Inhibition of tobacco NADH-hydroxypyruvate reductase by expression of a heterologous antisense RNA derived from a cucumber cDNA: Implications for the mechanism of action of antisense RNAs

Tobacco plants were genetically transformed to generate antisense RNA from a gene construct comprised of a full-length cucumber NADH-dependent hydroxypyruvate reductase (HPR) cDNA placed in reverse orientation between the cauliflower mosaic virus 35S promoter and a nopaline synthase termination/polyadenylation signal sequence. In vivo accumulation of antisense HPR RNA within eight independent transgenic tobacco plants resulted in reductions of up to 50% in both native HPR activity and protein accumulation relative to untransformed tobacco plants (mean transgenote HPR activity=67% wild type, mean transgenote HPR protein=63% wild type). However, in contrast to previous reports describing antisense RNA effects in plants, production of the heterologous HPR antisense RNA did not systematically reduce levels of native tobacco HPR mRNA (mean transgenote HPR mRNA level=135% wild type). Simple regression comparison of the steady-state levels of tobacco HPR mRNA to those of HPR antisense RNA showed a weak positive correlation (r value of 0.548, n=9 ; n is wild type control plus eight independent transformants; significant at 85% confidence level), supporting the conclusion that native mRNA levels were not reduced within antisense plants. Although all transgenic antisense plants examined displayed an apparent reduction in both tobacco HPR protein and enzyme activity, there is no clear correlation between HPR activity and the amount of either sense (r=0.267, n=9) or antisense RNA (r=0.175, n=9). This compares to a weak positive correlation between HPR mRNA levels and the amount of HPR activity observed in wild-type SRI tobacco plants (r=0.603, n=5). The results suggest that in vivo production of this heterologous HPR antisense RNA is inhibitory at the level of HPR-specific translation and produces its effect in a manner not dependent upon, nor resulting in, a reduction in steady-state native HPR mRNA levels. In this context, the observed antisense effect appears to differ mechanistically from most antisense systems described to date.