The carboxyl groups of a protein can be esterified with diazomethane, but more conveniently with methanolic hydrochloric acid. In the present research, however, neither reagent has proved to be entirely satisfactory, for in our hands the first has not given full esterification and the second would do this only at the expense of the amide N, which we were particularly anxious to keep intact. Fraenkel-Conrat & Olcott (1945) were the first to show that the concentration of mineral acid (0-02 to 01 N) required to catalyse the reaction with methanol was very much less than previous workers had employed for the purpose. They claimed that 97 % of the carboxyl groups of polyglutamic acid from Bacillu8 8ubtili8 could be esterifled by treatment with methanolic 0-05N-HCI at 22-24°for 24 hr. and that several proteins, including insulin, were fully esterified under similar conditions in the presence of 01IN-HCI. They mentioned that there was no loss of amide N under such treatment, though data in support of the assertion were not presented. Mommaerts & Neurath (1950) repeated the experiments with insulin, and confirmed that full esterification was apparently achieved with methanolic 0-1 N-HCI at 250 in 24 hr. They claimed that the ammonia liberated under these conditions amounted to not more than 2 % of the amide N.Our own observations do not support the contention that 0 1 N-HCI can be used as a catalyst in the esterification of proteins in this way without loss of amide groups. With insulin the ammonia produced in 24 hr. at 25°represents about 6-6 % of the amide N, a value much greater than the abovementioned workers seem to have suspected. Lowering the concentration of hydrochloric acid spares the amide N but unfortunately full esterification is not then achieved. We have thus been obliged to prepare the ester with methanolic 0 1 N-HCI, knowing that when this is used subsequently to determine inter alia the amide distribution of the protein, we had already set a limit to the accuracy with which this could be done. Another interesting side-reaction of protein esterification has been traced to N-O acyl migration in serine and particularly threonine residues. EXPERIMENTAL MateriakProtein samples, and methods for determining N, ammonia N and amide N were described in the preceding paper (Chibnall, Mangan & Rees, 1958b). Methoxyl. The Zeisel procedure of Pregl (1937) was used. Estimations on 20-30 mg. samples agreed to within 2%.Methanol. This was boiled under reflux for some hours with Mg turnings and then distilled.Methanolic hydrochloric acid. An approx. 2N-solution was prepared with dry HCI gas and stored at -150. As the acidity falls on keeping it is necessary to titrate a sample immediately before use. MethodsEstimation of the extent of esteriftcation. The extent of esterification of the total free carboxyl groups was followed by determination of methoxyl. As ,-lactoglobulin contains methionine, which yields methyl iodide on treatment with HI, the (apparent) methoxyl content of the protein itself was subtracted from that...

The authors express their sincerest gratitude to Professor J. Millot (Mus6um National d'Histoire Naturelle, Paris), without whose generous help this work could not have been undertaken. They thank Mr Anders Kallner (Karolinska Institute, Stockholm) for drawing their attention to and describing his unpublished experiments with the Raney nickel-cumene allomerization method.

The foaming properties of rumen liquor, bovine salivary mucoprotein, cytoplasmic protein from red clover, and saponins from various pasture legumes have heen examined.Foam strengths of the saponin foams have pH optima, those from red clover and lucerne being between pH 4.5 and 5.0. The presence of calcium in the saponin is necessary for the formation of rigid foams.Foams from red-clover cytoplasmic protein show maximum strens-ths at pH values ranging from 5.4 to 6.0. The foaming properties are markedly dependent on salt concentration.Salivary mucoprotein has greatest foam strength above pH 7.5, but some samples have a secondary peak at pH 6.5.Rumen liquor from cows fed on red clover has a pH optimum for foam strength at pH 5.4 to 5.7 similar to cytoplasmic protein. The particulate matter in rumen liquor has considerable anti-foaming properties.The evidence indicates that cytoplasmic protein is of major importance as a foaming agent in causing bloat in cattle. The implications of these results are discussed with reference to bloat.

C-TERMINAL RESIDUES IN PROTEINS 111 residues in proteins, the need for which was emphasized in the introduction, is still unsolved and remains a challenge to the analyst. SUMMARY 1. The carboxyl groups of C-terminal residues in proteins, after preliminary esterification, can be reduced with lithium borohydride, and on subsequent acid hydrolysis the modified residues are present in the hydrolysate as amino alcohols or hydroxyamino acids as the case may be.2. The suitability of the procedure for determining the C-terminal residues in proteins has been explored, and in agreement with Crawhall & Elliott (1955), it was found that the issue was complicated by the simultaneous reductive cleavage of peptide bonds which occurs, under the conditions chosen, to the extent of 1-2 % of the total peptide bonds.3. Data are presented showing that with a protein of low molecular weight such as insulin (5732) the interference is not serious and a satisfactory determination of the C-terminal residues can be made. With a protein of much higher molecular weight such as P-lactoglobulin (37 000), however, the interference is serious.4. The procedure is not recommended as a reliable one for proteins but it may be of use with peptides of low molecular weight.One of us (A. C. C.) would like to acknowledge the generous financial assistance provided by Eli Lilly and Co., Indianapolis, U.S.A.

No abstract