Profiling the omega-3 content of annual and perennial forages during growth and development and the relationship with other quality parameters

Soil

The soil used for the study was a Red Kandosol (Isbell 1996) with moderate phosphorus concentration (Colwell P 64 mg kg⁻¹, (Colwell 1963); Method 9b, (Rayment and Lyons 2011), and moderate cation exchange capacity (CEC, Gillman and Sumpter 1986), see Table 1) collected from the Wagga Wagga Agricultural Institute (WWAI, 35°02′51″S 147°20′17″E). Prior to commencement of the study, the field capacity (FC) of the soil was determined according to the method of Veihmeyer and Hendrickson (1931) as described previously (Twarakavi et al. 2009).

Treatments and allocation

The study design was a 4 × 2 factorial, with four species viz; oats, (Avena sativa L. cv. Eurabbie, Nutrien Ag, Holbrook), annual ryegrass (Lolium multiflorum L. cv. Rambo, Nutrien Ag, Holbrook), phalaris (Phalaris aquatica L. cv. Holdfast GT, Barenberg) and subterranean clover (Trifolium subterraneum L. cv. Coolamon, Nutrien Ag, Holbrook) and two fertiliser treatments (mono-ammonium phosphate (MAP) only or MAP plus urea, see Table 2) with four replicates per treatment combination. The allocation of species, fertiliser and replicates to pots was achieved using the ‘R’ statistical program (ver. 4.2.0) with the DiGGer package (Coombes 2002) as described previously (Cullis et al. 2006) and employed in similar pot trials (Li et al. 2008, 2010). Four blocks within the glass house were chosen for the four replicates.

Plant material and growth conditions

Dry soil was brought to 70% FC and mixed thoroughly with MAP (10% nitrogen, 21.9% phosphorus, applied at 100 mg kg⁻¹). Polyvinyl chloride pots (100 mm internal diameter, 300 mm height) were filled with the equivalent of 3118 g of oven-dried soil and additional water added to bring the soil to the mass calculated to be 80% of FC. The four forage species were sown at a rate of 15 seeds per pot. Clover was inoculated with an appropriate rhizobium strain (Group F, New Edge Microbial, Albury, NSW) prior to sowing. Due to the very low amount of biomass available for early harvests, the number of plants per pot was modified with stage of maturity by hand thinning, so that pots harvested at Growth Stages 1–3 had 12 plants pot⁻¹, those harvested at Growth Stages 4–5 had 6 plants pot⁻¹ and those harvested at Growth Stages 6–7 had 4 plants pot⁻¹. Pots were placed in a glasshouse (maximum 33°C, minimum 9.2°C) under natural light during July–December at the WWAI. Prilled urea (50 kg ha⁻¹) was added to pots by top-dressing 42 days after sowing when all seeds had germinated. Pots were maintained at approximately 80% field capacity by watering to weight three times per week.

Plant harvesting and preparation

Plants were harvested at seven stages of maturity according to the Zadoks scale (Zadoks et al. 1974) for oats, and stage of development in relation to vegetative or reproductive growth for phalaris and ryegrass (corresponding to early vegetative, late vegetative, stem elongation, boot, anthesis, soft dough and ripening); or relative to day of flowering in clover (Table 2). These stages of development were expected to reflect changes in quality and yield, particularly relating to potential grazing (Kaiser et al. 2007). The sub clover had not set seed at the time of the last harvest. All material from each pot was dried in an oven at 80°C for 24 h for proximate analysis or frozen and lyophilised in a freeze dryer prior to being ground through a 1 mm screen for fatty acid analysis. The total material in each pot, dry matter and water use (g water g DM⁻¹ produced) was recorded.

Proximate analysis

Oven-dried samples were analysed for quality parameters including acid detergent fibre (ADF), neutral detergent fibre (NDF), nitrogen (N), acid insoluble ash (Ash), dry matter digestibility (DMD), digestible organic matter in the dry matter (DOMD), total lipid (ether extract, EE) and water soluble carbohydrates (WSC) by Near Infrared Reflectance (NIR) Spectroscopy with a Bruker multi-purpose analyser (MPA, Bruker Optik GmbH, Ettlingen, Germany) and OPUS software (ver. 5.1) with calibrations developed by the NSW Department of Primary Industries Feed Quality Service (NSW DPI FQS) as described previously (Packer et al. 2011). Crude protein (CP, % DM) and metabolisable energy (M/D) content (MJ kg DM⁻¹) was calculated using the following equations (AFIA 2006):

Fatty acid analysis

Total fatty acids in freeze-dried samples were extracted and methylated using the one-step procedure of Lepage and Roy (1986) with modifications as described previously (Clayton et al. 2012). In brief, a 20 mg freeze-dried sample was added to a glass culture tube followed by 2.0 mL of methanol:toluene (4:1) containing C13:0 (10 μg mL⁻¹) and C19:0 (10 μg mL⁻¹) as internal standards. Methylation of the fatty acids was achieved by slowly adding 200 μL acetyl chloride while vortexing and heating at 100°C for 1 h. After methylation, samples were rapidly cooled and 5 mL of 6% K₂CO₃ added slowly while vortexting prior to centrifugation for 10 min at 1500g. A sub-sample of the toluene layer (approximately 250 μL) which contains the fatty acid methyl esters (FAME) was transferred into glass gas chromatography (GC) vials sealed with Teflon-lined screw-cap for subsequent analysis by GC.

Gas chromatography

Individual FAME were identified using an Agilent 7890A gas chromatograph with dual auto-injectors and dual flame ionisation detectors (FIDs). FAME were separated using two fused carbon–silica columns, coated with cyanopropylphenyl (BPX70, 30 m x 0.25 mm i.d. and 0.25 μm film thickness, SGE Analytical Science, Ringwood, Victoria, Australia, P/N 054622). The carrier gas was helium with a total flow rate of 12.4 mL min⁻¹, a split ratio of 10:1 and a column flow of 0.9 mL min⁻¹. The inlet temperature was 250°C, the inlet pressure was 107.8 kPa and the injection volume was 2 μL into a focus inlet liner (4 mm i.d. SGE Analytical Science, P/N 092002) that was changed at least every 150 injections.

The oven temperature was set at 150°C and held for 0.5 min, increased at 10°C min⁻¹ to 180°C, increased at 1.5°C min⁻¹ to 220°C and then increased at 30°C min⁻¹ up to 260°C and maintained for 5 min to give a total run time of 36.5 min. The FID temperature was set at 280°C with the following gas flow rates; hydrogen = 35 mL min⁻¹, instrument air = 350 mL min⁻¹ and nitrogen make-up gas = 30 mL min⁻¹. Sample FAME peaks were identified by comparing their retention times with those of a standard mixture of FAME (NuChek Prep, Elysian, MN USA; Sigma, St Louis, MO and Supelco, Bellefonte, PA, USA) as described previously (Clayton et al. 2012) and quantitated using Agilent Chemstation Ver. C.01 and Microsoft Excel using a three point standard curve for each FAME with C19:0 as the internal standard. Unknown fatty acids were also quantified and all results were calculated as g kg DM⁻¹. Total fatty acid concentration was also converted to an equivalent lipid content using a lipid conversion factor (LCF) of 0.883 (Weihrauch et al. 1977). The lower limit of detection (LoD) was less than 0.02 g kg DM⁻¹ for all FAME analysed and the intra and inter-assay coefficient of variation was 9.7 and 11.2%, respectively, for all FAME between 0.2 and 5.0% of total FAME; and 1.7 and 1.6%, respectively, for all FAME greater than 5% of total FAME.

Statistical analyses

Prior to analysis, data were assessed for assumptions of normal distribution by the evaluation of frequency histogram and quantile–quantile plots and the Kolmogorov–Smirnov goodness of fit statistic using the Univariate procedure in the SAS statistical program (SAS Institute Inc. 1997). Differences in the concentration of fatty acids between species over time were determined by repeated measures analysis using the Mixed Model procedure in SAS (SAS Institute Inc. 1997). The restricted maximum likelihood (REML) estimation used ‘Pot’ as the individual experimental unit (Littell et al. 1998). The most appropriate covariance structure for each analysis was determined by reference to the Schwarz’s Bayesian Information Criterion (BIC, Wang and Goonewardene 2004). The analysis examined the fixed effects of ‘Species’, ‘Fertiliser’ and ‘Stage of Growth’ as well as the interaction between fixed effects with ‘Range’ and ‘Row’ examined as random effects. As the fixed effect of ‘Fertiliser’ and the interaction between ‘Fertiliser’ with other fixed effects was not significant (P > 0.05) for any parameter analysed, results are presented for the fixed effects of ‘Species’ and ‘Stage of Growth’ only. The relationships between the analysis of proximate components of forage and the proportion of individual fatty acids or total fatty acid concentration were analysed by Pearson correlation also using SAS. Data are presented as least squares means ± s.e. of least squares means. An alpha of 0.05 was used for all statistical tests.

Plant growth, water use and quality

The amount of dry matter harvested (g dry matter plant⁻¹) differed between species (P < 0.001) and was higher for oats (10.35 ± 0.35 g dry matter [DM]) and ryegrass (9.68 ± 0.35 g DM) compared with phalaris (4.52 ± 0.45 g DM) and clover (3.21 ± 0.50 g DM) at the final harvest. Water use (g g DM⁻¹) was lowest for oats compared with any other species (Table 3) and generally decreased with increasing stage of maturity for ryegrass, phalaris and clover (Fig. 1b). The water use for phalaris declined initially and then increased towards later stages of maturity.

Fig. 1.

Mean change in (a) dry matter (DM, % as-is) and (b) water use (g g DM⁻¹) for oats (), ryegrass (), phalaris () and clover () over different stages of development. Values are least squares means ± s.e. of the least squares means.

The NDF content was highest for oats compared with any other species and lowest for clover (Table 3). The CP (Fig. 2a) and EE (Fig. 2b) content declined with increasing stage of maturity for all species and CP was highest for clover compared with any other species and lowest for oats (Table 3). The EE content of clover increased from Harvest 2 to 4, however, declined again with later maturity (Fig. 2b). The M/D content of oats increased at the ‘boot stage’ (Harvest 4), however, declined again with increasing stage of maturity (Fig. 2c). The M/D content of clover remained relatively constant with increasing stages of maturity (Fig. 2c).

Fig. 2.

Mean concentration of (a) crude protein, (b) ether extract, and (c) metabolisable energy (M/D) for oats (), ryegrass (), phalaris () and clover () over different stages of development. Values are least squares means ± s.e. of the least squares means.

The M/D of oats, ryegrass and phalaris (M/D = −0.18 × NDF + 20.3, r² = 0.83, P < 0.001) but not clover (M/D = −0.05 × NDF + 13.4, r² = 0.08, P = 0.854) was negatively correlated with the NDF content of forage at any stage of development. The EE content of oats, ryegrass and phalaris (EE = 0.85 × Ln(CP) + 1.42, r² = 0.84, P < 0.001) and clover (EE = 1.42 × Ln(CP) − 0.97, r² = 0.84, P < 0.001) was positively correlated with the CP content of forage across different stages of development.

Fatty acid changes over time

The proportion of C18:3n-3 (i.e. cis-9, cis-12, cis-15 octadecatrienoic acid, α-linolenic acid) and total omega-3 was highest in clover compared with any other species and higher in ryegrass and phalaris compared with oats (Table 4). Similarly, the proportion of the omega-6 fatty acid C18:2n-6 (i.e. cis-9, cis-12 octadecadienoic acid, linoleic acid) was also higher in clover compared with any other species (Table 4). The proportion of omega-3 decreased with increasing stage of maturity for oats (13.9 ± 0.86 vs 64.3 ± 0.68), ryegrass (17.3 ± 0.86 vs 68.2 ± 0.65) and phalaris (21.6 ± 1.63 vs 65.1 ± 0.79) but not to the same extent for sub clover (41.5 ± 1.38 vs 57.4 ± 0.79, Fig. 3a). The omega-6:omega-3 (n-6:n-3) ratio was highest for clover compared with any other species at early stages of growth, but similar to ryegrass and phalaris at later stages of growth (Fig. 3c).

Fig. 3.

Mean proportion of (a) C18:3n-3 (α-linolenic acid), (b) C18:2n-6 (linoleic acid), and (c) n-6:n-3 ratio for oats (), ryegrass (), phalaris (), and clover () over different stages of development. Values are least squares means ± s.e. of the least squares means.

The n-6:n-3 ratio increased markedly in oats after Zadocks 47 (Harvest 5) and continued to increase towards maturity (Fig. 3c). The proportion of C16:0 increased with increasing stage of maturity for all species, but declined again towards senescence for oats (Fig. 4a). The proportion of C18:1n-9 increased over time and was higher for oats than any other species in the ‘ripening’ stage of development (Fig. 4c).

Fig. 4.

Mean proportion of (a) C16:0 (palmitic acid), (b) C18:0 (stearic acid), and (c) C18:1n-9 (oleic acid) for oats (), ryegrass (), phalaris (), and clover () over different stages of development. Values are least squares means ± s.e. of the least squares means.

The proportion of total monounsaturated fatty acids (MUFA) was highest for oats compared with any other species (Table 4), which was mainly related to a higher proportion of C18:1n-9, particularly at later stages of development (Fig. 4c). The proportion of total saturated fatty acids (SFA) was highest for phalaris compared with any other species, however, the concentration of the major SFA C16:0 was highest for ryegrass (Table 4).

Correlation between forage quality and fatty acids

The proportion of omega-3 (g 100 g total fatty acids⁻¹) was positively correlated with the CP content of plant material (Fig. 5). The omega-3 content was lower for clover compared with oats, ryegrass and phalaris at similar CP concentrations (Fig. 5). For phalaris, the omega-3 content appeared to remain above 50% total fatty acids until the CP content decreased below approximately 6% DM and then declined markedly. The proportion of C18:3n-3 was also positively related to EE (r² = 0.77, 0.89, 0.43, 0.62 for oats, ryegrass, phalaris and clover, respectively) and M/D (r² = 0.89, 0.24, 0.18, 0.28 for oats, ryegrass, phalaris and clover, respectively). The total analysed concentration of FAME was significantly positively correlated with the total lipid concentration determined by EE (Fig. 6).

Fig. 5.

Relationship between crude protein (g 100 g DM⁻¹) and the proportion of C18:3n-3 (α-linolenic acid, g 100 g Total FAME⁻¹) in oats ( C18:3n-3 = 21.6 × Ln (CP) − 5.62, r² = 0.97, P = 0.000516), ryegrass ( C18:3n-3 = 21.0 × Ln (CP) − 3.57, r² = 0.97, P = 0.0000344), phalaris ( C18:3n-3 = 19.2 × Ln (CP) − 1.20, r² = 0.86, P = 0.0027) and clover ( C18:3n-3 = 16.8 × Ln (CP) − 0.90, r² = 0.90, P = 0.001) over different stages of development.

Fig. 6.

Relationship between total lipid determined from ether extract and total lipid calculated from total fatty acid concentration (total lipid = total FAME/0.883 g 100 g DM⁻¹) in oats (), ryegrass (), phalaris (), and clover () over different stages of development. Total fatty acid concentration = 1.88 × EE − 2.89, r² = 0.78, P = 0.0000000004165. Dashed line (– – –) denotes y = x.

Plant growth, water use and quality

The higher dry matter yield of oats and ryegrass compared with phalaris and clover was expected given previous assessments of the growth of annual and perennial forages (Muldoon 1986). The lower water use for phalaris compared with clover at early stages of growth, but higher water use at later stages, reflected the overall pattern of plant growth observed and previously established differences in water use efficiency between species (Johns and Lazenby 1973; Guobin et al. 1992).

The decrease in protein and energy content of all forages over time was typical of maturing forages (Cherney et al. 1993; Chaves et al. 2006). The negative relationship between NDF and M/D was expected and reflects the relatively low energy content of the fibre component of plant cell walls and the increase in structural carbohydrates at later stages of maturity (Cherney et al. 1993). The higher NDF content of oats compared with clover may indicate that the dry matter intake (DMI) of oats by livestock would be lower due to the negative relationship observed between NDF and potential intake, where DMI (% liveweight) = 120/NDF (% DM) (Van Soest 1965). The higher CP content of clover compared with other species at later stages of maturity was also expected given previous comparisons of forage species (Rattray and Joyce 1974). The higher M/D of oats approaching the ‘boot’ stage of development (Zadoks 47) and subsequent decline towards maturity was similar to previous results with cereal crops (Kaiser et al. 2007; Coblentz et al. 2018; Piltz et al. 2021) and corresponded to changes in water use observed in the current study.

Fatty acid changes over time

Although the omega-3 content was lower in phalaris than annual ryegrass at early stages of development, the consistent level of omega-3 over the first three stages of development in phalaris occurred despite the concentration of CP decreasing. The omega-3 content of phalaris remained higher than 50% of total fatty acids when CP was above 5% DM, indicating a preferential conservation of omega-3. The higher omega-6 content of oats at the ‘soft dough’ (Zadoks 85) and ‘ripening’ (Zadoks 92) stages was also expected given the development of seed material (Heppard et al. 1996). Desaturation of C18:1 fatty acids to C18:2 or C18:3 occurs after incorporation of C18:1 into glycerolipids following the action of acyl-CoA synthase on C18:0 (Browse and Somerville 1991). The desaturation of C18:1 to C18:2n-6 is facilitated by Δ−12 desaturase (FAD-2) found in higher abundance in seed compared with leaf material; and further desaturation to C18:3n-3 is facilitated by Δ−15 desaturase (FAD-3) found in higher abundance in leaf (Browse et al. 1986; Harwood 1996). The preferential action of FAD-2 on triacylglycerol (TAG) compared with FAD-3 acting on phosphatidylcholine (PC) or monogalactosyldiacylglycerol (MGDG) also contributes to the higher omega-6 fatty acid content of seeds, as the lipid content of seeds is predominantly TAG, whereas leaf material is predominantly PC and MGDG (Harwood 2021).

The higher proportion of C18:1n-9 in oats compared with other species was largely related to a significant increase in C18:1n-9 observed after Zadoks 65 (Harvest 5). Although LCn-3PUFA is considered to be healthy due to lowering the risk of CVD and other inflammatory disorders (Simopoulos 1999; Djuricic and Calder 2021), some research groups have claimed that higher MUFA, particularly C18:1n-9, may have beneficial effects on high density lipoprotein (HDL) cholesterol and triacylglycerols (TAGs) (Adams et al. 2010). The differential effect of these fatty acid profiles in early and later stages of maturity on indicators of human health outcomes should be examined, particularly for oats during seed head development.

Correlation between forage quality and fatty acids

The relationship between omega-3 content and CP was expected given the relationship between nitrogen in the leaf material and presence of omega-3 in leaf chloroplasts (Boufaïed et al. 2003). This relationship occurs as the rate limiting step for omega-3 production is the desaturation of C18:0–C18:1 in the chloroplast (Nagai and Bloch 1968). Following incorporation in glycerolipids, further desaturation to C18:3n-3 occurs (Browse and Somerville 1991) and, as indicated above, these processes are more abundant in leaf material, where growth is stimulated by higher nitrogen availability (Harwood 1980). The lower omega-3 content of clover at any CP concentration could have been related to a lower leaf:stem ratio compared with other annual or perennial species examined (Johns and Lazenby 1973; Khorasani et al. 1997). Unfortunately it was not possible to determine the leaf:stem ratio in the current study, however, future studies should determine whether maintenance of a higher leaf:stem ratio is associated with higher omega-3 at an equivalent CP content. Grazing management strategies that maximise the leaf:stem ratio could also be examined and may lead to improvements in overall forage quality (energy and protein) as well as omega-3 content.

The significant correlation between the total analysed concentration of fat (EE) with total fatty acid concentration indicates good recovery of fatty acids using the extraction and gas chromatography procedures employed, which was similar to previous results with sheep red blood cells (Clayton et al. 2012). The development of strategies to increase the amount of omega-3, such as plant breeding or grazing management, relies on the ability to be able to analyse fatty acid concentrations in large numbers of samples in a timely manner. Future studies should determine whether NIR technologies can be employed to determine changes in fatty acid profiles during growth and development of different plant species, thereby leading to rapid and cost-effective analysis.

Grazing management and maximising omega-3

Although forage yield (t DM ha⁻¹) could not be determined in the current study, the dry matter produced per plant was higher in all species at later stages of development (data not shown). Grazing management usually seeks to maximise production per hectare based on a compromise between high yield and high quality (Bell et al. 2014). The optimum time for grazing was likely to be just prior to boot stage for oats (Zadoks 47) or late stem elongation in ryegrass and phalaris, when development was changing from vegetative to reproductive growth (Piltz et al. 2021). At this stage of growth quality, including CP and M/D, was still high and the rapid increase in the n-6:n-3 ratio had not occurred. If a higher omega-3 concentration is desired in order to supply fatty acids needed to target meat that was a ‘good source’ of omega-3, forages could be grazed earlier if a lower forage yield was acceptable.

The nutrient profile of the soil used in the current study could have affected plant growth and, therefore, fatty acid profile. For example, the soil was relatively high in potassium, which could have lead to reduced plant uptake of magnesium (Haby et al. 1990). Differences in soil nutrient profile would affect plant growth and likely lead to different concentrations of omega-3, particularly changes in omega-3 over different stages of growth. Further research is required in order to determine whether relative differences between species are consistent in soils of different nutrient availability.

The lack of effect of urea on CP and fatty acid composition was unexpected given previously known effects of added nitrogen (Boufaïed et al. 2003; Witkowska et al. 2008). The absence of an effect from urea was likely due to the small physical amount of urea applied, which may not have been distributed evenly across the pot. Future studies could examine a higher dose of nitrogen or liquid application to improve distribution and uptake, or the timing of nitrogen application. A number of samples grew a significant amount of dry matter following harvesting and re-growth. Future studies should also determine the quality and fatty acid profile of re-growth material compared with initial growth to estimate changes following multiple grazings and second year growth for perennials. The effect of grazing new and re-growth material by sheep and cattle on the omega-3 content of beef and lamb meat should also be examined.