Alpha-linolenic acid (ALA; 18:3 n-3) is an important factor that contributes to the nutritional quality of milk fat (Barceló-Coblijn & Murphy, Reference Barceló-Coblijn and Murphy2009). Transfer of this fatty acid from feed to milk is therefore significant for the production of high-quality dairy products. Much of the ALA ingested is biohydrogenated to stearic acid in the rumen (cf. Chilliard et al. Reference Chilliard, Glasser, Ferlay, Bernard, Rouel and Doreau2007), which may explain why the concentrations of ALA in milk fat can be largely independent of the amount of ALA consumed by cows (Leiber et al. Reference Leiber, Scheeder, Wettstein and Kreuzer2004, Reference Leiber, Kreuzer, Nigg, Wettstein and Scheeder2005).
Dietary linoleic acid (LA; 18:2 n-6) is also subject to intensive ruminal biohydrogenation (Jouany et al. Reference Jouany, Lassalas, Doreau and Glasser2007; Moate et al. Reference Moate, Boston, Jenkins and Lean2008). However, as LA is a precursor of arachidonic acid (AA; 20:4 n-6), it is essential that a portion passes unaltered through the rumen. Biohydrogenation of LA not only decreases the amount that can bypass the rumen, but it also elevates the occurrence of a nutritionally relevant (Benjamin & Spener, Reference Benjamin and Spener2009) fatty acid (FA), rumenic acid (RA; cis-9, trans-11 18:2), in milk. This conversion occurs either directly (Jouany et al. Reference Jouany, Lassalas, Doreau and Glasser2007) or indirectly via the generation of vaccenic acid (VA; trans-11 18:1; Chilliard et al. Reference Chilliard, Glasser, Ferlay, Bernard, Rouel and Doreau2007).
When feeding effects on milk FA patterns are assessed, several external factors, in addition to the FA supply, that may influence the rate of biohydrogenation also need to be considered. Among these are the energy supply (Leiber et al. Reference Leiber, Kreuzer, Nigg, Wettstein and Scheeder2005), the dietary proportion of concentrates (Tsiplakou & Zervas, Reference Tsiplakou and Zervas2008), and the proportion of fibre (Dewhurst et al. Reference Dewhurst, Shingfield, Lee and Scollan2006). These are mutually dependent factors, since a higher concentrate proportion is usually associated with a higher energy density and lower fibre concentration. The use of concentrate may also change the FA profile of the entire diet, thus causing further direct and indirect effects on milk fat composition.
The aim of the present study was to determine the roles of the roughage to concentrate proportion and the concentrate type on FA profiles of milk fat under exclusion of alterations in dietary energy and protein. For this purpose, milk samples were obtained from an earlier dairy cattle experiment that had employed three different isoenergetic diets (based on ryegrass, barley or maize; Klevenhusen et al. Reference Klevenhusen, Bernasconi, Kreuzer and Soliva2010). These diets provided an opportunity to examine the influence of different proportions of concentrate (and fibre), as well as different intakes of LA and ALA from roughages and grains. Additional emphasis was placed on the prevalence of ALA, LA and their biohydrogenation products in milk, in relation to ALA and LA intake.
Materials and Methods
Animals, experimental design and diets
The experiment was approved by the Swiss governmental authority for animal welfare. The experiment was based on a completely randomised design and included 18 medium-to-late lactating dairy cows (Holstein Friesian and Brown Swiss breeds) allocated in a balanced manner for milk yield and stage of lactation to the three isoenergetic and isonitrogenous diets (n=6). These diets were fed for 22 d; data and milk samples were collected for the last 8 d (Klevenhusen et al. Reference Klevenhusen, Bernasconi, Kreuzer and Soliva2010). The concentrations of net energy for lactation (NEL) and crude protein (CP) were 5·4±0·05 MJ/kg dry matter (DM) and 213±3·5 g/kg DM, respectively (Table 1). One diet consisted of ryegrass hay (diet H). A second diet (diet M) comprised (expressed as g per kg DM) maize stover (444), whole maize pellets (368), maize gluten (151), molasses (20), and urea (14). The third diet (diet B) was composed of barley straw (459), crushed barley grain (266), soybean meal (238), molasses (21), and urea (8). While diet H represents a roughage-only diet, diets M and B represent roughage to concentrate diets with ratios of about 1:1. All diets were supplemented with 16 g/kg of a commercial dairy vitamin-mineral premix. Diets M and B additionally contained 2 and 1 g MgSO4/kg respectively. More details on the diets are described in Klevenhusen et al. (Reference Klevenhusen, Bernasconi, Kreuzer and Soliva2010). The total FA contents and the FA profiles showed that diets H and M represented a moderate lipid supply, with diet H being high in ALA and diet M being high in LA, while diet B represented a low lipid supply with proportionally high LA content.
Table 1. Ratio of dietary roughage:concentrate, nutrient content and energy content as well as fatty acid profile and intake of experimental dairy cow diets
Prior to the experiment, the cows had been fed hay ad libitum plus 3 kg barley grain/d. They were gradually switched, over 6 d, to the respective experimental diets. Milk samples were collected at the first day of this pre-experimental time. During the 22 d of the experiment (14 d of adaptation and 8 d of data collection), the cows were offered the roughages, i.e. ryegrass hay, maize stover or barley straw, ad libitum. The concentrates in diets M and B were supplemented in five equal portions over the day. Concentrate amounts were adjusted daily to the amount of roughage consumed on the day before, to maintain roughage to concentrate ratios at the desired NEL and CP proportions. Feed intake and milk yield were measured daily and cows were milked twice per day. During the 8 d data collection period, morning and evening milk samples from each cow were pooled daily and stored at −20°C. After thawing, milk samples from the 8 d were pooled to one sample per cow and refrozen at −20°C. Samples of the individual diet ingredients, except urea and molasses, were collected six times over the 8 d and frozen.
Nutrient composition of feeds
Feeds were analysed for DM (TGA-500, Leco Corporation, St. Joseph, Michigan, USA) and CP (N×6·25; C/N-analyser (Leco-Analysator Type FP-2000, Leco Instrumente GmbH, Kircheim, Germany). Neutral (NDF) and acid detergent fibre (ADF) were quantified according to Van Soest et al. (Reference Van Soest, Robertson and Lewis1991), correcting for ash content and using α-amylase and sodium sulphite for the NDF analysis. Contents of NEL were calculated from tabulated values (ALP, 2008).
Fatty acid analysis
Lipids were extracted from the feed samples by accelerated solvent extraction (ASE 200, Dionex Corporation. Sunnyvale, CA, USA) with hexane:2-propanol (3:2; v/v). An internal standard (19:0) was added and the solvent was evaporated under a stream of N2 gas. The residue was dissolved in dichloromethane and dried again under N2. The dried samples were then methylated using 2 m-methanolic sodium hydroxide and 1·3 m-boron trifluoride in methanol. Dyes and non-FA-methyl esters were removed by thin layer chromatography (Khiaosa-Ard et al. Reference Khiaosa-Ard, Bryner, Scheeder, Wettstein, Leiber, Kreuzer and Soliva2009). The FA composition was determined by gas chromatography (GC; model HP 6890 equipped with a FID detector, Hewlett-Packard, Palo Alto CA, USA) on a 30 m×0·32 mm Supelcowax-10IM capillary column (Supelco Inc., Bellefonte PA, USA). A mixed FA methyl ester (FAME) standard (Supelco 37 Component, Supelco Inc., Bellefonte PA, USA) was used for the identification of individual FAs and to quantify total FAs.
Frozen milk samples were thawed at 37°C in a water bath, then subjected to a base-catalysed trans-esterification, as described by Suter et al. (Reference Suter, Grob and Pacciarelli1997). Briefly, 0·5 ml milk was added to 5 ml mixed internal standard of trivalerin (5:0), nonanoic acid methyl ester (9:0), triundecanin (11:0), and 1-tetradecene (14:1) dissolved in 1,4 dioxane. Trans-esterification was carried out by adding 5 ml sodium methylate solution and vortexing for 3 s. The mixture was allowed to react for exactly 60 s, then 4 ml heptane and 10 ml 15% di-sodium citrate solution were added to stop the reaction. After phase separation had occurred, the upper phase was retrieved for FA analysis by GC and capillary column as described for FAs in feeds. The major FAs were identified based on a mixed FAME standard (Supelco 37 Component, Supelco Inc., Bellefonte PA, USA). Fatty acid concentration was calculated from the known amounts of the internal standard FAs (5:0 and 11:0). A second capillary column (200 m×0·25 mm; CP7421, Varian Inc., CA, USA) was used on an identical GC to distinguish between cis and trans 18:1 isomers. Hydrogen was used as the carrier gas.
Statistical analysis
Analysis of variance of the milk data was performed with Procedure GLM of SAS (version 9.1, SAS Institute Inc., Cary NC, USA) with diet type (fixed factor) as the source of variance. The pre-experimental values, except for the ratios of FA intake and secretion with milk, were included as covariates to correct for initial differences within cows. Tukey's method was used for multiple comparisons among means. Milk data presented in the tables are Least Square Means, standard errors of the means and P-values. One cow of group M had to be excluded from evaluation due to feed intake problems.
Results
Feed DM intake was similar among the three diets, at 13·5, 13·4 and 15·1 kg/day. The total amounts of ingested FAs were similar between diets H and M. However, in diet B, FA ingestion was about 40% lower than in the other treatments. Intake of LA was highest with diet M and lowest (−70%) with diet H, whereas ALA intake was ten times higher with diet H, compared with diets M and B (Table 1). With diet M, a considerable amount of oleic acid (cis-9 18:1) was ingested.
Milk and milk fat yield did not significantly differ among the three diets (Table 2). There was a trend toward a higher milk fat concentration with diet M compared with the other two diets (P<0·10). The proportions of all short- and medium-chain saturated FAs in milk fat were significantly affected by diet type, but not in any clearly systematic manner. Proportions of the two major medium-chain FAs, 14:0 and 16:0, in milk fat were lower when cattle were fed diets H and M compared with diet B. The two monounsaturated medium-chain FAs, 16:1 and 17:1, in milk fat were not affected by diet type.
Table 2. Effect of diet type on yield of milk and milk fat and on composition of short- and medium-chain fatty acids in milkFootnote †
† Means carrying no common superscript are different at P<0·05
The proportion of 18:0 was about 38% higher in milk fat of cows fed diet H compared with the other two treatments (Table 3). With diet M, milk fat proportions of trans 18:1 FAs were significantly higher than with diet B, but were similar for with diet H, except for trans-9 and trans-10 18:1 (both of these were significantly lower with diet H). For cis-9 to cis-13 18:1, concentrations in milk fat were highest with diet M; while proportions of cis-14 to cis-16 18:1 were highest with diet H. The concentration of LA in milk fat was 133% higher in diet M compared with H (P<0·05), whereas no statistically significant difference was found for LA between H and B. The milk fat proportion of RA (cis-9, trans-11 18:2) was significantly different among diets in the order of M>H>B. With diet H, ALA proportion in milk fat was threefold higher than with diets M and B. The longer chain n-6 FA, 20:3 and AA in milk fat occurred in significantly higher proportions with diets M and B compared with H, while for the n-3 category, 20:3 and 20:5 (eicosapentaenoic acid, EPA), higher levels were found with diet H than with diets M and B. Consequently, the ratio of n-6:n-3 was substantially higher with diets M and B than with diet H. The ratio of RA to VA (trans-11 18:1) in milk fat, partly reflecting the degree of mammary desaturation of VA, was significantly lower in diet H compared with the diets M and B.
Table 3. Effect of diet type on composition of long-chain fatty acids in milk (g/kg total FAME) and on selected milk fatty acid ratiosFootnote †
† Means carrying no common superscript are different at P<0·05
When compared with the amounts of ALA ingested daily, the secretion of ALA, RA, VA, EPA and total n-3 FA in milk fat was the lowest with diet H (Table 4), even though milk fat ALA and EPA concentrations were highest with this diet. When daily LA ingestion was compared, the relative secretion of LA, RA, VA, AA, and total n-6 FA was higher with diet H compared with diets M and B. The amounts of 18:0 secreted in milk were approximately six to seven times higher than the amounts ingested with diets B or H. For diet M, this ratio was significantly lower.
Table 4. Ratios of fatty acid amounts secreted with milk (g/d) to that of either 18:3 n-3 or 18:2 n-6 or 18:0 ingested with the diet (g/d)Footnote †
† Means carrying no common superscript are different at P<0·05
Discussion
Influences of dietary factors on the transfer of dietary ALA and LA from feed to milk fat
Comparison of the roughage only diet (hay, H) with the two mixed roughage and concentrate diets (maize, M; barley, B) enabled investigation of the effects of concentrates, independent of the energy and protein density, because these diets were isoenergetic and isonitrogenous. The lower LA and the higher ALA concentrations in milk fat of group H, compared with groups M and B, clearly corresponded to the different intake levels of the respective FAs with these diets. Thus, the dietary amount of individual FAs was a major factor in determining the ultimate ALA and LA concentrations in milk fat. However, the occurrence of ALA in milk, relative to its intake, was lowest in the group fed the hay diet. This finding was in agreement with experiments of Shingfield et al. (Reference Shingfield, Salo-Väänänen, Pahkala, Toivonen, Jaakkola, Piironen and Huhtanen2005) and of Van Dorland et al. (Reference Van Dorland, Kreuzer, Leuenberger and Wettstein2008) who also showed decreased ALA recovery rates in milk following increased ALA intakes. The dietary fibre concentration, which possibly could have played a role in the present study and in the study of Shingfield et al. (Reference Shingfield, Salo-Väänänen, Pahkala, Toivonen, Jaakkola, Piironen and Huhtanen2005), was not varied in the study of Van Dorland et al. (Reference Van Dorland, Kreuzer, Leuenberger and Wettstein2008) and was ruled out as an influencing factor by AlZahal et al. (Reference AlZahal, Or-Rashid, Greenwood, Douglas and McBride2009).
The differing levels of easily degradable starch could also have been a confounding factor in the present study. However, Leiber et al. (Reference Leiber, Scheeder, Wettstein and Kreuzer2004, Reference Leiber, Kreuzer, Nigg, Wettstein and Scheeder2005) and Van Dorland et al. (Reference Van Dorland, Kreuzer, Leuenberger and Wettstein2008) showed that lower ALA and LA intakes were always associated with higher relative recovery rates of these FA in milk, and vice versa. Differing dietary starch levels had not been applied in those studies indicating, that this effect occurs independently of starch.
In alpine studies (Leiber et al. Reference Leiber, Scheeder, Wettstein and Kreuzer2004, Reference Leiber, Kreuzer, Nigg, Wettstein and Scheeder2005) that compared roughages from lowland and alpine pastures, differences in FA transfer were confounded by differences in feed energy density. A lack of energy at the rumen level was hypothesised to have limited the capacity for microbial biohydrogenation (Leiber et al. Reference Leiber, Kreuzer, Nigg, Wettstein and Scheeder2005). In the present study, care was taken not to alter energy densities between the diets. Thus, the compensatory increase of LA or ALA recovery in milk could be clearly attributed to the lower intakes of these FAs.
Comparing the milk fat composition of the cows fed diets M or B revealed the same effect of ALA and LA intake on the respective recovery rates. With these diets, the roughage to concentrate ratio (an effect found to be influential by Tsiplakou & Zervas, Reference Tsiplakou and Zervas2008), the dietary fibre and the starch proportions were all constant. Again, the main factor altering LA and ALA transfer to milk was the amounts of these dietary FAs that were ingested.
Influences of dietary factors on biohydrogenation products in milk
Vaccenic acid is an intermediate product of ruminal biohydrogenation of both LA and ALA (Jenkins et al. Reference Jenkins, Wallace, Moate and Mosley2008). Vaccenic acid was significantly elevated in milk fat with diets H and M when compared with diet B, indicating that the amount of plant-derived PUFAs ingested, rather than the kind of dietary FA (LA or ALA) or the proportion of concentrate or fibre as such, determined the VA concentration in milk fat. This was underlined by the rather constant ratio of milk VA to the sum of LA and ALA ingested across all three diets. However, ruminal VA is also the major precursor for milk RA, which is partly synthesised by Δ9 desaturation in the mammary gland of lactating cows (Griinari et al. Reference Griinari, Corl, Lacy, Chouinard, Nurmela and Bauman2000). Therefore, milk VA levels do not reflect the entire amount of VA derived from ruminal biohydrogenation, and this limits use of VA as an indicator of ruminal processes.
Unlike the VA levels, RA proportion was higher in milk fat with diet M compared with diet H. In the rumen, RA is the first product of the ruminal biohydrogenation pathway of LA (Chilliard et al. Reference Chilliard, Glasser, Ferlay, Bernard, Rouel and Doreau2007; Jenkins et al. Reference Jenkins, Wallace, Moate and Mosley2008) and total RA may be hydrogenated to VA (Moate et al. Reference Moate, Boston, Jenkins and Lean2008). In contrast, ruminal biohydrogenation of ALA does not result in RA as a major intermediate (Jenkins et al. Reference Jenkins, Wallace, Moate and Mosley2008). Thus, the lower RA concentration found in milk fat following consumption of diet H, which had a high ALA content and less LA, suggests that not only Δ9 desaturation in the mammary gland but also ruminal RA levels are decisive in determining the final RA concentration in milk fat. A high dietary proportion of a roughage rich in ALA therefore does not necessarily lead to an increased RA concentration in milk even if it increases VA concentrations in milk. Rather, a high dietary intake of LA would appear to be required for RA increases to be observed.
The low proportion of trans-10 18:1 in milk fat with diets B and H (which both supplied lower amounts of LA) compared with diet M (which supplied a higher LA amount) suggests that this isomer was mainly a product of incomplete ruminal LA biohydrogenation. This suggestion is consistent with the literature (Jouany et al. Reference Jouany, Lassalas, Doreau and Glasser2007; Leiber et al. Reference Leiber, Tsvetkova, Petrova, Kreuzer and Scheeder2010). Comparison between diets B and M also showed that neither the roughage to concentrate ratio nor the fibre supply, as such, were relevant to the increased trans-10 18:1 concentrations noted in the present data. The high proportion of several other trans 18:1 isomers found in milk from cows fed diet M may have at least partially resulted from the high dietary 18:1 n-9 supply (Mosley et al. Reference Mosley, Powell, Riley and Jenkins2002). The high proportion of cis-15 18:1 in milk fat of cows fed diet H reflects that this FA is a typical product of ALA biohydrogenation (Jenkins et al. Reference Jenkins, Wallace, Moate and Mosley2008).
The generally high secretion of 18:0 (stearic acid) with milk in relation to its intake demonstrated that large amounts of this FA are produced along the transfer pathway from diet to milk, mainly in the rumen (Chilliard et al. Reference Chilliard, Glasser, Ferlay, Bernard, Rouel and Doreau2007). As indicated by the secretion to intake ratio, it seemed that diet M, when compared with diets H and B, in some way inhibited the terminal step of ruminal biohydrogenation, possibly due in some way to the high dietary LA concentrations (Moate et al. Reference Moate, Boston, Jenkins and Lean2008). Thus, in addition to plant secondary compounds (Khiaosa-Ard et al. Reference Khiaosa-Ard, Bryner, Scheeder, Wettstein, Leiber, Kreuzer and Soliva2009), increased dietary LA supplementation could be a tool for inhibiting the terminal biohydrogenation step and enhancing the VA and RA levels in milk.
The present study demonstrated that if constant dietary energy and protein levels are maintained, the rate of transfer of LA and ALA directly from the diet into the milk is reciprocal to the intake of the respective FA. The transfer is apparently independent of the concentrate and fibre proportions. Furthermore, the concentrations of VA and, to a lesser degree, of RA in milk fat will depend on the sum of ingested LA and ALA, regardless of their proportions in the diet. Finally, the conclusion can be drawn that the complete ruminal biohydrogenation of FAs to stearic acid may be partly inhibited by feeding elevated amounts of LA.
The authors acknowledge the financial support of the Vontobel Foundation, Switzerland, and the TUMSS (Towards an improved understanding of methane sources and sinks) project of the ETH Zurich. In addition, the authors thank Dr. H-R Wettstein and the employees of the ETH Research Station Chamau for their assistance and support during the dairy cow experiment.
