Introduction
Sunflower (Helianthus annuus L.) seeds are a prominent source of vegetable oil for human consumption and biofuel production (USDA, 2019). Sunflower seeds contain between 350 and 450 g oil/kg and 250 and 300 g hulls/kg (Finn et al., Reference Finn, Clark, Drackley, Schingoethe and Sahlu1985; Economides, Reference Economides1998; NRC, 2001). Sunflower meal (SFM) is a by-product obtained after oil extraction and contains 280–500 g crude protein (CP)/kg dry matter (DM) depending on cultivar, method of oil extraction and degree of seed decortication (Hesley, Reference Hesley1994; Canibe et al., Reference Canibe, Pedrosa, Robredo and Knudsen1999). Sunflower hulls have low CP and high level of lignified fibre (Arija et al., Reference Arija, Brenes, Viveros and Elices1998). Therefore, the variation in nutritional value of SFM is determined mostly by degree of seed decortication. However, non-decorticated SFM is commercialized, probably due to the extra cost of decortication.
SFM exhibits similar intestinal digestibility of undegradable protein to soybean meal (SBM), but with lower contents of lysine and threonine, and higher content of methionine (NRC, 2001; Branco et al., Reference Branco, Coneglian, Maia and Guimaraes2006; Rodriguez et al., Reference Rodriguez, Gonzalez, Alvir and Caballero2008). However, it has been demonstrated that non-decorticated SFM contains a greater proportion of rumen degradable protein (RDP) than traditional protein sources such as SBM, cottonseed and canola meal (Erasmus et al., Reference Erasmus, Botha, Cruywagen and Meissner1994; Branco et al., Reference Branco, Coneglian, Maia and Guimaraes2006; Rodriguez et al., Reference Rodriguez, Gonzalez, Alvir and Caballero2008), but lower ruminal DM degradation than SBM, possibly due to high content of hulls (Rodriguez et al., Reference Rodriguez, Gonzalez, Alvir and Caballero2008).
Previous studies using low-producing dairy cows (<20 kg of milk/day) showed no detrimental effect on milk production when 110 g/kg DM of low-oil non-decorticated SFM (Schingoethe et al., Reference Schingoethe, Rook and Ludens1977) or 150 g/kg DM of high-oil non-decorticated SFM (Silva et al., Reference Silva, Oliveira, Barbosa, Bueno and Mota2005) was included in the diet as a substitute for SBM and maize. In another study, replacing SBM with SFM reduced milk yield of high-producing dairy cows (Yildiz et al., Reference Yildiz, Todorov and Nedelkov2015). The current authors hypothesized that replacing SBM and wheat middlings (WM) with non-decorticated SFM would increase RDP supply and decrease diet digestible energy, which would consequently reduce the performance of dairy cows. Therefore, the current study evaluated the effects of replacing SBM and WM with non-decorticated SFM on milk production, total-tract diet digestibility and nitrogen (N) metabolism of lactating dairy cows.
Materials and methods
Cows and diets
Twelve multiparous Holstein cows (31 ± 4.1 kg milk/day; 128 ± 38.2 days in milk; and 627 ± 48.1 kg body weight [BW]) were blocked by days in milk and assigned randomly within squares to treatment sequences in three replicated 4 × 4 Latin squares. Treatment sequences within Latin squares were balanced for carry-over effects with four 21-day periods, which included 14 day for diet adaptation and 7 days for data and sample collection.
Cows were housed in individual tie-stalls (215 × 125 cm) with rubber beds and had free access to water. The chemical compositions of the maize silage (MS), SFM, SBM, WM and ground maize grain used in the current trial are shown in Table 1. SFM (BUNGE Alimentos, S.A. Brazil) was obtained after solvent oil extraction of non-decorticated whole sunflower seeds. Cows were fed with four isonitrogenous experimental diets as a total mixed ration (TMR) containing four levels of SFM (0, 70, 140 and 210 g/DM), partially or fully replacing a SBM–WM mixture (536 g SBM/kg and 464 g WM/kg DM of the mixture) (Table 2). All diets were formulated to meet nutrient requirements of 650 kg cows producing 30 kg/day of milk and 38 g/kg milk fat (NRC, 2001). The TMR was prepared by blending MS and concentrate mixtures. The concentrate mixtures were prepared for each 21 day period. Diets were offered twice daily at 07.00 and 16.00 h. Amounts of feed offered to the cows were adjusted daily to allow refusals equal to proportions of 0.05–0.10 of intake. DM content from weekly composites of the silage and concentrate mixture was used to adjust the as-fed TMR composition to maintain constant dietary nutrient supply throughout the trial.
Table 1. Chemical composition of MS and concentrate feeds
MS, maize silage; GMG, ground maize grain; SBM, soybean meal; WM, wheat middlings; SFM, sunflower meal.
a NDIN, neutral detergent insoluble nitrogen; ADIN, acid detergent insoluble nitrogen; aNDFom, neutral detergent fibre corrected for ash and nitrogen; iNDF, indigestible neutral detergent fibre obtained after in situ ruminal incubation for 264 h; iADF, indigestible acid detergent fibre obtained after in situ ruminal incubation for 264 h.
Table 2. Ingredient and chemical composition of experimental diets
a SFM replaced mixture of SBM and WM.
b Provided (per kg of DM): 383 g of Mg, 161 g of S, 30 mg of Zn, 9 mg of Mn, 7 mg of Cu, 0.4 mg of I, 0.2 mg of Se, 0.07 mg of Co.
c NDIN, neutral detergent insoluble nitrogen; ADIN, acid detergent insoluble nitrogen; aNDFom, neutral detergent fibre corrected for ash and nitrogen; iNDF, indigestible neutral detergent fibre obtained after in situ ruminal incubation for 264 h; iADF, indigestible acid detergent fibre obtained after in situ ruminal incubation for 264 h.
Animal measurements and sampling
Individual concentrate ingredients were sampled during each mixture preparation (21 days) and kept in a freezer (−15 °C) for subsequent grinding and chemical analysis. Daily DM intake (DMI) and diet component intakes were determined by differences between the weights of feed offered and feed refused. MS, concentrate mixture offered and diets refused were weighed twice daily for each cow. Approximately 100 g of the MS offered and refusal were sampled twice daily and stored (−15 °C). At the end of each collection period (7 days), the refusal samples from each animal were removed from the freezer, thawed at room temperature and blended manually to obtain a composite sample per animal for each period. The composite samples of MS (7 days) and refusal were pre-dried in a forced-air oven at 55 °C for 72 h.
Faecal samples were collected directly from the rectum once daily from day 15 to 19 of each period, at 08.00, 10.00, 12.00, 14.00 and 16.00 h, respectively. The daily faecal samples of each cow in each period were kept in a freezer (−15 °C) for later pre-drying in a forced-air oven at 55 °C for 72 h. After pre-drying and grinding, a single composite faecal sample was obtained per animal for each period.
Cows were milked twice daily (06.00 and 15.00 h) and milk yield recorded at each milking. Milk samples from the morning and afternoon milkings were collected on day 18 and 19 of each period. Composite samples were prepared daily according to milk production and three different aliquots were sampled. The first aliquot (60 ml) was stored at 6 °C with a preservative (bronopol-B2) for analysis of fat, lactose, solids and solid non-fat content. The second aliquot was analysed immediately for CP (N × 6.38). The third aliquot (10 ml) was deproteinized with 5 ml 250 g trichloroacetic acid/l and filtered on Whatman #1 filter paper; the filtrate was analysed for N content and the remainder stored at −15 °C for subsequent analysis of allantoin and urea. Fat-corrected milk (FCM; 3.5 g/100 g milk) was estimated according to the Gaines (Reference Gaines1928) model: FCM (kg/day) = 0.432 × milk yield (kg) + 0.1623 × milk fat concentration (g/100 g). BWs were measured in the morning and afternoon (after milking) on day 7 and 21 of each period.
Blood samples were taken in ethylenediaminetetraacetic acid tubes from the coccygeal vessels of each cow 4 h after feeding on day 19 of each period, centrifuged immediately (2300 g, 15 min, room temperature) and plasma was stored at −15 °C for urea analysis. Spot urine samples were obtained at approximately 0, 3 and 6 h post-feeding on day 17 of each period by manual stimulation of the vulva. After collection, 10 ml of urine was filtered and pipetted into a specimen container with 40 ml of 0.072 N sulphuric acid (H2SO4) and stored at −15 °C. Before urinary analysis, the urine samples for each time of collection from each cow were thawed, centrifuged at 2000 g for 15 min (room temperature) and combined into composite samples (10 ml for each time) for each cow in each period. These samples were analysed for N, urea, creatinine, allantoin and uric acid.
In situ ruminal degradability
In situ ruminal degradability of each dietary ingredient was obtained according to NRC (2001). Briefly, 5 g of pre-dried feed sample (2 mm) were added to 50-μm nylon bags (Tenyl Tecidos Tecnicos Ltda, Guarulhos, Brazil) measuring 16 × 8 cm2 in triplicate and placed into the rumen of two lactating dairy cows (BW 550 kg and 20 kg milk/day). The cows were managed under similar conditions as described previously and fed with the experimental diet containing no SFM ad libitum. CP and DM degradation were determined over incubation times of 0, 2, 4, 8, 16, 24 and 48 h; samples were further incubated for 72 and 96 h for WM, MS and SFM (NRC, 2001). An additional incubation time (240 h) was used for determination of ruminal neutral detergent fibre (NDF) degradability in all feeds. Upon removal from the rumen, bags were washed carefully in tap water and dried at 55 °C in a forced-air oven for 72 h: residues were analysed later to determine the concentrations of DM, CP and NDF.
Ruminal CP degradation kinetics were estimated by fitting degradation data to the exponential model proposed by Ørskov and McDonald (Reference Ørskov and McDonald1979):
$$Y = A + B \times (1 - \exp ^{(-k_{\rm d} \times t)})$$where Y = degradability at time t, A = soluble fraction, B = potentially degradable fraction and k d = rate of degradation of B(/h). Effective degradable (ED) fraction of the DM and RDP of each feed was calculated as:
$${\rm ED\ or\ RDP}={\rm } A + B{\rm} \;[k_{\rm d}/(k_{\rm d} + k_{\rm p})]$$where A is the soluble fraction and k p is the ruminal rate of passage (/h) (Ørskov and McDonald, Reference Ørskov and McDonald1979). The rate of passage (k p) was estimated using the equation of NRC (2001):
$$k_{\rm p}\;{\rm of\ wet\ forage} \;(\% /{\rm h}) = 3{\cdot}054 + 0{\cdot}614 \times {\rm DMI} \;({\rm g/100}\,{\rm g\ BW})$$
$$\eqalign{k_{\rm p}\;{\rm of\ concentrate} \;({\rm \%\ \!\!/h}){\rm} & ={\rm 2}{\rm. 904 + 1}{\rm. 375 } \cr & \times {\rm DMI} \;({\rm g/100}\,{\rm g\ BW}){\rm - 0}{\rm. 002} \cr & \times {\rm concentrate \; in\; diet}\ ({\rm g/kg\ DM}).} $$RDP intake was calculated as:
$$\eqalign{\hbox{RDP intake} ({\rm kg}/{\rm day}) & ={\rm DMI}\ ({\rm kg}/{\rm day}) \cr & \times \hbox{RDP in diet} ({\rm g}/{\rm kg}\,{\rm DM})/1000}.$$RDP in diet was calculated from RDP content of each feed.
Ruminal NDF degradation kinetics was estimated by fitting degradation data to the Mertens and Loften (Reference Mertens and Loften1980) exponential model:
$$Y = B \times \exp ^{((-k_{\rm d} \times (t-L))} +\, U$$where Y = residue remaining at time t, B = potentially degradable fraction, k d = rate of degradation of B(/h), L = discrete lag time (h) and U = undegradable fraction.
ED fraction of the NDF and RDP was calculated as:
$${\rm ED\ or\ RDP}={\rm } B \times [k_{\rm d}/(k_{\rm d} + k_{\rm p})]$$Chemical analysis
The pre-dried MS, refusals and faecal samples, and original samples of SFM, SBM and WM were ground in a Wiley mill (Arthur H. Thomas, Philadelphia, PA, USA) with 1-mm screen for chemical analysis and 2-mm screen for ruminal incubation in situ. Samples of feed, refusals and faeces were analysed for concentrations of DM (method no. 934.01), organic matter (OM, method no. 942.05), CP (method no. 954.01) and ether extract (EE, method no. 920.39) according to AOAC (2005). NDF was determined using heat stable amylase without sodium sulphite and corrected for residual ash (Mertens, Reference Mertens2002) and N (Licitra et al., Reference Licitra, Hernandez and Van Soest1996) (aNDFom). Both NDF and acid detergent fibre (ADF) (sequential) were analysed with an Ankom® fibre analyser (Ankom Technology, Fairport, NY, USA). Concentrations of non-protein N (NPN), neutral detergent insoluble N (NDIN) and acid detergent insoluble N (ADIN) were measured according to Licitra et al. (Reference Licitra, Hernandez and Van Soest1996). Lignin concentration was determined by solubilization of cellulose by hydrolysing the ADF residue with 72% H2SO4 (wt/wt) (Van Soest et al., Reference Van Soest, Robertson and Lewis1991). Non-fibre carbohydrate (NFC) was calculated by difference according to Hall (Reference Hall2000):
$$\eqalign{{\rm NFC} & = 100-[({\rm CP}{\rm -}{\rm CP\ derived\ from\ urea}) + {\rm urea} + {\rm EE} \cr & + \% \;{\rm ash} + {\rm aNDFom}]}$$Indigestible ADF was used as the internal marker to estimate apparent nutrient digestibility and faecal output (Cochran et al., Reference Cochran, Adams, Wallace and Galyean1986). Indigestible ADF in feeds, refusals and faeces was obtained after ruminal incubation in a polyester bag (Ankom®, filter bag 57) for 264 h (Casali et al., Reference Casali, Detmann, Valadares Filho, Pereira, Henriques, de Freitas and Paulino2008).
Total digestible nutrient (TDN) was obtained from digestion trial according to Weiss (Reference Weiss1998):
$$\eqalign{{\rm TDN}\,({\rm g}/100\,{\rm g}\,{\rm DM}) & = {\rm CP}\,({\rm g/100}\,{\rm g}\,{\rm DM}) \times {\rm dCP} \cr & + {\rm aNDFom}\,({\rm g/100}\,{\rm g}\,{\rm DM}) \times {\rm daNDFom} \cr & + {\rm NFC}\,({\rm g/100}\,{\rm g}\,{\rm DM}) \times {\rm dNFC} \cr & + {\rm EE}\,({\rm g/100}\,{\rm g}\,{\rm DM}) \times {\rm dEE} \times 2{\cdot}25}$$where dCP is the total-tract digestible CP, daNDFom is the total-tract digestible aNDFom, dNFC is the total-tract digestible NFC and dEE is the total-tract digestible EE, all expressed as coefficients.
Milk fat and lactose were analysed by infrared spectrophotometry (IDF, 1996). Nitrogen in milk and deproteinized milk were analysed by the micro-kjeldahl method (AOAC, 2005). Urea in milk, plasma and urine were measured using an enzymatic-colorimetric assay with urease (Urea CE Ref. 27, Labtest Diagnostica SA, Lagoa da Santa, Minas Gerais, Brazil). Urinary uric acid was quantified using the enzymatic-Trinder method (Ácido úrico Liquiform Ref. 73, Labtest Diagnostica SA, Lagoa da Santa, Minas Gerais, Brazil; Junge et al., Reference Junge, Wike, Halabi and Klein2004). Allantoin concentrations in milk and urine samples were determined by colorimetry (Young and Conway, Reference Young and Conway1942). Creatinine in urine was measured by an enzymatic-colorimetric assay (Creatinina Ref. 35, Labtest Diagnostica SA, Lagoa da Santa, Minas Gerais, Brazil). Total urine volume was estimated using creatinine concentration as a marker and assuming daily creatinine excretion of 24 mg/kg of BW (Cobianchi et al., Reference Cobianchi, de Oliveira, Campos, Guimarães, Valadares Filho, Cobianchi and de Oliveira2012). Excretion of purine derivatives (PD) was calculated as the sum of allantoin and uric acid excreted in urine, and allantoin secreted in milk. Excretion of PD per TDN intake was used as an index of energy efficiency, while PD excretion per intake of CP and RDP was used as an index of nitrogen efficiency for microbial protein synthesis (MPS) in the rumen. Milk N efficiency was calculated as the ratio of N in milk (g/day) to N intake (g/day).
Statistical analysis
Data were analysed using PROC MIXED in SAS (SAS Institute, 1999–2000) for a replicated 4 × 4 Latin square design. The following model was fitted to all variables:
$$Y_{ijkl} = \mu + S_i + P_j + C_{k(i)} + T_l + ST_{il} + E_{ijkl}{\rm,} $$where Y ijkl is the dependent variable, μ is the overall mean, S i is the effect of square i, P j is the effect of period j, C k(i) is the effect of cow k (within square i), T l is the effect of treatment l, ST il is the interaction between square i and treatment l, and E ijkl is the residual error (0; σ 2). All terms were considered fixed except for C k(i) and E ijkl, which were considered random. Significance was declared at P ⩽ 0.05, with trends at P > 0.05 and ⩽0.10. Dietary SFM levels were tested by partitioning degrees of freedom for diet into single degree of freedom variables corresponding to linear, quadratic and cubic effects. Cubic effects were not statistically significant for any of the variables and are not reported. All reported values were least squares means.
Results
The major change in the chemical composition of the diets with SFM inclusion was the increased concentration of aNDFom and lignin as a fraction of aNDFom (Tables 1 and 2). As a result, effective rumen degradation of DM in SFM was 15.8% lower than that of the SBM–WM mixture (0.517 v. 0.614, Table 3). As expected, SFM had greater RDP (0.687) than SBM (0.566), but lower RDP than WM (0.814) (Table 3). Potentially rumen-degradable NDF in SFM (0.452) was lower than those of SBM (0.978) and WM (0.673), probably due to its higher lignin concentration. Consequently, SFM had lower effective rumen degradable NDF (0.259) than SBM and WM (0.568 and 0.356 respectively; Table 3).
Table 3. In situ ruminal degradation kinetics of MS and concentrate feeds
MS, maize silage; GMG, grain maize ground; SBM, soybean meal; WM, wheat middlings; SFM, sunflower meal.
a A, soluble fraction; B, insoluble potential degradable fraction; U, undegradable fraction; k d, degradation rate of B fraction; ED, effective degradable fraction, with passage rate according to the NRC (2001); RDP, rumen degradable protein, with passage rate according to the NRC (2001); NDF, neutral detergent fibre; A, B, U, ED and RDP are expressed as coefficients.
Sunflower meal inclusion did not affect BW (P = 0.159), DMI (P = 0.118) or CP intake (P = 0.137). Increasing levels of SFM in the diet resulted in linear increases of RDP (P = 0.014) and aNDFom (P < 0.001) intakes, and linear reductions in NFC (P < 0.001) and TDN (P = 0.042) intakes (Table 4). In addition, SFM inclusion did not affect total-tract digestibility of CP (P = 0.112) and EE (P = 0.278); however, linear reductions in total-tract digestibility of DM (P = 0.036) and TDN (P = 0.031) were observed, as well as a tendency to reduce digestibility of OM (P = 0.075) and NFC (P = 0.076) (Table 4).
Table 4. Effects of replacement of SBM and WM mixtures with SFM in the diet on intake, total tract digestibility, milk production and composition of lactating dairy cows
a aNDFom, neutral detergent fibre corrected for ash and; BW, body weight.
b SFM replaced mixture of SBM and WM.
c s.e.d. = standard error of the least squares means.
d Probability of a significant effect linear (L) or quadratic (Q) of the SFM level.
e FCM = fat-corrected milk, estimated according Gaines (Reference Gaines1928): FCM (kg/day) = 0.432 × milk production (kg) + 0.1623 × milk fat concentration (g/100 g).
f Milk production/DMI.
Increasing levels of non-decorticated SBM in the diet linearly reduced milk production (P < 0.001), 3.5% FCM (P < 0.001), feed efficiency for milk production (P = 0.024) and milk component yields (P < 0.05) (Table 4). Concentrations of milk lactose (P = 0.213), NPN (P = 0.872) and fat (P = 0.229) were unaffected by dietary treatment (Table 4).
The concentrations of urea-N in milk (P = 0.339) and blood (P = 0.324) were not affected by SFM inclusion (Table 5). SFM inclusion did not affect milk allantoin secretion (P = 0.481) or urinary uric acid excretion (P = 0.946), but tended to reduce urinary allantoin excretion (P = 0.072) and PD excretion (P = 0.079) (Table 5). SFM inclusion did not affect PD excretion per TDN intake (P = 0.256), however, PD excretion as a proportion of CP intake (P = 0.069) tended to linearly reduce while PD excretion as a proportion of RDP intake (P = 0.026) linearly reduced with increasing levels of SFM in the diet. As a result, milk N production and milk N efficiency (milk N/N intake) were reduced (P ⩽ 0.001) (Table 5). Dietary treatment did not affect (P > 0.10) N intake, faecal N, urinary N or N balance of the dairy cows.
Table 5. Effects of replacement of SBM and WM mixtures with SFM on nitrogen (N) metabolism and efficiency of lactating dairy cows
a Total DP, Total PD, milk allantoin + urinary allantoin + uric acid; TDN, total digestible nutrient; CP, crude protein; nitrogen balance = N intake − (N milk + N faecal + N urine); RDP, rumen degradable protein; N milk efficiency = g N milk secreted/100 g N intake; N balance = g N balance/100 g N intake.
b SFM replaced combinations of SBM and WM.
c s.e.d. = standard error of the least squares means.
d Probability of a significant effect linear (L) or quadratic (Q) of the SFM level.
Discussion
The results of the current study confirmed the hypothesis that replacing the SBM and WM mixture (536 and 464 g/kg of the mixture, respectively) with non-decorticated SFM would increase intake of RDP and reduce intake of digestible energy, which would lead to reduced efficiency of N use and milk production of dairy cows. The lack of effect of SFM inclusion on DMI agrees with Schingoethe et al. (Reference Schingoethe, Rook and Ludens1977), who reported no change in DMI of dairy cows fed with a diet containing 110 g non-decorticated SFM/kg DM (370 g CP/kg DM and 212 g ADF/kg DM). Though SFM had greater concentrations of indigestible NDF and ADF, there is a potential for low ruminal fill due to high degradation rate of potentially degradable NDF, probably due to small particle size and high density of particles which reduce selective retention of feeds in the rumen (Lund et al., Reference Lund, Weisbjerg and Hvelplund2007). The reduced TDN intake observed in the current study was probably due to reduced total tract digestibility of NFC, which could be attributed to high level of hull seeds in non-decorticated SFM (Arija et al., Reference Arija, Brenes, Viveros and Elices1998).
The reduction in milk production and milk protein synthesis of the dairy cows with increasing levels of SFM in the diet was a result of reduced digestible energy intake and PD excretion. Because there is evidence that urinary PD excretion, mainly allantoin, has high correlation with intestinal flow of microbial nucleic acid (Perez et al., Reference Perez, Balcells, Guada and Castrillo1996; Valadares et al., Reference Valadares, Broderick, Valadares Filho and Clayton1999; González-Ronquillo et al., Reference González-Ronquillo, Barcells, Guada and Vicente2003), it was assumed that reduced PD excretion indicates reduced MPS in the rumen. According to NRC (2001), when RDP supply is not limited, the efficiency of energy use for MPS is fixed. This probably explains the lack of effect on the efficiency of energy utilization for MPS, measured by PD excretion per TDN intake. Reduced PD excretion per RDP intake indicates that SFM inclusion reduced RDP conversion to microbial protein probably due to reduced availability of carbon skeletons and energy for the rumen microbiota.
Digestible energy and MPS are the main factors contributing to duodenal flow of glucose and essential amino acids (EAA) such as lysine, methionine and histidine which are substrates that drive milk and milk protein synthesis in the mammary gland (Kronfeld, Reference Kronfeld1982; Schwab and Broderick, Reference Schwab and Broderick2017). In addition, lower lysine content of SFM (NRC, 2001; Branco et al., Reference Branco, Coneglian, Maia and Guimaraes2006; Rodriguez et al., Reference Rodriguez, Gonzalez, Alvir and Caballero2008) may have also caused reduced intestinal flow of lysine from the rumen. Therefore, the depressed milk production and milk protein synthesis with SFM inclusion is probably due to reduced glucose supply to the mammary gland and/or poor match between metabolizable protein supply and amino acid requirements for optimum milk production.
SFM inclusion did not affect indices associated with the pool of circulating urea, such as milk urea-N, blood urea-N and urinary urea-N and N balance. However, reduced milk N efficiency was observed with SFM inclusion, which is probably due to reduction in RDP conversion to microbial protein, efficiency of N captured by mammary gland and/or conversion of N captured into milk N (Lapierre et al., Reference Lapierre, Pacheco, Berthiaume, Ouellet, Schwab, Dubreuil, Holtrop and Lobley2006). This may also be as a result of poor match between metabolizable amino acid profile and requirements for lysine, methionine or histidine, and/or reduced glucose flows to mammary glands (Kronfeld, Reference Kronfeld1982; Schwab and Broderick, Reference Schwab and Broderick2017). However, efficiency of RDP conversion to microbial protein is much more likely to have caused the reduced milk N efficiency observed. Even though SFM appears to be a good source of methionine, increased RDP intake suggests duodenal supply of methionine may be inadequate. Further studies should examine how SFM inclusion affects duodenal flow of EAA, EAA in the plasma, and mammary gland metabolism of EAA.
In conclusion, replacement of SBM and WM mixture (536 and 464 g/kg of the mixture, respectively) with non-decorticated SFM in diet of lactating dairy cows does not affect DMI. However, due to reduced intake of digestible energy, efficiency of N use for milk production of dairy cows fed with SFM is depressed. It is important to note that the outcome of the current study is clearly a result of lower digestibility of fibre in non-decorticated SFM compared to that of the SBM : WM mixture. Further studies are needed to evaluate the use of decorticated SFM as a replacement for the SBM : WM mixture in diets for lactating dairy cows.
Author ORCIDs
A. S. Oliveira, 0000-0001-9287-0959
Acknowledgements
The authors wish to thank Professor Alberto Magno Ferreira Santiago, Ana Cristina Silva Souza, Gustavo Henriques Soares, Janaína Giordani, Janaína Paula do Carmo and Luciano Ferreira do Lago for their assistance during the animal trial; Dr Juliana Variz da Costa and Dr Shirley Motta de Souza for their assistance with allantoin analysis, and the Conselho Nacional de Pesquisa e Desenvolvimento Científico e Tecnológico (CNPq).
Financial support
Financial support from the Conselho Nacional de Pesquisa e Desenvolvimento Científico e Tecnológico (CNPq) is gratefully acknowledged.
Conflict of interest
None.
Ethical standards
All procedures were conducted according to the Guide for the Care and Use of Agriculture Animals in Research and Teaching (Federation of Animal Science Societies, 2010).
