in TBS and then blocked with TNB blocking buffer, supplied in the Tyramide Signal Amplification TSA™-Indirect kit (NEN, Boston, MA, USA) for 1 h at room temperature. After 2×5 min washing steps using TBS, sections were treated overnight with the primary antibodies diluted in TNB solution containing 0.05% Tween. Dilutions of the primary antibodies were 1:200 for mouse and 1:500 for mullet. Samples were then washed 2×5 min with TBS and incubated with the anti-rabbit biotinylated secondary antibody for 1 h and additionally with peroxidase-coupled extravidin (Sigma), according to the manufacturer’s instructions. Staining reaction was performed using a NovaRed kit (Vector, Burlingame, CA, USA).

Controls consisted of sections that were incubated in blocking solution instead of the primary antibody solution.