- Research article
- Open Access
Genome wide response to dietary tetradecylthioacetic acid supplementation in the heart of Atlantic Salmon (Salmo salar L)
© Grammes et al.; licensee BioMed Central Ltd. 2012
- Received: 30 June 2011
- Accepted: 13 April 2012
- Published: 11 May 2012
Under-dimensioned hearts causing functional problems are associated with higher mortality rates in intensive Atlantic salmon aquaculture. Previous studies have indicated that tetradecylthioacetic acid (TTA) induces cardiac growth and also stimulates transcription of peroxisome proliferator activated receptors (PPAR) α and β in the Atlantic salmon heart. Since cardiac and transcriptional responses to feed are of high interest in aquaculture, the objective of this study was to characterize the transcriptional mechanisms induced by TTA in the heart of Atlantic salmon.
Atlantic salmon were kept at sea for 17 weeks. During the first 8 weeks the fish received a TTA supplemented diet. Using microarrays, profound transcriptional effects were observed in the heart at the end of the experiment, 9 weeks after the feeding of TTA stopped. Approximately 90% of the significant genes were expressed higher in the TTA group. Hypergeometric testing revealed the over-representation of 35 gene ontology terms in the TTA fed group. The GO terms were generally categorized into cardiac performance, lipid catabolism, glycolysis and TCA cycle.
Our results indicate that TTA has profound effects on cardiac performance based on results from microarray and qRT-PCR analysis. The gene expression profile favors a scenario of ”physiological”lright hypertrophy recognized by increased oxidative fatty acid metabolism, glycolysis and TCA cycle activity as well as cardiac growth and contractility in the heart ventricle. Increased cardiac efficiency may offer significant benefits in the demanding Aquaculture situations.
- Atlantic Salmon
- Cardiac Hypertrophy
- Differentially Express
- Heart Ventricle
- Cardiac Metabolism
High levels of dietary lipids are used in commercial Atlantic salmon diets to promote rapid growth and as a inexpensive source of energy. These high lipid levels may promote excess lipid deposition in the viscera and the muscle, thereby reducing the market quality of the fish. Thus, tetradecylthioacetic acid (TTA: CH3-(CH2)13-S−CH2-COOH) has been tested for aquaculture nutrition, initially to increase lipid catabolism and thereby reducing lipid deposition. However, beneficial effects on cardiac growth and disease resistance have also been addressed.
TTA is a modified fatty acid (FA) that possesses a sulfur atom in the β position. Like a normal FA, TTA can be converted to co-enzyme A thioester, but further catabolism by β-oxidation does not occur. This lack of metabolism is likely to determine the biological effects of TTA. Biological effects of TTA have been the focus of extensive research in rodents and also in humans. The most important findings from these experiments are that TTA increases the mitochondrial and peroxisomal β-oxidation and possesses hypolipidemic effects. In addition, TTA acts as an antioxidant in vivo and can modulate the inflammatory response (reviewed in). Cell culture experiments demonstrated that TTA can act as a ligand for all Peroxisome proliferator activated receptors (PPARs)[3, 4], which are ligand-activated transcription factors. Upon ligand activation PPARs heterodimerizes with retinoic-x-acid receptor (RXR) and have been shown to regulate the expression of genes involved in fatty acid metabolism, cell differentiation, development and inflammation (reviewed in). Arguably most of the biological effects of TTA are mediated through activation of PPARs.
Studies addressing the biological effects of TTA in Atlantic salmon have demonstrated that TTA increases β-oxidation in the liver and white muscle. Furthermore, TTA reduces the secretion of triacylglycerides from Atlantic salmon hepatocytes in vitro and has been shown to increase the expression of genes associated with fat metabolism in the liver and the heart ventricle[1, 8]. Previous results also suggested that TTA stimulates the transcription of PPARα and β in the heart[8, 9], thus indicating that TTA affects the metabolism in Atlantic salmon through activation of PPARs, similar to the mechanism known from rodents. Interestingly, this activation of PPARs may have been related to increased survival after a natural outbreak of a heart related viral disease in Atlantic salmon[8, 9]. In mammals, cardiac activation of PPARs has yielded substantial attention due to the fact that PPARs have been proven to be major regulators of cardiac metabolism[10–13]. In addition, PPAR agonists have been reported to exert beneficial effects by attenuating the pathogenesis of heart failure and atherosclerosis[14, 15].
Poor development of the outer muscle layer, atherosclerosis and metabolic dysfunction have been related to under-dimensioned hearts and reduced cardiac function in Atlantic salmon aquaculture, consequently resulting in increased mortality. Therefore, methods to improve cardiac metabolism and performance in fish are needed; something which has been sparsely studied. It appears that, similar to the mammalian heart, the oxidative cardiac metabolism in fish depends on the metabolism of fatty acids and glucose.
This study aims to characterize the cardiac transcriptional response of Atlantic salmon to a TTA supplemented diet. A feeding trial was conducted in sea, feeding a control and TTA supplemented diet during the first 8 weeks and only control diet for the subsequent 9 weeks of the experiment. Fish were sampled both at the end of the TTA feeding period (8.weeks) and at the end of the experiment (17.weeks). Our results show that administration of TTA to Atlantic salmon resulted in a marked change of cardiac gene expression. The expression profile suggests that TTA induces cardiac fatty acid oxidation, glycolysis, TCA cycle and contractility as well as cardiac growth.
Atlantic salmon that were fed with 0.25% TTA had significantly lower fat content in the muscle at the 8.weeks sampling point, and showed a tendency for increased mean relative heart weight (Table1). No significant effect of TTA on fish weight was detected. During the experiment none of the dietary groups showed higher mortality than the control group. In the heart ventricles, 120.2 μgTTA/gTissue was detected at the 8.weeks sampling point in the TTA group, while 1.2 μgTTA/gTissue was detected in the control group. Based on the TTA measurements of a group fed a higher (0.5% w/w) TTA diet from the same trial, we can assume that the TTA levels in the heart ventricles at the 17.weeks sampling point were no different to the control group (see Additional file1: Table S1).
Effect of TTA on Atlantic salmon production parameters
102 ± 5
1.2 ± 0.1
1.1 ± 0.1
1.2 ± 0.1
1.2 ± 0.1
1.41 ± 0.04
0.103 ± 0.002
0.74 ± 0.18
0.41 ± 0.27
2 ± 0.23
2 ± 0.21
Muscle fat content4[%]
4.2 a ±0.1
3.9 b ±0.1
RNA cardiac samples from six individual fish from each dietary group and sampling point were used in the microarray analysis, utilizing the Atlantic salmon SIQ2 microarray in a one-color setup, resulting in a total of 24 arrays. After normalization and filtering, 13166 probes (63%) were classified as present. To obtain a global overview of the general structure of the dataset we applied correspondence analysis (CA) as an explorative technique. The first 2 components of the CA are displayed, together explaining 67% of the total inertia of the different samples (Figure1). The analysis shows a clear distinction between the two sampling points at 8.weeks and 17.weeks. It further shows a relatively dense cluster at the 8.weeks point with minor separation between samples from the TTA and control group. Even though it was impossible to draw a straight line to separate between the TTA and control samples at the 17.weeks sampling point, we observed a clear tendency of separation.
To identify differentially expressed (DE) probes in the data set we used moderated t-statistics, comparing samples from TTA to control fed Atlantic salmon for each sampling point.
Sampling point: 8.weeks, end of TTA feeding
At the 8.weeks sampling point, five genes were found to be DE between the TTA and the control fed group (Figure2). The genes were: Ephrin-b2, arf gtpase-activating protein (git2), f-box only protein 11 (fbx11), angiopoietin-related protein 4 (ANGPTL4) and sodium- and chloride-dependent creatine transporter 1 (sc6a8). Ephrin-b2 was the only gene found to be down-regulated in the TTA fed group.
Ephrin-b2 (Efnb2) in mammals has been reported to be highly expressed in the heart and serves also as a marker for angiogenesis. The protein GIT-2 participates in pleiotropic cellular processes like cell migration and T-cell activation; however, a function affecting the structure of the cytoskeleton may be relevant in our study. Cardiomyocytes rely solely on the creatine transporter sc6a8 for the uptake of creatine from the plasma. Over-expression of the creatinfabe transporter in mice has been reported to correlate with the myocardial creatine content, but also to be associated with cardiac hypertrophy. Angiopoietin-related protein 4 possesses a role in regulating angiogenesis and is also known as a target gene for PPARs and acts as an important stimulator of lipid metabolism.
Sampling point: 17.weeks, 9 weeks post TTA feeding
1198 probes (930 genes) were found to be DE between the TTA and the control fed group at the 17.weeks sampling point. In order to facilitate a functional interpretation of the vast number of DE genes, we tested them for enrichment (over-representation) of GO terms from the category “biological process” , using conditional hypergeometric testing . To ensure that one gene was represented by a maximum of one probe , probes matching the same gene were collapsed prior to hypergeometric testing. This step reduced the total number of probes in the data set from 11143 to 7659 probes and the number of DE probes from 1198 to 930 probes (930 genes). From these 930 genes, 90% showed higher gene expression in the samples from TTA fed fish.
Conditional hypergeometric testing revealed significant over-representation of 36 GO terms. To simplify interpretation, significant GO terms were grouped into five categories according to their function in the heart (Table2). The grouping was further supported by a strong gene overlap between the different GO terms within the categories (Additional file2: Figure S1 and Additional file3: Table S2 ). Overall, the results from the enrichment analysis suggests an increased capacity of heart ventricles from TTA fed Atlantic salmon to catabolize lipids and glycogen. Further, an increased capacity for cardiac contractility and cardiac tissue morphogenesis is indicated.
Gene Ontology enrichment analysis for sampling point 17.weeks
Gene ontology term
carboxylic acid catabolic process
cellular lipid catabolic process
regulation of heart contraction
cardiac muscle tissue morphogenesis
Citrate cycle (TCA)
co-enzyme catabolic process
acetyl-CoA metabolic process
tricarboxylic acid cycle
generation of precursor metabolites and energy
cellular catabolic process
energy reserve metabolic process
carbohydrate catabolic process
cellular glucan metabolic process
alcohol catabolic process
hexose catabolic process
regulation of protein metabolic process
cell cycle arrest
chordate embryonic development
positive regulation of growth rate
urogenital system development
protein localization in mitochondrion
cellular macromolecule catabolic process
mRNA catabolic process
The group fed TTA showed an up-regulation in the expression of genes encoding contractile proteins like myosin heavy chain 6 (MYH6), myosin light chain (MYL9), cardiac myosin binding protein (MYBPC3), cardiac troponin (TNNT2), myomesin-1 (Myom1) and actin (ACTA1, ACTA2).
Moreover, the same group showed increased expression of the cardiac homeodomain factor Nkx2.5 and the iroquois-related homeobox factors 3 and 5 (irx3, irx5). Nkx2.5 and the iroquois transcription factors have been reported to control cardiac morphogenesis and growth[27, 28]. Furthermore, we observed increased expression of FK506 binding protein 1A and 1B (FKBP1A, FKBP1B) and Na +/K +-transporting ATPase subunit α 3 (ATP2A2), encoding an ion-pump responsible for establishing and maintaining the electrochemical gradients at the plasma membrane of the cardiomyocyte. Decreased amounts of this transporter were found in biopsies from humans suffering heart failure. FKBP1A and B are known to interact with intracellular calcium-release channels. In cardiomyocytes FKBP1B is a binding partner for the major Ca2 + release channel ryanodine receptor 2 (RyR2). RyR2 is required for the Ca2 + -induced Ca2 + release from the sacroplasmatic reticulum (SR) causing activation of the contractile proteins. Binding of FKBP1B to RyR2 results in channel closure. Mice deficient for FKBP1B showed no divergence in the normal cardiac phenotype under normal conditions but showed exercise-induced arrhythmias. In relation to Ca2 + signaling, we also found the Na+ /Ca2 + exchanger SLC8A1 to be up-regulated. We also found an increased expression of the SR Ca2 + ATPase 2 (ATP2A2 also known as SERCA2), encoding an SR calcium pump that is a key component of the cardiac excitation-contraction mechanism.
The Kv channel interacting protein 1 (Kcnip1) was found to be down-regulated. The protein Kcnip1 is an integral part of the multimeric Kv4 channel complex, and important for modulating the K-flux across this channel by causing a shortening of the cardiac action potential. Prolongation of the cardiac action potential on the other hand, potentially caused by decreased Kcnip1 expression, is associated to cardiac hypertrophy.
In summary, the results suggest an increased cardiac hypertrophy together with increased potential for cardiac contractility, as indicated by the higher transcription of the various ion channels/pumps and contractile proteins.
Nearly all of the genes in this group were up-regulated (Figure3). We found up-regulation of the mitochondrial trifunctional protein HADHA, the mitochondrial fatty acid transporter carnitine palmitoyltransferase (Cpt1a), lipoprotein lipase (Lpl), the mitochondrial acyl-CoA dehydrogenases (ACADS, ACADV and ACADSB) and of peroxisomal multi-functional enzyme type 2 (Hsd17b4). We also observed an up-regulation of malonyl-CoA decarboxylase (MLYCD). Malonyl-CoA is a potent inhibitor of CPT1 and thus crucial in regulating the transport of fatty acids into the mitochondria for catabolism. Malonyl-CoA decarboxylase has been reported to function as a positive regulator of cardiac fatty acid oxidation by decreasing the levels of the CPT1 inhibitor malonyl-CoA. Thus, the results indicate increased fatty acid oxidation capacity in cardiac ventricles from TTA fed Atlantic salmon.
The genes of the six GO terms that were grouped together contained almost entirely genes encoding enzymes or subunits participating in glycolysis. We found increased expression of hexokinase 1 (HXK1), the phospho-fructokinases aldolase A (ALDOA) and 6-phosphofructokinase type C (K6PP), glycerol-3-phosphate dehydrogenase (GPDA/GAPDH) and the pyruvate dehydrogenases DLAT and OPDX. In accordance, we also observed an increased expression of MLX-interacting protein (MLXIP), which has been suggested to be an essential regulator of cellular glycolysis. All of the genes showed increased transcription, therefore clearly indicating increased glycolysis in the hearts from TTA fed Atlantic salmon.
Tricarboxylic acid (TCA) cycle
As in the previously described categories, almost all of the genes in this category showed an increased expression. The proteins encoded by nearly all of the genes in this category are part of the TCA-cycle. For a graphical representation of the genes within the TCA cycle, see Additional file4: Figure S2.
To validate the microarray results, six genes were analyzed by qRT-PCR between the TTA and control fed group for the 8.weeks and 17.weeks samples, using the same RNA samples that were used in the microarray experiment. The results showed a significant correlation between the logFCs obtained by qRT-PCR and those obtained by microarray (Pearson correlation r = 0.8314; p = 0.0008; Figure4).
In addition, we measured the gene expression of the three PPAR subtypes α,β and γ in control and TTA fed Atlantic salmon in the four different tissues: Heart, muscle, liver and gut (pyloric caeca) from both sampling points (Figure5A). Analyzing the expression levels using analysis of variance (ANOVA) showed that only the PPAR α expression in the heart was significantly increased in the TTA group. The gene expression levels of PPARγ in heart and muscle were too low to allow reliable quantification. Since the microarray data revealed quite clearly that TTA affected genes are involved in regulating the heart performance, we used qRT-PCR to measure the expression of the cardiac transcription factors GATA4, Mef2C and osteonectin (Osx). For all three transcription factors we observed a trend of higher mean transcription in the TTA group (Figure5B), however, only Mef2C showed statistical significance.
TTA has previously been reported to stimulate mitochondrial biogenesis in mammals. In this study TTA had no effect on mitochondrial biogenesis as measured by the ratio of mt/nDNA (Figure6). In the liver we found a significant interaction between dietary treatment and time for mitochondrial biogenesis.
In the present study we investigated the response of Atlantic salmon to TTA during the seawater phase. The results from our study show that feeding TTA had profound effects on the cardiac gene expression at sampling point 17.weeks, 9 weeks after TTA feeding ended. The level of TTA applied in the study was, with 0.25%, lower than previous studies that have been conducted in Atlantic salmon (compare[1, 6, 37]). The mortality rates previously observed in Atlantic salmon in response to high TTA levels was not observed in this study.
Gene set over-representation of the transcription profile at 17.weeks shows an increased capacity of fat catabolism, glycolysis and activity of the TCA cycle as well as cardiac contractility and cardiac hypertrophy. Overall, the results suggest a scenario where cardiac ventricles of TTA pre-fed fish are able to generate more energy via a TCA-cycle that is fueled by metabolites from fat catabolism and glycolysis. TTA functions as a ligand for all three PPAR subtypes[3, 4], which have crucial functions in the transcriptional regulation of cardiac metabolism. In mice the transcriptional effects of TTA in the heart have been shown to be mediated almost exclusively via PPARα. Gain-of-function and loss-of-function mutations have shown that PPARα is a crucial transcription factor in the cardiac metabolism, regulating mainly cardiac fatty acid uptake and oxidation[10, 11]. Furthermore, activation of PPARα has been demonstrated to shift cardiac energy utilization away from glucose and towards fatty acid oxidation, actually mimicking the cardiac phenotype observed in diabetic hearts. Interestingly the cardiac phenotype of PPARβ differs from that of PPARα, indicating that both transcription factors regulate, at least partly, different subsets of genes in the heart. PPARβ loss-of-function hearts suffer from myocardial lipid accumulation and cardiomyopathy. Gain-of-function mutations on the other hand clearly show that PPARβ positively regulates cardiac glucose utilization, and also stimulates cardiac growth. Thus, the significantly higher cardiac transcription of PPARα and the elevated mean transcription of PPARβ/ in concert with the activation of their down-stream pathways, fat catabolism and the glycolysis pathway suggest that cardiac effects of TTA in Atlantic salmon are mediated by both PPARα and PPARβ. Intriguingly, over-expression of a constitutively active form of PPARβ in murine skeletal muscle has been reported to mimic training-based muscle adaptation. Hence, it has been speculated, in accordance with the results from PPARβ over-expression in mice, that PPARβ causes “physiological” cardiac hypertrophy.
Between the 8.weeks and 17.weeks sampling points, the hearts grew by a considerable portion in absolute and relative terms. The gene expression profile in hearts of TTA fed fish at 17.weeks suggests that the cardiac growth of TTA fed fish is shifted towards “physiological” hypertrophy, which may translate to an increased cardiac output. This notion is supported by the expression profile found for the category “cardiac performance” at 17.weeks, unanimously pointing to an increased cardiac contractility and also showing up-regulation of crucial cardiac transcription factors. In particular the higher transcription of the cardiac transcription factors Nkx2.5 and Mef2C can be regarded as markers for cardiac hypertrophy/growth. It has been demonstrated in mice that over-expression of Mef2C is sufficient to induce cardiac hypertrophy. Furthermore, both Mef2C and Nkx2.5 have been shown, in vitro, to be regulated by PPARα in cardiomyocytes.
It should also be noted that although we did not find significant differences in relative heart weight in this study, in other studies we found that TTA significantly increases heart size in Atlantic salmon[8, 9], and that the effect seems to be correlated to the dose of TTA (Rørvik, unpublished data). Thus, it is tempting to speculate that the increase in relative heart weight may be related to the cardiac transcriptional changes induced by TTA. A “cardiac exercise” stimulating effect is of high relevance for salmonid aquaculture. Atlantic salmon, having a circulatory system that is naturally adapted to long migration routes and high activity, show alteration in cardiac morphology and a reduced relative heart weight in captivity. In addition, circulatory failure has been identified as an important cause of mortality in salmon farming. Thus, using TTA may be one way to support the cardiac performance of fish in captivity.
The highest tissue concentrations of TTA in Atlantic salmon, as well as in mice, can be found in the heart[1, 45]. In accordance, the heart was also the tissue where the strongest transcriptional response of PPARα was detected. The main transcriptional effects were found nine weeks after the TTA feeding stopped and where our data suggested that the cardiac tissue levels of TTA were neglectable. However, we have no information about the course of gene expression between both sampling points, thus it might very well be that the effects of sampling at 17.weeks are the remains of earlier, stronger transcriptional effects. It is remarkable that a similar, delayed response in expression of lipid metabolism related genes to TTA has been observed in our previous Atlantic salmon studies[8, 9], indicating a common underlying mechanism. It is possible that the delay in transcriptional response is caused by a common, yet unknown, mechanism.
In conclusion, based on results from microarray analysis, this study demonstrates that TTA increases cardiac fatty acid oxidation and glycolysis as well as contractility and cardiac hypertrophy in Atlantic salmon. The gene expression profiles further favor a scenario of “physiological” hypertrophy in response to TTA. This increased cardiac efficiency may offer significant benefits in situations with increased oxygen demand.
The experiment was conducted at Nofima Marin sea-water research station, Averøy, western Norway. Atlantic salmon used in this experiment were hatched at Nofima Marin research station (Sunndaløra, Norway) one year earlier (S1/1+ Salmon). The experiment started with the seawater transfer of the fish on the 15th of May 2007 and lasted until the 27th of September 2007. A randomized block design with triplicate seawater net-pens and 400 fish per pen (pen = 125 m3) was used for the experiment. Control and TTA diets (0.25% (w/w) TTA (Thia Medica, Norway)) were produced by Biomar (Biomar AS, Myre, Norway). Both TTA and control diets were fed to the fish until the 16th of July 2007, from this point until the end of the experiment only the control diet was fed to the fish. Low levels of TTA (0.25%) and a short feeding period were chosen in order to avoid negative TTA effects (mortality, altered kidney morphology[1, 46]). TTA was fed for the first eight weeks after sea transfer, a period where the physiology of the salmon alters due to changing from a fresh to a saltwater environment; and we speculated that an increased capacity for energy utilization may be beneficial. Fish were sampled from the cages to represent the average fish weight for the cage. Sampling was done on the following dates: 16-18th of July 2007 (sampling point: 8.weeks, end of the TTA feeding period) and 25-27th of September 2007 (sampling point: 17.weeks). For each sampling point, fish were sampled for: heart ventricle, liver, muscle and gut (pyloric caeca). The tissue samples were snap frozen in liquid nitrogen and stored at -80°C.
Fat content in the muscle (Norwegian quality cut–NQC, Norwegian standard procedure - NS 9401 1994) was measured in pooled samples (10 fish) from each net pen as described in. TTA was measured within the total cardiac lipids. For the analysis 10 ventricles from Atlantic salmon out of the same net pen were pooled. Total heart lipids were extracted with chloroform-methanol and fatty acid methyl esters (FAME) were obtained by heating of lipids with methanol at 90°C/1 hour, where H2S O4 was used as a catalyst. After extraction into an organic solvent, the FAME were analyzed by gas-liquid chromatography. A gas chromatograph GC 8000 TOP (Finnigan, USA) was equipped with a programmed temperature vaporization (PTV) injector, flame-ionization detector (FID), AS 800 autosampler and a fused silica capillary column coated with dimethylpolysiloxane stationary phase, DB1-ms (J & W Scientific, USA). Hydrogen was used as a carrier gas. Column temperature was programmed from 110 to 310°C with a gradient 2.5°C/min. The GC signal was acquired using Chromeleon software (Dionex, USA). Peaks were identified by means of known FA standards (Larodan Fine Chemicals, Sweden and Sigma-Aldrich, USA) and by means of mass spectra, obtained by GC/MS analysis (GCQ, Finnigan, USA) on the same column. An internal standard (C21:0) was used for quantitation after calibration with known mixtures of FA standards.
Two individual samples from each one of the 3 net-pens were samples for heart ventricle, liver, muscle and gut. The samples were randomly chosen (n=6 per dietary group and sampling point) and homogenized using a rotor tissue lyser (Precellys 24, Bertin technologies, France). Total RNA was extracted and purified using column purification (96 universal Tissue Kit, Qiagen, Hilden, Germany) according to the manufacturer’s instructions. Traces of genomic DNA in the samples were eliminated by on-column-DNase digestion (Qiagen). RNA concentrations were measured for all samples using a NanoDrop 1000 Spectrophotometer (Thermo Fisher Scientific, Wilmington, USA). RNA quality for samples later used in the microarray was determined using a Agilent 2100 Bioanalyzer (RNA 6000 NanoLabChip, Agilent, Waldborn, Germany).
A customized oligo (60-mer) Atlantic salmon microarray in the 4x44K format (Agilent,) was used to detect differential gene expression between samples from the heart ventricles of control and 0.25% TTA fed fish for the 8.weeks and 17.weeks sampling points. The array contained 21012 different probes spotted in duplicates. RNA samples from individual fish were hybridized to the microarray. 24 individual microarrays were performed using 12 fish (6 control and 6 TTA fish) at the 8 week sampling point, and similarly at the 17 week sampling point. All RNA samples used in the hybridization had RIN values ranging from 9.5 to 10. 500ng RNA were amplified and labeled with Cy3 using the Quick Amp Labeling Kit (One Color-Agilent). After purification the cRNA was quantified using NanoDrop. Subsequently 1.65μ g Cy-dye-labeled cRNA was fragmented (mean size, approximately 50-100 nucleotides) with fragmentation buffer (Agilent Technologies) at 60°C for 30 min; cRNA was subsequently hybridized to the microarray at 65°C for 17 h. All steps were conducted according to the Agilent protocol (One-Color Quick Amp Labeling, Version 5.7). The microarray chips were scanned using a Agilent Microarray Scanner (G2565CA) and analysis of the microarray images was done in Agilent’s Feature Extraction Software (Version 10.5.1.1) using the one-color (GE1_105_Dec08) protocol.
Normalization and analysis of the data was performed in R/Bioconductor[50, 51] using the “limma” package. The background corrected fluorescence signals (gProcessedSignal) were obtained from Feature Extraction (Agilent). Spots were filtered according to the following criteria provided by Feature Extraction: gIsFound, gIsPosAndSignif and gIsWellAboveBG (a description of the parameters can be found in the Feature Extraction Software Reference Guide). The mean signal of the duplicated probes was calculated and all control spots together with probe sets showing more than three missing values were removed from the dataset. The data was subsequently normalized using quantile normalization in order to adjust the scale of intensities across arrays. After normalization the signals were log2 transformed. The normalized/filtered dataset then contained 13166 probe sets (63% of the total). The raw and normalized data is publicly available at NCBI’s GEO repository (http://www.ncbi.nlm.nih.gov/geo/, AccNr: GSE25305). Differential expression of probe sets was assessed by fitting a linear model, including the effects of feeding (2 levels: Control and TTA) and the effects of sampling point (2 levels: 8.weeks and 17.weeks) and their interaction. The specific comparisons: TTA vs. Control at sampling point 8.weeks and TTA vs. Control at sampling point 17.weeks were made by extracting the appropriate contrasts from the linear model. For each contrast moderated t-statistics were calculated using an empirical Bayes method. Probes without annotation were removed from the dataset before controlling the false discovery rate simultaneously across probe sets and contrasts (method: “ global” in the limma function “decideTests”). Probe sets with a q-value ≤ 0.05 and a log2FC ≥ 0.5 were declared DE for the corresponding contrast. A comprehensive list of all DE probes for each contrast can be found in Additional file5: Table S3.
Probe annotation and GOs were retrieved using the top Blast function implemented in Blast2GO. Full length probe sequences were blasted against protein sequences from the Swissprot database in a BlastX search. The E-value cut off was set to 10−6. Hypergeometric testing for over-representation of GO terms from the category biological process among the genes DE for the contrast TTA vs. Control at the 17.weeks sampling point was conducted using the GOstats package. Before testing, probes matching to the same gene were collapsed to the probe showing the largest variance.
Correspondence analysis was conducted using the R package “made4”. Probe sets with missing values were removed from the dataset prior to correspondence analysis.
Single strand cDNA was synthesized from 500ng of total RNA using oligo dT primers and the Taq Man reverse transcription Kit (Applied Biosystems, CA, USA). qRT-PCR was performed on a Light-Cycler 480 (Roche, Switzerland). For the PCR reaction, 2x SYBR green I master Mix (Roche), 0.41nM of each primer and the cDNA template were mixed in a total reaction volume of 10μ l. Primer sequences are listed in Table3. A three step PCR protocol with 45 cycles (15s 95°C, 15s 60°C, 15s 72°C) was used. To verify specific amplification, a melting curve analysis step was done at the end of the program. In order to verify the results obtained through the microarray experiment, the same 24 samples used in the array were used in a qRT-PCR approach. Six genes were then randomly picked and samples were analyzed in duplicates. The expression level was calculated using the standard curve method (Applied Biosystems User Bulletin 2). The standard curve was produced from a serial dilution of a pool consisting of all cDNA samples. The expression levels were standardized to the expression of the housekeeping gene elongation factor 1 α(EF1 α,).
qRT-PCR primer sequences
Forward primer (3’ - 5’)
Reverse primer (5’ - 3’)
mt D Loop B (gDNA)
Genomic DNA was isolated from tissue samples from the same individuals as the ones used for the total RNA extraction (totally 96 samples, n=6). DNA was isolated using DNAeasy kit (Qiagen) according to the manual. The DNA quality for all samples was checked on a 1% agarose gel and concentration was measured using a NanoDrop Spectrophotometer. The MyoD gene (intron-exon spanning primers) and the mitochondrial D-loop were amplified by qRT-PCR. For the PCR reaction 1x SYBR green I master Mix (Roche), 0.41nM of each primer and the 6.4ng DNA template were mixed in a reaction volume of 10μ l. PCR amplification was conducted as described above. All reactions were run in duplicates. Absolute concentrations for mt- and nDNA samples were obtained using the standard curve method. The ratio was calculated by dividing the absolute mtDNA by the absolute nDNA concentration.
All data are presented as means ± SEM with an n value as stated. The effect of dietary treatment on the production parameters and qRT-PCR were analyzed by 2-way analysis of variance (ANOVA), using dietary treatment and sampling point as fixed factors and block as a random factor. TTA effects on gene expression and mt/nDNA ratio were calculated using EF1 α standardized expression values in a 2-way ANOVA with dietary treatment and time as fixed factors. Unless otherwise stated the statistical unit is the individual fish. All analyses were conducted using R, plots were produced using the R package ggplot and the heatmaps were produced with the R package lattice/latticeExtra.
HT,KR and MT conceived the study and designed the experiment. RB provided the TTA measurements. FG conducted the lab experiment, data analysis and drafted the manuscript. All authors read and approved the final manuscript.
We would like to thank Dr. Aleksei Krasnov for designing the Atlantic salmon microarray. We would also like to thank Thia Medica for supplying the TTA used in this experiment.
- Moya-Falcón C, Hvattum E, Dyrøy E, Skorve J, Stefansson SO, Thomassen MS, Jakobsen JV, Berge RK, Ruyter B: Effects of 3-thia fatty acids on feed intake, growth, tissue fatty acid composition, beta-oxidation and Na+,K+-ATPase activity in Atlantic salmon. Comp Biochem Physiol B , Biochem Mol Biol. 2004, 139 (4): 657-68. 10.1016/j.cbpc.2004.08.009.View ArticlePubMedGoogle Scholar
- Berge RK, Skorve J, Tronstad KJ, Berge K, Gudbrandsen OA, Grav H: Metabolic effects of thia fatty acids. Curr Opin Lipidol. 2002, 13 (3): 295-304. 10.1097/00041433-200206000-00010.View ArticlePubMedGoogle Scholar
- Westergaard M, Henningsen J, Svendsen ML, Johansen C, Jensen UB, Schrøder HD, Kratchmarova I, Berge RK, Iversen L, Bolund L, Kragballe K, Kristiansen K: Modulation of keratinocyte gene expression and differentiation by PPAR-selective ligands and tetradecylthioacetic acid. J Invest Dermatol. 2001, 116 (5): 702-12. 10.1046/j.1523-1747.2001.01329.x.View ArticlePubMedGoogle Scholar
- Forman BM, Chen J, Evans RM: Hypolipidemic drugs, polyunsaturated fatty acids, and eicosanoids are ligands for peroxisome proliferator-activated receptors alpha and delta. Proc Natl Acad Sci USA. 1997, 94 (9): 4312-7. 10.1073/pnas.94.9.4312.PubMed CentralView ArticlePubMedGoogle Scholar
- Berger JP, Akiyama TE, Meinke PT: PPARs: therapeutic targets for metabolic disease. Trends Pharmacol Sci. 2005, 26 (5): 244-51. 10.1016/j.tips.2005.03.003.View ArticlePubMedGoogle Scholar
- Rørvik KA, Alne H, Gaarder M, Ruyter B, Måseide NP, Jakobsen JV, Berge RK, Sigholt T, Thomassen MS: Does the capacity for energy utilization affect the survival of post-smolt Atlantic salmon, Salmo salar L., during natural outbreaks of infectious pancreatic necrosis?. J Fish Dis. 2007, 30 (7): 399-409. 10.1111/j.1365-2761.2007.00823.x.View ArticlePubMedGoogle Scholar
- Vegusdal A, Gjøen T, Berge RK, Thomassen MS, Ruyter B: Effect of 18:1n-9, 20:5n-3, and 22:6n-3 on lipid accumulation and secretion by Atlantic salmon hepatocytes. Lipids. 2005, 40 (5): 477-86. 10.1007/s11745-005-1407-z.View ArticlePubMedGoogle Scholar
- Alne H, Thomassen MS, Takle H, Terjesen BF, Grammes F, Oehme M, Refstie S, Sigholt T, Berge RK, Rørvik KA: Increased survival by feeding tetradecylthioacetic acid during a natural outbreak of heart and skeletal muscle inflammation in S0 Atlantic salmon, Salmo salar L. J Fish Dis. 2009, 32 (11): 953-61. 10.1111/j.1365-2761.2009.01078.x.View ArticlePubMedGoogle Scholar
- Grammes F, Rørvik KA, Takle H: Tetradecylthioacetic acid modulates cardiac transcription in Atlantic salmon, Salmo salar L., suffering heart and skeletal muscle inflammation. J Fish Dis. 2012, 35 (2): 109-17. 10.1111/j.1365-2761.2011.01326.x.View ArticlePubMedGoogle Scholar
- Sambandam N, Morabito D, Wagg C, Finck BN, Kelly DP, Lopaschuk GD: Chronic activation of PPARalpha is detrimental to cardiac recovery after ischemia. Am J Physiol Heart Circ Physiol. 2006, 290: H87—95-PubMedGoogle Scholar
- Finck BN, Lehman JJ, Leone TC, Welch MJ, Bennett MJ, Kovacs A, Han X, Gross RW, Kozak R, Lopaschuk GD, Kelly DP: The cardiac phenotype induced by PPARalpha overexpression mimics that caused by diabetes mellitus. J Clin Invest. 2002, 109: 121-30.PubMed CentralView ArticlePubMedGoogle Scholar
- Burkart EM, Sambandam N, Han X, Gross RW, Courtois M, Gierasch CM, Shoghi K, Welch MJ, Kelly DP: Nuclear receptors PPARbeta/delta and PPARalpha direct distinct metabolic regulatory programs in the mouse heart. J Clin Invest. 2007, 117 (12): 3930-9.PubMed CentralPubMedGoogle Scholar
- Cheng L, Ding G, Qin Q, Huang Y, He N, Evans RM, Schneider MD, Brako FA, Xiao Y, Chen YE, Yang Q: Cardiomyocyte-restricted peroxisome proliferator-activated receptor-[delta] deletion perturbs myocardial fatty acid oxidation and leads to cardiomyopathy. Nat Med. 2004, 10 (11): 1245-50. 10.1038/nm1116.View ArticlePubMedGoogle Scholar
- Chinetti G, Fruchart JC, Staels B: Peroxisome proliferator-activated receptors (PPARs): nuclear receptors at the crossroads between lipid metabolism and inflammation. Inflamm Res. 2000, 49 (10): 497-505. 10.1007/s000110050622.View ArticlePubMedGoogle Scholar
- Finck BN: The PPAR regulatory system in cardiac physiology and disease. Cardiovasc Res. 2007, 73 (2): 269-77. 10.1016/j.cardiores.2006.08.023.View ArticlePubMedGoogle Scholar
- The health situation in Norwegian aquaculture 2009. Tech. rep., Norwegian Veterinary Institute 2009.Http://www.vetinst.no/eng/Research/Publications/Fish-Health-Report,
- Driedzic WR: Cardiac energy metabolism. Fish Physiology, Volume 7. Edited by: Hoar WS, Randall DJ, Farrell AP. 1992, Academic Press, New York, 219-266.Google Scholar
- Krasnov A, Timmerhaus G, Afanasyev S, Jørgensen SM: Development and assessment of oligonucleotide microarrays for Atlantic salmon (Salmo salar L.). Comp Biochem Physiol Part D Genomics Proteomics. 2010Google Scholar
- Fellenberg K, Hauser NC, Brors B, Neutzner A, Hoheisel JD, Vingron M: Correspondence analysis applied to microarray data. Proc Natl Acad Sci USA. 2001, 98 (19): 10781-6. 10.1073/pnas.181597298.PubMed CentralView ArticlePubMedGoogle Scholar
- Smyth GK: Limma: linear models for microarray data. Bioinformatics and Computational Biology Solutions using R and Bioconductor. Edited by: Gentleman R, Carey V, Dudoit S, R Irizarry WH. 2005, Springer, New York, 397-420.View ArticleGoogle Scholar
- Gerety SS, Anderson DJ: Cardiovascular ephrinB2 function is essential for embryonic angiogenesis. Development. 2002, 129 (6): 1397-410.PubMedGoogle Scholar
- Hoefen RJ, Berk BC: The multifunctional GIT family of proteins. J Cell Sci. 2006, 119 (Pt 8): 1469-75.View ArticlePubMedGoogle Scholar
- Wallis J, Lygate CA, Fischer A, ten Hove, Schneider JE, Sebag-Montefiore L, Dawson D, Hulbert K, Zhang W, Zhang MH, Watkins H, Clarke K, Neubauer S: Supranormal myocardial creatine and phosphocreatine concentrations lead to cardiac hypertrophy and heart failure: insights from creatine transporter-overexpressing transgenic mice. Circulation. 2005, 112 (20): 3131-9. 10.1161/CIRCULATIONAHA.105.572990.View ArticlePubMedGoogle Scholar
- Mandard S, Zandbergen F, van Straten E, Wahli W, Kuipers F, Müller M, Kersten S: The fasting-induced adipose factor/angiopoietin-like protein 4 is physically associated with lipoproteins and governs plasma lipid levels and adiposity. J Biol Chem. 2006, 281 (2): 934-44.View ArticlePubMedGoogle Scholar
- Ashburner M, Ball CA, Blake JA, Botstein D, Butler H, Cherry JM, Davis AP, Dolinski K, Dwight SS, Eppig JT, Harris MA, Hill DP, Issel-Tarver L, Kasarskis A, Lewis S, Matese JC, Richardson JE, Ringwald M, Rubin GM, Sherlock G: Gene ontology: tool for the unification of biology. The Gene Ontology Consortium. Nat Genet. 2000, 25: 25-9. 10.1038/75556.PubMed CentralView ArticlePubMedGoogle Scholar
- Falcon S, Gentleman R: Using GOstats to test gene lists for GO term association. Bioinformatics. 2007, 23 (2): 257-8. 10.1093/bioinformatics/btl567.View ArticlePubMedGoogle Scholar
- Christoffels VM, Habets PE, Franco D, Campione M, de Jong, Lamers WH, Bao ZZ, Palmer S, Biben C, Harvey RP, Moorman AF: Chamber formation and morphogenesis in the developing mammalian heart. Dev Biol. 2000, 223 (2): 266-78. 10.1006/dbio.2000.9753.View ArticlePubMedGoogle Scholar
- Chen JN, Fishman MC: Zebrafish tinman homolog demarcates the heart field and initiates myocardial differentiation. Development. 1996, 122 (12): 3809-16.PubMedGoogle Scholar
- Schwinger RHG, Bundgaard H, Müller-Ehmsen J, Kjeldsen K: The Na, K-ATPase in the failing human heart. Cardiovasc Res. 2003, 57 (4): 913-20. 10.1016/S0008-6363(02)00767-8.View ArticlePubMedGoogle Scholar
- Wehrens XHT, Lehnart SE, Huang F, Vest JA, Reiken SR, Mohler PJ, Sun J, Guatimosim S, Song LS, Rosemblit N, D’Armiento JM, Napolitano C, Memmi M, Priori SG, Lederer WJ, Marks AR: FKBP12.6 deficiency and defective calcium release channel (ryanodine receptor) function linked to exercise-induced sudden cardiac death. Cell. 2003, 113 (7): 829-40. 10.1016/S0092-8674(03)00434-3.View ArticlePubMedGoogle Scholar
- Frank KF, Bölck B, Erdmann E, Schwinger RHG: Sarcoplasmic reticulum Ca2+-ATPase modulates cardiac contraction and relaxation. Cardiovasc Res. 2003, 57: 20-7. 10.1016/S0008-6363(02)00694-6.View ArticlePubMedGoogle Scholar
- Beck EJ, Bowlby M, An WF, Rhodes KJ, Covarrubias M: Remodelling inactivation gating of Kv4 channels by KChIP1, a small-molecular-weight calcium-binding protein. J Physiol. 2002, 538 (Pt 3): 691-706.PubMed CentralView ArticlePubMedGoogle Scholar
- Wickenden AD, Kaprielian R, Kassiri Z, Tsoporis JN, Tsushima R, Fishman GI, Backx PH: The role of action potential prolongation and altered intracellular calcium handling in the pathogenesis of heart failure. Cardiovasc Res. 1998, 37 (2): 312-23. 10.1016/S0008-6363(97)00256-3.View ArticlePubMedGoogle Scholar
- Dyck JR, Barr AJ, Barr RL, Kolattukudy PE, Lopaschuk GD: Characterization of cardiac malonyl-CoA decarboxylase and its putative role in regulating fatty acid oxidation. Am J Physiol. 1998, 275 (6 Pt 2): H2122—9-PubMedGoogle Scholar
- Sans CL, Satterwhite DJ, Stoltzman CA, Breen KT, Ayer DE: MondoA-Mlx heterodimers are candidate sensors of cellular energy status: mitochondrial localization and direct regulation of glycolysis. Mol Cell Biol. 2006, 26 (13): 4863-71. 10.1128/MCB.00657-05.PubMed CentralView ArticlePubMedGoogle Scholar
- Totland GK, Madsen L, Klementsen B, Vaagenes H, Kryvi H, Frøyland L, Hexeberg S, Berge RK: Proliferation of mitochondria and gene expression of carnitine palmitoyltransferase and fatty acyl-CoA oxidase in rat skeletal muscle, heart and liver by hypolipidemic fatty acids. Biol Cell. 2000, 92 (5): 317-29. 10.1016/S0248-4900(00)01077-7.View ArticlePubMedGoogle Scholar
- Kleveland EJ, Ruyter B, Vegusdal A, Sundvold H, Berge RK, Gjøen T: Effects of 3-thia fatty acids on expression of some lipid related genes in Atlantic salmon (Salmo salar L.). Comp Biochem Physiol B , Biochem Mol Biol. 2006, 145 (2): 239-48. 10.1016/j.cbpb.2006.07.012.View ArticlePubMedGoogle Scholar
- Hafstad AD, Khalid AM, Hagve M, Lund T, Larsen TS, Severson DL, Clarke K, Berge RK, Aasum E: Cardiac peroxisome proliferator-activated receptor-alpha activation causes increased fatty acid oxidation, reducing efficiency and post-ischaemic functional loss. Cardiovasc Res. 2009, 83 (3): 519-26. 10.1093/cvr/cvp132.View ArticlePubMedGoogle Scholar
- Wagner N, Jehl-Piétri C, Lopez P, Murdaca J, Giordano C, Schwartz C, Gounon P, Hatem SN, Grimaldi P, Wagner KD: Peroxisome proliferator-activated receptor beta stimulation induces rapid cardiac growth and angiogenesis via direct activation of calcineurin. Cardiovasc Res. 2009, 83: 61-71. 10.1093/cvr/cvp106.View ArticlePubMedGoogle Scholar
- Narkar VA, Downes M, Yu RT, Embler E, Wang YX, Banayo E, Mihaylova MM, Nelson MC, Zou Y, Juguilon H, Kang H, Shaw RJ, Evans RM: AMPK and PPARdelta agonists are exercise mimetics. Cell. 2008, 134 (3): 405-15. 10.1016/j.cell.2008.06.051.PubMed CentralView ArticlePubMedGoogle Scholar
- Xu J, Gong NL, Bodi I, Aronow BJ, Backx PH, Molkentin JD: Myocyte enhancer factors 2A and 2C induce dilated cardiomyopathy in transgenic mice. J Biol Chem. 2006, 281 (14): 9152-62.View ArticlePubMedGoogle Scholar
- Sharifpanah F, Wartenberg M, Hannig M, Piper HM, Sauer H: Peroxisome proliferator-activated receptor alpha agonists enhance cardiomyogenesis of mouse ES cells by utilization of a reactive oxygen species-dependent mechanism. Stem Cells. 2008, 26: 64-71. 10.1634/stemcells.2007-0532.View ArticlePubMedGoogle Scholar
- Poppe TT, Johansen R, Gunnes G, Tørud B: Heart morphology in wild and farmed Atlantic salmon Salmo salar and rainbow trout Oncorhynchus mykiss. Dis Aquat Org. 2003, 57 (1-2): 103-8.View ArticlePubMedGoogle Scholar
- Tørud B, Hillestad M: “Hjerte-rapporten” Rapport om hjertelidelser hos laks og regnbueørret. 2004, 1-69. [www.fiskerifond.no/files/projects/attach/hjerterapporten.pdf]Google Scholar
- Asiedu DK, Frøyland L, Vaagenes H, Lie O, Demoz A, Berge RK: Long-term effect of tetradecylthioacetic acid: a study on plasma lipid profile and fatty acid composition and oxidation in different rat organs. Biochim Biophys Acta. 1996, 1300 (2): 86-96. 10.1016/0005-2760(95)00235-9.View ArticlePubMedGoogle Scholar
- Gjøen T, Kleveland EJ, Moya-Falcón C: Effects of dietary thia fatty acids on lipid composition, morphology and macrophage function of Atlantic salmon (Salmo salar L.) kidney. Comp Biochem Physiol B , Biochem Mol Biol. 2007, 148: 103-11. 10.1016/j.cbpb.2007.04.021.View ArticlePubMedGoogle Scholar
- Alne H, Oehme M, Thomassen M, Terjesen B, Rørvik KA: Reduced growth, condition factor and body energy levels in Atlantic salmon Salmo salar L. during their first spring in the sea. Aquacult Res. 2010, 42 (2): 248-259.View ArticleGoogle Scholar
- Bligh EG, Dyer WJ: A rapid method of total lipid extraction and purification. Can J Biochem Physiol. 1959, 37: 911-917. 10.1139/o59-099.View ArticlePubMedGoogle Scholar
- Dates M: General analytical procedures. Techniques in Lipidology. Edited by Kates M. Elsevier; 1986Google Scholar
- R Development CoreTeam: R: A Language and Environment for Statistical Computing. 2009, R Foundation for Statistical Computing, Vienna, http://www.R-project.org. [ISBN 3-900051-07-0]Google Scholar
- Gentleman RC, Carey VJ, Bates DM: Bioconductor: Open software development for computational biology and bioinformatics. Genome Biology. 2004, 5: R80-10.1186/gb-2004-5-10-r80.PubMed CentralView ArticlePubMedGoogle Scholar
- Bolstad BM, Irizarry RA, Astrand M, Speed TP: A comparison of normalization methods for high density oligonucleotide array data based on variance and bias. Bioinformatics. 2003, 19 (2): 185-93. 10.1093/bioinformatics/19.2.185.View ArticlePubMedGoogle Scholar
- Smyth GK: Linear models and empirical bayes methods for assessing differential expression in microarray experiments. Stat Appl Genet Mol Biol. 2004, 3: Article3-PubMedGoogle Scholar
- Benjamini Y, Yosef H: Controlling the false discovery rate: a practical and powerful approach to multiple testing. J Roy Stat Soc B. 1995, 57: 289-300.Google Scholar
- Conesa A, Götz S: Blast2GO: a universal tool for annotation, visualization and analysis in functional genomics research. Bioinformatics. 2005, 21 (18): 3674-6. 10.1093/bioinformatics/bti610.View ArticlePubMedGoogle Scholar
- Culhane AC, Thioulouse J, Perrière G: MADE4: an R package for multivariate analysis of gene expression data. Bioinformatics. 2005, 21 (11): 2789-90. 10.1093/bioinformatics/bti394.View ArticlePubMedGoogle Scholar
- Olsvik PA, Lie KK, Jordal AEO, Nilsen TO, Hordvik I: Evaluation of potential reference genes in real-time RT-PCR studies of Atlantic salmon. BMC Mol Biol. 2005, 6: 21-10.1186/1471-2199-6-21.PubMed CentralView ArticlePubMedGoogle Scholar
- Wickham H: Ggplot: Elegant graphics for data analysis. 2009, Springer, New YorkView ArticleGoogle Scholar
- Sarkar D: Lattice: Multivariate Data Visualization with R. 2008, Springer, New YorkView ArticleGoogle Scholar
- Moriya Y, Itoh M, Okuda S, Yoshizawa AC, Kanehisa M: KAAS: an automatic genome annotation and pathway reconstruction server. Nucleic Acids Res. 2007, 35 (Web Server issue): W182-5.PubMed CentralView ArticlePubMedGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.