The N(2)O : N(2) product ratio of denitrification is negatively correlated with soil pH, but the mechanisms involved are not clear. We compared soils from field experiments where the pH had been maintained at different levels (pH 4.0-8.0) by liming (> or = 20 years), and quantified functional gene pools (nirS, nirK and nosZ), their transcription and gas kinetics (NO, N(2)O and N(2)) of denitrification as induced by anoxic incubation with and without a carbon substrate (glutamate). Denitrification in unamended soil appeared to be based largely on the activation of a pre-existing denitrification proteome, because constant rates of N(2) and N(2)O production were observed, and the transcription of functional genes was below the detection level. In contrast, glutamate-amended soils showed sharp peaks in the transcripts of nirS and nosZ, increasing the rates of denitrification and pH-dependent transient accumulation of N(2)O. The results indicate that the high N(2)O : N(2) product ratio at low pH is a post-transcriptional phenomenon, because the transcription rate of nosZ relative to that of nirS was higher at pH 6.1 than at pH 8.0. The most plausible explanation is that the translation/assembly of N(2)O reductase is more sensitive to low pH than that of the other reductases involved in denitrification.