Although the isoleucyl-tRNA synthetase from Escherichia coli (IRS) does not catalyze the overall mischarging of tRNAIle with valine, it does undergo the first step of the reaction, the formation of an IRS-Val-AMP complex. The addition of tRNAIle to this complex leads to its quantitative hydrolysis and the IRS acts as an ATP pyrophosphate in the presence of valine and tRNAIle (Baldwin, A.N., and Berg, P. (1966), J. Biol. Chem. 241, 839). It is shown that during the ATP pyrophosphatase reaction: (a) IRS forms an IRS-Val-AMP complex; (b) the turnover number of the ATP pyrophosphatase reaction is the same at the rate constant for the transfer of isoleucine from IRS-Ile-AMP to tRNAIle over a wide range of temperature and pH; (c) mischarged Val-tRNAIle is hydrolyzed by IRS with a turnover number of 10 s-1 at pH 7.78 and 25 degrees C, compared with a value of 1.2 s-1 for the transfer of isoleucine from IRS-Ile-AMP to tRNA or for the ATP pyrophosphatase reaction. Although this appears to be consistent with an editing mechanism in which there is a slow transfer of the valine from the IRS-Val-AMP to tRNAIle follwed by the rapid hydrolytic step, as recently found for the rejection of threonine by the valyl-tRNA synthetase, there is an inconsistency. This scheme predicts that on mixing IRS.[14C]Val-AMP with tRNAIle there should be a transient misacylation of the tRNA such that about 10% of the [14C]Val is present as [14C]Val-tRNAIle at the peak. But 0.8% or less is found. This could possibly be caused by the IRS having a higher hydrolytic activity during the mischarging reaction than is measured on mixing the unligated enzyme with performed Val-tRNAIle. Alternatively, a two-stage editing mechanism must be considered in which the majority of the Val-AMP is destroyed before the transfer to tRNA in the major editing step, while the hydrolytic activity of the IRS towards Val-tRNAIle is a second editing step to mop up any mischarged tRNA formed by the Val-AMP escaping the first editing step. It is shown that the "kinetic proofreading" mechanism of Hopfield is not consistent with the experimental data.