Th e number of fish species exceeds 32400 [1] and most of them swim by undulations of the whole body [2]. Fish locomotion is powered by at most three (but mostly only two) morphologically diff ering myotomal muscle fi ber types. Red oxidative muscles provide generally maximally 10% of the whole body musculature and are responsible for continuous cruise swimming, while white muscle fi bers provide ca. 50% of body mass, being myoglobin-poor and thus powering primarily short term, burst-type locomotion [3]. Interestingly, in most fishes these fi bers show spatial separation within myotomes and fi ber types, while in cyprinids (the subject of this study) they form mixed type (type IIx) pink muscles [4,5] Repeated exercise leads to muscle morphological adaptation, namely increased red muscle proportion (with increased capillarization), with a shift away from the type IIb fi bers [6]. Concerning the relationship of exercise intensity and adaptation, moderate aerobic load appears to be of fi rst importance [7]. By fish exercise the sequence of muscle fuels is led by phosphagenes [8], followed by glycogen and oxidative metabolism is the subsequent source. In fish, muscle lactate is not an oxidative substrate of recovery but is retained in the muscle for glycogenesis. Thus, besides the notable anaerobic process, intramuscular lipids also fuel exercise, but interestingly mostly the recovery from an exhaustive bout [9].

Th e metabolic consequences of regular training include the changes of substrate metabolism [10], but induce as well the compositional alteration of the muscle cell. During submaximal exercise, augmented fat oxidation signifi cantly contributes to energy turnover [11], and it is possible that a recurring preferential recruitment and oxidation of fatty acids [12] could modify the fatty acid composition of adipose tissue and intramuscular fat, and subsequently aff ect muscle membrane phospholipid fatty acid composition. In numerous homeothermic vertebrate species [13-15] this was experimentally proven, namely regular swimming exercise markedly altered the fatty acid profi le of skeletal muscle phospholipids (PL). Th is seems to be a consequence of regular exercise and the coupled altered insulin action and oxidative stress. However, mechanisms underlying the alteration of the muscle membrane FA composition are not fully understood and those have mostly been investigated in homeothermic vertebrates, while those are mostly less elucidated in fish. Accordingly, this study aimed to analyse the muscle membrane compositional adaptation (and the coupled lipid peroxidation rate) of a freshwater fish (common carp, Cyprinus carpio L.) induced by a regular and controlled training protocol.

Training settings, handling of fish, sampling

Two types (scaled and mirror) of one-summer-old common carps (average weight: 50.4 ± 20.1 g) were introduced into 500 l fish tanks and overwintered in the Fish Laboratory of the Kaposvár University (Hungary). During the overwintering period fish could acclimate to laboratory conditions. During the conditioning and the experimental period a commercial feed (Aller Aqua, fatty acid profi le is shown in Table 1) was fed ad libitum. In the experiment both carp types were sub-divided for two groups, experimental (trained) and control groups.

Th e swimming facility was self-constructed, and installed to a recirculating system. It was a lengthwise halved plastic tube placed in a bigger trough. Th e ends of the tube were closed with a tightly woven mesh. Water velocity was adjusted by changing the level of the raceway and the volume of the infl uent water. Water velocity was measured with FP311 Global Water Flow Probe propeller-based current measuring instrument. Experimental fish groups were exercised daily in a 35-day period, 30 minutes every day, at constant velocity (0.6 m/s).

On the 35^th day of the experiment 12 male fish from each treatment (i.e. trained and control) were selected and over-anaesthetised with clove oil (dose 0.025 mL/L, 2 min). A fast-twitch type muscle part of the left fillet was dissected freshly, washed in ice-cold physiological saline, wiped dry and stored frozen (-70°C) until analysis (as well as the plasma samples).

Lipid extraction, fractionation, fatty acid analysis

Sample total lipid content was extracted [16], while lipid fractionation was performed [17]. In brief, extracted total lipids were transferred to glass chromatographic columns, containing 300 mg silicagel (230-400 mesh) for 10 mg of total lipids. Neutral lipids were eluted with 10 mL chloroform for the above fat amount, then 15 mL acetone:methanol (9:1 vol/vol) was added, while 10 mL pure methanol eluted the total phospholipids. The derivatisation of this latter fraction for subsequent gas chromatographic analysis was performed with the base-catalyzed NaOCH₃ method [18]. Gas liquid chromatography was performed on a Shimadzu 2100 apparatus, equipped with a SP-2380 type capillary column (30 m x 0.25 mm ID, 0.20 micrometer film, 24110-U, Supelco, USA) and flame ionisation detector (FID 2×10^-11). Characteristic operating conditions were: injector temperature: 270°C, detector temperature: 300°C, helium flow: 28 cm/sec. The oven temperature was graded: from 80 to 205°C: 2.5°C/min, 5 min at 205°C, from 205 to 250°C 10°C/min and 5 min at 250°C. To identify individual FA, an authentic FA standard (Mixture Me100 (90-1100, Larodan Fine Chemicals AB, Sweden) was used. Fatty acid results were expressed as weight % of total fatty acids. Unsaturation Index (UI) was defined as the number of double bonds in 100 fatty acyl chains.

Determination of tissue malondialdehyde concentration

The malondialdehyde concentration was determined from homogenized (Ultra Thurrax, Donau Lab AG, Linz), frozen stored samples, after the addition of 9 ml of physiological saline (0.9 % w/vol NaCl) per 1 g of tissue. Estimation of Thiobarbituric Reactive Substances (TBARS) levels was performed [19]. The assay procedure was calibrated using tetra-ethoxypropane (Fluka, Buchs, Switzerland) as a malondialdehyde source, and levels of tissue sample malondialdehyde were calculated as μmol per gram of wet tissue.

Body weight

Body weight was measured at the start and end of treatment.

Statistical analysis

From the basic dataset outlier values [20] were filtered and the remaining data were tested for normality (Shapiro-Wilk test). Between-group differences were analysed by independent samples t-test at the significance level of 0.05 [21].

Growth

Neither the initial (51.4 ± 26.5 vs. 49.7 ± 17.9 for control and trained, resp.), nor the final (59.8 ± 37.4 vs. 52.7 ± 20.8, resp.) bodyweight was different between groups.

Phospholipid fatty acid composition

The scaled and mirror genotypes did not differ in their reaction, thus these groups were pooled and only trained vs. untrained fish groups were compared.

The total lipid and polar lipid fatty acid composition of the diet is given in Table 1. The compared (control vs. trained) data on the fillet phospholipid fatty acid composition is given in Table 2.

Table 1: Total and phospholipid fatty acid composition of the diet.

Table 2: Fillet phospholipid fatty acid composition of the trained and the control groups.

The 5 week regular training significantly decreased the proportion of myristic (C14:0), margaric (C17:0) and arachidonic (C20:4 n6) acids, while increased the proportion of behenic acid (C22:0). Interestingly, in the calculated fatty acid groups only the total n6 proportion showed significant proportional modification (decrease) as a response to regular swimming exercise.

Fillet malondialdehyde concentration

The fillet malondialdehyde concentration of the trained group decreased significantly (Figure 1).

Figure 1: The malondialdehyde concentration of the fillet in the trained and control fish.

The novelty of our results lies in the fact that training-induced adaptations of fish muscles to regular swimming exercise have not yet been investigated. Here slight but systematic results are reported which are consonant with the relevant homeothermic vertebrate literature data.

Arachidonic acid and total n6 proportion

Arachidonic Acid (ARA) is one of the dominant n6 FAs of the membrane lipids and is generally a specific indicator of physiological alterations. Similarly to our results [13] provided evidence that rat skeletal muscle membrane lipids react to regular training with lower ARA proportions. In human muscle PLs [22] found that training reduced n6 FAs, particularly C18:2 n6 and ARA. We as well gained supportive results in regularly treadmill-exercised rabbit muscles (longissimus dorsi and vastus lateralis; [23], but significance was only proven in the thigh, a muscle more intensively exposed to exercise. Ayre et al. [24] exercised obese and normal rats and reported that muscle membrane arachidonate proportion is lowered by submaximal aerobic exercise, in particular in muscles substantially recruited during exercise (i.e. fast-twitch pink). This is consonant with our finding, since we also sampled fast-twitch, pink muscles. Interestingly, for fish no relevant literature data were found, while in another vertebrate group, birds decreased ARA proportion was proven during migration [15] in the flight muscle and blood plasma phospholipids.

The exercise physiological background of the decreasing muscle membrane ARA proportion is not fully clear, but exercise studies in mammals suggest, in fact, that endurance performance is enhanced by high relative amounts of n-6 FA [25], and that training causes a selective depletion of these FA from membranes [13,26,27] This was experienced as well in our study, as linoleic (C18:2 n6) and eicosatrienoic (C20:3 n6) acids, and so the total n6 FA proportion also provided slight (and for n6, significant) proportional decreases in the trained muscle. A further possible mechanism underlying the lowered proportion of ARA may be the activation of phospholipase A2, suggesting membrane structure changes [28] in the muscle cell. This has indeed been proven in rats [29] and more recently [30] it was found that calcium-independent PLA2 gamma activity is positively related to exercise endurance capacity and is thus inversely related to membrane arachidonate proportion, at least in mice.

Albeit the relevant literature is scarce, the systematically experienced, training-induced, preferred selective release of ARA from the skeletal muscle membrane lipids may thus be a pre-requisite of the synthesis of prostacyclin [31], a vasoactive compound which as well prolongs exercise duration. Moreover, [32] reported that ARA is preferentially hydrolyzed from ceramides and sphingomyelins in rat muscles by exhausting exercise of moderate intensity, suggesting the exercise-induced activation of the sphingomyelin-signalling pathway.

However, for fish this is the first report on the training induced liberation of ARA from the fish muscle membrane lipids, thus, further speculation is void.

Margaric acid

Margaric acid is an odd carbon number fatty acid and is thus not synthesized by vertebrates. Its tissue presence has a dietary background, such as in our case (Table 1), but its exercise physiological role is less clear. In rabbit muscles total lipids and also in polar lipids we described findings of identical tendency (without proven significance), irrespective of muscle fiber type [14,23]. According to [33], in canine myocardium labelled margaric acid is preferentially taken up by phospholipids (as compared to triglycerides), and its fate happens relatively slowly via β-oxidation from this fraction. We thus supposed a dietary uptake of this acid and a moderate oxidation rate, which is supported by the finding that margaric acid undergoes β-oxidation with a preference similar to palmitic acid (C16:0) [33].

Behenic acid

The proportion of behenic acid was minor and would hypothetically preclude its discussion. However, this saturated long chain FA is an important component of muscular sphingomyelins, located mostly in the outer layer of the plasma membrane. This fraction is hydrolyzed by the enzyme neutral Mg²⁺-dependent sphingomyelinase to phosphorylcholine and ceramide. Ceramide may be further converted to sphingosine and a long-chain fatty acid by the enzyme ceramidase [32]. Ceramide has been shown to act as a second major messenger and its production is triggered by stimuli and agents such as damaging agents or inflammatory cytokines. Indeed, [32] found that single-bout exhaustive exercise significantly increases behenic acid (percentage) proportion of rat muscle ceramides and sphingomyelins, irrespective of fiber type. Our results suggest an identical trend, namely a slight but statistically significant accretion of this acid in the total muscle phospholipid fraction, including both of the aforementioned polar lipid fractions.

Fillet malondialdehyde concentration

Non-enzymatic lipid peroxidation is often characterized with the determination of Malondialdehyde (MDA), a cytotoxic and mutagenic end-product formed from FAs with over three double bonds [34]. Regular aerobic exercise augments free-radical production and oxidative stress [35], while low oxidative stress stimulates the expression of certain antioxidant enzymes, which is mediated by the activation of redox-sensitive signalling pathways [36]. Ultimately, this may lead to the lowered rate of lipid peroxidation in regularly exercised muscles (and lower free radical production), as found in chronically trained eels [37], rabbits longissimus dorsi [14] and in human blood [38]

Based on the above results of this pilot study it was concluded that common carp is able to perform regular swimming exercise which slightly but definitely influences the phospholipid fatty acid composition of its fast-twitch muscle. Alterations of the FA profile (reduced n6 and ARA proportions) echo those found in other homeothermic vertebrates and might be associated with altered cytokine production and the activation of the sphingomyelin- signalling pathway. In addition, the in vivo lipid peroxidation of fillet lipids decreased due to the training.

The study was supported by the Hungarian Research Fund (OTKA, 83150), by the Hungarian Academy of Sciences (Bolyai János Research Grant, BO/26/11/4, Sz.A.) and by the TÁMOP 422B project (D.V.).

The experiment was approved by the Animal Experimentation Ethics Committee of the University of Kaposvár, as allowed by the Somogy County Animal Health and Food Control Authority (allowance no.: 1151/006/SOM/2005).