This study tests the purported signal amplification capability of the glutamine synthetase (GS) regulatory cascade in Escherichia coli. Intracellular concentrations of the pivotal regulatory protein GlnB were modulated by varying expression of its gene (glnB). Neither glnB expression nor P(II)* (i.e. the sum of the concentration of the P(II)-like proteins GlnB and GlnK) had control over the steady-state adenylylation level of GS when cells were grown in the presence of ammonia, in which glnK is not activated. Following the removal of ammonia, the response coefficient of the transient deadenylylation rate of GS-AMP was again zero with respect to both glnB expression and P(II)* concentration. This was at wild-type P(II)* levels. A 20% decrease in the P(II)* level resulted in the response coefficients increasing to 1, which was quite significant yet far from expected for zero-order ultrasensitivity. The transient deadenylylation rate of GS-AMP after brief incubation with ammonia was also measured in cells grown in the absence of ammonia. Here, GlnK was present and both glnB expression and P(II)* lacked control throughout. Because at wild-type levels of P(II)*, the molar ratio of P(II)*-trimer/adenylyltransferase-monomer was only slightly above 1, it is suggested that the absence of control by P(II)* is caused by saturation of adenylyltransferase by P(II)*. The difference in the control of deadenylylation by P(II)* under the two different growth conditions indicates that control of signal transduction is adjusted to the growth conditions of the cell. Adjustment of regulation rather than ultrasensitivity may be the function of signal transduction chains such as the GS cascade. We discuss how the subtle interplay between GlnB, its homologue GlnK and the adenylyltransferase may be responsible for the 'redundant', but quantitative, phenotype of GlnB.