Colonization of C. michiganensis and symptoms developed in tomato cotyledons
No symptoms were detected within one day post inoculation (dpi). At 3 dpi, small yellow spots and bumpy surfaces were detected. At 6 dpi, severe disease symptoms, including chlorosis and imbibition, were observed all over the cotyledons (Fig. 1a). The bacterial biomass in the cotyledons after infection was measured by quantitative polymerase chain reaction (qPCR) analysis (Fig. 1b). Because the amount of bacteria inoculated was small, C. michiganensis DNA was detected at low levels from the cotyledons immediately after inoculation. At 1 dpi, C. michiganensis DNA was detected from all samples tested despite no obvious symptoms. From 3 to 6 dpi, the bacterial biomass increased dramatically with the spread of disease symptoms.
Transcriptome profiling of tomato in response to infection by C. michiganensis
The analysis generated 21.8 to 28.5 million raw reads for each sample, and 97.3–98.5% of the obtained reads were properly mapped to the S. lycopersicon reference genome (SL4.0) and International Tomato Annotation Group S. lycopersicon gene annotation (ITAG4.0) (https://solgenomics.net/organism/Solanum_lycopersicum/genome), which contained 34,075 annotated genes (Table S1). Through differentially expressed gene (DEG) analysis, 1,788 and 540 genes were found to be upregulated or downregulated DEGs after C. michiganensis inoculation, respectively (Table S2 and S3).
The expression pattern of the upregulated DEGs was divided into five clusters by hierarchical clustering based on the time-series TPM values. Most of the upregulated DEGs (1,646 genes) were specified as cluster 1, in which expression levels dramatically increased at 6 dpi (Fig. 1c, d). Among the remainder, 86, 8, 43, and 5 genes were assigned to cluster 2, 3, 4, and 5, respectively.
In the upregulated DEGs, 46 Gene Ontology (GO) terms were over-represented (Table S4). These included immune-related GO terms, such as defense response to fungus (GO:0050832), plant-type hypersensitive response (GO:0009626), response to biotic stimulus (GO:0009607), and regulation of systemic acquired resistance (GO:0010112). In addition, GO terms associated with signaling pathways, such as the hormone-mediated signaling pathway (GO:0009755), regulation of SA biosynthetic process (GO:0080142), regulation of SA biosynthetic process (GO:0080142), calcium signaling (GO:0009931), redox regulation (GO:0006749), and protein phosphorylation (GO:0006468), were over-represented. Moreover, GO terms associated with the extracellular (GO:0005615) and cell surface receptor signaling pathway (GO:0007166) were over-represented. In the downregulated DEGs, 15 GO terms were over-represented (Table S4), which included GO terms associated with photosystems (GO:0015979), response to high light intensity (GO:0009644), and response to light stimulus (GO:0009416).
Expression of PR genes after infection with C. michiganensis in tomato
After C. michiganensis infection of tomato cotyledons, the expression of 40 PR genes belonging to six classes was induced (Table S2). The gene-set hypergeometric enrichment test demonstrated that the PR gene homologs were significantly (p-value < 0.05) over-represented in the upregulated DEGs. As shown in Fig. 2, quantitative reverse transcription-PCR (qRT-PCR) analysis confirmed the expression data of PR genes by RNA-seq analysis.
Expression of the genes involved in defense signaling after infection with C. michiganensis in tomato
After C. michiganensis infection, 186 RGAs, including 24 NBSs, 25 RLPs, 15 TM-CCs, and 122 RLKs, were transcriptionally upregulated (Table 1). NBSs, RLPs, TM-CCs, and RLKs were significantly (p-value < 0.05) over-represented in upregulated DEGs. Based on grouping by Sakamoto et al. (2012) [22], C. michiganensis-responsive RLK genes were categorized into various groups, such as LRR, RLCK, and LysM. Upregulated DEGs included SlSERK3s, TFT1/TARK1, SOBIR/EVR, and SlLYKs [11, 13, 15]. Upregulated DEGs also included the receptor-like cytosolic kinase (RLCK) gene ACIK1 [38] and two FER-like genes; FER4 and FER10 [39].
Overall, 90 TF genes in PlantTFDB 5.0 [40] were upregulated after infection, of which 22 WRKYs, 14 NACs, and 5 HSFs were significantly (p-value < 0.05) over-represented (Table 1). The 22 C. michiganensis-responsive WRKYs consisted of all six groups, Group I, II-a, II-b, II-c, II-d, and III [41, 42]. When applied to the phylogenetic classification of Jensen et al. (2010) [43], the 14 C. michiganensis-responsive NAC genes were classified into the seven following groups: I, II, III, VI, V, IV, and IX. They included the four NAC genes encoding SlNACMTF3, 8, 11, and 12, which each have a membrane binding domain [44]. Furthermore, five of the 12 genes encoding the CBP60-like TF, which is not included in PlantTFDB 5.0, were significantly (p-value < 0.05) over-represented in the upregulated DEGs.
Increase of SA levels and the expression of the SA-associated genes after infection with C. michiganensis in tomato
As described above, SA-associated GO terms, including regulation of SA biosynthetic process (GO:0080142) and regulation of systemic acquired resistance (GO:0010112), were over-represented in the upregulated DEGs (Table S4). The expressions of Solyc06g071280, Solyc10g054100, and Solyc02g032850, which are tomato orthologous genes for Arabidopsis EDS1 [45], EDS5/SID1 [46], and PAD4/EDS9, [47] respectively, were induced after infection (Table S2). We named these three genes SlEDS1, SlEDS5, and SlPAD4, respectively, and their expression after infection with C. michiganensis was validated by qRT-PCR (Fig. 3a).
As shown in Fig. 3b, SA levels in tomato cotyledons were lower than 40 ng/gfw under normal conditions and were not changed by buffer treatment. C. michiganensis infection increased SA levels in cotyledons within 1 dpi. SA levels showed a similar pattern to the expression of defense-associated genes, increasing substantially on 6 dpi to approximately 370 ng/gfw. The SA levels in C. michiganensis-inoculated cotyledons were significantly higher than those in the buffer-treated cotyledons at 1, 3, and 6 dpi (p < 0.005, t-test). On the other hand, the levels of JA in the cotyledons were below the detection limit regardless of the presence or absence of C. michiganensis infection (data not shown).
We attempted to identify candidate genes involved in the regulation of SA levels from upregulated DEGs. In the PAL pathway, one of the candidate SA synthesis pathways, four PAL genes (SlPAL2/Solyc09g007900, SlPAL4/Solyc09g007920, SlPAL5/Solyc09g007910, and SlPAL6/Solyc05g056170), three 4CL homologs (Sl4CL/Solyc03g117870, Solyc06g068650, and Solyc12g042460), and three AIM1 homologs (Solyc07g019670, Solyc12g007170, and Solyc08g068390) were identified in the upregulated DEGs (Fig. 3c and Table S2). Whereas there were no enzyme genes of the ICS pathway in the upregulated DEGs (Table S2).
Effect of SA on the colonization of C. michiganensis and defense-associated genes in tomato
Disease symptoms in seedlings treated with SA were less severe than those of seedlings without SA (data not shown). qRT-PCR analysis revealed that the bacterial biomass in SA-treated cotyledons was significantly (p-value < 0.01) smaller than that in non-treated cotyledons (Fig. 4a). To examine the effect of SA on the immunity of tomato plants, qRT-PCR analysis was conducted to determine the expression of disease-associated genes of upregulated DEGs. Interestingly, the expressions of four WRKY genes; SlWRKY45, SlWRKY51, SlWRKY80, and SlWRKY81, were significantly (p-value < 0.01) upregulated by SA treatment (Fig. 4b).