Se is an essential trace element involved in antioxidant defence of the cells and participates in animal and human health(Reference Schrauzer1, Reference Rayman, Infante and Sargent2). This trace element plays a major role in enzymatic antioxidant systems, such as glutathione peroxidases, thioredoxine reductases and other selenoproteins(Reference Hawkes and Alkan3). Mineral sources like sodium selenite (SS) or sodium selenate are the most common sources of Se added to livestock feeds to ensure an optimal supply. However, during the last decade, new organic sources have been proposed to feed manufacturers, such as Se-enriched yeasts(Reference Surai4, Reference Yoon, Werner and Butler5) or Se chelates(Reference Reis, Vieira and Nascimento6), providing an improved bioavailability compared with mineral sources.
Besides the bioavailability improvement from organic Se forms, Se metabolism pathways have gained a lot of interest. Selenomethionine metabolism is closely linked to its sulphur homologue and can be incorporated into proteins in place of methionine(Reference Schrauzer1), leading to Se-containing proteins, representing a pool of Se(Reference Surai7). Thus, selenocysteine (SeCys), recognised as the twenty-first amino acid, represents the active form of Se through selenoproteins(Reference Arnér8). For instance, glutathione peroxidase, which catalyses the hydroperoxide detoxification with glutathione, contains one SeCys on its catalytic site(Reference Epp, Ladenstein and Wendel9). In human subjects, twenty-five selenoproteins have been identified and are involved in various antioxidant functions(Reference Hawkes and Alkan3, Reference Pappas, Zoidis and Surai10, Reference Allmang, Wurth and Krol11), and the particularly complex and cell energetically costly Se incorporation as SeCys is still intriguing(Reference Allmang, Wurth and Krol11).
Based on the similarities between selenomethionine and methionine, a new organic Se source called Selisseo® (SO) has been recently developed (Adisseo France S.A.S.), which is a selenomethionine hydroxyanalogue, 2-hydroxy-4-methylselenobutanoic acid or HMSeBA (patent no. US2006Ò105960, WO2006008190) (Fig. 1).
Fig. 1 Molecular formula of 2-hydroxy-4-methylselenobutanoic acid.
The aim of the present study was to compare the bioavailability of HMSeBA with mineral and organic sources. The relative bioavailabilities of the different Se sources were assessed through muscle transfer efficiency and apparent digestibility in broiler chickens. Moreover, the metabolic transformation of this source, a selenomethionine precursor, was investigated through seleno-amino acid speciation analyses.
Materials and methods
Experimental context
Two experiments (Expts 1 and 2) were conducted on male broiler chickens (Ross PM3) fed standard maize–soyabean meal starter diets (Table 1). The experiments were conducted in the facilities of Adisseo, CERN, 03 100 Commentry, France, under the quality standards of ISO 9001. These facilities are in accordance with the agreement no. C 03 159 4 of 6 November 2008, relative to experimentation on vertebrate living animals (European regulation 24/11/86 86/609 CEE; Ministerial decree of 19 April 1988). All employees were qualified for experimental animal manipulation (internal file 3H2OG001). Mineral and organic Se sources used to supplement diets were SS (Microgan™ Se 1 % BPM; DSM Nutritional Product AG), seleno-yeast (SY) (Sel-Plex® 2000; Alltech) and SO (Selisseo®, Adisseo).
Table 1 Feed ingredients and composition of the basal diets
* The additive premixture was composed of a vitamin mix and a Se-free trace element mix. The vitamin mix provided (per kg of feed): vitamin A from vitamin A acetate, 12 000 IU; vitamin D3 from cholecalciferol, 2000 IU; vitamin E from α-tocopheryl acetate, 30 mg; vitamin K3 from menadione dimethylpyrimidinol bisulphite, 4·84 mg menadione equivalent; vitamin B1 from thiamine mononitrate, 1·97 mg; vitamin B2 from riboflavin, 6 mg; vitamin B5 from pantothenate calcium, 14·85 mg; vitamin B6 from pyridoxine, 2·97 mg; vitamin B12 from cyanocobalamin, 0·02 mg; vitamin B3 from nicotinic acid, 29·85 mg; vitamin B9 from folic acid, 0·95 mg; vitamin H from biotin, 0·1 mg. The trace element mix provided (per kg of feed): Fe from iron carbonate, 120 mg; Cu from copper sulphate, 12 mg; Zn from zinc oxide, 61 mg; Mn from manganese oxide, 94 mg; I from calcium iodide, 1·2 mg; Co from cobalt carbonate, 0·6 mg.
Expt 1: selenium transfer efficiency in muscle
A total of 816 day-old chicks (average day (d) 0 body weight (BW): 41 g) obtained from a commercial hatchery were allocated to eight treatments, with six pen replicates of seventeen birds. The eight starter diets used in the experiment were supplemented with different Se sources and levels, as follows: negative control (NC) (not supplemented with Se); SS-0·1 and SS-0·3 supplemented with SS at 0·1 and 0·3 mg Se/kg feed, respectively; SY-0·1 and SY-0·3 supplemented with SY at 0·1 and 0·3 mg Se/kg feed, respectively; SO-0·1, SO-0·2 and SO-0·3 supplemented with SO at 0·1, 0·2 and 0·3 mg Se/kg feed, respectively. The feed was provided as crumbles from d0 to d14 and fed in pellets during the rest of the experiment. Feed and water were provided ad libitum throughout the experiment. Each pen of approximately 1·25 × 1·25 m was covered with wood shavings and contained one hanging bell drinker and one hanging tube feeder. The house was maintained at 32 ± 2°C from d0 to d6, at 28 ± 2°C from d6 to d11 and at 25 ± 2°C from d11 to d21. The lighting schedule followed 23 h light–1 h dark cycle throughout the experiment.
At the end of the experiment (d21), birds were fasted for 15 h before being weighed and euthanised. Feed intake (FI) and feed conversion ratio were recorded once for the period 0–21 d. A total of twelve birds (two per pen replicate) with BW as close as possible to the average pen weight were euthanised by CO2 inhalation, and the muscle samples (pectoralis major) were collected and stored at − 20°C until analysis.
Expt 2: apparent digestibility of selenium
The present experiment aimed to compare the apparent digestibility of Se (ADSe) from different Se sources. The study was divided into an adaptation phase and an excreta collection phase. During the adaptation phase, a total of 90 day-old chicks obtained from a commercial hatchery were allocated to three treatments, with three pen replicates of ten birds from d0 to d13.
The three starter diets used in the present experiment (SS-0·3, SY-0·3 and SO-0·3) were supplemented with different Se sources SS, SY and SO at a concentration of 0·3 mg Se/kg feed.
The feed was provided ad libitum as crumbles from d0 to d13. Each pen of approximately 3 × 2 m was covered with wood shavings and contained one hanging bell drinker and one hanging tube feeder. The house was maintained at 32°C from d0 to d6 and at 28°C from d6 to d13. The lighting schedule followed 23 h light–1 h dark cycle during this period. On d13, twelve birds per treatment were selected (average weight 347 (sd 21) g) and placed in individual cages kept in an environment-controlled room following a complete block design and factorial arrangement, with a 14 h light–10 h dark cycle lighting schedule. Average BW was kept constant among the treatments. During d13 to d24, each bird had free access to water and the same experimental diet as during d0 to d13, which was supplemented as pellets. Individual FI and bird BW were determined. During the period d20 to d23, a digestibility procedure was carried out according to a modified method of Bourdillon et al. (Reference Bourdillon, Carré and Conan12). For each cage, total excreta were weighted and collected daily during the 3 d collection period and stored at − 18°C before being pooled and freeze dried. Dried faeces were ground and homogenised after a 24 h water recovery period before analysis. Individual FI data and faeces collection data were used to determine ADSe according to the following calculation:
$$\begin{eqnarray} AD_{Se}\,(\%) = (Se\,input\,(mg) - Se\,output \ (mg))/Se\, input \ (mg), \end{eqnarray}$$
where
$$\begin{eqnarray} Se\,input = collection\,period\,FI\,(g)\times analytical\,feed\,Se\,concentration\,(mg\,Se/kg\,feed)/1000, \end{eqnarray}$$
$$\begin{eqnarray} Se\,output = collection\,period\,faeces\,weight\,(g\,DM)\times faeces\,Se\,concentration\,(mg\,Se/kg\,dry\,product)/1000. \end{eqnarray}$$
After collecting the faeces at d24, all birds were euthanised by CO2 inhalation and the muscle samples (pectoralis major) were collected and stored at − 20°C before analysis.
Selenium analysis
The measurement of total Se in each feed and faeces sample was carried out on ‘as is’ basis, whereas muscles were freeze dried, mixed and sieved before analysis, with results being reported on a DM basis.
Total selenium measurement
Total Se measurements were performed according to the method of Mester et al. (Reference Mester, Willie and Yang13). Briefly, total Se concentration in feed samples was determined by mineralisation of 1 g of sample in a mixture of 4 ml of 69–70 % HNO3 and 2 ml of 35 % H2O2 at 85°C for 4 h within a closed vessel heating block system (DigiPrep; SCP Science). For tissue samples, the mass uptake was reduced to 250 mg, digested by 2 ml of HNO3 and 1 ml of H2O2. The solution was further diluted with water and the total Se content subsequently determined by inductively coupled plasma MS (Agilent 7500cx). Isotopes 76, 77 and 78 were used for quantification. The standard addition method was used.
Tissue speciation measurement (selenomethionine, selenocysteine)
Speciation of SeMet and SeCys amino acids was carried out according to the method described by Bierla et al. (Reference Bierla, Dernovics and Vacchina14). Briefly, SeCys was reduced and alkylated to be stabilised. It was subjected to proteolytic digestion to release free amino acids that were purified by size-exclusion liquid chromatography. Both amino acids were then quantified by reversed phase HPLC–inductively coupled plasma MS.
Tissue 2-hydroxy-4-methylselenobutanoic acid measurement
In order to assess the conversion of the HMSeBA (the active compound of SO) to SeMet and further to SeCys, a method was implemented to detect HMSeBA in muscle samples.
A measure of 0·5 g of tissue sample was extracted by 5 ml of phosphate buffer (pH 7·5) under mechanical shaking (200 rpm, 1 h). The supernatant was separated by centrifugation (2093 g, 5 min) and analysed by reverse phase HPLC–inductively coupled plasma MS, according to the method described in detail elsewhere(Reference Vacchina, Moutet and Yadan15).
Statistical analyses
All results (growth performance and Se content) were analysed using SAS 9.1.3 (Copyright ©) 2002–3 by SAS Institute, Inc. All Rights Reserved).
The Se dose–response for total muscle Se content was tested for linearity for the three Se sources with the F test.
The growth performance data were analysed with PROC GLM and least square means were grouped using the Adjust = Tukey option.
In all analyses, the normality and homoscedasticity were checked on Studentised residuals. If needed, logarithmic transformation was performed on the data.
The relative bioavailability of SO v. SY in Expt 1 was evaluated using a five-point slope ratio design (NC, SY-0·1, SY-0·3, SO-0·1 and SO-0·3), according to Finney(Reference Finney16). As stated by Littell et al. (Reference Littell, Henry and Lewis17), a non-linear model (in the parameters) was fitted to the data using the PROC NLIN from SAS. The model was as follows:
$$\begin{eqnarray} Se\lowbar dry\lowbar muscle = a + a _{0}\times X _{0} + b _{S}\times ( b _{TS}\times Dose_{SO} + Dose_{SY}), \end{eqnarray}$$
where Se_dry_muscle is the content of Se in the muscle (in mg/kg of dry product), a is the intercept (a 0× X 0 is a correction for the NC), doses SO and SY are the Se amount from SO and SY sources, b S is the slope for the effect of SY on the response and b TS is the ratio between b T (the slope for the effect of SO) and b S. This allows obtaining directly an estimate of the relative biological value (i.e. the ratio between slopes b S and b T) and its CI.
Results
Selenium forms and concentrations in the different diets
Feed Se analyses indicated concentrations slightly above the expected levels (Table 2). Only the SO-0·3 starter diet of Expt 1 had a lower concentration than the added value; all other higher Se contents can be explained based on Se present naturally in the feed ingredients, as determined in the NC group diet. The results did not indicate major discrepancies between treatments of the same targeted Se level for different Se sources.
Table 2 Feed selenium source, level and analysis for the different diets (Mean values and 95 % confidence intervals)
NC, negative control; SS, sodium selenite; SS-0·1, SS at 0·1 mg Se/kg feed; SS-0·3, SS at 0·3 mg Se/kg feed; SY, seleno-yeast; SY-0·1, SY at 0·1 mg Se/kg feed; SY-0·3, SY at 0·3 mg Se/kg feed; SO, Selisseo®; SO-0·1, SO at 0·1 mg Se/kg feed; SO-0·2, SO at 0·2 mg Se/kg feed; SO-0·3, SO at 0·3 mg Se/kg feed.
* Feed Se level assessed on one sample aliquot (CI: 2 × standard deviation (sd) of the duplo analytical measurement).
† Feed Se level assessed on four sample aliquots (CI: t× sd/n 0·5; where t= 3·18 or t (P= 5 %; n− 1 df)) with n 4).
Expt 1
Performance parameters (final BW, FI, feed conversion ratio) given in Table 3 were not affected by treatments (P>0·05). However, the different Se sources and levels induced different muscle Se concentrations (P< 0·05) (Fig. 2). As expected, the NC treatment induced the lowest Se concentration and all 0·3 mg Se/kg feed treatments resulted in higher muscle Se concentrations than 0·1 mg Se/kg feed treatments.
Table 3 Effect of the different selenium sources and levels on growth performances of broiler chickens (Expt 1)
NC, negative control; SS-0·1, sodium selenite at 0·1 mg Se/kg feed; SS-0·3, sodium selenite at 0·3 mg Se/kg feed; SY-0·1, seleno-yeast at 0·1 mg Se/kg feed; SY-0·3, seleno-yeast at 0·3 mg Se/kg feed; SO-0·1, Selisseo® at 0·1 mg Se/kg feed; SO-0·2, Selisseo® at 0·2 mg Se/kg feed; SO-0·3, Selisseo® at 0·3 mg Se/kg feed; RSD, residual standard deviation; BW, body weight; d21, day 21; d0, day 0; FI, feed intake; FCR, feed conversion ratio.
* Not subjected to statistical analysis.
Fig. 2 Mean values for selenium in muscles at day 21 of broiler chickens fed different selenium sources and levels. a,b,c,d,e,f,gMean values with unlike letters were statistically different (P< 0·05) and determined after logarithmic transformation to fulfil the variance homogeneity requirement (Expt 1). The error bar represents the standard error of the mean. NC, negative control; SS-0·1, sodium selenite at 0·1 mg selenium/kg feed; SS-0·3, sodium selenite at 0·3 mg selenium/kg feed; SY-0·1, seleno-yeast at 0·1 mg selenium/kg feed; SY-0·3, seleno-yeast at 0·3 mg selenium/kg feed; SO-0·1, Selisseo® at 0·1 mg selenium/kg feed; SO-0·2, Selisseo® at 0·2 mg selenium/kg feed; SO-0·3, Selisseo® at 0·3 mg selenium/kg feed.
Whatever the doses considered, birds fed organic Se sources (SY and SO) exhibited a significantly higher muscle Se concentration than mineral Se source-fed birds. Moreover, a significant difference (P< 0·05) in muscle Se concentration was observed between organic sources, with SO inducing higher muscle Se concentration than SY, when supplemented at both 0·1 and 0·3 mg Se/kg feed. The intermediate dose of 0·2 mg Se/kg feed from SO resulted in equivalent muscle Se concentration than SY supplemented at 0·3 mg Se/kg feed (P>0·05).
Seleno-amino acid speciation results (Fig. 3) indicated that muscle Se is mainly present as SeMet and SeCys, as total Se concentrations were not different from SeMet+SeCys summation concentrations (P>0·05). Within each Se source, SeMet and SeCys amounts were not significantly different (P>0·05); however, at 0·3 mg Se/kg feed, SO induced 54 % more SeCys than SY (P< 0·05). Moreover, residual HMSeBA in muscle from animals fed SO was not detected, revealing a concentration below the quantification limit of 0·01 mg/kg (Se equivalent) dry product for the six tested samples.
Fig. 3 Muscle total selenium, selenomethionine (SeMet) and selenocysteine (SeCys) concentrations of broiler chickens fed a control diet (NC) or supplemented with 0·3 mg selenium/kg feed from seleno-yeast or Selisseo®. a,b,c,d,e,fMean values with unlike letters were significantly different (P< 0·05). Error bars represent the standard deviation (n 6) (Expt 1). SY-0·3, seleno-yeast at 0·3 mg selenium/kg feed; SO-0·3, Selisseo® at 0·3 mg selenium/kg feed. □, Total selenium; 
, SeMet; 
, SeCys; ■, SeMet+SeCys.
The relative bioavailability of the different Se sources can be assessed by the efficiency of SY and SO to increase muscle Se concentration relative to feed Se supply. First, the deviation to linearity was tested for the three dose–responses of the Se sources studied with the F test. The results obtained showed that linearity is verified for SY and SO (P< 0·05), but not for SS (P>0·05). Differences within slope ratio indicated an increased muscle Se concentration of 1·39 (95 % CI 1·28, 1·49) between SO and SY, showing a bioavailability improvement of 39 % with SO compared with SY.
Expt 2
Final BW, FI and feed conversion ratio parameters were not affected by the different Se sources (P>0·05) between d13 and d23 (data not shown). Muscle Se concentrations indicated significant improvement with organic Se sources (SY and SO) compared with SS, with an additional increase with SO (P< 0·05) (Table 4). Convergently, faeces Se outputs were significantly lower for SY and SO compared with the SS Se source (P< 0·05). Notably, the highest ADSe was obtained for SO. Then, SY and SO induced significantly higher ADSe compared with the SS (P< 0·05).
Table 4 Mean values for body weight, apparent digestibility of selenium and muscle selenium concentration of Expt 2*
SS-0·3, sodium selenite at 0·3 mg Se/kg feed; SY-0·3, seleno-yeast at 0·3 mg Se/kg feed; SO-0·3, Selisseo® at 0·3 mg Se/kg feed; BW, body weight; d23, day 23; FI, feed intake; d20, day 20.
a,b,cMean values between columns with unlike superscript letters were significantly different (P< 0·05) and determined after logarithmic transformation to fulfil the variance homogeneity requirement.
* Data for each treatment are the mean of twelve measurements.
Discussion
Se is an essential element for antioxidant enzymes important for detoxifying reactive oxygen species and/or lipid peroxide originating from oxidative stress(Reference Tsai18, Reference Reeves and Hoffmann19). These results are consistent with most of the studies comparing the effect of SS and SY on growth performances showing no effect caused by Se supply(Reference Yoon, Werner and Butler5, Reference Payne and Southern20–Reference Juniper, Phipps and Bertin22). Other studies containing deficient or non-Se-supplemented diets indicated that Se supplementation is required for optimal growth performances(Reference Upton, Edens and Ferket23–Reference Ševčíková, Skrivan and Dlouhá26). In the present study, the non-Se-supplemented diet did not result in altered growth performances compared with the ROSS PM3 broiler guide, meaning that the Se level in the NC diet cannot be considered as deficient. Moreover, depending on the breeder selenium status, considered as standard (0·3 mg/kg feed from SS), the duration of the experiment and the controlled environmental conditions, the present experiments did not induce performance modifications.
As expected, muscle Se concentrations appeared to be significantly increased by different Se additives, and a consistent improvement was observed from organic Se sources compared with SS source or control diet. These results agree with previous results from different authors(Reference Choct, Naylor and Reinke27–Reference Pan, Huang and Zhao29) who reported similar muscle Se concentration enhancement with SS and SY sources and corroborate other comparable studies(Reference Wang and Xu24, Reference Payne and Southern25, Reference Kuricova, Levkut and Boldizarova30–Reference Dlouhá, Ševčíková and Dokoupilová32).
Over the muscle Se concentration, measurements showed a significantly lower ADSe for SS compared with the organic Se sources at 0·3 mg Se/kg feed. The different absorption routes involved mainly explain the digestibility discrepancy between mineral and organic Se sources. Indeed, results from Wolffram et al. (Reference Wolffram, Ardüser and Scharrer33) indicated a passive intestinal absorption of selenite, whereas SeMet absorption from the intestine followed an active transport, common with methionine(Reference Wolffram, Berger and Grenacher34). These results are supported by other studies(Reference Ševčíková, Skrivan and Dlouhá26), obtaining comparable results for muscle and excreta Se concentrations in broiler chickens. Results from Yoon et al. (Reference Yoon, Werner and Butler5) showed Se retentions of 79·0 or 72·1 % for two SY sources and 68·7 % for SS in broiler chicks from d0 to d21 when fed with 0·3 mg Se/kg feed. In their study, the authors showed significant differences for Se retention between mineral and one of the SY tested. These results demonstrated a higher level for Se retention than in our work, but tend to confirm the higher Se retention for SY compared with SS. Due to limited information on the retention calculation method used in the study of Yoon et al. (Reference Yoon, Werner and Butler5), further comparisons are limited. Choct et al. (Reference Choct, Naylor and Reinke27) also determined muscle and faeces Se concentrations on chicks fed organic and mineral Se sources. These authors also concluded with higher muscle Se concentration and lower faeces Se concentration from organic Se source compared with mineral sources. The difference on ADSe between our values and the literature could be explained by the duration of retention test, the analytical method used for Se measurement and/or the retention calculation method used. However, the present results agree with lower excretion obtained with organic Se sources compared with mineral Se sources and confirm higher Se retention, as demonstrated by the ADSe measurements. Considering these results, the difference between mineral and organic Se sources can be partly explained through better apparent digestibility of organic Se sources. However, the significant difference for muscle Se concentration between SY and SO cannot be totally explained through better digestibility, since no significant differences were noted on that parameter. Suzuki(Reference Suzuki35) described Se metabolism through the metabolomic approach and indicated that mineral sources (e.g. selenate and selenite) were able to form SeCys through selenide compound (HSe−). This selenide appeared as a metabolic crossroad between organic (e.g. SeMet) and mineral Se for SeCys formation through selenophosphate and Serine-tRNA. However, if mineral Se sources are able to form selenoproteins, they cannot revert back to SeMet. An enrichment of the muscle SeMet does not affect protein structure or properties and represents an endogenous Se pool available in challenging conditions due to environmental or physiological stress(Reference Schrauzer1, Reference Surai7). As an example, a study with broiler chickens fed organic or mineral Se(Reference Payne and Southern25) demonstrated that endogenous Se could be released from tissues, and, thus, that organic Se sources were more efficient in maintaining the glutathione peroxidase level. The muscles are mainly made of structural protein accretion and no distinction is made by the cell between methionine and SeMet during protein synthesis(Reference Schrauzer1); thus, SeMet from organic Se sources can be readily incorporated into structural proteins. In addition, Thiry et al. (Reference Thiry, Ruttens and De Temmerman36) reported that part of the Se coming from SeMet is excreted as a volatile compound (e.g. dimethyl selenol) through breath. These Se losses were not taken into account in the present study; organic Se sources and HMSeBA, like 4-hydroxy-2-methylthiobutanoic acid, could be more rapidly oriented to SeCys and hence be less exposed to excretion routes. Taking into account these differences between Se sources, the difference of muscle Se concentration and ADSe between mineral and organic forms appeared as the resultant of that specific metabolism and/or absorption, leading to lower storage and higher excretion. Moreover, SY and SO sources induced a linear response for muscle Se concentration, indicating large seleno-amino acid retention, while this retention was very limited for mineral sources. Our observations are in good agreement with the different metabolic pathways between mineral and organic Se sources(Reference Rayman, Infante and Sargent2, Reference Suzuki and Ogra37).
When comparing the two organic Se sources, no significant differences were obtained on ADSe. Separately carried out tolerance and toxicological studies (data not shown) did not indicate particular threat from the new Se source. However, Se muscle concentrations significantly improved with SO, increasing the relative bioavailability for total Se by 39 % compared with SY. From a practical point of view, the better bioavailability of SO compared with SY can be explained through the different product forms. Indeed, Se from SY is mainly present as organic Se, but only 60 % of the total Se is present as SeMet, whereas SO is an almost pure product of selenomethionine hydroxyanalogue (purity >99 % HMSeBA). This difference between products partly explains the variations of bioavailability observed between those two organic Se sources. Se and seleno-amino acid measurements from Vignola et al. (Reference Vignola, Lambertini and Mazzone38) obtained on lambs indicated a higher efficiency of SY to improve muscle Se concentration, but also that Se enrichment from SY was mainly due to the SeMet part in the SeMet:SeCys ratio compared with the mineral source. Hence, the present results confirmed these observations and also indicated additional bioavailability improvement with SO compared with SY. In the present study, the better efficiency of SO to improve SeCys may be related to the overall higher efficiency of SO to improve muscle Se content or to a specific pathway that may favour higher proportions of SeCys when Se originates from HMSeBA compared with SeMet from yeasts. The results obtained with HMSeBA can be linked to previous results obtained with its sulphur analogue, 4-hydroxy-2-methylthiobutanoic acid, which was demonstrated to be more oriented to trans-sulphuration pathway than dl-methionine(Reference Martín-Venegas, Geraert and Ferrer39). The difference of affinity for trans-sulphuration pathway between methionine sources led to obtain higher synthesis of cysteine and taurine compared with l-methionine source(Reference Martín-Venegas, Geraert and Ferrer39). Hence, according to the obtained results, it can be speculated that HMSeBA and SeMet follow similar conversion pathways than 4-hydroxy-2-methylthiobutanoic acid and methionine, as described by Dibner & Knight(Reference Dibner and Knight40), and that SeMet obtained from HMSeBA can be more efficiently oriented to trans-selenation pathway, allowing higher SeCys species in the muscle. However, it is not clear whether the conversion of SeMet to SeCys results from direct trans-selenation of SeMet or rather from an indirect conversion through selenophosphate. Further work is evidently needed in order to elucidate the specific metabolic pathway of HMSeBA and its selenised compounds' fate. However, the improved SeCys level in skeletal muscle could be indicative of an improvement of selenoprotein status available for antioxidant functions. Indeed, the SeCys status can be considered as a global indicator of the immediate antioxidant status, because it represents the proportion of active selenoproteins. SeCys plays a major role in the human and animal health; it is incorporated specifically in selenoprotein, inducing particular functions to the protein(Reference Allmang, Wurth and Krol11). Some of the most studied selenoproteins, like glutathione peroxidases(Reference Pappas, Zoidis and Surai10, Reference Rotruck, Pope and Ganther41) and thioredoxine reductases(Reference Lu, Berndt and Holmgren42), are known for their involvement in the animal antioxidant status. Hence, the SeMet, and particularly the SeCys, improvement observed with SO should benefit to maintain animal redox homeostasis.
Further investigations are needed to elucidate the pathways of the different Se sources, and to study the interest of an improved SeMet or SeCys seleno-amino acid form in muscle structural protein or other tissues. To conclude, the organic Se source SO participates efficiently in broilers' muscle Se enrichment, which improves the oxidative stress resistance of the bird.
Acknowledgements
The present study was supported by ADISSEO France S. A. S., which set-up the experiment in its own facilities. M. B., Y. M., F. R. and P.-A. G., contributed to the conception and design of the study and participated in the final interpretation of the data. M. B. and Y. M. were responsible for the planning and data collection. V. V., was responsible for the selenium analysis of the samples. F. R. was responsible for the statistical analysis of the data. All authors participated in the writing of the final draft of the manuscript and agreed with the final format. The authors greatly acknowledge Dr Alain Lescure, from the University of Strasbourg, for precious advice on the manuscript. The authors state that there are no conflicts of interest.
