The production of acetate, propionate and butyrate in the rumen of sheep: fitting models to 14C- or 13C-labelled tracer data to determine synthesis rates and interconversions

Abstract

A procedure is described for solving an open, fully exchanging, three-compartment model representing ruminal volatile fatty acids (VFA) kinetics in sheep. This model was solved using results from a published study in which labelled VFA, viz. 1-14C-acetate (Ac), 1-14C-propionate (Pr) and 1-14C-butyrate (Bu), were infused individually at a constant rate for 240 min into the rumen of sheep on different occasions. During the tracer infusions, the sheep were given 75 g of lucerne (Medicago sativa) every hour. The patterns of increasing specific radioactivity (SA) during the infusions were described by differential equations that were solved using the computer software, WinSAAM. This linear kinetic analysis gave estimates of the rates of synthesis and absorption of Ac, Pr and Bu and the carbon interconversions between each acid. The sizes of the Ac, Pr and Bu compartments (10.5, 3.9 and 2.1 g C respectively), were also estimated, which is not possible with commonly used algebraic analyses. The model output showed that tracer : tracee equilibrium (plateau SA) had not been reached in the Ac, Pr and Bu compartments during the 240 min of tracer infusion and the algebraic method of analysis used in the original study was therefore compromised. The procedures described here eliminated this source of error; in addition, confidence in the model solution was improved because all data representing the build-up to plateau SA were used, rather than just estimates of ‘plateau’ SA. After accounting for VFA interconversions, the rates of VFA absorption (or incorporation into other materials such as microbial polymers or methane) were 85, 48 and 49 g C/day, or 78%, 91% and 73% of the total production of each VFA, respectively. The variability in the observed SA responses to the tracer infusions was relatively large and this is discussed. Model solutions using the linear kinetic analysis and the previously widely used algebraic analyses are compared. The effect of positional labelling in tracer VFA is also discussed in the light of evidence that the rate of absorption of Pr may be overestimated when [1-14C]-Pr rather than [2-14C]-Pr or uniformly labelled [14C]-Pr is used as the Pr tracer.