The phosphorylation by T4 polynucleotide kinase of various double-stranded DNAs containing defined 5''-hydroxyl end group structures was studied. Particular emphasis was placed on finding conditions that allow complete phosphorylation. The DNAs employed were homodeoxyoligonucleotides annealed on the corresponding homopolymers, DNA duplexes corresponding to parts of the genes for alanine yeast tRNA and a suppressor tyrosine tRNA from Escherichia coli. The rate of phosphorylation of DNAs with 5''-hydroxyl groups in gaps was approximately 10 times slower than for the corresponding single-stranded DNA. At low concentrations of ATP, 1 .mu.M, incomplete phosphorylation was obtained, whereas with higher concentrations of ATP, 30 .mu.M, complete phosphorylation was achieved. In the case of DNAs with 5''-hydroxyl groups at nicks, approximately 30% phosphorylation could be detected using 30 .mu.M ATP. A DNA containing protruding 5''-hydroxyl group ends was phosphorylated to completion using the same conditions as for single-stranded DNA, i.e., a ratio between the concentrations of ATP and 5''-hydroxyl groups of 5:1 and a concentration of ATP of approximately 1 .mu.M. For a number of DNAs containing protruding 3''-hydroxyl group ends and 1 DNA containing even ends, incomplete phosphorylation was found under similar conditions. For all these DNAs a plateau level was observed varying from 20-45% of complete phosphorylation. At 20 .mu.M and higher ATP concentrations, the phosphorylation was complete also for these DNAs. With low concentrations of ATP a rapid production of Pi was noted for all the latter DNAs. The apparent equilibrium constants for the forward and reverse reaction were determined for a number of different DNAs, and these data revealed that the plateau levels of phosphorylation obtained at low concentrations of ATP for DNAs with protruding 3''-hydroxyl group and even ends is not a true equilibrium resulting from the forward and reverse reaction. The plateau levels may be due to formation of inactive enzyme-substrate and enzyme-product complexes. For all double-stranded DNAs tested, except DNAs containing protruding 5''-hydroxyl group ends, addition of KCl to the reaction mixture resulted in a drastic decrease in the rate of phosphorylation and in the maximum level phosphorylated. Spermine had little influence. Both of these agents were previously shown to activate T4 polynucleotide kinase using single-stranded DNAs as substrates. The inhibition of phosphorylation of double-stranded DNAs by salt might be the result of stabilization of the 5''-hydroxyl group regions of these DNAs.