ROUND TABLE DISCUSSION 513 
From what has already been said, it would appear that a reasonable step toward the solution 
of this problem would be to devise a means that would permit the protein molecule to be broken 
down by the usual enzymatic means and would, at the same time, permit the complete hydrolysis 
of those highly resistant peptides which incorporate p-amino acid residues. We became interested 
in this problem several years ago, not for the purpose of ascertaining whether or not there were 
bD-amino acid residues in proteins, but because we sought a means that would allow us to determine 
the optical purity of small synthetic peptides. We were, at the time, engaged in the synthesis of a 
number of peptides of the D- and L-configuration for ultimate use as peptidase substrates, and we 
became increasingly concerned with the question as to whether the synthetic procedures we 
employed were accompanied by racemization. Thus, for example, after preparing the four stereo- 
isomers of leucylleucine, how would it be possible to determine to what extent, if any, the L-L 
isomer was contaminated with its stereoisomeric D-L, L-p and p-p forms? Measurement of the 
optical rotation of these different isomers certainly couldn’t provide the answer, because not only 
are such measurements of insufficient sensitivity to detect trace contamination, but they have 
little meaning in the case of molecules which contain two or more asymmetric centers. We there- 
fore instituted the search for a method that would enable us to determine whether a given optically 
active peptide was contaminated with any of its stereoisomers to the extent of only a fraction of a 
per cent. 
In this connection, it should be pointed out that some years ago GREENSTEIN isolated an 
aminopeptidase from hog kidney which exhibited an entirely different specificity than leucine 
aminopeptidase and which he unequivocally demonstrated was not identical with this latter 
enzyme. GREENSTEIN went on to demonstrate further that the p- and L-form of a large variety 
of dipeptides, all of which possessed an N-terminal glycine residue, were hydrolyzed at nearly 
equally rapid rates by this aminopeptidase, but the lack of suitable peptide substrates at that 
time did not permit him to pursue these specificity studies further. When we recently extended 
these specificity studies to each of the four stereoisomers of a large number of dipeptides, we found 
that not only were the L-L and L-p isomers rapidly hydrolyzed, as would be expected from the 
results observed earlier with the glycine-containing peptides, but the D-L and D-D isomers were 
also hydrolyzed at quite appreciable although somewhat slower rates. Even the stereoisomers of 
valylvaline, which are so slowly hydrolyzed by acid or alkali because of their high degree of 
steric hindrance, were readily cleaved by this enzyme. In any case, the observation of a peptidase 
activity toward which smaller peptides would be susceptible, irrespective of the configuration of 
the component amino acid residues, permitted us to develop a method for determining the optical 
purity of small peptides with a high degree of sensitivity. 
The method briefly is as follows: too wmoles of an aqueous solution of the pertinent dipeptide, 
L-leucyl-L-leucine for example, in borate buffer at pH 8.5 is placed in the conventional Warburg 
vessel and treated with the renal aminopeptidase. An aqueous solution of hog renal D-amino 
acid oxidase is placed in the side arm of the vessel and the whole allowed to come to temperature 
equilibrium at 37°. Then the oxidase in the side arm is added to the peptide solution and the 
hydrolytic and oxidative reactions permitted to take place simultaneously. Oxygen consumption 
is recorded until it stops. Under these conditions, 1 wzmole of a D-amino acid would consume 11.2 sul 
of oxygen so that the complete hydrolysis of each micromole of the contaminating optical anti- 
pode, p-leucyl-p-leucine in this instance, should yield 2 «moles of the D-amino acid and con- 
sequently cause the consumption of 22.4 wl of oxygen. A control is always run which is identical 
to the above but to which, in addition, is added 1 wmole of the D-p isomer of the dipeptide. Of 
course, the optical purity of a dipeptide of the p-p configuration can be ascertained in the same 
manner, except that it is obviously necessary here to use L-amino acid oxidase to detect contami- 
nating L-residues. In any event, the method is capable of permitting the detection of optical 
contamination with a sensitivity of less than 1 part in 100. 
Until Dr. RosENBERG raised the problem yesterday, I had not considered the application of this 
method to the detection of D-amino acid residues in proteins. In this case, it might prove of interest 
to hydrolyze a given protein with the usual proteolytic enzymes, treat the enzymic digest further 
with hog renal aminopeptidase in order to break down any small resistant peptides containing 
p-residues, and finally subject the whole mixture to the action of D-amino acid oxidase in a 
Warburg apparatus. 
While we are in the realm of speculation, there is an alternative method that might be used 
here that does not involve the prior isolation of the aminopeptidase, since it is an 77 vivo method. 
I offer this method because of our experiences with chemically defined diets. Over the past 6 years, 
we have formulated and studied the effect of some 200 different chemically defined diets, com- 
posed of pure crystalline L-amino acids, vitamins, glucose, the pertinent salts and essential fatty 
acids, on growth, reproduction, lactation and pathologic conditions of various types. Now it has 
long been known that certain of the p-amino acids, such as p-methionine, can be used instead of 
the corresponding essential L-amino acid for growth and that, on the basis of in vitro studies, this 
is presumably due to the conversion of the p-amino acid to its L-antipode by the successive en- 
References p. 524 
