Animals, diets and general experimental design

A total of 15 Yorkshire barrows were obtained from the Arkell Swine Research Station at the University of Guelph (Guelph, ON, Canada) and randomly assigned to one of three dietary treatments. The first group (Study 1; two experimental periods, partial-crossover design) consisted of six barrows with an initial BW of 19.9 kg (s.d.=0.99). The second group (Study 2; one experimental period) consisted of nine barrows with an initial BW of 22.4 kg (s.d.=0.90). All pigs were housed in smooth-sided metabolic crates in a temperature controlled room at 21°C (Libao-Mercado et al., Reference Libao-Mercado, Zhu, Fuller, Rademacher, Seve and de Lange2009). The three dietary treatments were (Table 1): a cornstarch–soybean meal-based control diet (CON), a high fibre diet in which 12% pectin was added at the expense of cornstarch (HF) or a low-protein diet in which soybean meal was partially replaced by cornstarch (LP). The diets were formulated to meet or exceed nutrient requirements according to National Research Council (NRC, 1998) and contained titanium dioxide as an indigestible marker (Zhu et al., Reference Zhu, Lapierre, Rademacher and de Lange2003).

Table 1

Ingredient composition (as-fed, %) as well as calculated and analysed nutrient content (adjusted to 90% DM basis) of experimental diets

DM=dry matter; CON=control diet; HF=high fibre diet; LP=low-protein diet; DE=digestible energy; NE=net energy; CP=crude protein; SID=standardized ileal digestible.

¹ Supplied per kg of complete diet: vitamin A, 10,000 IU as retinyl acetate (2.5 mg) and retinylpalmitate (1.7 mg); vitamin D3, 1000 IU as cholecalciferol; vitamin E, 56 IU as dl-α-tocopherol acetate (44 mg); vitamin K, 2.5 mg as menadione; choline, 500 mg; pantothenic acid, 15 mg; riboflavin, 5 mg; folic acid, 2 mg; niacin, 25 mg; thiamine, 1.5 mg; vitamin B₆, 1.5 mg; biotin, 0.2 mg; vitamin B₁₂, 0.025 mg; Se, 0.3 mg from Na₂SeO₃; Cu, 15 mg from CuSO₄·5H₂O; Zn, 104 mg from ZnO; Fe, 100 mg from FeSO₄; Mn, 19 mg from MnO₂; and I, 0.3 mg from KI (DSM Nutritional Products Canada Inc., Ayr, ON, Canada).

² Nutrient contents of diets were estimated based on nutrient contents of feed ingredients according to NRC (1998).

Pigs on CON and LP were fed at 90% of energy intake according to NRC (1998) and based on individual pig BW. Due to the reduced net energy content of the HF diet, daily feed allowance in the HF-fed pigs was increased to maintain daily net energy intake across treatments (Table 2). To maintain isonitrogenous intake between the CON and HF treatments, nutrient content in the HF diet was decreased to account for the increase in total dietary intake (Libao-Mercado et al., Reference Libao-Mercado, Zhu, Fuller, Rademacher, Seve and de Lange2009). Pigs were given their daily ration as a wet mash (water : feed 2 : 1) in two equal meals per day during the adaptation period and eight meals per day (every 3 h starting at 0800 h) during the infusion period in order to maintain metabolic steady state conditions (Libao-Mercado et al., Reference Libao-Mercado, Zhu, Fuller, Rademacher, Seve and de Lange2009). Pigs were allowed to adapt to their diets for at least 1 week before the infusion of isotopes.

Table 2

BW, nutrient intake and digestibility, ileal digesta characteristics and nitrogen (N) flows in growing pigs after a 4-day continuous infusion of ¹³C-urea and ¹⁵N-ammonium chlorideFootnote ^{1 ,} Footnote ²

CON=control diet, HF=high fibre diet, LP=low-protein diet; DM=dry matter.

^a,bValues with different letters are significantly different (P<0.05); *P<0.05.

¹ Pigs were fed one of three diets: CON, 18.6% CP; HF, 17.7% CP, 12% added pectin; LP, 13.2% CP.

² Main effect of treatment analysed using PROC MIXED with treatment as a fixed effect and pig within period and period within study as random effects. Pig within period was the experimental unit.

³ Expressed per litre of digesta fluid.

Surgery and isotope infusion

Two weeks before the start of infusion, pigs were fitted with a simple T-cannula at the distal ileum (Libao-Mercado et al., Reference Libao-Mercado, Zhu, Fuller, Rademacher, Seve and de Lange2009) for collection of ileal digesta and a catheter (Microrenathane implantation tubing, 2.4 mm OD and 1.7 mm ID; Braintree Scientific Inc., Braintree, MA, USA) in the stomach for infusion of ¹⁵N-ammonium chloride. One week later, pigs were implanted with a jugular catheter (Libao-Mercado et al., Reference Libao-Mercado, Zhu, Fuller, Rademacher, Seve and de Lange2009) to allow for infusion of ¹³C-urea.

¹⁵N-ammonium chloride and ¹³C-urea were obtained from ACP Chemicals Inc. (Montreal, QC, Canada). A 24-h infusion of saline solution (day 0) was followed by a 4-day infusion of saline solutions containing ¹⁵N-ammonium chloride (intragastrically) and ¹³C-urea (intravenously) infused at a rate of 0.5 ml/min to provide 0.8 and 0.2 mmol/kg BW/day of ¹⁵N-ammonium chloride and ¹³C-urea, respectively.

Sample collection, processing and analysis

Urine was collected quantitatively each day into containers with 15 ml of concentrated sulphuric acid to maintain pH below 3 (Zhu et al., Reference Zhu, Lapierre, Rademacher and de Lange2003). Total urine output was weighed on a daily basis and a subsample stored at 4°C until further processing. Ileal digesta samples were collected into bags attached to the cannula for a 6-h period on days 0, 3 and 4 (0800 to 1400 h) and for a 24-h period on day 4 of the infusion period. The first 6-h digesta collection on day 4 (0800 to 1400 h) was subsampled for comparison to day 3 digesta samples before generating a pooled 24-h collection sample. Bags were changed at least every 2-h and collected digesta was placed on ice until further processing. Pooled digesta samples were divided into several aliquots, acidified with 10% formic acid (1 ml/100 ml digesta), and stored at −20°C until further processing (Libao-Mercado et al., Reference Libao-Mercado, Zhu, Fuller, Rademacher, Seve and de Lange2009). In the first study, isotope infusion was repeated after a 9-day washout period during which pigs were randomly reassigned to different diets according to a crossover design. In the second study, additional digesta samples were taken at the end of the infusion. Pigs were euthanized at the end of the collection period (Euthansol; 340 mg/ml sodium pentobarbital; Schering-Plough Animal Health; Kirkland, QC, Canada) and the entire gastrointestinal tract was removed. Digesta samples were collected from the stomach, proximal duodenum and mid-jejunum (1 m segments) in order to determine rate of disappearance of the labelled ammonia. Samples were immediately frozen.

DM content of digesta was determined by oven drying at 103°C for 24 h. Urinary and ileal digesta nitrogen content were determined using a LECO-FP 428 automatic analyzer (Leco Instruments Ltd, Mississauga, ON, Canada). Titanium dioxide in feed and digesta was determined according to Zhu et al. (Reference Zhu, Lapierre, Rademacher and de Lange2003). Total ammonia concentration was determined in digesta fluid using a commercially available kit (AA0100; Sigma-Aldrich Corporation, St. Louis, MO, USA). Urinary urea concentration was determined using a Roche Urea Reagent kit (UREAL 04460715190) analysed on a Hitachi 911 Chemistry Analyzer (Roche Diagnostics GmbH; Indianapolis, IN, USA).

All isotope enrichment analysis was performed in the laboratory of Metabolic Solutions Inc. (Nashua, NH, USA; Libao-Mercado et al., Reference Libao-Mercado, Zhu, Cant, Lapierre, Thibault, Seve, Fuller and deLange2007). Samples of digesta, plasma and urine from day 0 were used to generate background enrichment values. Digesta samples were processed and analysed according to Libao-Mercado et al. (Reference Libao-Mercado, Zhu, Cant, Lapierre, Thibault, Seve, Fuller and deLange2007). In brief, ammonia from digesta fluid was released by fluid aeration in a sealed tube containing a trapping well of 2% sulphuric acid after adjusting the pH to greater than 12 by adding 5 M NaOH. The ¹⁵N enrichment of ammonia sulphate was analysed using a Europa 20/20 Continuous Flow Isotope Ratio Mass Spectrometer (CF-IRMS) interfaced with the Europa ANCA Solid/Liquid Preparation Module (Europa Ltd, Crewe, UK) using Callisto IRMS Software Ver. 5.2.6 (Sercon Ltd, Crewe, UK). Urea in urine and plasma was prepared for enrichment analysis according to the method of Nelson and Ruo (1988). In brief, ¹⁵N¹⁴N and ¹⁵N¹⁵N enrichment of urea in urine was determined on an Agilent 5973 Mass Selective Detector (MSD) with Agilent 6890 Gas Chromatograph using MSD Chemstation software Ver. D.01.00 (Agilent Technologies Inc., Palo Alto, CA, USA) using chemical ionization detection with methane gas under SIM mode monitoring ions 293, 294 and 295. The ¹⁵N¹⁴N urea enrichment was corrected for ¹³C urea enrichment by subtracting the directly determined ¹³C urea enrichment from the ¹⁵N¹⁴N urea enrichment. The ¹³C enrichment in urinary and plasma urea was performed according to Tuchman et al. (Reference Tuchman, Caldovic, Daikhin, Horyn, Nissim, Nissim, Korson, Burton and Yudkoff2008) using a urea-specific urease to liberate CO₂. The liberated CO₂ was analysed for ¹³C enrichment with a Europa 20/20 ABCA isotope ratio mass spectrometer.

Calculations and model development

It is generally assumed that the isotopic enrichment of urinary urea is identical to the enrichment of plasma urea and therefore use of enrichment in either pool will yield the same results (Jackson et al., Reference Jackson, Picou and Landman1984) and has been demonstrated using similar infusion and analysis procedures (Columbus et al., Reference Columbus, Lapierre, Htoo and de Lange2014; Mansilla et al., Reference Mansilla, Columbus, Htoo and de Lange2015). However, in the current study there was a discrepancy between plasma and urine urea ¹³C enrichment (Table 3), for reasons that could not be identified. It was determined that urea kinetics calculated using urinary urea enrichment resulted in more reasonable values based on comparison with previously reported values in pigs (Thacker et al., Reference Thacker, Bowland, Milligan and Weltzien1982; Mosenthin et al., Reference Mosenthin, Sauer and de Lange1992a and 1992Reference Mosenthin, Sauer, Henkel, Ahrens and de Langeb; Zhu et al., Reference Zhu, Lapierre, Rademacher and de Lange2003) and humans (Jackson et al., Reference Jackson, Picou and Landman1984; Mariotti et al., Reference Mariotti, Pueyo, Tome and Benamouzig2001). Therefore, urea isotopic enrichment was used to estimate plasma isotopic enrichment in all calculations.

Table 3

Infusion rate and isotopic enrichment of urea in plasma and urine and ammonia in digesta in growing pigs after a 4-day continuous infusion of ¹³C-urea and ¹⁵N-ammonium chlorideFootnote ¹

CON=control diet; HF=high fibre diet; LP=low-protein diet; MPE=moles per cent excess.

^a,bValues with different letters are significantly different (P<0.05); *P<0.05.

¹ Pigs were fed one of three diets: CON, 18.6% CP; HF, 17.7% CP, 12% added pectin; LP, 13.2% CP.

² Main effect of treatment analysed using PROC MIXED with treatment as a fixed effect and pig within period and period within study as random effects. Pig within period was the experimental unit.

³ n=6 for all treatments.

⁴ Calculated as [(¹⁴N¹⁵N-urea MPE×1)+(¹⁵N¹⁵N-urea MPE×2)]/2.

⁵ CON (n=4), HF (n=2), LP (n=3) for stomach, duodenum and jejunum ammonia isotopic enrichment.

Plasma urea flux was calculated using the conventional isotope dilution equation according to Matthews and Downey (Reference Matthews and Downey1984):

$${\rm Plasma}\,{\rm urea}\,{\rm flux}\,{\rm (mmol}/{\rm kg}\,{\rm BW}/{\rm day)}={\rm IR}{\times}({\rm EI}/{\rm EU}-{\rm 1})$$

where IR is ¹³C-urea infusion rate and EI and EU are ¹³C-urea enrichment in infusate and urinary urea, respectively. Urea recycling was then calculated as the difference between plasma urea flux and urinary urea excretion according to Mosenthin et al. (Reference Mosenthin, Sauer and de Lange1992a and 1992Reference Mosenthin, Sauer, Henkel, Ahrens and de Langeb).

The equations used in the determination of nitrogen and carbon flows through the two-pool model (Figure 1) consisting of a plasma urea pool (p) and an ileal digesta ammonia pool (i) are given below, where F _xy is flux of nitrogen or carbon from pool x to pool y, c _x is carbon enrichment in pool x, n _x is nitrogen enrichment of pool x and I is infusion of isotope-labelled ammonia or urea. The measurements used in these calculations are: (1) urinary urea ¹⁵N and ¹³C enrichment, (2) total urinary urea-nitrogen excretion and (3) ileal digesta ammonia ¹⁵N enrichment.

Figure 1

Two-pool model for nitrogen flow. Where F _xy is the movement of nitrogen from pool x to pool y, I _Ammonia and I _Urea are the infusions of ¹⁵N-ammonium chloride and ¹³C-urea, and mAA, dAA and eAA, are microbial, dietary and endogenous amino acids, respectively.

Entry rate of nitrogen and ¹³C from infusion of urea can be calculated based on the measured infusion rate of the tracer urea:

(1) $$F_{{Ip}} \,({\rm mmol}\,{\rm N}/{\rm day})=I_{{{\rm Urea}}} $$

(2) $$F_{{Ip}} ^{{{\rm 13}}} ({\rm mmol}\,^{{{\rm 13}}} {\rm C}/{\rm day})=c_{I} F_{{Ip}} /{\rm 2}$$

Excretion of total urea-nitrogen, ¹³C- and ¹⁵N-urea in urine is measured, therefore:

(3) $$\eqalign{F_{{pu}} \,({\rm mmol}\,{\rm N}/{\rm day})=\,&#x0026;{\rm urine}\,{\rm volume}{\times}{\rm urea}{\hbox{-}}{\rm nitrogen}\cr &#x0026;\,\ {\rm concentration}$$

(4) $$F_{{pu}} ^{{{\rm 13}}} ({\rm mmol}\,^{{{\rm 13}}} {\rm C}/{\rm day})=c_{u} F_{{pu}} /{\rm 2}$$

(5) $$F_{{pu}} ^{{^{{{\rm 15}}} }} ({\rm mmol}\,^{{{\rm 15}}} {\rm N}/{\rm day})=n_{p} F_{{pu}} $$

Assuming that infused ¹³C-urea that is not excreted in the urine enters the small intestine gives:

(6) $$F_{{pi}} ^{{{\rm 13}}} ({\rm mmol}\,^{{{\rm 13}}} {\rm C}/{\rm day})=F_{{Ip}} ^{{{\rm 13}}} -F_{{pu}} ^{{{\rm 13}}} $$

Assuming that ¹⁵N-urea is partitioned between urine and the small intestine by the same proportions as for ¹³C-urea, then:

(7) $$F_{{pi}} ^{{{\rm 15}}} ({\rm mmol}\,^{{{\rm 15}}} {\rm N}/{\rm day)}=F_{{pi}} ^{{{\rm 13}}} {\times}{\rm 2}n_{p} /c_{p} $$

Determination of nitrogen flow to the small intestine can be obtained by adjusting for the plasma urea ¹⁵N enrichment, which was assumed to be identical to urinary ¹⁵N-urea enrichment:

(8) $$F_{{pi}} ({\rm mmol}\,{\rm N}/{\rm day})=F_{{pi}} ^{{{\rm 15}}} /n_{p} $$

Given that the entry of ¹⁵N into the plasma urea-nitrogen pool must equal the amount exiting the pool allows for calculation of ¹⁵N entry into the plasma urea-nitrogen pool from the small intestine:

(9) $$F_{{ip}} ^{{{\rm 15}}} ({\rm mmol}\,^{{{\rm 15}}} {\rm N}/{\rm day})=F_{{pu}} ^{{{\rm 15}}} {\plus}F_{{pi}} ^{{{\rm 15}}} $$

Total nitrogen entry into the small intestine nitrogen pool can be calculated based on the measured infusion rate of tracer ammonia:

(10) $$F_{{Ii}} ({\rm mmol \,N}/{\rm day})=I_{{{\rm Ammonia}}} $$

Assuming complete absorption of the infused ¹⁵N, the difference between ¹⁵N absorption and infusion is the amount of label that is reabsorbed from recycled urea:

(11) $$F_{{ip}} ^{{{\rm 15R}}} ({\rm mmol}\,^{{{\rm 15}}} {\rm N}/{\rm day})=F_{{ip}} ^{{{\rm 15}}} -F_{{Ii}} $$

Given that ammonia flow and ammonia enrichment at the ileum are known, the amount of nitrogen and ¹⁵N leaving the small intestine can be calculated:

(12) $$F_{{il}} ({\rm mmol}\,{\rm N}/{\rm day})={\rm ileal}\,{\rm ammonia}\,{\rm flow} $$

(13) $$F_{{il}} ^{{{\rm 15}}} ({\rm mmol}\,^{{{\rm 15}}} {\rm N}/{\rm day})=n_{i} {\times}{\rm ileal}\,{\rm ammonia}\,{\rm flow}$$

The remaining ¹⁵N disappearance not yet accounted for is assumed to be incorporated into microbial protein production:

(14) $$F_{{im}} ^{{{\rm 15}}} ({\rm mmol}\,^{{{\rm 15}}} {\rm N}/{\rm day})=F_{{pi}} ^{{{\rm 15}}} -F_{{ip}} ^{{{\rm 15R}}} -F_{{il}} ^{{{\rm 15}}} $$

Rearranging the above equations, with the condition that all nitrogen flows must be greater than or equal to zero, yields minimum and maximum possible nitrogen fluxes out of the ileal pool due to microbial protein production and absorption of ammonia-nitrogen into the plasma urea pool, plasma urea entry due to endogenous AA catabolism and entry of nitrogen into the ileal pool due to FAAC, which are outlined below.

The incorporation of nitrogen into microbial protein production has the following range (mmol N/kg per day):

(15) $$\openup 4pt\eqalignno{&#x0026;\left( {[10]{\plus}[8]{\minus}[12]} \right)\left( {1{\minus}{{[7]{\plus}[5]} \over {[10]{\plus}[7]{\plus}[13]}}} \right)\cr &#x0026;\leq F_{{im}} \leq \left( {[3]{\plus}[8]{\minus}[1]} \right)\left( {{{[10]{\plus}[7]{\plus}[13]} \over {[7]{\plus}[5]}}{\minus}1} \right)$$

Given this range, the amount of ammonia-nitrogen that is absorbed from the upper gut can be calculated (mmol N/kg per day):

(16) $${{[\min \,15]} \over {\left( \displaystyle{{{[10]{\plus}[7]{\minus}[13]} \over {[7]{\plus}[5]}}{\minus}1} \right)}} \leq F_{{ip}} \leq {{[\max \,15]} \over {\left( \displaystyle{{{[10]{\plus}[7]{\minus}[13]} \over {[7]{\plus}[5]}}{\minus}1} \right)}}$$

The remaining plasma urea-nitrogen production is from endogenous AA catabolism (mmol N/kg per day):

(17) $$ {{2{\times}[2]} \over {[4]{\minus}[1]{\minus}[\max \,16]}}\leq F_{{ep}} \leq {{2{\times}[2]} \over {[4]{\minus}[1]{\minus}[\min \,16]}}$$

Finally, the minimum and maximum amount of small intestine ammonia-nitrogen production that is due to FAAC (mmol N/kg per day) is determined by

$$\eqalign{&#x0026;[{\rm min}\,{\rm 16}]{\plus}[{\rm min}\,{\rm 15}]{\plus}[{\rm 12}]-[{\rm 1}0]-[{\rm 8}]\leq F_{{oi}} \leq \cr &#x0026;\quad[{\rm max}\,{\rm 16}]{\plus}[{\rm max}\,{\rm 15}]{\plus}[{\rm 12}]-[{\rm 1}0]-[{\rm 8}]$$

Statistical analysis

All data were analysed using the mixed model procedure (PROC MIXED) of the SAS statistical program (SAS 9.1; SAS Institute Inc., Cary, NC, USA). Diet was included in the model as a fixed effect and period within study and pig within period were included as random effects. Pig within period was the experimental unit for all measures. Differences between treatment means were determined using the Tukey test and were considered statistically significant at P⩽0.05.