Farmed Atlantic cod (Gadus morhua) with an average initial weight of 534 g were obtained from AKVAFORSK Sunndalsøra and kept in four separate sea net pens at Averøy, Norway and fed either a standard fish meal diet(two net pens) or a diet in which 24% of the protein was replaced with SBM (two net pens). The average water temperature during the 84 days of feeding was 10.0 °C, with a maximum temperature of 13.7 °C at the start of the trial, and ending at 7.0 °C at the day of sampling. The fish were fed daily for 1 h at dawn. Three randomly selected fish were sampled from each pen (six fish fed FM and six fish fed SBM in total). All fish were handled in accordance with “Regulation on Animal Experimentation” (Norwegian ministry of agriculture). Further details on fish and rearing condition are given in (Refstie et al., 2006a).

Three diets were manufactured by high-pressure moist extrusion technology at Nutreco Technology Centre in Stavanger, Norway. Formulation of the two diets used in our studies, are given in Table 1 . The FM diet contained 100% crude protein (CP) from low-temperature dried FM (LT-FM, Norse LT-94) obtained from Vedde Herring Oil Factory (Egersund, Norway), whereas in the SBM diet 24% of crude protein was replaced by dehulled and extracted SBM obtained from Hamlet Protein (Horsens, Denmark). Both diets were formulated to contain 55% CP and 20% lipid (DM basis), and to be iso-energetic on a gross energy basis. The diets were supplemented with D, l-methionine to contain similar amounts of (calculated) methionine. Both diets contained 100 mg yttrium oxide (Y₂O₃, Sigma, St. Louis, Mo, USA) kg^− 1 dry mix as an inert marker to permit apparent digestibility measurements. Further details on formulation and chemical composition of the diets are given in (Refstie et al., 2006a).Table 1

Diet formulations (g/kg)

Atlantic cod were anaesthetized in tricain methanesulfonate (MS222, Argent Chemical Laboratories Inc., Redmont, WA, USA), and subsequently killed with a sharp blow to the head. The whole gastro-intestinal tract was then immediately dissected out, cut open and intestinal contents carefully and thoroughly removed. Subsequently the intestine was divided into the following five segments. The intestine between the ring of pyloric caeca and the valve separating the most distal segment was defined as the mid-intestine (MI), and divided into four equal sections (MI1, MI2, MI3 and MI4). The section between the valve separating the most distal segment and anus was defined as the distal chamber (DC). Samples of about 300 mg were taken from MI4, and then quickly rinsed in sterile phosphate-buffered saline (PBS), before immediate transfer to ten times the volume of RNAlater® (Ambion, Inc., Austin, USA), and storage at − 20 °C until further analysis.

Total RNA was isolated from the 12 (six fish fed FM and six fish fed SBM in total) MI4 tissue samples by use of Trizol (Invitrogen) according to the manufacturer's protocol. High RNA integrity was verified using a 2100 Bioanalyzer from Agilent Technologies, Inc., Palo Alto, USA.

For the suppression subtractive hybridization, only one fish from each feeding treatment was used in an effort to obtain as many different clones as possible. cDNA was synthesized from 1 μg of total RNA from one fish fed FM and one fish fed SBM by use of the BD Clontech SMART^™ PCR cDNA synthesis kit (BD Clontech, Cat. no. 634902/K1052-1), part VII, according to the manufacturer's protocol (VIIA1-B13). This kit was used to produce high quality cDNA from a small sample. Part VII and VIII of this protocol are designed for synthesizing cDNA for applications such as PCR-Select^™ cDNA subtraction or Virtual Northern blots. Briefly, 1 μg of total RNA was combined with 12 μM of the primers 3′ BD SMART CDS primer II A (a modified oligo dT primer) and BD SMART II A oligonucleotide, respectively, and incubated at 70 °C for 2 min, before addition of first-strand buffer, DTT, dNTPs, and BD Powerscript Reverse Transcriptase according to the manufacturer's protocol. The RNA was then reversely transcribed for 1 h at 42 °C to synthesize single stranded (ss) cDNA. After this, the ss cDNA was diluted in 40 μl of TE buffer (10 mM Tris (pH 7.6), 1 mM EDTA), and heated at 72 °C for 7 min. Subsequently, 1 μl of the diluted ss cDNA was used in long distance (LD) PCR, while the remaining (49 μl) was stored at − 20 °C until further use. LD PCR was performed according to the ma-nufacturer's protocol, and an optimal number of cycles determined for each sample. Each step of the protocol was monitored by gel electrophoresis of aliquots of samples. The LD PCR was terminated by the addition of 5 μl of 20X EDTA/glycogen Mix (0.2 M EDTA; 1 mg/ml glycogen) from the Clontech PCR-Select^™ cDNA subtraction kit (BD Clontech, Cat. no. 637401). Instead of performing column chromatography according to part VIII of the SMART cDNA synthesis kit protocol, we performed precipitation of the LD PCR (approximately 86 μl after gel electrophoresis) according to the PCR-Select^™ cDNA subtraction kit proto-col, part IV D7.

The terminated LD PCR synthesized by the SMART cDNA synthesis kit was precipitated according to the PCR-Select^™ cDNA subtraction kit protocol (BD Clontech, Cat. no. 637401), part IV-D7 by adding 100 μl of phenol:chloroform:isoamyl alcohol (25:24:1), vortexing thoroughly, and centrifugation at 13,000 ×g for 10 min at room temperature. The steps following precipitation were also performed exactly according to the protocol. Briefly, the cDNAs were subsequently digested with Rsa I, precipitated again, and dissolved in H₂O, and stored at − 20 °C until further use. For the SSH procedure, cDNA from FM-fed fish was defined as the “driver” cDNA, and cDNA from SBM-fed fish defined as the “tester” cDNA. 1 μl of Rsa I-digested tester cDNA (from SBM-fed fish) was then ligated to either adaptor 1 or adaptor 2R at 16 °C overnight (ON), and then terminated by addition of EDTA/glycogen and heating at 72 °C for 5 min. Subsequently the tester cDNA with adaptor 1 was hybridized with an excess of the driver cDNA (from FM-fed fish) in tube 1, while tester cDNA with adaptor 2R was hybridized with an excess of driver cDNA in tube 2 at 68 °C for 8 h (first hybridization), immediately followed by the second hybridization in which tube 1 and 2 from the first hybridization was combined with addition of fresh denatured driver cDNA (subtraction kit protocol, part IV-H), and incubated further at 68 °C ON. The following day 200 μl of dilution buffer was added to the sample, followed by heating to 68 °C for 7 min. The diluted sample was subsequently stored at − 20 °C until further analysis. Differentially expressed cDNAs were selectively amplified by two subsequent PCR steps, the first using a PCR primer 1 complementary to both adaptors, and the second using two nested primers complementary to each of the two adaptors (nested PCR primer 1 and nested PCR primer 2R) respectively, according to the manufacturer's protocol (section IV, I-1 to I-17). In this way the final PCR mixture was enriched for differentially expressed cDNAs. A subtraction efficiency test was performed as a control according to the manufacturer's protocol.

The enriched PCR product mixture from SSH was ligated into the pCR^®2.1-TOPO^® vector using the TOPO TA Cloning^® kit from Invitrogen life technologies, California, USA, according to the manufacturer's protocol and transformed into chemically competent E. coli. Optimalization of the amount of PCR product ligated with pCR^®2.1-TOPO^® vector was performed to ensure an optimal number of transformed colonies. Transformation of E. coli was also performed according to the manufacturer's protocol. A total number of 192 clones were picked at random and PCR was performed with vector specific primers, M13F and M13R to ensure a single PCR product per clone, and thus one clone picked per well. PCR was performed as follows: 30 cycles at 94 °C for 30 s, 60 °C for 30 s, and 72 °C for 1 min.

Subsequently the clones were sequenced, using M13R as sequencing primer. Sequences were subsequently trimmed to remove primer- and vector-sequences and 159 successful clones were then subjected to translated BLAST search (blastx or tblastx) at NCBI BLAST. 33 clones failed to pass quality control after trimming, because of sequences either being too short, or consisting of only poly A- or vector-sequences.

On the basis of their possible relevance to changes in amino acid and lipid digestion, some clones were then selected for real-time PCR (qPCR). In addition, some randomly selected clones were included. The selected clones were sequences showing similarity to the following genes: fatty acid binding protein (FABP, clone ID GH4A-F142), putative transmembrane 4 superfamily member protein (TM4, clone ID GH4A-F103), polypeptide N-acetylgalactosaminyltransferase (ppGaNTase, clone ID GH4A-F107), Aminopeptidase M (Alanyl aminopeptidase)(CD13)/Aminopeptidase N (APN, clone ID GH4A-F77), transcobalamin I precursor (TCI, clone ID GH4A-F84), Sec61-alpha (SEC61, clone ID GH4A-F137), F-box protein 44 (F-BOX, clone ID GH4A-F138), glutathione peroxidase (GPx, clone ID GH4A-F127), peroxiredoxin 4 (Prx4, clone ID GH4A-F167), cytochrome P450 3A40 (CYP3A40, clone ID GH4A-F166), ras-related nuclear protein (Ran, clone ID GH4A-F53), and 14-3-3B2 protein (14-3-3, clone ID GH4A-F159).

Prior to reverse transcription, total RNA from all samples intended for real-time PCR were subjected to DNase treatment using a TURBO DNA-free^™ kit in accordance with the manufacturer's recommendations (Ambion, Inc., Austin, USA).

Real-time PCR was performed on samples from all twelve individuals sampled (6 fish fed FM and 6 fish fed SBM). Each sample was reversely transcribed in duplicates by PowerScript^™ Reverse Transcriptase (BD Biosciences, Franklin Lakes, NJ, USA) (RT) from 740 ng of total RNA isolated from MI4 using a mixture of 250 ng oligo (dT) (Invitrogen Ltd, Paisley, UK) and 25 ng random primers (Invitrogen Ltd, Paisley, UK) according to the manufacturer's protocol (PowerScript^™ Reverse Transcriptase kit, BD Biosciences, Franklin Lakes, NJ, USA).

Real-time PCR amplifications were performed to examine the relative expression of selected genes (Table 2, Table 4 ) in MI4 in a LightCycler 2.0 (Roche Diagnostics, Mannheim, Germany). Duplicates of 2.5 μl cDNA from each reverse transcription reaction were used as PCR templates. PCR reactions were performed in a total volume of 10 μl using the LightCycler FastStart DNA MasterPLUS SYBR GREEN I kit (Roche Diagnostics, Mannheim, Germany) and a modified protocol with 4.5 μl PCR-grade water, 0.5 μl of each PCR primer (10 μM) and 2 μl master mix. The following LightCycler thermocycling program was used: denaturation program (10 min at 95 °C), amplification and quantification program repeated 40 times (10 s at 95 °C, 15 s at the appropriate annealing temperature for the gene specific primers (Table 2) and 10 s at 72 °C with a single fluorescence measurement), melting curve program (60 °C to 99 °C with a heating rate of 0.1 °C/s) and cooling program down to 40 °C. Fluorescence was monitored (excitation at 470 nm and emission at 530 nm) at the end of the annealing phase (the LightCycler F1 channel). For determination of the crossing point (CP) the “second derivative maximum method” measuring maximum increase rate of newly synthesized DNA per cycle was used on the basis of LightCycler software 4.0 (Roche Diagnostics, Mannheim, Germany). To confirm amplification specificity the PCR products from each primer pair were subjected to melting curve analysis and subsequent agarose gel electrophoresis.Table 2

Primer sequences used for real-time PCR

The relative expression ratio of target mRNAs was calculated using the LightCycler software 4.0 (Roche Diagnostics, Mannheim, Germany). The following calculation method was employed: calibrator-normalized relative quantification using PCR efficiency correction based on a linear regression fit. Elongation factor 1 alpha (ELF1a) and 18S rRNA were used as reference genes and cDNA from a FM fed fish was used as the calibrator. Relative standard curves were generated on the basis of cDNA from the calibrator sample diluted in 5-fold or 10-fold dilution steps to cover the expected detection range of target genes and the reference genes. Triplicates of the respective dilution steps were used to determine the standard curves. Results are presented as a normalized calibrated ratio. Hence the concentration ratios for each sample are calibrated to the calibrator sample so that the quantification results are reported as a normalized ratio with the calibrator sample as the denominator:

Relative mRNA level = ratio of sample (target / reference) / ratio of calibrator (target / reference).

Elongation factor 1 alpha, 18s rRNA and beta-actin were evaluated for use as reference genes. All samples from MI4 were subjected to PCR amplification in the LightCycler with primers specific for the 3 reference genes (Table 2) as described above. By use of the geNorm software (Vandesompele et al., 2002), the most stable reference genes were identified. GeNorm calculates the stability value “M” of the reference genes by comparing the variation in expression for all the genes. Based on the M value, ELF1a and 18S rRNA were selected as the reference genes in this study.

All tests were carried out two-tailed, with a significance level of 5%. The Shapiro–Wilk W test was used to test conformity with the normal distribution. Differences in the relative expression of selected genes were calculated using the pair wise, student's t-test. Prior to the statistical analyses the relative expression data from each gene was calibrated to the mean of the FM group of the same gene. Hence the mean normalized calibrated ratio of the FM group is 1 for all the genes. Results are presented as mean values with 95% confidence intervals. The statistical analyses were performed using JMP 5.0.1 software package (SAS Institute Inc. Cary, NC, USA). Calculated means ± SD for each selected clone is given in Table 4. Fig. 2A and B were made on the basis of statistical reports in the JMP 5.0.1 software package.

Sequencing of 192 randomly picked clones from the Atlantic cod SSH library showed that 33 clones were either too short or consisted of only vector- or poly A-sequences. The remaining 159 clones were subjected to translated blast (NCBI blastx). Of the 159 clones, two virtually translated sequences showed low similarity to other protein sequences, and was shown to be 18S rRNA by nucleotide blast search (NCBI blastn). Another 44 clones gave no or low similarity to other known protein sequences. Virtual translation of the remaining 113 clone sequences showed protein similarities to known sequences from other fish species or mammals (Table 3 ). Blastx E values were between 8.00E− 04 and 3.00E− 91 as shown in Table 3. The 157 blasted clones were submitted to GenBank, and GenBank accession nos. for the clones are given in Table 3. A grouped overview of the sequenced clones is showed in Fig. 1 . The major identified groups were structural proteins, membrane associated proteins, enzymes, and proteins involved in protein metabolism. The majority of these clones (88 out of 113, shaded grey in Table 3) represented new partial sequences that, to our knowledge, have not previously been described nor annotated for Atlantic cod in GenBank. Predictions of similarity were based on NCBI BLAST searches.Table 3

Sequenced clones from the suppression subtracted library

Mean normalized calibrated ratios from real-time PCR of 12 clones, selected based on their similarity to genes involved in processes such as protein- and lipid metabolism, growth and antioxidant functions, showed that expression of 4 out of the 12 clones tested were significantly up regulated (P  < 0.05) in intestine from cod fed SBM compared to intestines from cod fed FM. The 4 clones were: Aminopeptidase M (Alanyl aminopeptidase)(CD13)/Aminopeptidase N (APN, clone ID GH4A-F77, GenBank acc.no. EB677077), transcobalamin I precursor (TCI, clone ID GH4A-F84, GenBank acc.no. EB677078), cytochrome p450 3A40 (CYP3A40, clone ID GH4A-F166, GenBank acc.no. EB677101) and ras-related nuclear protein (RAN, clone ID GH4A-F53, GenBank acc.no. EB677072) (Table 4 , Fig. 2Aa, c, d, and Bb). Mean normalized calibrated ratios of TCI, APN, CYP3A40, and RAN were up regulated by factors of 1.48, 2.06, 1.95, and 1.49, respectively, in fish fed SBM compared to fish fed FM (Table 4). Fatty acid binding protein (FABP, clone ID GH4A-F142, GenBank acc.no. EB677063) showed a trend towards up regulation in SBM fed fish compared to FM fed fish (P  < 0.10) (Table 4 and Fig. 2Ab).Table 4

Quantified gene expression of selected clones

No significant difference in expression was observed for the seven other clones studied: putative transmembrane 4 superfamily member protein (TM4, clone ID GH4A-F103, GenBank acc.no. EB677017), polypeptide N-acetylgalactosaminyltransferase (ppGaNTase, clone ID GH4A-F107, GenBank acc.no. EB677083), Sec61-alpha (SEC61, clone ID GH4A-F137, GenBank acc.no. EB677091), F-box protein 44 (F-BOX, clone ID GH4A-F138, GenBank acc.no. EB677092), glutathione peroxidase (GPx, clone ID GH4A-F127, GenBank acc. no. EB677081), peroxiredoxin 4 (Prx4, clone ID GH4A-F167, GenBank acc.no. EB677102), and 14-3-3B2 protein (14-3-3, clone ID GH4A-F159, GenBank acc.no. EB677100). The average relative expression ratios of all clones evaluated by qPCR, except for GPx, were numerically higher in the SBM group compared with the FM group. For GPx the average relative expression ratio was numerically lower in the SBM group compared to the FM group (Table 4).