| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
,2,3
* Department of Animal and Avian Sciences University of Maryland, College Park 20742
Bovine Functional Genomics Laboratory USDA-Agricultural Research Service, Beltsville, MD 20705
4 Corresponding author: rkohn{at}umd.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Key Words: rumen-degraded protein • nitrogen efficiency • protein requirement • milk urea nitrogen
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
One of the first steps in diet formulation for lactating dairy cows is to provide sufficient RDP to meet the requirements of rumen microorganisms. The total metabolizable protein requirement of the cow is met by supplementing RUP when microbial protein synthesis alone is insufficient to meet the metabolizable protein requirements. Because excess protein in the ration of dairy cows is excreted, excess dietary protein may contribute to N pollution of the environment. Improving diet formulation to meet but not exceed the RDP requirement of microbes will optimize microbial growth, reduce N excretion, and improve overall N use by the cow.
The NRC (1989) requirements for RDP suggested 10.4% RDP as the upper minimal dietary concentration required for microbial growth in high-producing cows. The most recent NRC publication (2001) ties RDP requirements to dietary energy intake where microbial N (g) is equivalent to 20.8 x total digestible nutrients (TDN). Assuming the maximal efficiency of RDP use for microbial N synthesis is 85%, the RDP requirement would be 24.5 g per g of TDN intake (NRC, 2001). Other research indicates that microbial synthesis may be improved when RDP is greater than 10.4% (Stokes et al., 1991a, b); however, no previous research has evaluated the effect of feeding ruminally degraded protein in decreasing gradient levels. Feeding recommendations for RDP have been based on in vitro and in situ studies and theoretical calculations, and recommendations need to be tested in animal feeding experiments. Furthermore, the risk of economic loss due to decreased milk production from underfeeding RDP needs to be balanced against the potential for environmental damage due to overfeeding RDP. It is therefore necessary to determine how much milk production is likely to be lost from underfeeding RDP.
In research trials, often the ratio of RDP to RUP is changed while the CP content remains constant. Results from these experiments are difficult to interpret because the increasing concentration of RDP is confounded with the decreasing concentration of RUP. The effects of RDP deficiency can be masked by RUP excess. For example, reduced microbial protein from lack of RDP may not influence production if RUP substitutes for the microbial protein lost and more RUP allows for greater recycling of N back to the rumen. The current study was designed to test the effects of reducing RDP on ruminal fermentation and milk production, and therefore we intended to change only RDP concentration.
The objectives of this experiment were to: 1) determine the effects of feeding RDP below predicted requirements on milk production, milk composition, DMI, feed efficiency, N use efficiency, and N excretion, 2) compare NRC (1989 and 2001) models with observed data from this experiment, and 3) quantify the cost in lost milk production from underfeeding RDP and compare that with the decreased feed cost. Results from this experiment will help determine optimal RDP concentrations of diets for lactating dairy cows to optimize milk production and milk components while reducing N excretion to the environment.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Each experimental period consisted of 21 d of which the first 14 d were for adaptation. Data from d 15 to 21 were used to compare treatment effects. Cows were housed in tie-stalls, milked twice daily at 0530 and 1730 h, and fed once daily at 0800 h. Cows treated with bST received injections on d 8 of period 1 of the study and continued to receive bST every 14 d. Therefore, cows received bST once during periods 1 and 3 (d 8), and twice during periods 2 and 4 (d 1 and 15). Because the design was a balanced 4 x 4 Latin square, an equal number of observations were made for each dietary treatment during periods in which bST was injected on d 8 vs. d 1 and 15. Two cows were removed from the study due to illness.
Diets were formulated to meet requirements for NEL, RUP, minerals, and vitamins of a midlactation dairy cow (120 DIM) weighing 615 kg, producing 41 kg of milk with 3.5% fat (NRC, 1989). Diets contained 50% corn silage and 50% concentrate (DM basis). Ingredients of the diets were equal across treatments except for changes in ground corn, solvent-extracted soybean meal, and nonenzymatically browned soybean meal (Soy Pass; Lignotech USA, Rothschild, WI). Ration formulation and composition are shown in Table 1
and ingredient composition is shown in Table 2. Diets provided 4 concentrations of dietary RDP (% of DM) while RUP was formulated to remain constant at 5.8% of DM: 1) 6.8% RDP, 12.3% CP; 2) 8.2% RDP, 13.9% CP); 3) 9.6% RDP, 15.5% CP; and 4) 11.0% RDP, 17.1% CP.
|
|
were fitted to a nonlinear model using the Marquardt iterative method as described previously by Erdman et al. (1987). Predicted CP degradation was calculated according to the NRC (2001) using feed analysis and estimated passage rates for the cows and rations in this study.
|
. The lowest RDP diet was estimated to provide 69 or 68% of required RDP, and the highest RDP diet was estimated to provide 114 and 111% of required RDP, according to the NRC 1989 and 2001 models, respectively. All diets were formulated to provide 98 or 127% of the RUP requirement for lactating cows for the NRC 1989 and 2001 models, respectively.
|
Data Analysis
Prior to the start of the experiment, treatments were randomly assigned to cows blocked into squares by parity and management. Because the row effect of the Latin square was period and was shared by all squares, the experiment was statistically analyzed as a Latin rectangle (Mead et al., 1993). Data were analyzed using the mixed model procedure of SAS (SAS Institute, 1996). The statistical model was Y = treatment + parity + management + (treatment x parity) + (treatment x management) + (parity x management) + (treatment x parity x management) + period + cow (parity, management). The random effect was cow nested within parity and management. Interactions were not significant and were dropped from the model in a stepwise manner. Effects from the pattern of bST injections are part of the period effect. Unless noted in the tables and text, the management effect, which includes the effect of bST injections, was not significant. When the treatment effect was significant (P < 0.10), treatment least squares means were separated using the PDIFF test (SAS Institute, 1996) where P < 0.05, and orthogonal polynomials were used to test linear (weighted 3, 1, –1, –3), quadratic (weighted 1, –1, –1, 1), and cubic responses (weighted –1, 3, –3, 1) of increasing concentrations of RDP in the diet (Gill, 1978). Cubic responses were not significant (P > 0.1), and therefore were not included in the tables or text.
In practice, diet formulation requires the prediction of both DMI and the requirement and supply of nutrients so that target feed nutrient concentrations can be determined. Our evaluation of the NRC models (1989 and 2001) separated the DMI prediction from that of protein requirement and supply. The predicted DMI was determined for each individual cow using the observed milk production. Then, the protein requirement, supply and amount of milk (allowable milk) that could be produced by this amount of protein were determined using the individually measured DMI. Both of these predictions were evaluated against the measured DMI or milk production. The average difference in prediction vs. observed was taken as the mean bias. The root mean square prediction error (RMSPE) was calculated as: 
{
[(predicted – observed)2]/n} (Bibby and Toutenburg, 1977). Residuals (predicted – observed) were plotted against predicted values of DMI to identify slope biases. Residuals were plotted against observed values of milk production because allowable-milk predictions were made by reversing the equations developed by regression within the NRC models. Therefore, although it is customary to plot residuals vs. predicted values to avoid detection of unmeaningful slope bias, the opposite is true when the equations have been reversed (Kohn et al., 1998). The residuals for allowable milk were plotted against observed milk minus the milk production for the high RDP diet (control). The control milk was predicted for individuals fed lower RDP diets by using the prediction coefficients for the cow and period of those individuals and the coefficient for the control diet.
An economic analysis was conducted to compare the increased cost of feed for higher RDP diets to the additional value of milk components produced on those diets. Five-year average prices (2000 to 2004) were used for soybean meal ($0.22/kg), corn grain ($0.097/kg), and milk ($0.28/kg). Grain prices were obtained indirectly from the Chicago Board of Trade from Capitol Commodity Services, Inc. (2005). Soy Pass was estimated to be $0.03/kg higher than that of soybean meal. Milk component prices were calculated as the average price for Class II components published by the USDA (2005). Five-year average price of milk fat was $3.33/kg, protein was $4.68/kg, and price of other solids was $0.15/kg.
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
, both NRC models (1989 and 2001) predicted similar deficiencies, and RDP concentrations ranged from 68 to 111% of NRC (2001) requirements. Both models also predicted RUP concentrations to be similar across treatments, but the NRC 1989 model predicted RUP to average 98% of requirement whereas the 2001 model predicted RUP to average 127% of requirement. Both models predicted similar RUP supply, but NRC 2001 predicted a lower RUP requirement. Thus, NRC 2001 did not predict the 9.6 or 11.0% RDP diets to be deficient in metabolizable protein but NRC 1989 predicted the 9.6% RDP diet to be slightly deficient and the higher RDP diet was predicted to be adequate in total protein. The NRC 1989 model predicted greater losses in milk production for the deficient diets.
Dry matter intake was not affected by dietary RDP concentration (Table 5), but multiparous cows consumed more than primiparous cows (21.7 vs. 20.5 kg/ d; P < 0.05). When nutrients like protein are provided in insufficient quantities for microbial fermentation, rate of digestion and DMI can be reduced (Faverdin, 1999). In corn silage-based diets, cows fed increasing levels of protein and RDP, provided by soybean meal or urea, resulted in greater DMI in some studies (Wohlt and Clark, 1978; M’hamed et al., 2001), but not others (Armentano et al., 1993). In the present study, feed intake may not have been limited by rumen bulk because of the high-energy diet, and therefore, potentially reduced microbial fermentation from low RDP diets may not have affected intake (Allen, 1999). In agreement with treatment formulation, intake of RDP increased from 1.41 to 2.38 kg/d for the lowest to the highest RDP diets. Intake of RUP and NEL also increased slightly as DMI increased nonsignificantly with increasing RDP.
|
). Cows fed the lowest RDP diet produced 2.1 kg/d less milk than cows fed the highest RDP diet indicating that RDP was deficient in the low RDP diet. Although differences in milk production were not significant between any diet and the next closest treatment, the linear trend extended to the full range of the treatments. The primary evidence for feeding dietary RDP at concentrations greater than 9.5% is from in vitro continuous culture experiments (Stokes et al., 1991b). As long as carbohydrates were not limiting, bacterial N and bacterial efficiency continued to increase as RDP increased in the diets (Stokes et al., 1991b). In a metabolism study, cows fed diets formulated to contain 11.8 or 13.7% RDP supported greater microbial protein synthesis than diets containing 9% RDP (Stokes et al., 1991a), lending support to the theory that greater RDP will support greater microbial protein synthesis and consequently, greater milk production. However, in a production study with lactating dairy cows (Armentano et al., 1993), increasing RDP from 9.5 to 11.7% RDP did not result in increased milk production, indicating that there was no benefit to feeding diets with RDP greater than 9.5%. In the present production study, milk production was highest for cows fed the highest RDP diets, but the study was not designed to determine if even greater milk production could have been obtained from feeding even more RDP. Fat-corrected (4.0%) milk improved from 30.3 to 33.1 kg/d as RDP increased in the diets. Cows fed the lowest RDP diet produced less FCM than cows fed the 2 higher RDP diets (P < 0.05). Cows fed to meet RDP requirements produced the greatest concentrations and yields of milk fat. Fat percentage was 3.70% for cows fed the lowest RDP diets and increased linearly to 3.86% as RDP increased in the diets (P < 0.01). This observation supports research which demonstrated that providing additional RDP increases fat percentage and yield (M’hamed et al., 2001), although others have not observed fat percentage or yield differences when dietary RDP was increased (Armentano et al., 1993). In a summary of previous literature, the NRC (2001) reported higher milk fat percentage with improved methionine and lysine nutrition; however, results were variable. Methionine and lysine may play a role in milk fat synthesis through increased de novo synthesis of short-and medium-chain fatty acids or through increased synthesis of chylomicrons and very low-density lipoproteins (NRC, 2001); however, limited data are available. In the current study, increased AA absorbed from the small intestine may have increased the supply of precursors available to increase milk fat percentage for cows fed increasing concentrations of RDP.
Increasing RDP from 6.8 to 11.0% in the diets of lactating cows increased milk CP concentration from 2.95 to 3.11% and protein yield from 0.94 to 1.05 kg/d (Table 5). Cows fed to meet RDP requirements produced the greatest amount of protein in milk. Cows fed the lowest RDP produced milk with a lower protein percentage than cows fed the other diets (P < 0.001). The relationship between CP supplied in the diet, mainly as RDP, and milk protein production is illustrated in Figure 1. Increased milk protein percentage and milk protein production in response to increasing RDP in the diet may be a result of providing additional N for ruminal microbial protein synthesis.
|
indicates that 19% of CP added to the diet was eventually converted to milk CP. The NRC (2001) estimates that when RDP is limiting to microbial growth, 85% of RDP can be converted to microbial CP. This microbial CP is 64% metabolizable, and the metabolizable CP is used for lactation at 67% efficiency after accounting for maintenance, lactation, and growth. These coefficients factor together to 36%. The previous NRC (1989) calculations assumed 90% efficiency of RDP use for microbial growth, plus 15% of CP intake recycled to the rumen, 64% digestibility of bacterial CP, and 70% efficiency of use of absorbed protein for lactation. These coefficients factor together to an assumed efficiency of conversion of 47%. Both estimates are greater than 19% observed in this study; however, our slope may be skewed because some cows may have been fed RDP in excess of requirements. When we excluded the data from the highest RDP diet, the slope was unaffected (0.22; SE = 0.021), suggesting a linear response to RDP deficiency at a lower magnitude than predicted by the NRC 1989 and 2001 models. Multiparous cows produced 1.8 kg/d more milk than primiparous cows (P < 0.08), but there was no parity by treatment interaction, indicating that the RDP requirements for primiparous and multiparous cows are similar once feed intake and animal performance are taken into account. Primiparous and multiparous cows were not significantly different in fat percentage, fat yield, protein percentage, protein yield, lactose percentage, or lactose yield.
Body weight did not differ by dietary treatment group (Table 5). Multiparous cows weighed more than primiparous cows in this study (607 vs. 562 kg; P < 0.01). There was a treatment by lactation interaction (P < 0.001) for BCS (data not shown); for primiparous cows, BCS declined as RDP increased in the diet (3.37, 3.27, 3.16, and 3.14 for cows fed the 4 diets in order of increasing RDP concentration), but for multiparous cows, BCS increased as RDP increased in the diet (2.74, 2.82, 2.85, and 2.89 for cows fed diets increasing in RDP concentration). It is not clear why BCS would decline for primiparous, but not for multiparous cows as RDP increases in the diet.
Maximizing synthesis of microbial protein as a relatively inexpensive source of readily digestible protein in the small intestine is desirable; however, inefficiency of protein use within the animal increases as RDP increases, causing concern of increased N excreted as waste. In this experiment, N efficiency declined from 36.5 to 28.2% as RDP increased from 6.8 to 11.0% of DM (Table 6; Figure 2). Within the treatment in which cows were fed the lowest RDP diet, N efficiency decreased as CP intake increased (P < 0.001); however, this response within treatment was not observed (P > 0.10) when cows were fed the diets with higher concentrations of RDP. The high efficiency of N use for the low RDP diets must be largely attributed to the highly efficient use of RUP in the base diet. As RDP was added to the diet, with only 19% efficiency of use for milk protein, the efficiency of CP use was diluted even as milk protein increased.
|
|
. In agreement with Kalscheur et al. (2000) and Hristov et al. (2004), predicted urinary N excretion increased linearly as RDP increased in the diet. In disagreement with Kalscheur et al. (2000) and Hristov et al. (2004), predicted fecal N estimate, which includes retained N, increased significantly as RDP increased in the diet. Most total collection experiments (Dinn et al., 1998; Kalscheur et al., 2000) show protein retention increases with protein concentration in the diet. However, this increase may be a result of measurement error rather than true retention (Spanghero and Kowalski, 1997). If N retention increased from 25 to 45 g/d as RDP increased as indicated by Kalscheur et al. (2000) using similar diets, then fecal N would be similar across treatments consistent with Kalscheur et al. (2000).
To accurately formulate diets for dairy cattle, the requirements for the expected level of milk production need to be balanced against the expected supply of nutrients from the feeds available. Predicting DMI is a critical component to predicting supply of nutrients. Although DMI can be measured for groups of cows on a farm, diet formulation requires knowledge of the DMI of individual cows when they are fed the new diet. Therefore, accurate DMI predictions are essential for accurate diet formulation. Figure 3 shows the individual cow DMI predicted by each model minus that observed in this study vs. the predicted DMI. The NRC 1989 model predicted an average of 1.0 kg/d lower DMI than observed, whereas the 2001 model predicted an average of 1.2 kg/d greater DMI than observed. Thus, the different predictions of DMI explain a part of the reason the NRC 1989 model formulated for a higher RUP content than the NRC 2001 model. The NRC 1989 estimates DMI as the amount needed to match energy requirements per day with the energy supplied by feeds: DMI (kg/d) = energy required per day/energy provided per kilogram of feed. When a feed of greater estimated energy content is offered, the DMI is expected to be lower. The NRC 2001 model estimates DMI using an empirical model that includes milk yield, BW, and DIM. This prediction is insensitive to energy content of the feed.
|
. Negative values on the x-axis represent losses in milk production compared with that expected for that cow and period if fed the 11% RDP diet. The actual milk production and DMI were used to predict allowable milk for both NRC models (1989 and 2001). A perfect model would accurately predict milk yield when protein is limiting production (0 on the y-axis when values on the x-axis are negative), and predict equal or greater milk yield than observed when protein may not be limiting milk production (positive values on the y-axis when values on the x-axis are positive). For the 2 dietary treatments with the greatest losses in milk production (6.8 and 8.2% RDP), both NRC models (1989 and 2001) overestimated milk production loss to a similar extent. However, the NRC 2001 model failed to predict that a loss would occur on the 9.6% RDP diet, and the 1989 model predicted the slight loss accurately. The NRC 2001 model also predicted greater DMI than the 1989 model so that assuming predicted DMI would have further supported the apparently incorrect assumption that the 9.6% RDP diet was adequate (Table 4).
|
To optimally formulate diets for dairy cows, the risks to the environment from overfeeding N need to be balanced against the risks of lost milk production from underfeeding. The results from the present study help quantify the effects of underfeeding RDP on milk production and urinary N excretion. In this study, RDP was increased by replacing 5% of the diet DM supplied as corn grain and protected soybean meal with regular soybean meal. The DMI did not change and averaged 21 kg/d. Thus, 0.7 kg/d corn grain and 0.47 kg/d protected soybean meal (3 or 2% substitution x 21 kg/d / 90% DM) were replaced with 1.17 kg of regular soybean meal between each treatment. The cost of this substitution would be $0.072/d per cow (1.17 x $0.220–0.7 x $0.097–0.47 x $0.25). The average change in milk yield between treatments was 0.7 kg/d, average change in milk fat was 43.3 g/d, and change in milk protein was 36.7 g/d. Thus, the average change in value of milk between treatments was: $0.32/d per cow ($3.33/kg x 0.0433 kg fat/d per cow + $4.68/kg x 0.0367 kg of protein/ d per cow + $0.15 x 0.7 kg of milk/d per cow x 0.057 kg of other solids/kg of milk). Feeding below RDP requirement by substituting corn and protected soybean meal for regular soybean meal resulted in 4.4 times greater loss in milk income as would have been saved in feed cost.
The loss in milk production currently far outweighs the benefits of reduced N excretion. However, further research into balancing rations for groups of cows may find it acceptable to risk some loss in milk production of the cows with the greatest requirements to reduce N excretion of an entire group. To proceed with developing models to balance risks and benefits of protein feeding levels, particularly for groups of animals, the losses from underfeeding CP need to be quantified.
This study demonstrates that even when feeding below CP requirements, as dietary CP increases, the efficiency of N use declines and the amount of urinary N losses increases. Thus, diets formulated for maximal milk production may not be optimal to minimize N excretion per unit of milk produced.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
FOOTNOTES |
|---|
2 Current address: Dairy Science Department, South Dakota State University, Brookings 57007.
3 Current address: Biotechnology Industry Organization, 1225 I Street, NW, Ste 400, Washington, DC 20005-5958.
Received for publication March 17, 2005. Accepted for publication September 2, 2005.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  L. Li, J. Cyriac, K. F. Knowlton, L. C. Marr, S. W. Gay, M. D. Hanigan, and J. A. Ogejo Effects of Reducing Dietary Nitrogen on Ammonia Emissions from Manure on the Floor of a Naturally Ventilated Free Stall Dairy Barn at Low (0-20{degrees}C) Temperatures J. Environ. Qual., October 29, 2009; 38(6): 2172 - 2181. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. A. Broderick, M. J. Stevenson, and R. A. Patton Effect of dietary protein concentration and degradability on response to rumen-protected methionine in lactating dairy cows J Dairy Sci, June 1, 2009; 92(6): 2719 - 2728. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Cyriac, A. G. Rius, M. L. McGilliard, R. E. Pearson, B. J. Bequette, and M. D. Hanigan Lactation Performance of Mid-Lactation Dairy Cows Fed Ruminally Degradable Protein at Concentrations Lower Than National Research Council Recommendations J Dairy Sci, December 1, 2008; 91(12): 4704 - 4713. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. A. Broderick, A. F. Brito, and J. J. O. Colmenero Effects of Feeding Formate-Treated Alfalfa Silage or Red Clover Silage on the Production of Lactating Dairy Cows J Dairy Sci, March 1, 2007; 90(3): 1378 - 1391. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
), 8.2 (
), 9.6 (
), and 11.0 (x) % of DM. The equation is Y (milk protein, g/d) = 425 (SE = 55) + 0.19 (SE = 0.018) x (CP, g/d) with R2 = 0.39.
