Abstract: 【Objective】 We aimed to determine the role of bacterial toxin in pine wilt disease by examining the changes in ultrastructure of 3-year-old Pinus thunbergii branches inoculated with different materials. 【Method】The inoculation materials were suspension of Bursaphelencus xylophilus, non-cell culturing liquid of the pathogenic bacteria GcM5-1A, a mixture of aseptic B. xylophilus + the pathogenic bacteria GcM5-1A, flagellin toxin of the pathogenic bacteria GcM5-1A, aseptic B. xylophilus + non-pathogenic bacteria Staphylococcus lentus AM2C, and suspension of aseptic B. xylophilus. Healthy P. thunbergii was used as control. For the diseased pines, samples were taken from different locations at the second or third stage after inoculation. For the non-diseased pines, samples were collected after inoculating for 11 days. The samples were sliced into thin sections and ultrastructural changes were observed with a transmission electron microscope(TEM).【Result】The ultrastructures of examined tissues were severely damaged in four treatment groups, namely wild B. xylophilus, non-cell culturing liquid of the pathogenic bacteria GcM5-1A, the mixture of aseptic B. xylophilus + the pathogenic bacteria GcM5-1A, and flagellin toxin of the pathogenic bacteria GcM5-1. The tendency of their pathological variations was consistent. However, in the treatment of non-cell culturing liquid of the pathogenic bacteria GcM5, the ultrastructures were more severely damaged in the samples close to the inoculation point than in the samples taken far away from the inoculation point. In wild B. xylophilus treatment, the results were reversed. Notably, the tissues and ultrastructure showed almost no damage in aseptic B. xylophilus + non-pathogenic bacterial S. lentus AM2C and suspension of the aseptic B. xylophilus treatments, and the ultrastructure in these two treatments were similar to those in healthy pine.【Conclusion】 The results indicate that the second stage of pine wilt disease is a critical period. The toxins produced by the pathogenic bacteria carried by B. xylophilus result in pine death, rather than the aseptic B. xylophilus itself. Our results support the hypothesis that pine wilt disease is caused by both the nematode and its associated pathogenic bacteria. Comparison of wild B. xylophilus and non-cell culturing liquid of the pathogenic bacteria GcM5-1A treatments showed that the movement of PWN in the pine tree increased the rate of pathogenic bacteria dispersal, which in turn resulted in acceleration of toxin production and diffusion, hastening the death of pines.