**3. Discussion**

BSR1, a RLCK-VII member, has a protein kinase activity and is important for the initiation of defense responses against chitin oligomers, known as a fungus-derived MAMP [39,40]. BSR1 is implicated in resistance to bacteria, as well as fungi, in wild-type and overexpressing rice lines [36,38,40], suggesting that BSR1 is also involved in responses triggered by bacterium-derived MAMPs. In this study, our experiments on suspension-cultured rice cells showed a correlation between BSR1 and the response to two bacterial elicitors, peptidoglycan and LPS. Knocking out *BSR1* significantly suppressed the elicitation of oxidative bursts and the transcript levels of defense-related genes caused by these bacterial MAMPs (Figures 1 and 2). In some experiments, variations of the H2O2 concentration and the transcript levels were observed among knockout lines (Figures 1 and 2). Since the variations

among knockout lines were not reproducible, they were considered as the influence of the experimental manipulation. The suppression of immune responses by *BSR1* knockout is in accordance with the significant contribution of BSR1 to chitin-induced responses [39]. Rice recognizes chitin through receptor complexes containing OsCERK1 and CEBiP [6]. OsCERK1, but not CEBiP, possesses a protein kinase activity to phosphorylate cytoplasmic signaling factors [42]. The perception of peptidoglycan and LPS was mostly mediated by OsCERK1 complex [4,7,8]. Thus, in response to peptidoglycan and LPS exposure, as well as chitin, OsCERK1 would transmit a signal directly or indirectly to BSR1 to regulate its protein kinase activity (Figure 6).

**Figure 6.** Proposed model in which BSR1 regulates defense responses, such as oxidative bursts, after the perception of MAMPs in wild-type (WT; **left**) and BSR1-overexpressing rice lines (BSR1-OX; **right**). PGN, peptidoglycan; LPS, lipopolysaccharide; RLPs, receptor-like proteins; ROS, reactive oxygen species; MTI, MAMP-triggered immunity.

Knocking out *BSR1* did not make rice cells nonresponsive to MAMPs (Figures 1 and 2), indicating the existence of functionally redundant factor(s) for BSR1. In *A. thaliana*, RLCK-VII members function in MAMP-induced defense responses with a robust functional redundancy [43]. The participation of other rice RLCK-VII members in PTI have been studied [27]. OsRLCK57, OsRLCK107, OsRLCK118, OsRLCK176, and OsRLCK185 positively regulate chitin- and peptidoglycan-induced responses [7,32,42]. No rice RLCKs, except for BSR1, have been reported to mediate LPS-induced oxidative bursts. However, known interactors of the LPS-(co)receptor OsCERK1, such as OsRLCK176 and OsRLCK185, may mediate LPS-signaling. To take into consideration of functional redundancy for BSR1, these RLCK-VII members could act downstream of LPS as well as peptidoglycan.

We compared BSR1-HPB:OX lines with GUS-HPB:OX control line to assess the effects of BSR1 overexpression on MAMP-triggered responses. In the absence of MAMPs, H2O2 levels in cell cultures and leaf strips derived from BSR1-HPB:OX lines were slightly greater than those of the control (Figures 3–5). Where this H2O2 originates from is unknown. Under peptidoglycan-, LPS-, and chitin-treated conditions, BSR1-HPB:OX suspension-cultured cells produced a greater amount of H2O2 than control cells (Figures 3 and 4; Supplementary Materials Figures S2 and S3), while the overexpression of BSR1 facilitated the transcriptional activation of *PAL1* but not *PBZ1* and *KSL4* (Figures 3b and 4b). There were variations in transcript levels of *PBZ1* and *KSL4* between two BSR1 overexpression lines, BSR1-HPB:OX#17 and BSR1-HPB:OX#39. Since transcript levels of defense-marker gene *PBZ1* in MAMP-treated and untreated BSR1-HPB:OX#39 were even lower than those in the control line, the line may have contained mutations which decrease the transcript levels of these defense-related genes. Alternatively, the condition to culture BSR1-HPB:OX#39 line may have given stress to the cells, resulting in slight increase in the expression of internal control *RUBQ1*, which encodes protein turnover

factor. That is because the environment surrounding the cells, such as cell concentration, cannot be completely uniformized. In accordance with results using cell culture, BSR1-HPB:OX leaf blade tissues displayed remarkably greater oxidative bursts against MAMPs extracted from autoclaved conidia (Figure 5a). In rice and *Arabidopsis*, OsRLCK118 and *A. thaliana* BIK1, two RLCK-VII members, directly and positively regulate RBOH proteins whose NADPH oxidase activities generate ROS and cause oxidative bursts [30,31,33]. Highly expressed BIK1 leads to enhanced ROS production in response to MAMP [44]. Excess BSR1 protein also could hyperactivate RBOHs, directly or indirectly, resulting in the enhancement of oxidative bursts. Recent work showed that *BIK1* overexpression does not enhance fungal disease resistance, although deletions of negative regulators for PTI signaling result in the accumulation of BIK1 and the strong disease resistance in *A. thaliana* [45]. Unlike BIK1, BSR1 overexpression confers the robust disease resistance in rice and *A. thaliana* [36], indicating that functions of BSR1 would be quite different from those of BIK1.

A time course of the H2O2 levels under co-cultivation conditions revealed how the overexpression of BSR1 acts during the early phase of host–microbe interactions. Under our co-cultivation conditions, suspensions of living conidia of the blast fungus did not elicit host-derived H2O2 production in control leaf strips (Figure 5b; Supplementary Materials Figure S5b). These results reconfirmed previous reports that the supernatants of conidial suspensions contain ROS-degrading activities that mostly depend on CPXB, a catalase-peroxidase secreted by the blast fungus [19]. Under the same co-cultivation conditions, the overexpression of BSR1-HPB resulted in leaf blade tissues producing detectable amounts of H2O2 (Figure 5), indicating that oxidative bursts in BSR1-overexpressing plants are intense enough to overcome the inhibition caused by the infecting blast fungus. The enhanced responses against peptidoglycan and LPS, as well as chitin resulting from the overexpression (Figures 3 and 4; Supplementary Materials Figure S3) strongly suggested that pathogenic bacterial challenges would elicit the same responses.

In host–microbe interactions, host-derived ROS is regarded as an antimicrobial substance and a diffusible second messenger that contributes to immunity [11,16]. Indeed, host-derived ROS detoxification should be required for pathogenicity. For example, ROS-degrading activities are present in the supernatants of *P. oryzae* conidial suspensions and contribute to lesion formation when exogenously added [20]. The deletion of *P. oryzae DES1*, which is required for extracellular peroxidase activity, causes the accumulation of host-derived ROS and the induction of defense-related genes, resulting in non-pathogenicity [23]. Pathogens could not completely abolish overproduced ROS (Figure 5b), and therefore do not show full virulence in plants overexpressing BSR1 (Figure 6). Our data support the model which host-derived ROS is critical for plant interactions with pathogens.

In conclusion, we propose that the broad-spectrum disease resistance could be achieved by the enhancement of MAMP-triggered oxidative bursts and following transcriptional activation. To date, many RLCKs have been characterized as signaling factors in PTI [27]. However, no RLCK, other than *BSR1*, can enhance oxidative bursts and disease resistance when overexpressed in rice. It is unclear what allows the functional enhancement. This report clearly showed that hyperactivated MAMP-triggered immune responses could be used for broad-spectrum disease control.
