The abundance of different tRNAs in Escherichia coil at different grow
th rates correlates strongly with the usage of the corresponding cogna
te codons. On the assumption that the investment in the translation sy
stem is optimised to provide a maximal growth rate, the relationship b
etween tRNA levels and codon usage can be predicted. When the complica
tions due to different degeneracies and different association rate con
stants for the different tRNA-codon combinations are accounted for, re
cent data from the literature indicate that the predicted relations ho
ld up very well: the tRNA levels correlate with codon frequencies in a
way that would support a maximal growth rate. The relations can also
be used to predict the association rate constant between an A-site cod
on and the cognate ternary complex. Ln the cases where they can be com
pared, the results agree reasonably well with experimental results fro
m the literature. (C) 1997 Academic Press Limited.