Strongly in favor of this conception of a difference in reactivity of G units from the terminal methyl end vs. all others derived from the fatty acid chain are the data of Geyer and his coworkers (16-18) with tissue slices. They have measured the C*0/C*OOH ratio of acetoacetate formed from carboxyl-labeled fatty acids as a function of chain length, with the finding that the value for this ratio approaches unity with increasing chain length, as would be expected by the theory (Table 1). These workers have also measured the distribution of isotope in acetoacetate derived from carboxyl-labeled fatty acids with an odd number of carbon atoms, and found C*O/ C*OOH ratios as low as those of the even-numbered acids. These results indicate that the 3-carbon " propionic acid " residue formed from the terminal 3-carbon atoms of the odd-numbered fatty acids must give rise to a C2 unit of the " omega " type, and thus must be different from pyruvic acid. It has also been found by Geyer, Cunningham, and Pendergast (17) that the presence of malonate tends to equalize the C*O/C*OOH ratio. They have explained this effect in terms of the two species theory and have shown that this effect is to be expected if the " carboxyl" type of fragment is the species preferentially fed into the citric acid cycle.

All of the experimental work cited above concerning acetoacetate formation from fatty acids must be interpreted in the light of the important results reported by Stadtman, Doudoroff, and Lipmann (58). In experiments with unlabeled acetate plus labeled acetyl phosphate, in the presence of the enzyme phosphotransacetylase from C. kluyveri and the acetoacetate-synthesizing enzyme from pigeon liver, acetoacetate synthesis could be demonstrated. When the acetoacetate so synthesized was degraded, it was found that two moles of acetyl phosphate were incorporated into each mole of acetoacetate, while free acetate was not significantly active in forming this product. These results are interpreted as indicating the formation of acetoacetate by the condensation of two moles of acetyl-CoA, formed from acetyl phosphate by the reaction catalyzed by the enzyme phosphotransacetylase (59).

It is clear from these experiments that both halves of the acetoacetate molecule are derived from acetyl-CoA. The two-species theory thus demands that two different forms of acetyl-CoA be generated, one form arising exclusively from the terminal 2 carbons of the fatty acid chain and the other type from the other carbons of the chain. It has been pointed out above that the data upon which the two species theory is based require that the two types of C2 units be interconvertible, but at a rate slower that the rate of acetoacetate synthesis. Stern, Shapiro, Stadtman, and Ochoa have suggested that the properties of acetyl-CoA in displaying both carboxyl and methyl activation may be due to a tautomeric equilibrium between enolized and non-enolized forms of the acetyl coenzyme (61).

On the basis of this suggestion, it is possible that the " carboxyl " fragments represent the enolized form, while the " omega " fragments are of the non-enolized type. Present hypotheses concerning the splitting off of the acetyl-CoA units from the fatty acid molecule, based on formulations presented by Kennedy and Barker (25) and by Lieberman and Barker (3), involve a thioclastic addition of CoA at the β position.

If it is assumed that this thioclastic reaction involves the splitting off of the acetyl-CoA in the enolized form, this assumption would be consistent with the more rapid condensation of this type of G unit with oxalacetate. Since the terminal 2-carbon units of the fatty acid chain could not undergo such a thioclastic reaction at the f$ position it would be generated as the non-enolized form. If the tautomeric equilibrium visualized by Stern et al. really exists, the kinetics of this reaction must be such as to preclude the complete equilibration of the methyl hydrogen with water of the body fluids, as might be expected to occur during the course of such an enolization. Bloch and Rittenberg (6,53) have performed experiments, with acetate labeled with both O3 and deuterium, which have an important bearing on this point. When a preparation of acetate was fed in which deuterium and the carbon isotope were present in a ratio of 8/1, the fatty acids isolated from the liver contained the isotopes in a ratio of 2.5/1. In building up a long chain fatty acid from acetic acid, there should be introduced for every two hydrogens of the methyl carbon of acetate, at least two hydrogens from the body fluids, which would yield fatty acids containing a deuterium: C1' ratio of 4.0/1. The greater loss of deuterium actually observed may be the result of a limited enolization which occurs somewhere in the process. Bloch (5) has suggested that this may be due to a keto-enol tautomerism of an intermediary keto acid. On the other hand, it may be the result of the equilibration of acetyl-CoA suggested by Stern et al. It is also of interest that in other experiments (6) in which doubly labeled acetate was fed and acetyl phenylamino-butyric acid was isolated, the ratio of isotopes in the acetylated product was the same as that of the acetate originally fed, a result indicating that no exchange of methyl hydrogen with body fluids had taken place. This is consistent with the view that C2 units of the acetylating or " omega " type are of the non-enolized form.

To some extent representing an exception to the " two species theory" is the very recent work of Chaikoff, Goldman, Brown, Dauben, and Gee (10), who have measured C*0/C*OOH ratio of labeled acetoacetate formed during the incubation of liver slices removed from rats which had been given, a short time previously, intravenous infusions of three different finely emulsified tripalmitin samples, containing palmitic acid labeled respectively at the carboxyl, 5-, and 11-carbon atoms. This technique was employed to assure that the labeled fatty acid was brought to the interior of the liver cell in physiological form. The data collected are also shown in Table 1. It is seen that the carboxyl-labeled palmitate yielded acetoacetate with symmetrical isotope distribution, as expected from work already outlined. However, palmitate samples labeled at the 5 and 11 positions yielded, instead of the expected C*0/C*OOH ratio of 1.0, ratios of 1.2 to 1.3. This unexpected finding is not consistent with either completely random condensation or with the two species hypothesis described above. To explain it, the authors have abandoned the " two species " concept entirely and have proposed that the major part of acetoacetate formation occurs by completely random condensation, as proposed by Weinhouse et al. (63), whereas the rest may arise by an interruption in this process whereby C2 units already removed oxidatively may condense back (" oriented condensation ") with oxidized residues of fatty acids already shortened by one or more units. Subsequent cleavage of a 4-carbon (acetoacetate) molecule from the carboxyl end of the chain may lead to acetoacetate in which the carbonyl half of the molecule contains statistically more C2 units from the middle part of the chain. It is obvious that recondensation of this type may occur during fatty acid synthesis. In any event, the work of the Chaikoff group has demonstrated the importance of more study of the long-chain acids of physiological importance, since so much of the tracer work has been done in the past with the short-chain acids and may not represent the metabolism of the long-chain acids accurately. It is hardly warranted at the present time to abandon the " two species hypothesis " on the basis of this work, since in a long-chain acid the special properties of the single terminal CH3CO- unit in contrast to the 7 -CH2CO-units could on statistical grounds be expected to show only small differences in the C*O/C*OOH ratio, a fact which is mentioned by the California workers. It may therefore be concluded that the " two-species " theory presented by Crandall, Brady, and Gurin does not explain the important results recently presented by Chaikoff et al. (10); but on the other hand, it is unlikely that an oriented recondensation such as that postulated to explain the results obtained with palmitic acid can have an important bearing on previous findings with 3-labeled octanoate or carboxyl-labeled hexanoate.