Plant growth-promoting rhizobacteria, in association with plant roots, can trigger induced systemic resistance (ISR). Considering that low-molecular weight volatile hormone analogues such as methyl jasmonate and methyl salicylate can trigger defense responses in plants, we examined whether volatile organic compounds (VOCs) associated with rhizobacteria can initiate ISR. In Arabidopsis seedlings exposed to bacterial volatile blends from Bacillus subtilis GB03 and Bacillus amyloliquefaciens IN937a, disease severity by the bacterial pathogen Erwinia carotovora subsp. carotovora was significantly reduced compared with seedlings not exposed to bacterial volatiles before pathogen inoculation. Exposure to VOCs from rhizobacteria for as little as 4 d was sufficient to activate ISR in Arabidopsis seedlings. Chemical analysis of the bacterial volatile emissions revealed the release of a series of low-molecular weight hydrocarbons including the growth promoting VOC (2R,3R)-(Ϫ)-butanediol. Exogenous application of racemic mixture of (RR) and (SS) isomers of 2,3-butanediol was found to trigger ISR and transgenic lines of B. subtilis that emitted reduced levels of 2,3-butanediol and acetoin conferred reduced Arabidopsis protection to pathogen infection compared with seedlings exposed to VOCs from wild-type bacterial lines. Using transgenic and mutant lines of Arabidopsis, we provide evidence that the signaling pathway activated by volatiles from GB03 is dependent on ethylene, albeit independent of the salicylic acid or jasmonic acid signaling pathways. This study provides new insight into the role of bacteria VOCs as initiators of defense responses in plants.

One hundred isolates of Phytophthora infestans collected from 10 provinces in China between 1998 and 2004 were analyzed for mating type, metalaxyl resistance, mitochondrial DNA (mtDNA) haplotype, allozyme genotype, and restriction fragment length polymorphism (RFLP) with the RG-57 probe. In addition, herbarium samples collected in China, Russia, Australia, and other Asian countries were also typed for mtDNA haplotype. The Ia haplotype was found during the first outbreaks of the disease in China (1938 and 1940), Japan (1901, 1930, and 1931), India (1913), Peninsular Malaysia (1950), Nepal (1954), The Philippines (1910), Australia (1917), Russia (1917), and Latvia (1935). In contrast, the Ib haplotype was found after 1950 in China on both potato and tomato (1952, 1954, 1956, and 1982) and in India (1968 and 1974). Another migration of a genotype found in Siberia called SIB-1 (Glucose-6-phosphate isomerase [Gpi] 100/100, Peptidase [Pep] 100/100, IIa mtDNA haplotype) was identified using RFLP fingerprints among 72% of the isolates and was widely distributed in the north and south of China and has also been reported in Japan. A new genotype named CN-11 (Gpi 100/111, Pep 100/100, IIb mtDNA haplotype), found only in the south of China, and two additional genotypes (Gpi 100/100, Pep 100/100, Ia mtDNA haplotype) named CN-9 and CN-10 were identified. There were more diverse genotypes among isolates from Yunnan province than elsewhere. The SIB-1 (IIa) genotype is identical to those from Siberia, suggesting later migration of this genotype from either Russia or Japan into China. The widespread predominance of SIB-1 suggests that this genotype has enhanced fitness compared with other genotypes found. Movement of the pathogen into China via infected seed from several sources most likely accounts for the distribution of pathogen genotypes observed. MtDNA haplotype evidence and RFLP data suggest multiple migrations of the pathogen into China after the initial introduction of the Ia haplotype in the 1930s.

Ellipsis and co-reference are common and ubiquitous especially in multi-turn dialogues. In this paper, we treat the resolution of ellipsis and co-reference in dialogue as a problem of generating omitted or referred expressions from the dialogue context. We therefore propose a unified end-to-end Generative Ellipsis and CO-reference Resolution model (GECOR) in the context of dialogue. The model can generate a new pragmatically complete user utterance by alternating the generation and copy mode for each user utterance. A multi-task learning framework is further proposed to integrate the GECOR into an end-to-end task-oriented dialogue. In order to train both the GECOR and the multitask learning framework, we manually construct a new dataset on the basis of the public dataset CamRest676 with both ellipsis and co-reference annotation. On this dataset, intrinsic evaluations on the resolution of ellipsis and co-reference show that the GECOR model significantly outperforms the sequenceto-sequence (seq2seq) baseline model in terms of EM, BLEU and F 1 while extrinsic evaluations on the downstream dialogue task demonstrate that our multi-task learning framework with GECOR achieves a higher success rate of task completion than TSCP, a state-of-theart end-to-end task-oriented dialogue model .

No abstract