Nitrogen absorbed postruminally by cattle may be made available to ruminal microbes via N recycling. Recycled N can be incorporated into microbially synthesized AA, which may be absorbed by cattle and used for metabolic processes such as anabolism. This is an advantage when dietary protein content is decreased or when ruminally available N (RAN) is limited by ruminal protein degradation. Work has been conducted recently (Archibeque et al., 2001; Marini and Van Amburgh, 2003; Wickersham et al., 2008a,b; Huntington et al., 2009) to better quantify the capability of cattle to recycle N.
Accurately predicting the amount of recycled N that reaches the rumen is important because of the variety of supplemental protein sources fed to cattle. From a survey of consulting feedlot nutritionists, Vasconcelos and Galyean (2007) reported increasing use of coproducts from ethanol production (83% of all clients reported using grain coproducts in finishing diets). Ruminal degradability of protein in distillers grains is estimated to be only approximately one-fourth of the total protein (NRC, 1996). Thus, N recycling may be of greater relative importance when distillers grains are used for N supplementation of cattle.
Little work has quantified urea recycling in cattle fed high-concentrate diets. The goal of our study was to better predict the amount of N recycled by growing cattle fed corn-based diets supplemented with distillers dried grains with solubles (DDGS) or urea and to quantify the use of recycled N by ruminal microbes.
MATERIALS AND METHODS
All procedures involving the use of animals were approved by the Institutional Animal Care and Use Committee at Kansas State University.
Six ruminally and duodenally cannulated steers (initial BW = 244 ± 33 kg) of British breeding were used in 2 concurrent 3 × 3 Latin squares, with the treatment sequence reversed between squares to balance for carryover effects. Treatments were 3 corn-based diets (Table 1) fed as total mixed rations: control (10.2% CP), urea (13.3% CP), and DDGS (14.9% CP). Treatments delivered 3 different amounts of CP, which resulted from inclusion rates of urea and DDGS that were similar to those commonly used in corn-based diets fed to finishing cattle (Galyean, 1996; Vasconcelos and Galyean, 2007). The supplemental DDGS was from a single source (Dakota Gold, Poet Nutrition, Sioux Falls, SD). Distillers dried grains with solubles was selected as a supplemental protein source because of its relatively large undegradable intake protein (UIP) content, and urea was selected as a supplemental N source that is completely ruminally available.
Experimental periods were 14 d; each period consisted of 9 d for adaptation to treatment diets and 5 d for sample collection. Steers were housed continuously in metabolism crates to allow for total collection of urine and feces. Steers were allowed ad libitum access to water and were fed equal amounts of the total mixed ration twice daily at 0500 and 1700 h. The amount of feed offered was near ad libitum intake, as determined for each individual steer before the experiment. Five grams of Cr2O3 was manually mixed into diets at each feeding (10 g/d) beginning on d 7 and continuing through the end of the period to serve as an indigestible marker of nutrient flow to the duodenum.
A clean urine collection vessel containing 900 mL of 10% (wt/wt) H2SO4 was placed under each steer at 0530 h for daily collection of urine from d 10 through 13 of each period. Feces were collected from d 10 through 13 of each period in pans located behind the steers as part of the metabolism crates. Blood (10 mL) was collected by jugular venipuncture into heparinized (143 USP units of heparin) Vacutainer tubes (Becton Dickinson, Franklin Lakes, NJ) 4 h after feeding (0900 h) on d 10. Samples were placed in ice water immediately after collection and centrifuged at 1,200 × g at 4°C for 15 min within 1 h of collection. Plasma was isolated and frozen for later analysis of plasma urea-N (PUN), glucose, and creatinine.
On d 10, a temporary indwelling catheter was placed into an ear vein for infusion of 15N15N-urea. Indwelling jugular catheters were used to deliver the continuous infusions in 2 steers during period 2 and in 3 steers during period 3 because of an inability to place ear catheters. Sterile saline solution (0.9% NaCl) was infused continuously after catheters were placed until 0530 h on d 11 of each period. Continuous infusion of the 15N15N-urea solution (4.16 mL/h) began at that time and continued through the end of each period. The infusion of the 15N15N-urea solution delivered 0.48 mmol of urea-N/h via a programmable syringe pump (BS-9000 Multi-Phaser, Braintree Scientific Inc., Braintree, MA). The 15N15N-urea solution was prepared using sterile techniques in a laminar flow hood by combining 3.6 g of 15N15N-urea (99%, Medical Isotopes Inc., Pelham, NH) with 1 L of sterile saline solution (0.9% NaCl). The solution was passed through a 0.22-μm filter (Sterivex, Millipore Corporation, Billerica, MA) into a sterilized glass container. A sterilized rubber septum was crimped onto the container after filtration, and the solution was stored at 4°C until use. The 15N15N-urea solution was prepared immediately before infusion for each period.
Diets were sampled (100 g/d) as they were weighed, and samples were frozen (−20°C). If any orts were present, they were removed at 0455 h daily, weighed, and frozen (−20°C). Collection vessels for urine and feces were removed at 0530 h daily and weighed. Urine samples were mixed thoroughly, and then 1% of daily output was sampled and frozen. At the same time, a representative portion of urine was mixed with 0.05 M H2SO4 (1 part urine with 4 parts H2SO4) such that the final solution weight was equal to 1% of the daily urinary output and was frozen for analysis of urinary purine derivatives and creatinine. Fecal samples were mixed thoroughly by hand, and 5% was sampled and frozen. Samples of feces and urine from d 10 through 13 were pooled by steer and used to measure N balance. Feed and ort samples collected from d 9 through 12 corresponded to urine and fecal samples collected from d 10 through 13. Urine and fecal samples used for 15N determinations were samples from the total daily collection from d 10 for measuring background 15N and samples from the total daily collection from d 13 for measuring enriched amounts of 15N. Urine (100 mL) and wet feces (470 mL) were sampled and subsequently frozen (−20°C) for analysis of 15N enrichments. At the same times, 20 mL of urine was diluted with 80 mL of 0.05 M H2SO4 and frozen (−20°C) for analysis of 15N enrichment of purine derivatives.
On d 14 of each period, ruminal bacterial samples were collected for measurement of 15N enrichment. Approximately 400 mL of ruminal digesta was collected from the dorsal and ventral rumen through the ruminal cannula 1, 3, 5, 7, 9, and 11 h after feeding. The digesta was immediately strained through 4 layers of cheesecloth, and the liquid portion was analyzed for pH. Immediately, 10 mL of strained ruminal fluid was mixed with 1 mL of 6 M HCl and frozen at −20°C for analysis of ruminal NH3. Another 8 mL of the strained ruminal fluid was mixed with 2 mL of 25% (wt/wt) metaphosphoric acid and frozen at −20°C for analysis of ruminal VFA. The remaining strained ruminal fluid and ruminal contents were blended (1 min; NuBlend, Waring Commercial, Torrington, CT) with 0.5 L of saline solution (0.9% NaCl) to isolate ruminal bacteria. After blending, the liquid fraction isolated by filtration through 4 layers of cheesecloth was immediately frozen (−20°C) and the remaining particulate matter was replaced in the rumen. On d 14, approximately 300 mL of duodenal digesta was collected from the duodenal cannula 1, 3, 5, 7, 9, and 11 h after feeding and frozen (−20°C).
Within period, feed samples were pooled across day on an equal weight basis. Ort samples were composited by steer within period. Feed and ort samples and subsamples of feces were dried at 55°C in a forced-air oven for 72 h, air-equilibrated for 24 h, and weighed to determine partial DM. Duodenal digesta samples were freeze-dried. Once dried, all samples were ground to pass a 1-mm screen (Thomas-Wiley Laboratory Mill Model 4, Thomas Scientific USA, Swedesboro, NJ). The DM of feed, ort, fecal, and duodenal samples was determined by drying for 24 h at 105°C in a forced-air oven. The OM was determined by ashing for 8 h in a muffle oven at 450°C. The N content of feed, ort, duodenal digesta, wet feces, and urine samples was determined by combustion (Nitrogen Analyzer Model FP-2000, Leco Corporation, St. Joseph, MI), and CP was calculated as N × 6.25. Chromium concentration of fecal and duodenal samples was determined by atomic absorption after samples were prepared as described by Williams et al. (1962). Ruminal bacteria were isolated by thawing samples of ruminal contents and then centrifuging samples at 500 × g for 20 min at 4°C. Supernatants were centrifuged at 20,000 × g for 20 min at 4°C to form a bacterial pellet. The pellet was resuspended with saline (0.9% NaCl) and centrifuged again at 20,000 × g for 20 min at 4°C. The bacterial pellets were frozen and freeze-dried.
Concentrations of allantoin, uric acid, and creatinine were determined in pooled (d 10 to 13) urine samples by reverse-phase HPLC (adapted from Shingfield and Offer, 1999). Samples were analyzed on a Hewlett-Packard 1050 Ti-Series liquid chromatography system (Hewlett-Packard, Palo Alto, CA) equipped with an UV-visible detector set at 218 nm (Acutect 500 UV/VIS, Thermo Fisher Scientific Inc., Waltham, MA) and autosampler (AS 1000 SpectraSYSTEM, Thermo Fisher Scientific Inc.). Separation of the sample components was achieved with a 5-µm Discovery BIO Wide Pore C18 column (250 × 4.6 mm i.d., Sigma-Aldrich, St. Louis, MO) with a 5-µm Discovery BIO Wide Pore C18 (20 × 4.6 mm i.d., Sigma-Aldrich) guard column. The mobile phase was prepared by dissolving 1.01 g of sodium 1-heptane sulfonic acid and 0.86 g of ammonium phosphate into 1 L of deionized H2O with 35 mL of methanol and 70 µL of triethylamine added. The pH was adjusted to 3.2 with HCl, and the entire solution was filtered (0.45-µm Magna-R, MSI, Westboro, MA) and degassed with He. Urine samples were diluted to be within the linear range of the standards (20:1) with a diluent that was prepared by dissolving 0.86 g of ammonium phosphate and 1.01 g of sodium 1-heptane sulfonic acid into 1 L of H2O (pH adjusted to 2.1 with HCl). Diluted samples were filtered (0.45-µm Syringe Filter Fisherbrand, Fisher Scientific, Pittsburgh, PA) and stored at 4°C. Sample injection volume was 5 µL. Chromatography at room temperature (approximately 24°C) was achieved at a flow rate of 0.5 mL/min (10 min), then 1.5 mL/min (29 min), and then 0.5 mL/min (1 min), with a total run time of 40 min.
Dried bacterial, duodenal, and fecal samples were analyzed for 15N enrichment with a stable isotope elemental analyzer (ThermoFinnigan Delta Plus, Thermo Electron Corporation, Waltham, MA). Ruminal VFA were determined by GLC as described by Vanzant and Cochran (1994). Colorimetric determination of ruminal NH3 (Broderick and Kang, 1980) was completed with an AutoAnalyzer (Technicon Analyzer II, Technicon Industrial Systems, Buffalo Grove, IL). Starch concentrations of feed, orts, and feces were determined using the procedures of Herrera-Saldana and Huber (1989) with glucose measured according to Gochman and Schmitz (1972). Diet samples were analyzed for NDF (Ankom-Fiber Analyzer 200, Ankom Technology, Fairport, NY; with amylase and without ash correction), soluble N (Licitra et al., 1996), and ether extract (gravimetrically after extraction with diethyl ether).
Urinary urea and NH3 concentrations were determined colorimetrically with an AutoAnalyzer (Technicon Analyzer II) according to the methods of Marsh et al. (1965) and Broderick and Kang (1980). Measurement of 15N enrichment of urinary urea was conducted on N2 samples produced from Hoffman degradation of urinary urea by using techniques similar to those described by Wickersham et al. (2009b). Samples were analyzed with an online gas introduction system (ThermoFinnigan Gas Bench) and an isotope ratio elemental analyzer (ThermoFinnigan Delta Plus) for 28N2, 29N2, and 30N2.
Purine derivatives in urine were analyzed for total 15N enrichment with a stable isotope elemental analyzer (ThermoFinnigan Delta Plus). Purine derivatives were isolated from the diluted urine samples by using a modification of the methods of Chen et al. (1998). Urine (6 mL) was combined with 3 mL of 6 M NH4OH and vortexed. This solution was pipetted over a column containing 2 mL of an anion exchange resin (Dowex 1 × 8 chloride form, 100 to 200 mesh, Sigma Chemical, St. Louis, MO) and rinsed with 12 mL of double-deionized water. The effluent was discarded. A final rinse of the columns was performed with 4 mL of 0.1 M HCl, and the effluent was collected for analysis. Then 20 µL of 40% (wt/wt) NaOH was added to the final effluent and the samples were vortexed for 10 s. Samples were then pipetted to deliver 0.1 mg of N into microcentrifuge tubes and dried at 90°C within a dry block heater for 6 h. The dried purine derivatives were resolubilized in 150 µL of double-deionized water and vortexed. The solution was transferred into pressed tin capsules (5 × 9 mm), which were placed into a 96-well microtiter plate and dried at 63°C for 4 h in a dehydrator (Snackmaster Dehydrator Model 2200/FD-30, American Harvest, Chaska, MN). Plasma urea (Marsh et al., 1965), plasma creatinine (Chasson et al., 1961), and plasma glucose (Gochman and Schmitz, 1972) were measured with an AutoAnalyzer (Technicon Analyzer II).
Urea kinetics were calculated according to the methods described by Lobley et al. (2000). Duodenal flows were calculated by dividing the fecal output of Cr by the Cr concentration of duodenal digesta. Bacterial and duodenal 15N enrichments were calculated as 15N/total N and were corrected for values in the background fecal samples (Wickersham et al., 2009a). Bacterial N flow was calculated by multiplying duodenal N flow by the ratio of duodenal 15N enrichment to bacterial 15N enrichment. The flow of bacterial N derived from recycled urea-N (Wickersham et al., 2009a) was calculated by multiplying bacterial N flow by the ratio of bacterial 15N enrichment to 15N enrichment of urinary urea (calculated as one-half the 14N15N-urea enrichment plus the 15N15N-urea enrichment). Duodenal flow of UIP was calculated by subtracting microbial N flow from total duodenal N flow and therefore would include endogenous N. Microbial N supply was also calculated from urinary excretion of purine derivatives using the procedures of Chen and Gomes (1992); the BW of the steers at the initiation of the trial were used for these calculations. Ruminal microbial capture of recycled N was alternatively calculated by replacing measured bacterial N flow with estimates of bacterial N flow derived from urinary excretion of purine derivatives and by substituting bacterial-N enrichments with urinary purine derivative enrichments, assuming that microbial enrichments were the same as enrichments for urinary purine derivatives (Hristov et al., 2005). Renal clearances of urea and creatinine were calculated as the rates of urea or creatinine excretion in urine divided by the concentration of the corresponding metabolite in plasma.
All data from 1 steer from all periods were removed because this steer did not exhibit normal digestive function. This steer had low ruminal (−1%) and total tract (71%) DM digestibilities, increased ruminal pH (6.4), and reduced ruminal VFA concentration (69 mM) compared with other steers in this experiment. Data related to duodenal flow for 1 steer fed the control diet in a single period were excluded because of an apparent marker failure (negative ruminal digestion). All observations from another steer fed the control diet in a single period were excluded because of negative ruminal DM digestion as well as unusual urea kinetics (very large urea entry rates).
Data were analyzed using the MIXED procedure (SAS Inst. Inc., Cary, NC). For variables without repeated sampling, terms in the model included treatment and period, and steer was included as a random effect. Model terms for fermentation profile variables were treatment, period, hour, and hour × treatment, and steer was included as a random term. The repeated term was hour with steer × period serving as the subject, and compound symmetry was used for the covariance structure. For comparisons of methods related to estimation of bacterial N flow to the duodenum, the model included period, treatment, method, and treatment × method and steer and steer × period × treatment were included as random terms. The LSMEANS option was used to calculate treatment means. Significance among treatments was declared at P ≤ 0.05 and tendencies were declared at 0.05 < P ≤ 0.15. Means were separated using pair-wise t-tests with P = 0.05 when F-tests were significant, and with pair-wise t-tests with P = 0.15 when F-tests tended to be significant.
RESULTS AND DISCUSSION
Intake, Digestibility, and Nutrient Flow
Dry matter intake (6.04 kg/d) and OM intake (5.77 kg/d) did not differ among treatments (Table 2), but they were numerically 7% less than for the control diet when steers consumed the urea diet. Total starch intake (Table 2) was decreased by DDGS in the diet (P = 0.05), which was expected because the DDGS contained less starch than the corn it replaced. Calculated ME intakes (NRC, 1996) were 18.4, 17.0, and 18.3 Mcal/d for the control, urea, and DDGS diets, respectively.
Ruminal digestibilities of DM and OM (true and apparent) did not differ among treatments (Table 2). Total tract digestibilities of DM and OM (Table 2) tended to be greater (P < 0.11) for the urea diet than for the DDGS and control diets. These increases in DM and OM digestibilities may be explained by urea-N stimulating microbial fermentation, although we were unable to detect differences among treatments for ruminal digestion of DM and OM (P > 0.64). The numerically smaller DMI for the urea diet may have contributed to the increases in total tract digestibilities. Ruminal apparent digestibility of dietary N tended to be less (P = 0.06) for the control diet than for the DDGS and urea diets.
Total tract starch digestibility tended (P = 0.11; Table 2) to be greater for the DDGS diet than for the control diet. This coincided with a numerically greater MP supply (Table 3), and it is possible that the greater amounts of UIP flowing to the hindgut increased pancreatic amylase activity (Richards et al., 2003). Decreased starch intakes for DDGS may also have contributed to the trend toward increased digestibility if digestion was limited by amylase activity in the intestines.
As expected, N intake increased with increasing N concentration in the diet (P < 0.01; Table 2); consumption of N was greatest for the DDGS diet, least for the control diet, and intermediate for the urea diet. Additionally, UIP (Table 3) tended (P = 0.12) to be greater for DDGS than for the other treatments, although microbial N flowing to the duodenum and microbial efficiency did not differ among treatments. Total tract digestion of N was different (P < 0.01; Table 2) among treatments, which was expected because of the differences in dietary N concentrations. As dietary N increases, endogenous fecal N losses, if proportional to indigestible DM excretion (NRC, 1985), would represent a smaller proportion of N intake.
Increases in N intake led to increased N output. Fecal N output (Table 2) was greater (P = 0.05) for the DDGS diet than for the control or urea diet. Urinary N excretion was greater (P = 0.02) for the DDGS and urea diets than for the control diet, but the DDGS and urea diets did not differ from each other. Nitrogen retention was greater (P = 0.02) for the DDGS diet than for the urea or control diets. These treatment effects on N retention may have been a response to an increasing MP supply. When Wessels and Titgemeyer (1997) limit fed steers of similar BW (254 kg) to gain 1 kg/d with increasing CP and MP, they observed linear increases in N retention with increasing CP and MP. Gleghorn et al. (2004) fed cattle of slightly heavier initial BW (357 or 305 kg) diets based on steam-flaked corn over the course of 2 experiments. They observed an increase in ADG during the initial 56 d on feed as N inclusion (and presumably MP) increased, regardless of the N source or degradable intake protein (DIP):UIP. Cole et al. (2006) fed heavier steers (315 kg), and N retention during the initial 112 d and the final 56 d on feed was not improved when they fed as much CP as that provided by our DDGS diet. Thus, improvements in N retention by steers in the current study when DDGS was supplemented were likely a reflection of the relatively young age and light BW of the steers.
Ruminal NH3 concentration was greater (P = 0.05) for the urea diet than for the control diet, and the DDGS diet did not differ from the control (Table 4). Although the diet × time interaction was not significant for ruminal NH3, concentrations were increased at times near feeding for the urea and DDGS diets, but not the control diet (Figure 1). For the DDGS and control diets, ruminal NH3 would be limiting for microbial growth, whereas concentrations for steers fed the urea diet would be near the requirement to optimize microbial growth (Satter and Slyter, 1974). Ruminal pH demonstrated a treatment × time interaction (P = 0.02; Figure 2), with steers fed urea tending to have pH at 1 and 3 h after feeding greater than in the other treatments, whereas the urea diet tended to be less than the DDGS diet at 11 h after feeding. Zinn et al. (2003) similarly observed that increases in dietary urea resulted in increases in ruminal pH 1 h after feeding. Ruminal concentrations of acetate, propionate, butyrate, and isovalerate did not differ among treatments (Table 4), nor were treatment × time interactions observed (P > 0.10). Ruminal isobutyrate tended (P < 0.15) to be greater for the control diet than for the DDGS or urea diet, and ruminal valerate concentrations were greatest (P < 0.05) for the control diet and least for the DDGS diet.
Urea kinetics are presented in Table 5. Urea entry rate tended to be greater (P = 0.09) for the DDGS diet than for the control diet, whereas the urea diet was not different from either the DDGS or control diet. Urinary urea-N excretions were greater (P = 0.03) for both the DDGS and urea diets than for the control diet. Gastrointestinal entry of urea (GER) did not differ among treatments (P = 0.25), but there were large numerical differences that corresponded to the pattern of urea entry rate. The DDGS diet yielded the numerically greatest GER, the control diet yielded the least, and the urea diet was intermediate. The amount of urea-N returned to the ornithine cycle and reincorporated into urea tended to be greater (P = 0.09) for the DDGS diet than for the control or urea diet. We did not observe statistical differences among treatments in urea-N losses to the feces; amounts were small, representing not more than 4 g/d.
When expressed as a fraction of urea entry rate, transfers of urea-N to the urine, gut entry, and return to the ornithine cycle were not different among treatments (Table 5). However, the partitioning of GER was affected by treatment. Expressed as a fraction of GER, urea-N returning to the ornithine cycle tended (P = 0.11) to be greater for the DDGS diet (0.63) than for the control diet (0.37), with the urea diet (0.49) being intermediate and not different from either of the other diets. Urea-N used for anabolism (as a fraction of GER) tended (P = 0.14) to be less for the DDGS diet (0.31) than for the control diet (0.56), with the urea diet (0.45) being intermediate and not different from the other diets. This may reflect the greater UIP from DDGS leading to less dependence of the animal on microbially produced AA. Measured as a proportion of GER, urea-N lost in the feces was greatest (P = 0.02) for the control diet and least for the DDGS and urea diets.
When feeding diets similar to those in our experiment, Reynolds et al. (1991) and Huntington et al. (1996) observed that urea-N appearance across the liver and urea-N disappearance across the portal drained viscera were generally similar to our estimates of urea entry and GER, respectively. This suggests that the method of Lobley et al. (2000) allows for an adequate evaluation of urea kinetics in cattle fed grain-based diets.
Previous work has demonstrated that urea entry rates increase with dietary N intake (Archibeque et al., 2001; Marini and Van Amburgh, 2003; Wickersham et al., 2008a,b, 2009b), which would be expected because of the important role of urea synthesis in the detoxification of NH3. These same studies demonstrated that the proportion of urea production that is recycled decreases as dietary N intake increases. In most cases, the amount of urea-N recycled has also increased with dietary N intake (Archibeque et al., 2001; Wickersham et al., 2008a,b, 2009b). Thus, our increases in urea entry rate were expected when DDGS or urea was supplemented, as were the numerically greater GER and numerically smaller GER as a proportion of urea production.
Wickersham et al. (2008a,b) used methods similar to those in the current study to measure the effects of increasing DIP on urea kinetics in steers consuming prairie hay. When casein was supplemented ruminally in amounts that led to N intakes similar to those in the current study, urea entry rates were generally similar to those for the control and urea diets in our study. However, Wickersham et al. (2008a,b) reported a greater proportion of urea production being recycled to the gut (78 to 99%), likely because of the inherently greater endogenous N recycling that is characteristic of low-quality forage diets (Huntington et al., 1996). This may be a result of greater salivary transfer of urea-N for forage diets than for grain-based diets.
Wickersham et al. (2008a) measured increases in urea entry rate and GER as DIP supplementation increased, but increases in urea entry rate and GER were greater when UIP was supplemented (Wickersham et al., 2009b) than when DIP was provided. Thus, we expected that urea entry rate and GER might be greater for the DDGS diet than for the urea diet. Although these expectations were reflected in the data, the greater concentration of dietary N for the DDGS diet than for the urea diet makes it difficult to separate the effects of protein source from those of dietary N concentration. However, we did observe that urea entry rate expressed as a percentage of N intake was numerically greater for the DDGS diet (83%) than for the urea diet (73%), which is consistent with greater increases in urea entry rate and GER when UIP is provided (i.e., DDGS) than when DIP is provided (i.e., urea).
Concentrations of PUN did not differ among treatments but increased numerically with N intake (Table 6). Others have reported that PUN is closely related to protein intake (Thornton, 1970a,b). In the present study, urea excreted in the urine was greater for the DDGS and urea diets than for the control diet. Cocimano and Leng (1967) observed that as PUN increased in lambs, so did the amount of urea excreted. Moreover, these authors reported that the relationship between PUN and urea excretion rates followed a sigmoidal curve whereby minimal amounts of urea were excreted via urine at decreased N intakes (dietary CP ≤9.0% of DM) and maximum amounts were reached at increased N intakes, when sheep were limited by their capacity to eliminate urea to urine. Thornton (1970a,b) reported a relationship between PUN and urea excretion rates in cattle similar to that in sheep but stated that greater differences among urea excretion rates were detected at less PUN with cattle than with sheep (Cocimano and Leng, 1967; Thornton, 1970b). Thus, when N intake is restricted, ruminants conserve and recycle urea more efficiently than when N intake is in excess. The PUN in the current study is among the smaller values reported by Cocimano and Leng (1967) for sheep but is similar to values observed by Thornton (1970a,b) for cattle. According to Cocimano and Leng (1967), the range of PUN in the current study should have led to only small differences among treatments in urea excreted in the urine. We observed, however, significant increases in urinary urea with only small increases in PUN, which was similar to the results of Thornton (1970a) with cattle. Thus, urinary urea excretion appears to be more responsive to changes in PUN in cattle than in sheep. More work is needed to better quantify the relationship between PUN and urinary urea excretion by cattle.
Renal clearance of creatinine (Table 6) did not differ (P = 0.34) among treatments. However, renal clearance of urea and the proportion of urea that was filtered by the kidneys and subsequently excreted in the urine (i.e., urea clearance/creatinine clearance; Table 6) tended (P < 0.15) to be greater for the DDGS diet than the control diet, with the urea diet being intermediate but not different from the others. Marini and Van Amburgh (2003) reported that as N intake increased in cattle, renal clearance of creatinine was not different, but they observed linear increases in renal clearance of urea with increases in N intake. Thornton (1970a) reported that as N intake of cattle increased, the percentage of filtered urea subsequently excreted in the urine also increased. Endogenous mechanisms that support urea recycling (i.e., reabsorption of urea by the kidney tubules) may be reduced as N intake increases.
Microbial Capture of Endogenously Produced Urea
Capture of recycled urea-N by ruminal microbes (Table 5) was not different among treatments (P = 0.28) but was numerically greater for the DDGS treatment (30 g/d) than for the control (17 g/d) or urea treatment (18 g/d). The total amount of urea-N captured by ruminal microbes depends on the quantity of urea-N recycled to the rumen and on the proportion of the recycled urea-N that is captured by the ruminal microbes. Ruminal microbial capture of recycled N as a percentage of urea entry rate was greatest (P = 0.03) for the control diet compared with both the DDGS and urea diets; the DDGS and urea diets did not differ. Microbial capture of recycled N as a percentage of GER tended to be greater (P = 0.11) for the control diet than for the DDGS or urea diet. The efficiency with which ruminal microbes capture recycled urea-N is related to the availability of fermentable energy to the ruminal microbes, the proportion of GER that is recycled to the rumen, and the availability of competing N sources within the rumen. The opportunity of ruminal microbes to capture recycled urea-N theoretically must increase as the proportion of GER that is returned to the rumen (as opposed to postruminal segments of the gut) increases. In addition, ruminal microbes are more dependent on recycling mechanisms to meet their needs for N when RAN is limiting. The fraction of recycled urea captured by ruminal microbes (i.e., microbial capture as a percentage of GER) appeared to be related to ruminal NH3; as ruminal NH3 increased, the efficiency of recycled-N capture by microbes decreased.
Steers fed the DDGS diet tended to derive (P = 0.10) a greater proportion of their microbial N from recycled urea-N (35%) than did steers fed the urea (22%) or control diet (17%). Ruminal microbial capture of recycled urea-N is related to microbial N needs and to amounts of recycled urea-N. Wickersham et al. (2009b) reported increases in the amount of recycled urea-N captured by ruminal microbes when UIP was provided, and these increases corresponded to increased urea entry rate and GER. Moreover, Wickersham et al. (2009b) observed that an increasing proportion of microbial N was derived from recycled urea-N when UIP was supplemented, but no difference in the proportion of microbial N from recycled urea-N was observed in response to DIP supplementation (Wickersham et al., 2008a). Our observation that microbial capture of recycled urea-N, expressed as a proportion of total microbial N, tended to be greater for the DDGS diet than for the urea diet agrees with the results of Wickersham et al. (2008a, 2009b) and suggests that recycled N is more important for ruminal microbes when UIP, rather than DIP, is provided to cattle. This greater importance is reflective of both greater GER and less RAN when UIP, rather than DIP, is supplemented.
Gastrointestinal entry of urea as a proportion of GER plus measured DIP (data not presented tabularly) tended (P = 0.13) to be greater for the DDGS treatment (0.58) than for the urea (0.49) or control treatment (0.40). If all GER entered the rumen, it would be expected that GER:(GER+DIP) would be similar to the fraction of microbial N that was derived from urea recycling. The observed proportions of microbial N derived from recycled urea (0.17 to 0.35) were approximately one-half (ranging from 42 to 60% among treatments) the values for GER:(GER+DIP), suggesting that only approximately one-half the GER entered the rumen and completely mixed with the ruminal NH3 pool. The difference between GER:(GER+DIP) and the proportion of microbial N derived from recycled urea-N may be due to urea-N that was recycled to the rumen but did not mix with the ruminal NH3 pool before absorption (accounted for as urea-N returned to the ornithine cycle by our methodology) as well as to urea recycled to the postruminal gut.
Methods for Estimating Microbial Capture
Chen and Gomes (1992) provided a simple approach to predict microbial N flow from urinary purine derivative excretion, obviating the need for cannulated cattle. They described microbial N flow to the duodenum (g of N/d) as 0.727 × microbial purines absorbed (mmol/d), where absorption of microbial purines was calculated as a function of excretion of urinary purine derivatives and metabolic BW. Predictions of microbial N flow based on urinary purine derivative excretion and the equations of Chen and Gomes (1992) led to values that were 30% greater (P < 0.01) than those directly measured at the duodenum (Table 3). Differences likely arose because microbial purine:N ratios in our study (not measured) differed from the average values used by Chen and Gomes (1992).
Marini and Van Amburgh (2003) used 15N enrichments of plasma urea and bacteria in conjunction with urinary excretions of purine derivatives to estimate microbial capture of recycled urea-N. We directly measured microbial capture of recycled urea-N using the methods of Wickersham et al. (2009a), and we also estimated microbial capture of recycled urea-N using methods similar to those reported by Marini and Van Amburgh (2003), except that we used the 15N enrichment of urinary urea and of urinary purine derivatives to determine the proportion of microbial N derived from recycled urea. When calculated with an approach similar to that of Marini and Van Amburgh (2003), microbial capture of recycled urea-N averaged 35% less (P < 0.01) than the measured values. The difference reflects the view that the equation of Chen and Gomes (1992) overpredicted microbial N by 30% (as discussed above) and that the 15N enrichment of purine derivatives averaged only 63% of that of rumen bacteria (Table 5). By modifying the equation of Chen and Gomes (1992), as discussed above, and by multiplying the 15N enrichment of purine derivatives by 1.58 (to account for the smaller enrichment in urinary purine derivatives compared with ruminal bacteria), microbial capture of recycled urea-N could be adequately predicted (i.e., no differences between measured and predicted values, P = 0.76). We suggest caution when applying these values to other experiments unless experimental methods and conditions closely mimic those presented herein.
Hristov et al. (2005) observed that when 15N was ruminally infused for 8 d, 15N enrichment of urinary purine derivatives was similar to that of ruminal bacteria. The observations of Hristov et al. (2005) differed from those of other researchers (Orellana Boero et al., 2001; Gonzalez-Ronquillo et al., 2003), whose enrichments of urinary purine derivatives were less than those of duodenal purine bases. This difference may have arisen because labels were provided for either 3 (Gonzalez-Ronquillo et al., 2003) or 4 d (Orellana Boero et al., 2001), which did not allow enough time for the purine derivatives in urine to become equally enriched. Although Wickersham et al. (2009a) reported that 15N enrichments of ruminal microbes reached a plateau after 48 h of infusion in cattle consuming poor-quality forages, we observed differences (P < 0.01; Table 5) between 15N enrichments of ruminal microbes (0.068 atom percent excess) and urinary purine derivatives (0.043 atom percent excess), suggesting that 2 d of adaptation does not allow equilibration of these 2 pools. It is also possible that the enriched urinary purine derivatives of microbial origin may have been diluted by unlabeled purines of endogenous origin, although this is doubtful because Hristov et al. (2005) reported that urinary purine derivatives were predominantly of microbial origin (93.4%). Calculated using the equations of Chen and Gomes (1992), endogenous purine derivatives represented 15% of total urine purine derivatives. Additional work is needed to verify the minimum length of adaption to obtain equal enrichments of 15N in urinary purine derivatives and ruminal microbes.
Urea entry rate and GER were related to N intake in cattle consuming corn-based diets. Efficiency of microbial capture of recycled N (as a fraction of either urea entry rate or GER) increased as ruminal NH3 decreased.
Increasing the accuracy with which we can estimate urea recycling and subsequent microbial capture of recycled N in cattle consuming corn-based diets will allow more precise diet formulation and reductions in wasteful nitrogenous excretions. The ability to predict microbial capture of recycled N in noncannulated cattle will reduce the cost of obtaining these data and thereby increase the number of observations on which predictions of microbial capture of recycled N can be based. However, adaptation times longer than those used in our experiment will be required for the methods to be fully valid. Advances in the measurement of microbial capture of recycled N in cattle, in combination with less invasive measures of urea kinetics, may allow for robust predictive models of urea recycling.