Cuticle is the first barrier for rice to resist blast fungus on the surface of the leaf. Studies on how the rice leaf cuticle responds to rice blast and attempts to perform early detection of rice blast are limited, and these two issues were explored in this study via depth-profiling Fourier transform infrared photoacoustic spectroscopy (FTIR-PAS). Rice leaves with four different scales of injury (healthy leaves as CK, asymptomatic leaves from mildly diseased seedlings as S1, infected leaves with fewer than five lesions as S2, and infected leaves with more than 10 lesions as S3) were scanned by three moving mirror velocities 0.32, 0.47, and 0.63 cm/s for the depth profiling of the rice leaf surface. The response patterns were acquired via chemometrics to analyze the variations of the chemical group absorptions in the different layers of a sample and in the same layer between different samples. Results showed that the leaf cuticle tended to be thicker and the relative content of fatty alcohols and cutin, unsaturated compounds, and aromatics in the cuticle increased when rice seedlings were infected by blast fungus. Together with the principal component analysis, the probabilistic neural network was applied to identify the samples in early stages (CK and S1), which reached an accuracy of 90% for the samples in the greenhouse and 82% for the samples in the field. Thus, depth-profiling FTIR-PAS was good at analyzing the variation in cuticle layers and showed great potential in the early detection of rice blast or other diseases in different species.
Keywords: cuticle; infrared photoacoustic spectroscopy; probabilistic neural network; rice blast.