Sera of atopic dogs and controls
To examine IgE reactivity in atopic dogs, we collected surplus sera from 179 dogs that were clinically diagnosed with atopic dermatitis from among dogs visiting the Fujimura Animal Hospital (Osaka, Japan) based on the criteria proposed by Willemse  and Prelaud et al. , which is clinically compatible with the symptoms of atopic dermatitis in humans . Twenty samples collected from laboratory dogs were used as negative controls. The dogs were housed indoors as experimental laboratory animals and did not have prior exposure to fish antigens. None of the laboratory dogs exhibited signs of atopic dermatitis. All sera samples were stored at −80 °C until used. Verbal informed consent was provided by the dog owners. All experimental procedures were performed in accordance with Japanese law and were approved by the animal care and user committee of Azabu University.
Food elimination and oral provocation tests
The elimination diet trials were performed for 6–8 weeks using commercial hydrolyzed diets on 179 dogs that exhibited clinical symptoms of atopic dermatitis. The dog owners fed commercial hydrolyzed diets that contained ingredients that the dog had no prior exposure to. The ingredients were confirmed based on information provided by the dog owners. When clinical signs were resolved (disappearance of pruritus and hair regrowth), food provocation was performed for one week by administering the original diet. The dogs were checked for clinical symptoms by a veterinary physician every two weeks or once per month. Among the 144 dogs that exhibited clinical improvement in response to the food elimination trials, 31 dogs were observed to exhibit cod reactivity, as evidenced by the oral provocation tests. These dogs were selected based on the diet history, admitted to the animal hospital, and were challenged with various food products containing cod components, including grilled cod meat and cod-containing dog foods (Select Protein Cod and Rice Dry, Royal Canin, Aimargues, France). The provocation trials were conducted by one of the authors (MF), following which the complete resolution of the clinical signs occurred after conducting the food elimination tests. The cod provocation tests were discontinued immediately upon the relapse of the clinical signs that included vomiting, diarrhea, erythema, pruritic, urticaria, and conjunctival hyperemia. We obtained verbal consent for the food elimination trials and provocation tests from the dog owners, as previously described . Supplementary Figure 1, Additional file 1 presents the flow diagram for study participants.
Preparation of crude cod extracts
Pacific cod was purchased from a fish market in Japan. The fresh, raw meat of four fishes (500 µg) was homogenized in 500 µl of phosphate-buffered saline (PBS, 10 mM pH 7.2) overnight at 4 °C under rotating conditions. After centrifugation at 21500 x g for 5 min at 4 °C, the supernatant was collected, and the protein was quantified using the BCA protein assay (Bio-Rad, Hercules, CA, USA).
Purification of parvalbumin, collagen, and tropomyosin
Fish parvalbumin  and collagen  were purified as described previously. Tropomyosin was purified from the freeze-dried powder using Bailey’s method with slight modifications . Briefly, freeze-dried fish powder was stirred in a beaker with 75 ml extraction buffer containing 15 mM Tris HCl pH 7.6 (Sigma Aldrich, St Louis, MO, USA), 1 M KCl (Kanto Kagaku, Tokyo, Japan), and 2 mM dithiothreitol (Sigma Aldrich) overnight at 4 °C. The extract was collected by centrifugation at 5400 x g for 10 min at 4 °C. The pH of the supernatant was adjusted to 4.5 using 1 N HCl to precipitate tropomyosin, and the precipitate was collected by centrifugation at 5400 x g for 10 min at 4 °C. The isoelectric precipitation process was repeated once and the precipitate was dissolved in the extraction buffer. The supernatant obtained after the extraction was collected by centrifugation and fractionated using ammonium sulfate to a concentration of 50%. The sample precipitated by ammonium sulfate fractionation was dissolved and dialyzed against PBS. The obtained protein extracts were identified by sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE).
SDS-PAGE and immunoblotting
SDS-PAGE was performed according to the method described by Laemmli . Precision plus protein standards (Bio-Rad, Hercules, CA, USA) were used as molecular-mass markers. The crude cod extracts were separated electrophoretically using 5–20% gradient polyacrylamide gels, and the proteins were visualized either by Coomassie brilliant blue R250 (Bio-Rad) staining or by transferring onto polyvinylidene difluoride membranes (GE Healthcare, Chicago, IL, USA). Immunoblotting was performed as described previously . IgE from patient dog sera were used as primary antibodies, and were diluted at a 1:10 ratio in Tris buffered saline containing 0.1% Tween-20 and 5% nonfat dried milk (pH 7.4). Mouse monoclonal anti-dog IgE antibodies (0.5 μg/ml, clone D9) were used as secondary antibodies . Horseradish peroxidase-conjugated goat anti-mouse IgG (0.05 μg/ml) (Bio-Rad Laboratories) were used as tertiary antibodies. An enhanced chemiluminescence immunoblotting detection reagent (GE Healthcare) and the ImageQuant LAS 4000mini (GE Healthcare) were used for detection and visualization, respectively.
Liquid chromatography-tandem mass spectrometry (LC-MS/MS)
For LC- MS/MS, the protein bands formed after the immunoblotting experiment were excised, and in-gel digestion was performed using 0.5 mg N-tosyl-L-phenylalanine chloromethyl ketone-treated trypsin (Promega, Madison, Wisconsin) at 37 °C for 16 h. The tryptic digests were acidified using formic acid (pH < 2.0) and centrifuged at 21500 x g for 15 min. The supernatants were analyzed using a high-performance liquid chromatograph (Advance System; AMR, Tokyo, Japan) connected to an electrospray ionization triple quadrupole mass spectrometer (4000 QTRAP; AB Sciex, Framingham, MA, USA). The extracts were injected into a reversed-phase column (electrospray ionization column [octa decyl silyl]; particle inner diameter, 75 mm; length, 100 mm; diameter 3 mm; LC Assist, Tokyo, Japan) that was eluted with a 5–45% acetonitrile gradient containing 0.1% formic acid for 60 min at 300 nl/min. Ionization was performed at an ion-spray voltage of 2000 V and a capillary temperature of 200 °C. The mass spectrometer was operated in the positive ion mode between 450–1200 m/z. The MS/MS spectra were obtained in the enhanced production scan mode, and two higher-intensity peaks in each mass spectrometry scan were selected for collision-induced dissociation.
The MS/MS data were used to search the entries under the Liza aurata category in the UniProt database with the Mascot peptide search engine. An MS tolerance of 1.0 Da for precursor ion and an MS/MS tolerance of 0.8 Da were set as the windows of processing parameters for matching the peptide mass values.
Fluorometric enzyme-linked immunosorbent assay (ELISA) for the detection of allergen-specific serum IgE
The levels of specific IgE to cod crude extracts and individual purified allergens were measured using fluorometric ELISA, as previously described . A microplate (NUNC Immuno Plate Maxisorp F96; Nalge Nunc International, Roskilde, Denmark) was coated with the crude cod extracts (10 µg/ml) or the purified allergens (parvalbumin, collagen, or tropomyosin; 1 µg/ml) and maintained overnight at 4 ºC. After washing, the plate was incubated with diluted sera (1:10) in PBS with 10% (v/v) fetal calf serum and 0.05% (v/v) Tween 20 at room temperature for 3 h. The plate was then washed and incubated with mouse monoclonal anti-dog IgE antibody (0.5 µg/ml, clone D9) overnight at 4 ºC . After washing with PBS containing 0.1% (v/v) Tween 20 (PBS-T), the plate was incubated with biotinylated rat monoclonal anti-mouse IgG1 (Zymed Laboratories, San Francisco, CA, USA) for 1 h at room temperature. After washing, the plate was incubated with b-D-galactosidase-conjugated streptavidin (Zymed Laboratories) for 1 h at room temperature. After the final wash, the plate was incubated with 0.1 mM 4-methylumbelliferyl- b-D-galactopyranoside (Sigma Aldrich) for 2 h at 37 ºC. The enzymatic reaction was terminated by adding 0.1 M glycine-NaOH (pH 10.2). The fluorescence intensity was measured in terms of fluorescence units (FU) using a microplate fluorescence reader (Fluoroskan; Flow Laboratories, McLean, VA, USA). The absorbance was measured at 355 nm using a 460 nm reference filter. All the washing steps were performed thrice for 5 min in PBS-T. The cutoff value was determined as the average + three standard deviations (SDs) of FU in the sera from the 20 dogs that served as negative controls in the analysis of the levels of specific IgE to cod crude extracts. ELISA for parvalbumin or collagen from salmon (Atlantic salmon; Salmo salar), sardine (Japanese pilchard; Sardinops melanostictus), and mackerel (Chub mackerel; Scomber japonicus) were performed using individual sera, which ensured that a substantial quantity of sera was available (See Supplementary Figure 1, Additional File 1). All tests were performed in triplicate.