This thesis outlines studies of the structure of milk triglycerides. Initially the best conditions were found for the separation of milk triglycerides by thin layer chromatography (TLC) techniques into classes according to number and geometry of their double bonds. Prior separation of the sample into high, medium and low molecular v;eight triglyceride fractions by silicic acid column chromatography was necessary to obtain a satisfactory resolution of bands. The structure of the triglyceride classes was determined by a combination of fatty acid composition and a knowledge of the fatty acid double bond geometry obtained by silver ion adsorption TLC, The major triglyceride classes were identified as SSS, SSMt, SSMc, SMcMt and SMcMc (where S = saturated fatty acids; M = monoene fatty acids; c = cis; t = trans), A number of more polar unsaturated triglycerides were grouped as 'other'. There was excellent agreement between fatty acid and triglyceride molecular species composition of the original sample and values reconstructed by proportional summation of fatty acid and triglyceride molecular species composition for the classes of the three molecular weight fractions.
Triglyceride classes, separated by silver ion adsorption TLC from eight samples of milk fat, had a low correlation between softening point and triglyceride class content. When the molecular species composition of the triglyceride classes was determined and correlated with softening point, the best correlations were between softening point and some low and some high molecular weight triglycerides of the intact fat and the trisaturated (SSS) triglyceride class. Correlation coefficients were determined between softening point and mole percentage of individual fatty acids and groups of acids. The highest correlations were between softening point and individual short-chain length fatty acids (C4:0-C10:0) and groupings of short-chain length acids (C4:0- C10:0) and groupings of long-chain length saturated acids (C16t0- C20:0). Interesterification increased the softening point of milk fat from 31.6 to 36.3°C and produced increased amounts of both low and high and decreased amounts of medium molecular weight molecular species in all triglyceride classes. The increases in the amounts of high molecular weight triglycerides were several times greater than the increases in the amounts of low molecular weight triglycerides.
A sample of milk fat was fractionated into high, medium and low molecular weight triglycerides by silicic acid column chromatography. Each fraction was further separated by silver ion adsorption TLC into triglyceride classes which were subjected to stereospecific analysis, trans-monoene fatty acids were found to be incorporated into triglycerides in the same manner as their cis-isomers. The specific positional distribution for fatty acids in intact milk triglycerides was not always maintained in the molecular weight fractions or in the various classes of unsaturation.
A set of mammary gland secretions from 24 days pre-partum to 28 days post-partum were subjected to fatty acid and stereospecific analysis. Pre-partum mammary gland secretion differed from normal milk in that it contained approximately twice the amount of 16:0 and larger than normal amounts of 14:0. This was compensated for by lesser amounts of other medium-chain, short-chain, 18:0 and 18:1 acids. Post-partum milk collected at 1 to 14 days contained less 6:0 to 14:0 and more 18:0 and 18:1 acids than milk collected at 21 and 28 days. The positional distribution of fatty acids in milk from one day post-partum onwards generally followed the specific pattern normally encounted for milk triglycerides. With the exception of the first sample at 24 days pre-partum, pre-partum mammary gland secretion had the medium-chain acids 8:0, 10:0, 12:0 and sometimes 14:0 preferentially esterified at the sn-3-position rather than the sn-2-position. Linear relationships were found between the content of fatty acids at the three stereospecific positions and the content of the same acid in the triglycerides. These relationships are presented by regression equations and were generally highly significant. With varying triglyceride fatty acid composition the fatty acid composition at the three positions did not change at the same rate.
Milk triglycerides from the echidna, koala, Tammar wallaby, guinea pig, dog, cat, seal, horse, pig and cow were subjected to fatty acid and stereospecific analysis. While milk from the various species provided a wide range in type and composition of fatty acids, all species except the echidna exhibited a similar fatty acid distribution pattern.
Triglyceride molecular species composition of milks from the above mammals v;as compared with that predicted by the 1-random-2-random-3-random fatty acid distribution hypothesis. There was excellent agreement between predicted and experimental compositions for echidna, koala, Tammar wallaby, guinea pig and pig and reasonable agreement for dog, cat, horse and cow milks.