When rabbit reticulocyte lysates were gel-filtered on Sephadex (3-25 or (3-50, the rate of protein synthesis in the gel-filtered lysate was the same as in the parent lysate provided appropriate concentrations of ATP, G T P and spermidine or spermine were added. The polyamines were found to increase the rate of synthesis and to lower the Mg2 + optimum. Although gel-filtered lysates prepared in this way syntheses protein at a high initial rate, this rate is maintained for only about 20 min, after which it declines rapidly to a very low rate, even though optimal concentrations of haemin are present. This shut-off was completely prevented by the addition of low concentrations of sugar phosphates capable of acting as NADPH generators: glucose 6-phosphate, 2-deoxyglucose 6-phosphate, ribulose 5-phosphate and ribose 5-phosphate. Although the addition of isocitrate resulted in the generation of NADPH, the stimulation of protein synthesis was normally less than was achieved with glucose 6-phosphate. Dithiothreitol also promoted only partial activation, as did fructose 1,6-bisphosphate, a sugar phosphate which does not generate NADPH in this system. Full stimulation was observed, however, when both fructose 1,6-bisphosphate and either dithiothreitol or isocitrate were added. It is argued that the maintenance of maximum rates of protein synthesis in gel-filtered lysates requires the presence of both a sugar phosphate and a reducing agent, which can be dithiothreitol or an NADPH-generating system. Low concentrations of thioredoxin were shown to be required for the stimulatory effect of NADPH-generating systems, but not for stimulation by dithiothreitol. On the basis of these studies, a recommended procedure is given for the preparation of highly active gel-filtered lysates and gel-filtered nuclease-treated lysates.The rabbit reticulocyte lysate is unique amongst mammalian cell-free systems in that it synthesises protein at the same rate as the intact cell for periods of 60 min or more [I]. Almost all the work that has been done with this system has used crude undialysed lysates for the reason that dialysis or gel-filtration generally results in very considerable loss of translation activity. For many years this has been a challenging problem, since the ability to make gel-filtered lysates as active as the parent lysate would not only extend the utility of the lysate as a system for translation of heterologous mRNA [2], but would also allow one to define what low-molecular-weight cofactors are required to maintain a protein synthesis rate comparable to that in the intact cell. In fact, we and others have discovered some conditions which allow gel-filtered lysates to synthesise protein at high rates [3-91, but the exact nature of the requirements for low-molecular-weight cofactors, and the reasons underlying these requirements have remained obscure and controversial. In this and the following papers we describe recent work which clarifies and largely solves these outstanding problems.Although the data are extensive and the ar...