Recently we have demonstrated that guanosine 5'-triphosphate (GTP) is required for the formation of the first peptide bond in protein synthesis by ain E. coli cell-free system." 2 A preliminary investigation of the role of GTP in this reaction indicated that the GTP-dependent step occurred after the binding of formylmethionyl-tRNA (F-met-tRNA) to the ribosome. This conclusion was based on experiments which showed only a relatively small stimulation by GTP of the rate of the binding reaction.' However, a thorough investigation of this point hasrevealed a more complicated situation; it is now apparent that the effect of GTP on the binding reaction may vary markedly, depending upon the Mg++ concentration and the nature of the ribosome preparation used (that is, the relative proportions of 70S, 50S, and 30S subunits). Under certain conditions GTP is absolutely required for binding, whereas under others it merely stimulates the rate of the binding reaction. Our result agrees with recent reports from other research groups.3-7 In addition we have observed that guanylyl-5'-methylene diphosphonate (GMPPCP), a competitive inhibitor of peptide synthesis," 2, 8 also stimulates the binding reaction.6 The F-met-tRNA bound in the presence of GMP-PCP is unreactive with puromycin, whereas that bound with GTP reacts completely. This result suggests that the role of GTP in initiation may be complex, involving a series of distinct steps. The first is thought to be a binding of F-met-tRNA to a ribosome; the second is an activation step which renders the bound F-met-tRNTA reactive to puromycin or a second aminoacyl-tRNA. The fact that GMP-PCP cannot substitute for GTP in the second step suggests that hydrolysis of the 3-'y anhydride bond is involved, whereas this is probably not the case in the first step. Methods.-All preparative methods were similar to those previously reported,' except for the washing of ribosomes and the preparation of initiation factors. Ribosomes were pelleted from an S-30 extract (E. coli strain 1113) and resuspended and stored overnight in Buffer A [10 mM TrisHCI, pH 7.8, 40 mM Mg(OAc)2, 10 mM j3-mercaptoethanol, 2 mM EDTA, and 1 M NH4C1 . Ribosomes were pelleted from this solution by centrifugation for 6 hr at 105,000 X g (the supernatant is saved as a source of initiation factors), resuspended in Buffer A, clarified by low-speed centrifugation, and then repelleted. Ribosomes were resuspended in Buffer B [10 mM Tris-HCl, pH 7.8, 10 mM Mg(OAc)2, 6 mM a-mercaptoethanol, and 50 mM NH4Cl], clarified, and repelleted. This final pellet was resuspended in Buffer B, clarified, and stored at 00. Ribosomes prepared by this method are virtually free of initiation factors; they consist of about 90% 70S ribosomes and 10% 50S subunit. There are no detectable 30S subunits present (see Table 1). Initiation factors were prepared from the first ribosomal wash by filtration (Millipore filter, type HAWP) and precipitation with ammonium sulfate (65% saturated). The precipitate was redissolved in Buffer C (50 mM Tris-HCl, pH 7.8, 10 mM 0-mercaptoethanol) and passed over a G-25 sephadex column equilibrated in the same buffer. The breakthrough peak of the eluate was charged onto a DEAE-cellulose column equilibrated with Buffer C containing 0.25 M NH4Cl, and eluted with the same buffer. Protein was precipitated from the eluate with ammoniumi sulfate (80% saturated). The initiation factors prepared by this techiiique are virtually free of