Estimation of the metabolizable energy equivalence of dietary proteins

Protein contributes significantly to the human daily energy budget. The high diversity in composition, digestibility and dietary proportion complicates the estimation of its actual energy contribution. In practical terms we continue using the energy equivalents estimated by Atwater. This results in a persistent source of imprecision in the calculation of dietary energy that at least can be partially corrected.

We used experimentally obtained data to compute an algorithm that will allow to estimate the gross energy content of a protein which composition is known. The relationship between gross energy (i.e. bomb calorimeter-derived) of protein is not a direct correlate of its metabolic efficacy as energy supplier. Thus we estimated the metabolic energy yield (i.e. ATP equivalents) of amino acid residues, using the data to compute the estimated protein metabolic energy yield. Both approaches were to be used to propose a corrected protein energy equivalence factor that will increase the precision in the calculation of dietary protein energy, especially when information on protein composition is available.

The gross energy content of amino acids was measured with a bomb-calorimeter, and compared with that of glucose. Amino acid estimated metabolizable energy yield, in moles of ATP per mol of amino acid residue, was also calculated. The net heat yield of all amino acids were used to compute the theoretical heat production of albumin, collagen, gelatin and whole rat protein, which gross energy was also measured experimentally. The mean estimated energy yield (both gross and metabolizable) for each amino acid residue were used to compute the theoretical energy of a number of dietary protein sources which composition was available in the literature.

Calculated energy yield of a few selected proteins coincided with the data directly measured in the bomb calorimeter. The computed gross energy yield and metabolizable energy yield for a number of dietary protein sources was estimated. There were minor differences between both parameters: the proportion of aromatic and branched chain amino acids was the main factor affecting the gross energy yield of a given protein; conversely, the higher proportion of nitrogen, especially, but not exclusively, related to arginine and glycine content correlated with lower metabolizable energy. These parameters, corrected by the gross and metabolizable energy of glucose were used to compute a mean energy equivalence for dietary protein: 19 kJ/g protein (i.e. 4.55 kcal/g protein). This value, higher than the current Atwater factor, does not include protein digestibility (as Atwater value did), but included the cost of nitrogen excretion.

The methodology presented allows the approximate calculation of both the purported heat production of a protein (pure or mixture) for which we know its amino acid composition (and even get a good estimate if we only know its proportion of nitrogen), and its metabolic energy equivalence. We also propose the use of a new energy correlate of dietary protein; this can be further tuned if the proportion of nitrogen in the protein is known, and even further if its amino acid composition is available. As a consequence of its application to dietary proteins, their energy yield may be higher than usually considered, which may influence the calculations of energy balance.