| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
,3
* Department of Dairy Science, University of Wisconsin, Madison, 53706
USDA, ARS, US Dairy Forage Research Center, Madison, WI 53706
3 Corresponding author: gbroderi{at}wisc.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Key Words: omasal flow • microbial protein • dairy cow
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
The objectives of this study were to evaluate the effects of different dietary AS:CS ratios on RDP supply and omasal flows of RUP, microbial protein, individual AA, and other nutrients in lactating dairy cows. It is important to emphasize that when AS was replaced with CS, SSBM was increased and rolled high-moisture shelled corn (HMSC) was decreased in the diet. This approach was adopted because diets were formulated to be isonitrogenous.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Sampling and Laboratory Analyses
Spot samples of omasal digesta leaving the rumen were obtained from the 8 ruminally cannulated cows using the omasal sampling technique described by Huhtanen et al. (1997), Ahvenjärvi et al. (2000) and Reynal and Broderick (2005). These omasal digesta markers were used: indigestible NDF [(INDF); Huhtanen et al., 1994] for the large particle phase (LP), YbCl3 (modified from Siddons et al., 1985) for the small particle phase (SP), and CoEDTA (Udén et al., 1980) for the fluid phase (FP). A marker solution containing YbCl3, CoEDTA, and 15NH4SO4 with 10 atom% excess (APE) 15N (Isotec; Miamisburg, OH) was prepared as described by Reynal and Broderick (2005). Just before beginning marker infusion, samples of whole ruminal contents were taken from each cow to determine the background abundance of 15N. The marker solution was continuously infused into the rumen for approximately 135 h at a constant rate of 2.76 L/d (providing 2.33 g of Co, 3.35 g of Yb, and 0.22 g of 15N per day) from d 21 to 26 using 2 syringe pumps (model no. 33; Harvard Apparatus, Inc., Holliston, MA). Omasal sampling began approximately 64 h after beginning marker infusion; samples were taken 4 times daily at 2-h intervals over 3 consecutive days to represent the 24 h-feeding cycle: 0000, 0200, 0400, 0600 (d 24), 0800, 1000, 1200, 1400 (d 25), and 1600, 1800, 2000, and 2200 h (d 26). Before each sampling, it was necessary to confirm the location of the sampling tube, and occasionally it had to be repositioned in the omasal canal. Sometimes it also was necessary to unplug the holes in the end of the sampling tube because of the presence of coarse digesta. At each sampling time, 300 mL of omasal sample was collected and split into 2 subsamples (of 100 and 200 mL). The four 100-mL subsamples were pooled and stored on ice to yield one daily composite of 400 mL from each cow, which was used for bacterial isolation. Therefore, 3 daily composites were obtained over the 3 sampling days for each cow. The four 200-mL subsamples were pooled and stored at –20°C over all 3 d to obtain a single 2.4-L composite from each cow in each period for later separation into the 3 omasal phases. On the second sampling day, an extra 500 mL of omasal digesta (total sample, 800 mL) also was collected at the last daily sampling time and used for protozoal isolation.
Protozoa were isolated using a modification of the procedure of Hristov et al. (2001). The 500-mL samples were squeezed through 2 layers of cheesecloth and the retained solids were washed with 500 mL of McDougall’s (1948) buffer containing 0.5 g of glucose and 50 mg of cysteine-HCl/100 mL that had been warmed to 39°C. Filtrates were transferred to 1-L separatory funnels and placed in a water bath at 39°C. After approximately 45 min, separatory funnels were removed from the water bath and the distinct white layers of sedimented protozoa that had formed on the bottom of each separatory funnel were carefully drawn off into vials and stored on ice until transported to the laboratory. At the laboratory, 20 mL of sucrose solution (30% wt/ vol) was added to 50-mL centrifuge tubes. The protozoal sediments were poured on top of the sucrose solution, and the tubes were centrifuged at 150 x g for 3 min at 4°C. Supernatants were discarded and the resulting pellets were washed 3 times with saline (0.85% wt/vol of NaCl), followed by centrifugation (1,200 x g for 5 min at 4°C). These pellets were frozen for about 12 h, then transferred from centrifuge tubes to vials and stored at –20°C until freeze-dried. The fluid-associated bacteria (FAB) and particle-associated bacteria (PAB) were isolated from the daily 400-mL subsamples from each cow on each of the 3 sampling days using filtration and differential centrifugation as described earlier in detail (Reynal and Broderick, 2005). Both the FAB and PAB pellets were stored at –20°C until freeze-dried. After freeze-drying, both bacterial samples and the protozoal sample were ground with a mortar and pestle and equal DM from each was pooled by cow per period for later analysis.
The 2.4-L pooled omasal composites were thawed at room temperature, separated into the 3 omasal phases (LP, SP, and FP) as described by Reynal and Broderick (2005), and stored at –20°C until freeze-dried. After freeze-drying, these samples were ground through a 1-mm screen (Wiley mill; Arthur H. Thomas, Philadel-phia, PA) and then analyzed for markers. Concentrations of Co, Yb, and INDF in LP and SP and Co and Yb in FP were determined using the methods detailed earlier (Reynal and Broderick, 2005). Marker concentrations were used to physically recombine DM from the freeze-dried FP, SP, and LP in the correct proportions to reconstitute the omasal true digesta (OTD) flowing out of the rumen using the triple-marker method of France and Siddons (1986). Concentrations of Co, Yb and INDF were distinctly greater in, respectively, the FP, SP and LP, thus allowing for successful application of the triple-marker method. Dry matter from SP and LP also were mixed in the correct proportions based on the markers to yield a 2-g sample that was ground through a 0.5-mm screen (Udy cyclone mill; Udy Corporation; Fort Collins, Co) and defined as small plus large particle phase (SP + LP).
Omasal true digesta samples were analyzed for total N (Leco 2000; Leco Instruments, Inc., St. Joseph, MI), absolute DM (105°C), ash and OM (AOAC, 1980); sequentially for NDF and ADF using heat stable 
-amylase and Na2SO3 (Hintz et al., 1995); and for neutral detergent insoluble nitrogen (NDIN) and ADIN. Extracts were also prepared from OTD samples as follows: 10 mL of Na-citrate buffer (pH 2.2; 77.5 mM Na citrate) was added to 0.5 g of dry sample and then vortexed. After 30 min in a warm room (39°C), extracts were centrifuged at 15,000 x g, for 15 min at 4°C and the supernatants stored at –20°C for later analysis of ammonia and total free AA using assays (Broderick and Kang, 1980) adapted to flow-injection (Lachat Quik-Chem 8000 FIA; Lachat Instruments; Milwaukee, WI). For determination of individual AA analysis, OTD samples were hydrolyzed for 24 h at 110°C in sealed vials under N2 atmosphere containing 6 N HCl with 0.1% wt/vol phenol (Mason et al., 1979). The ratio of sample N to acid was about 1 mg of total N per 5 mL of 6 N HCl. After hydrolysis, samples were cooled, HCl removed by evaporation, and the residue was redissolved in pH 2.2 sample buffer containing norleucine as the internal standard. Analysis of individual AA was conducted using ion-exchange chromatography with ninhydrin detection (Beckman 6300 amino acid analyzer; Beckman Instruments, Inc., Palo Alto, CA).
Samples of FAB, PAB, protozoa, FP, and SP + LP were prepared for total NAN and 15N analyses as follows: approximately 100 µg of N from each sample was weighed into tin cups (Elemental Microanalysis Limited, Okehampton, UK) followed by addition of 50 µL of 72 mM K2CO3. The tin capsules were placed in 96-well microtiter plates and dried in a 60°C oven overnight to volatilize ammonia. Samples of ruminal digesta collected prior to infusion for determination of background abundance of 15N were freeze-dried and ground to pass through a 1-mm Wiley mill screen and then a 0.5-mm Udy mill screen. These samples also were prepared for NAN and for 15N analyses as described above. Nitrogen and 15N were determined using a Carlo-Erba instrument interfaced with an isotope ratio mass spectrometer (University of California–Davis Stable Isotope Facility). Bacteria (FAB and PAB), protozoa, FP, and SP + LP also were analyzed for absolute DM, ash, and OM (AOAC, 1980).
Calculations
Omasal flow of NAN was determined by the difference between total N and ammonia N flows. Total NAN flowing past the omasum was assumed to be composed of PAB NAN, FAB NAN, and nonammonia nonmicrobial N (NANMN). In the calculations, 15N enrichment was defined as 15N APE above the background 15N measured in ruminal samples from individual cows collected in each period. The mean (± SD) background 15N during trial was 0.36823 (± 0.00027) atom%. The 15N APE above background (15N enrichment) for samples from each cow in each period was calculated as
 |
Assuming that FAB and PAB were representative of bacteria flowing with the FP and the SP + LP, respectively, FAB NAN, PAB NAN, total microbial NAN, NANMN, and RUP flows at the omasal canal, RDP supply, OM truly digested in the rumen (OMTDR), and efficiency of microbial NAN synthesis were all calculated as follows:
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
where flows and intakes are in grams per day or kilograms per day and NAN concentrations are in grams per gram of OM.
Statistical Analysis
Data were analyzed using the mixed procedures of SAS (SAS Institute, 1999–2000) for a replicated 4 x 4 Latin square design according to the following model:
 |
where Yijkl is the dependent variable, µ is the overall mean, Si is the effect of square i, Pj is the effect of period j, Ck(i) is the effect of cow k (within square i), Tl is the effect of treatment l, STil is the interaction between square i and treatment l, and Eijkl is the residual error. All terms were considered fixed, except Ck(i) and Eijkl, which were considered random. The interaction term was removed from the model when P > 0.25. Significance was declared at P 
0.05 and trends at 0.05 < P 0.10. All reported values are least squares means, which were separated using the PDIFF test in SAS. Because the proportions of AS and CS among treatments were not equally spaced, linear, quadratic, and cubic effects of the AS:CS ratio were tested by partitioning the degrees of freedom for diet into single degrees of freedom corresponding to linear, quadratic, and cubic effects. The polynomial regression model included square, period, linear effect of forage, quadratic effect of forage, and cubic effect of forage. Forage is used to indicate either the decreasing levels of AS or the increasing levels of CS in the diet. Cubic effects were not statistically significant (P 
0.10) for any variable and are therefore not reported.
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
). Cows fed diets B and C had the highest DMI, those fed diet A were intermediate, and those fed diet D were lowest. A quadratic effect (P = 0.01) also was observed with maximum DMI at 35% dietary AS (Table 2). The pattern observed in the companion production study was somewhat different in that DMI was highest on diets A and B, intermediate on diet C, and lowest on diet D, with the quadratic maximum found at 51% AS (Brito and Broderick, 2006). Dry matter flow at the omasum decreased linearly (P < 0.01) from 17.6 kg/d on diet A to 15.9 kg/d on diet D (Table 1). No significant difference in DM flow was observed among diets A, B, and C but DM flow was lower (P < 0.01) on diet D (Table 1). A quadratic effect (P = 0.04) with maximum at 41% AS (Table 2) also was observed for DM flow. Dry matter apparently digested in the rumen did not differ, averaging 9.2 kg/d among treatments (Table 1). However, a significant quadratic response (P = 0.05) with maximum at 31% dietary AS (Table 2) was observed for this variable. Similarly, DM apparently digested in the rumen, expressed as a percentage of DMI, did not differ and averaged 35% across diets. Apparent digestibility of DM in the total tract (Brito and Broderick, 2006) also was similar among diets, with an overall mean of 66%. Therefore, 53% of apparent total tract digestion of DM occurred in the rumen.
|
|
Organic matter intake did not differ (P = 0.06; Table 1
) among diets, but a quadratic effect (P = 0.01) with maximum at 33% AS (Table 2) was found for this variable. The pattern of omasal OM flow was similar to that for DM; OM flows for diets A, B, and C averaged 14.4 kg/d and were 8% higher (P < 0.01) than that on diet D (Table 1). However, OM apparently digested in the rumen, expressed either as an amount or as a percentage of OM intake, did not differ with changing AS:CS ratio and averaged 10.2 kg/d and 42% (Table 1). A quadratic effect (P = 0.04) was observed for amount of ruminal OM digestion with maximum at 31% dietary AS (Table 2). Organic matter truly digested in the rumen, expressed either as an amount or as a percentage of intake, was not different among treatments and averaged 15.6 kg/d and 64% (Table 1). Again, a quadratic effect (P = 0.02) was detected for amount of OM digested with maximum at 33% AS (Table 2), and a quadratic trend was detected for the proportion of OM digested (P = 0.09).
Apparent total tract digestibility of OM also was similar among diets, with an overall mean of 67% (Brito and Broderick, 2006). Thus, an estimated 63% of the apparent total tract digestion of OM occurred in the rumen. According to Titgemeyer (1997), expected values for OM apparently digested in the rumen and OM truly digested in the rumen should vary from 30 to 60% and from 40 to 70% of OM intake, respectively, and are diet dependent. Because wide ranges were given to prevent exclusion of valid data, results from a flow study could fall within them but still be inaccurate (Titgemeyer, 1997). In the current study, OM apparently and truly digested in the rumen averaged, respectively, 42 and 64% of OM intake and fell within the ranges suggested by Titgemeyer (1997). Reynal and Broderick (2005) reported means of 44 and 65%, and values from the Finnish group (Ahvenjärvi et al., 2000; Ahvenjärvi et al., 2002; Korhonen et al. 2002) averaged 56 and 72% for, respectively, apparent and true ruminal OM digestibilities. In the Finnish studies, apparent total tract OM digestibility averaged 73%; thus, 77% of the apparent total tract OM digestion occurred in the rumen. The difference of 14 percentage units between this value and the 63% we observed may be attributable to DMI being 25% greater in our study.
Quadratic effects (Table 1) were observed for intake of NDF (P = 0.03) and ADF (P = 0.05), which reflected the quadratic response observed for DMI. Maximal NDF and ADF intakes were predicted at, respectively, 33 and 51% dietary AS (Table 2). Neutral detergent fiber entering the omasal canal and apparently digested in the rumen did not differ among diets, and an average 31% of dietary NDF was digested in the rumen. Total tract NDF digestion averaged 36% of intake (Brito and Broderick, 2006); thus, about 86% of total NDF digestion occurred in the rumen. Previously, an overall mean of 81% of total NDF digestibility was found to occur in the rumen (Brito, 2004). According to Titgemeyer (1997), for ruminants fed mixed diets, ruminal digestion usually accounts for more than 80% of the total fiber digestion. Unlike NDF, ADF intake decreased linearly (P < 0.01) with increasing dietary CS plus SSBM (Table 1) because of the reduction of DMI (Table 1) and dietary ADF content (Brito and Broderick, 2006) and was lowest (P < 0.01) on diet D. The flow of ADF decreased linearly (P < 0.01) from diet A to diet D, being lowest (P < 0.01) on diet D (Table 1). The amount of ADF apparently digested in the rumen also declined linearly (P < 0.01) with increasing CS, which supports the hypothesis that fiber digestion was reduced because of greater fluctuation of ruminal pH when CS plus SSBM replaced AS plus HMSC in the diet (Brito and Broderick, 2006).
Intake of CP decreased linearly (P < 0.01) from 4.53 kg/d on diet A to 3.98 kg/d on diet D (Table 1). No differences in CP intake were observed among cows fed diets A, B, or C but those offered diet D consumed less (P < 0.01). Crude protein entering the omasal canal was greatest on diets B and C, intermediate on diet A, and lowest on diet D (Table 1). Quadratic responses were detected for CP intake (P = 0.02) and omasal CP flow (P = 0.01) with the respective maxima at 39 and 34% dietary AS (Table 2). When omasal CP flow was expressed as a percentage of CP intake, a linear increase (P < 0.01) was observed from diet A to diet D, which was consistent with greater RUP flow as AS CP was replaced by CP from SSBM. Significant differences among treatments also were observed (P = 0.03) for this variable.
Microbial Protein Synthesis and Omasal Nitrogen Flows
The supply of RDP decreased linearly (P < 0.01) from 3,068 g/d (diet A) to 2,469 g/d (diet D) when AS plus HMSC was replaced with CS plus SSBM (Table 3). A trend (P = 0.08) for a quadratic effect also was observed. On average, the RDP supply to cows fed diet D was 599, 673, and 340 g/d lower (P < 0.01) than for cows fed, respectively, diets A, B, or C. The linear decrease in RDP resulted partly from N intake declining from 724 g/d on diet A to 637 g/d on diet D (Table 3). Brito (2004) observed a greater RDP supply on higher NPN diets containing AS than on lower NPN diets containing red clover silage. The RDP supply decreased linearly from 11.7 to 10.1% of DMI from diet A to diet D (Table 3). The NRC (2001) model predicted that RDP would decline from 12.9 to 10.2% of DMI (Table 3), indicating a small discrepancy of RDP supply in the diet with the greatest proportion of AS (Table 4). This suggested that the NRC model yielded reliable predictions for RDP supply under the conditions of this study.
|
|
). However, when expressed as a percentage of DMI, there was a linear increase (P = 0.03) from 5.59% on diet A to 6.13% on diet D. This increase was expected because CP from SSBM is more resistant to ruminal degradation than that from AS. Similarly, RUP predicted by the NRC (2001) model increased from 4.36% of DMI on diet A to 6.08% of DMI on diet D (Table 3). Therefore, underpredictions of RUP flow by the NRC model ranged from 21% (diet A) to 0.7% (diet D; Table 4).
Flow of both NDIN and ADIN at the omasal canal decreased linearly (P < 0.01) when CS plus SSBM replaced AS plus HMSC (Table 3), which was expected because the concentrations of both declined sharply with reduced dietary AS (Brito and Broderick, 2006). Acid detergent insoluble nitrogen, fraction C in the Cornell Net Carbohydrate and Protein System, is composed of heat-damaged proteins such as the Maillard products as well as proteins associated with tannins and lignin (Sniffen et al., 1992). This fraction is considered to be indigestible in the rumen and intestines. Although ADIN flow on diet A was 16, 45, and 65% greater than on diets B, C and D, overall ADIN flow was probably too small to affect N utilization. Subtracting ADIN from NDIN yields fraction B3 of the Cornell model. The flow of B3 did not differ among diets, averaging 15 g/d (Table 3). Fractional degradation rates reported for this fraction range from 0.06 to 0.55%/h and are much slower than those reported for fractions B1 (120 to 500%/h) and B2 (4 to 40%/h) according to the NRC (1996).
Total NAN flows differed (P = 0.03) and were greatest on diets B and C, intermediate on diet A, and lowest on diet D (Table 3). A quadratic effect was found (P = 0.01) with maximum at 34% dietary AS (Table 2). Hristov and Broderick (1996) also observed greater NAN flow for cows fed an all-AS diet than for those fed an all-CS diet with supplemental CP added as urea. When omasal NAN flow was expressed as a percentage of N intake, a linear increase (P < 0.01) was observed from diet A to diet D. Treatments also differed (P = 0.02) and values were greatest on diet D, intermediate on diet C, and lowest on diets A and B. Omasal flow of NANMN did not differ, averaging 238 g/d among diets (Table 3). As a proportion of total NAN flow, omasal NANMN was greatest (P = 0.05) on diet C, intermediate on diets B and D, and lowest on diet A. Similarly, a significant dietary effect was observed for omasal flow of NANMN as a percentage of N intake, which was greater on diets C and D (P = 0.02) than on diets A and B (Table 3). Although omasal NANMN, either as a percentage of total NAN flow or as a percentage of N intake, increased on diets with greater proportions of CS plus SSBM, production was not improved (Brito and Broderick, 2006). In fact, yields of milk and milk protein were lowest on diet D. However, low production on this diet could also have been the result of excessive starch intake and of lower ruminal pH and fiber digestion, leading to reduced DMI as suggested previously (Brito and Broderick, 2006).
The NAN contributed by FAB and PAB entering the omasal canal did not differ with changing dietary ratios of AS:CS averaging, respectively, 203 and 254 g/d among diets, and corresponding to 45 and 55% of total microbial NAN flow (Table 3). Brito (2004) observed that FAB and PAB accounted for, respectively, an average of 29 and 71% (AS diets) and 41 and 59% (red clover silage diets) of total microbial NAN flow. Reynal et al. (2005) also found that PAB accounted for a greater proportion of the total microbial NAN flow. It is possible that the greater contribution from PAB was due to the large proportion of ruminal microorganisms associated with particulate matter. Craig et al. (1987), using 15N as a microbial marker, observed that 70 to 80% of microbial OM in whole ruminal contents was bound to the particulate phase. However, Hristov and Broderick (1996) found that, although PAB averaged 78% of the ruminal pool of microbial NAN, FAB and PAB contributed about equally to total microbial NAN flow because the liquid flow rate was more than 3 times the particulate flow rate. Conversely, Olmos Colmenero and Broderick (2006) reported that FAB was 56% and PAB was 44% of the total microbial NAN flow, which was consistent with the findings of Ahvenjärvi et al. (2002). Regardless of any apparent disagreement among studies, literature data and present results show clearly that FAB contribute a large proportion of the total microbial NAN flow.
Total microbial NAN flow, both the amount and proportion of total NAN flow, differed significantly among diets and decreased linearly from diet A to diet D (Table 3). On average, the microbial NAN flow in cows offered diet D was 42, 56, or 37 g/d lower (P < 0.01) than in cows offered, respectively, diets A, B, or C. A quadratic effect (P = 0.02) with maximum at 38% dietary AS (Table 2) also was observed. Hristov and Broderick (1996) reported similar microbial NAN flows averaging 239 g/d in cows fed diets composed of either all AS or all CS. However, the total microbial NAN flow averaged 457 g/d in the present study, in which cows yielded about 2 times more milk and ate 44% more Mcal/d of NEL than in the trial of Hristov and Broderick (1996). It is important to emphasize that the total microbial NAN flow contributed to the greater proportion of the total NAN flow, averaging 66% across diets (Table 3).
In 29 comparisons from 15 metabolism trials, microbial N flow to the duodenum was significantly decreased by feeding high RUP sources in 10 comparisons and was numerically increased by feeding soybean meal in 25 comparisons, indicating that replacing soybean meal with sources rich in RUP resulted in a RDP shortage (Santos et al., 1998). In a meta-analysis of literature reports, Ipharraguerre and Clark (2005) found an overall 7% depression of microbial NAN flow to the small intestine in response to RUP supplementation. It is possible that replacing AS plus HMSC with CS plus SSBM resulted in an inadequate supply of RDP as amino N. Literature data (Chikunya et al., 1996; Griswold et al., 1996; Atasoglu et al., 1999; Carro and Miller, 1999) have clearly shown that amino N supplied as free AA and peptides increased microbial growth or fiber digestion or both. Moreover, many 15N studies (Hristov and Broderick, 1996; Atasoglu et al., 1999; Carro and Miller, 1999; Russi et al., 2002) have shown decreased microbial N derived from ammonia when there was increased availability of amino N. There was a linear reduction in ruminal concentrations of total free AA from diet A to diet D (Brito and Broderick, 2006), and concentrations of total free AA on diet D remained lower (about 2 mM) for longer periods of time compared with the other diets (Figure 1). Thus, total microbial NAN flow on that diet (Table 3) might have been limited by the supply of amino N to the ruminal microbes. Furthermore, isobutyrate decreased linearly (Brito and Broderick, 2006) with the reduced RDP and may have contributed to the linear decline in total microbial NAN flow.
|
). However, no significant differences were observed among diets, and ammonia N flow averaged 6.6 g/d (Table 3). As expected, ammonia flow paralleled ammonia concentration in the rumen, which decreased linearly (P < 0.01) from diet A to diet D (Brito and Broderick, 2006). Brito (2004) also found that ammonia N flow increased on diets having greater ruminal ammonia. Ammonia is often the main N source for microbial protein synthesis (Nolan, 1975; Aharoni et al., 1991) and is essential for the growth of several species of ruminal bacteria (Blackburn, 1965; Allison, 1970; Bryant, 1973). Therefore, a low ruminal concentration of ammonia could depress microbial yield.
Mixed populations of ruminal bacteria are reported to reach maximum protein yields when ruminal fluid has a minimum concentration of 5 mg of ammonia N/ dL (Satter and Slyter, 1974). Kang-Meznarich and Broderick (1980), feeding incremental amounts of urea to nonlactating dairy cows, concluded that microbial protein synthesis was maximized at a ruminal ammonia N concentration of 8.5 mg/dL. Balcells et al. (1993) observed that 5 mg/dL of ruminal ammonia N was needed to avoid significant reduction in the DMI and fermentation rate but microbial protein synthesis was maximized at 11 mg/dL when ewes were fed alkali-treated barley straw alone or supplemented with increasing levels of urea. Results from Reynal and Broderick (2005) suggested that omasal microbial NAN flow was maximized when the ruminal ammonia N concentration was above 11.8 mg/dL in lactating dairy cows fed increasing dietary levels of RDP. In the present trial, ruminal ammonia N on diet D fell below 5 mg/dL for at least 8 h during the day, was significantly lower at 4, 8, and 12 h postfeeding than that on diets A and B, and was lower at 4 h postfeeding than that on diet C (Figure 2). Thus, an inadequate concentration of ruminal ammonia may explain the reduced total microbial NAN flow on diet D.
|
). Thus, the greater apparent N efficiency (milk N:N intake) on diets with lower proportions of AS (Brito and Broderick, 2006) reflected the linear decrease in N intake and reduced RDP supply. Brito (2004) observed that microbial efficiency, but not dietary N efficiency, was greater on AS diets, suggesting that N intake and RDP supply had a greater impact on N utilization by the animal.
Differences between mean total microbial NAN flows measured in this trial and flows predicted by the NRC (2001) model were quite large: 110, 109, 96, and 87 g/ d for diets A, B, C, and D, respectively (Table 4). The NRC model also underpredicted microbial NAN flow on AS diets in other studies (Reynal et al., 2003; Brito, 2004; Reynal and Broderick, 2005). It is possible that underprediction of microbial protein synthesis by the NRC model might be related to the bacterial marker used in its measurement. Most of the data used in the NRC prediction of microbial protein synthesis came from studies in which microbial protein synthesis was quantified using the purines assay of Zinn and Owens (1986). However, Obispo and Dehority (1999) and Makkar and Becker (1999) reported low recoveries of total purines using the Zinn and Owens (1986) assay vs. quantitation using HPLC. Therefore, underestimation of microbial flow may occur in cases of disproportionate recoveries of purines between bacterial isolates and duodenal samples. Reynal et al. (2003) adopted a modified Zinn and Owens (1986) assay that gave improved recovery of purines, and 15N was used in the present trial as well as in the studies of Brito (2004) and Reynal and Broderick (2005). Furthermore, many factors that influence microbial protein synthesis, such as forms of N available in the rumen, ruminal pH, ruminal dilution rate, and source and amount of dietary carbohydrate, fat, and protein, are not considered by the NRC model, which may cause discrepancies between NRC predictions and in vivo measurements of microbial protein synthesis.
No significant differences among diets were observed when microbial protein synthesis was computed from urinary allantoin excretion; this value averaged 261 g N/d across diets in the present trial (Tables 3 and 4), which underestimated microbial flow by an average of 43% compared with total microbial NAN determined using 15N as the microbial marker. Variation in these estimates also was substantially greater than for measurements based on omasal sampling. Krause and Combs (2003) also found no significant differences in microbial protein synthesis estimated from urinary allantoin when comparing a diet with AS as the only forage to a diet with AS:CS at a 50:50 ratio. Perez et al. (1997) reported that urinary excretion of purine derivatives consistently underestimated microbial N flows compared with abomasal flows determined using purines or 15N. Purine derivative excretion yielded underestimates of 29% in sheep (Perez et al. 1996) and 51% in lactating dairy cows (Reynal et al., 2005). However, regressing purine derivative estimates on 15N measurements indicated a good relationship between methods, and Perez et al. (1997) concluded that purine derivatives would serve as a useful indirect marker for detecting treatment effects under practical feeding conditions.
Omasal flows of individual AA are given in Table 5. Omasal Ala flow was lowest (P = 0.05) on diet A, intermediate on diets B and D, and greatest on diet C, indicating a quadratic effect (P = 0.02) with maximum at 27% dietary AS (Table 2). Leucine and total branched-chain AA entering the omasal canal were highest on diets B and C, intermediate on diet A, and lowest in cows fed the diet with the most CS and SSBM (Table 5); the quadratic effects (P < 0.01) had maxima at 33% AS (Table 2). Although omasal Met flow did not differ, averaging 69 g/d across diets (Table 5), a quadratic effect (P = 0.03) with maximum at 32% dietary AS (Table 2) was detected. Duodenal flow of Met predicted by the NRC (2001) model averaged 62 g/d across diets (Table 4), indicating an overall discrepancy of 10% between predicted values and in vivo measurements. Studies have shown that Met and Lys are likely the first and second limiting AA for yield of milk and milk protein secretion of cows on typical North American diets (King et al., 1990; Schwab et al., 1992a; Schwab et al., 1992b). Milk and milk protein yields were maximized at dietary AS:CS ratios of 37:13 and 31:19, respectively (Brito and Broderick, 2006), which were close to the 32:18 AS:CS ratio that maximized omasal Met flow. Flow of Lys at the omasal canal tended (P = 0.07) to be lower on diet D (Table 5), which was expected because of the low Lys content of corn protein. A trend (P = 0.08) for linear reduction of omasal Lys flow also was observed as CS plus SSBM replaced AS plus HMSC in the diet. Duodenal Lys flows predicted by the NRC (2001) model averaged 215 g/d across diets (Table 4), indicating an overall difference of 9% between measured and predicted data. The greatest discrepancy was observed for diet D (15%), suggesting that the NRC model was less able to account for the altered Lys supply with when more CS and SSBM were fed. According to the NRC (2001) model, optimal utilization of MP for maintenance and milk protein production requires Lys and Met concentrations in MP that approximate 7.2 and 2.4% or an optimum ratio of Lys:Met in MP of 3.0:1.0. Across diets, the Lys:Met ratio in digesta passing the omasum averaged 2.85:1, whereas that predicted by the NRC model averaged 3.48:1 (Table 4). This suggested that, although the NRC model predicted a Lys:Met ratio indicating that Met was limiting, results based on Lys and Met flows measured at the omasum suggested that the Lys supply limited production. The Lys:Met ratios calculated from omasal flows of Lys and Met plus intestinal digestibility coefficients from the NRC model were nearly identical to those obtained directly from omasal Lys and Met flows (Table 4).
|
) on diet D compared with diets B and C, respectively, which may further account for the poorer milk yield on diet D. Omasal flow of essential AA, nonessential AA, and total AA all responded quadratically (Table 5) with maxima at dietary AS:CS ratios of, respectively, 34, 33, and 33% dietary AS (Table 2). Moreover, Glu, Ile, and Phe flows to the omasum also responded quadratically, with maxima at 29, 33, and 36% dietary AS. Overall, these data showed that AS and CS are complementary forages with maximal AA flows occurring at dietary AS:CS ratios ranging from 32:18 to 34:16.
Microbial Composition
The OM content of both FAB and PAB did not differ significantly among diets and averaged, respectively, 83 and 87% (Table 6). However, the OM content of PAB tended (P = 0.07) to differ with the changing dietary AS:CS ratio and a linear effect (P = 0.01) also was observed for this variable. This was in agreement with previous findings (Martin et al., 1994; Rodriguez et al., 2000). The contents of NAN in both FAB and PAB increased linearly (P< 0.01) when CS plus SSBM replaced AS plus HMSC and was greatest on diet D (Table 6). In addition, the NAN contents of FAB and PAB were very similar, averaging 7.1 and 7.4%, respectively, across diets in the current study. These findings were similar to an earlier trial (Brito, 2004). Conversely, Martin et al. (1994) and Martin-Orue et al. (1998) reported a greater N content in FAB than in PAB.
|
). The decrease in the ruminal ammonia pool as CS plus SSBM increased (Brito and Broderick, 2006) would result in greater 15N enrichment of the ammonia utilized by both FAB and PAB. Microbial enrichment may also reflect greater uptake of ammonia by the ruminal microbes. Hristov and Broderick (1996) reported that 82% of microbial NAN were derived from ammonia when cows were fed an all-CS diet with supplemental CP from urea compared with 57% when fed an all-AS diet. Feeding red clover silage diets containing lower NPN also increased 15N enrichment of both FAB and PAB compared with feeding AS diets with greater NPN (Brito, 2004). The 15N enrichment of PAB was greatest on diets C and D, intermediate on diet B, and lowest on diet A. On average, 15N enrichment of FAB was 7% greater than that of PAB in the present trial. Brito (2004) found that 15N enrichment of FAB was 18% greater than that of PAB. Others (Perez et al., 1996; Rodriguez et al., 2000; Ahvenjärvi et al., 2002; Reynal et al., 2005) also reported greater 15N enrichment of FAB. Because this finding has been consistent across many studies, different isotopic enrichments of FAB and PAB must be accounted for when estimating microbial NAN flows from the rumen. This difference in isotopic enrichment may be attributed to the differential utilization by PAB and FAB of N from the free ammonia pool. According to Wallace (1997), estimates of contributions from ammonia and pre-formed AA to the synthesis of microbial protein by ruminal microbes are highly variable. The proportions of microbial N arising from ammonia have ranged from 18 to 100% (Salter et al., 1979), 50% or more (Nolan, 1975; Leng and Nolan, 1984), and 26 to 100% (Wallace, 1997). This variation has been attributed to the availability of N sources other than ammonia to the ruminal microbes (Salter et al., 1979; Wallace, 1997). Therefore, microbes associated with particulate matter may utilize more N from AA and peptides than microbes present in the liquid phase, which may dilute the 15N isotope. In fact, Brito (2004) observed that PAB captured, on average, 15% more 15N coming from the nonammonia 15N pool than did FAB, strongly indicating a greater utilization of preformed AA and peptides by the microbes associated with particulate matter.
The OM content of protozoa decreased linearly (P < 0.01) when AS plus HMSC was replaced with CS plus SSBM and was lowest (P < 0.01) on diet D (Table 6); a quadratic effect (P = 0.02) also was detected. Conversely, the protozoal NAN content increased linearly (P < 0.01) from diet A to diet D. However, the NAN content of protozoa was much lower than that of bacteria in the current study. Ahvenjärvi et al. (2002) reported greater N content in PAB compared with protozoa but no difference between protozoa and FAB. According to Ahvenjärvi et al. (2002), protozoal pellets were always contaminated with fine feed particles. Therefore, lower protozoal NAN in the present study might have been the result of contamination with feed particles, as well as dilution from glycogen accumulation used in the protozoal isolation. Ahvenjärvi et al. (2002) used saline to wash the protozoal pellets through a polyester fabric, which may have reduced NAN dilution by feed residues. Reynal et al. (2005) observed an average of 3.1% NAN in protozoal samples isolated using the same method as the present study. According to Reynal et al. (2005), contamination with feed N probably did not account for low protozoal NAN because 15N enrichment of protozoa was only slightly lower than FAB and PAB. These authors suggested that N in protozoal isolates was diluted with nonnitrogenous compounds, such as the glycogen and sucrose deriving from isolation. That could also explain the low protozoal NAN observed on diets A, B, and C in the current trial. However, the protozoal NAN content on diet D was 45% greater (P < 0.01; Table 6) than the average of the other 3 diets. Because isotopic enrichment of protozoa was lowest (P = 0.01) on diet D, it can be speculated that contamination with feed N diluted protozoal 15N. In fact, diet D contained the greatest proportion of SSBM and it is possible that protozoa engulfed more SSBM particles, resulting in increased NAN content. In addition, protozoa:FAB and protozoa:PAB ratios were lowest on diet D, further indicating that protozoal 15N APE was depressed because of contamination with feed N.
Enrichment of protozoa would be expected to be lower than that of bacteria because engulfment of feed particles by protozoa will dilute their 15N. The following protozoal:bacterial 15N enrichment ratios were reported in the literature: 0.40 (Ahvenjärvi et al., 2002), 0.63 (Hristov and Broderick, 1996), 0.69 (Martin et al., 1994), 0.75 (Cecava et al., 1991), and 0.69 (Firkins et al., 1987). In the present trial, mean protozoal:FAB and protozoal:PAB 15N enrichment ratios were, respectively, 0.89 and 0.95. Using the same method for protozoal isolation, Brito (2004) found 15N enrichment ratios of 0.83 and 1.00 for protozoal:FAB and protozoal:PAB. Contamination from more highly enriched bacteria could have resulted in greater apparent 15N content of the protozoa, resulting in protozoal:bacterial ratios that were higher than literature reports. Protozoa are indirectly enriched with 15N through bacterial predation and preferentially engulf bacteria associated with the FP. The FAB were more enriched than PAB, which would also help explain why the protozoal:PAB ratio was greater than the protozoal:FAB ratio in this trial.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
|
ACKNOWLEDGEMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
|
FOOTNOTES |
|---|
2 Current address: Agriculture and Agri-Food Canada, PO Box 90, 2000 Route 108 East, Lennoxville, Quebec, Canada.
Received for publication December 7, 2005. Accepted for publication April 21, 2006.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
