Coronatine is more potent than jasmonates in regulating Arabidopsis circadian clock

Recent studies establish a crucial role of the circadian clock in regulating plant defense against pathogens. Whether pathogens modulate host circadian clock as a potential strategy to suppress host innate immunity is not well understood. Coronatine is a toxin produced by the bacterial pathogen Pseudomonas syringae that is known to counteract Arabidopsis defense through mimicking defense signaling molecules, jasmonates (JAs). We report here that COR preferentially suppresses expression of clock-related genes in high throughput gene expression studies, compared with the plant-derived JA molecule methyl jasmonate (MJ). COR treatment dampens the amplitude and lengthens the period of all four reporters tested while MJ and another JA agonist JA-isoleucine (JA-Ile) only affect some reporters. COR, MJ, and JA-Ile act through the canonical JA receptor COI1 in clock regulation. These data support a stronger role of the pathogen-derived molecule COR than plant-derived JA molecules in regulating Arabidopsis clock. Further study shall reveal mechanisms underlying COR regulation of host circadian clock.


Results
COR is a phytotoxin produced by pathovars of P. syringae and is important for the pathogenesis of the bacteria. The lack of COR makes P. syringae less virulent under a diurnal light and dark (LD) cycle 15,16 . In continuous light (LL), a free-running condition often used to test clock activity, we found that compared with P. syringae strain DC3000, the isogenic P. syringae strain DC3118 that does not produce COR, grew much less and induced less chlorosis and lesion in the infected Arabidopsis leaves ( Figure S1). One way that COR promotes bacterial virulence is through interfering with JA signaling. A number of studies demonstrated crosstalk between JA signaling and the circadian clock [3][4][5]17 . We are interested in elucidating in this report whether P. syringae-derived COR can regulate plant circadian clock. Toward this goal, we first compared expression of a set of circadian genes (Table S1), using two sets of time-series RNA-seq data from samples treated with 100 µM MJ or 5 µM COR 18,19 . Heatmap analysis showed that COR induced more changes in expression of the circadian genes than MJ (Fig. 1A,B). We estimated the relative number of affected gene expression by normalizing the total number of affected expression events with the number of time points. While the number of  www.nature.com/scientificreports/ induced circadian genes was similar with the two treatments, COR showed a stronger suppression of circadian genes than MJ, including some core clock genes ( Fig. 1C; Figure S2). We performed a similar analysis with a set of defense genes (Table S2). Interestingly, MJ and COR showed less difference in affecting defense gene expression ( Fig. 1C; Figure S3). While these analyses suggest that COR exerts a stronger suppression of circadian genes than MJ, we recognized that the two RNA-seq experiments were conducted in two different laboratories that used plants grown in different conditions 18,19 . In addition, both experiments were conducted under diurnal cycles. To reconcile these differences and examine the role of COR in clock regulation, we grew seedlings for 7 days under a 12 h light and 12 h dark (LD) cycle for entrainment. We then transferred the seedlings to continuous light (LL) for 1 day and treated them with MJ (100 μM) or COR (10 μM) for gene expression analysis. qRT-PCR results showed that both MJ and COR suppressed expression of selected core clock genes ( Figure S4 and 5 ), supporting that both MJ and COR regulate clock activity.
To further test the clock regulatory role of COR, we performed the luciferase (LUC) assay with plants expressing the LUC gene driven by promoters of different clock genes, including CCA1, TOC1, PRR7, and GRP7. Similar to MJ, we found that COR suppressed seedling growth in our clock assay condition ( Figure S5 and 5 ). After normalizing the LUC amplitude to relative leaf area of seedlings, we observed that COR dampened the amplitude and lengthened the period of all four reporters largely in a dosage dependent manner, regardless COR was applied at 25 or 37 h after light onset (subjective dawn or subjective dusk, respectively) ( Fig. 2). COR further induced a lagging phase with the TOC1:LUC and GRP7:LUC reporters ( Fig. 2D2 and 2D4), suggesting a higher sensitivity of these two reporters to COR than other reporters tested. To test if this clock regulatory role of COR requires intact JA signaling, we used the JA receptor mutant (coi1-17) expressing CCA1:LUC 5,20 . COR did not affect seedling growth and rhythmicity of the CCA1:LUC reporter in coi1-17 ( Fig. 2A5, 2B5, 2C5, 2D5; Figure S5A). Thus, these results support that the role of COR in regulating clock activity requires a functional JA receptor.
Like COR, MJ also affects seedling growth ( Figure S5B). We previously showed that MJ only affects the amplitude but not the period and phase of the CCA1:LUC reporter 5 . We report here that three additional reporters (TOC1:LUC, PRR7:LUC, and GRP7:LUC) showed an amplitude dampening in the presence of MJ (Fig. 3). The PRR7:LUC and GRP7:LUC reporters also displayed period lengthening, depending on MJ dosages. Furthermore, MJ induced phase lagging in TOC1:LUC and GRP7:LUC. Unlike COR, MJ did not affect the period of CCA1:LUC and TOC1:LUC. These results suggest a stronger effect of COR than MJ in regulating clock activity, at least for some clock genes. They also illustrate differential sensitivity of different clock reporters to COR, MJ, and JA-Ile.
We previously reported that another plant-derived jasmonate, JA-Ile, acts through COI1 to suppress the amplitude, lengthen the period, but not affect the phase of CCA1:LUC and GRP7:LUC reporters in Col-0 5 . Seedling growth was not affected by JA-Ile ( Figure S5C and 5 ). We confirmed these results with the PRR7:LUC reporter ( Figure S5C and S6). Interestingly, the TOC1:LUC reporter showed less sensitivity to JA-Ile than other reporters tested, only showing a dampened amplitude but no change in the period and the phase. These results support a stronger role of COR than JA-Ile in clock regulation and differential sensitivity of clock reporters to COR and plant JA derivatives.

Discussion
Growing evidence indicates that pathogens can reprogram the circadian clock of the host. For instance, the bacterium P. syringae, the oomycete Hyaloperonospora arabidopsidis, and the fungus Botrytis cinerea were shown to manipulate the circadian clock of Arabidopsis 9,21-23 . Even gut microbiota in the animal host are capable of reprograming their animal host clock 24,25 . The key question remains how pathogens affect host clock activity and defense responses. Pathogens are known to secrete a vast range of molecules to interfere with host immunity. Studies just begin to reveal that some signals emanating from pathogens modulate the circadian clock of the host. Pathogen associated molecular patterns (PAMPs), including bacteria lipopolysaccharide (LPS) and flg22, were shown to affect the circadian system of animals and Arabidopsis, respectively 21,26 . We report here that the P. syringae-produced toxin molecule COR exerts a stronger influence on Arabidopsis clock than some plantderived JA molecules. Our conclusion is strongly supported by experimental evidence. First, large-scale gene expression analysis showed a stronger suppression of circadian genes by COR than by MJ ( Fig. 1; Figures S2 and  S3). Second, luciferase assays using marker gene promoters fusing to the luciferase reporter showed stronger effect of COR than JA-Ile and MJ in regulating clock activity (Figs. 2, 3; Figure S6). Third, COR also exerted a stronger effect than MJ and JA-Ile on seedling growth in LL ( Figure S5). And finally, we found that COR is critical for pathogen virulence in LL ( Figure S1). These various biological processes impacted by COR, JA-Ile, and/ or MJ are all regulated by the circadian clock.
Such a stronger role of COR in clock regulated events than that of JA-Ile and MJ is consistent with previous studies that show more potent effect of COR than some JA molecules on other physiological processes 14,27,28 . It is possible that pathogens use COR through a specific mechanism(s) to hyperactivate the JA signaling. Indeed, COR was shown to bind with a higher affinity to the JA receptor COI1 than plant-derived JA molecules 29,30 . Downstream of COI1, the JA signaling is highly modular; both the JA signaling repressors (JAZ proteins) and activators (MYC proteins) belong to protein families, members of which interact with different proteins to influence multiple biological processes 31 . Therefore, it is possible that the COR-COI1 complex could selectively target some JAZ proteins for degradation, leading to a stronger or differential impact on MYC proteins and other signaling targets, such as the circadian clock. In addition to a differential perception of COR and JA molecules that could cause differences in regulating the circadian clock and other biological processes, the different efficacy between COR and other JA molecules could also be due to the solubility, uptake efficiency, stability, and catabolism of each compound in plants. www.nature.com/scientificreports/ Our data further demonstrate that the four clock reporters used in this study showed different responses to COR, MJ, and JA-Ile treatments. Such a differential response of clock reporter genes to external treatments has been reported previously in response to nutrient status, ROS, phytohormones, temperature, and photoperiod [6][7][8][32][33][34][35] . The differential response of these reporters may reflect tissue specific gene expression that allows differential clock response to the chemicals in separate tissues 36 . Alternatively, there may be different clocks functioning simultaneously with different rhythms in the same tissue or even in the same cell 37 . Together they support the plasticity of the circadian clock that may create flexibility for plants to respond to various external stimuli 38 .

Scientific RepoRtS
Manipulation of host circadian clock may represent a common strategy of microbes to suppress host immunity. How pathogen-produced specific molecules modulate host clock activity and defense responses still remains largely unknown. Our finding of the role of the pathogen-derived molecule COR in modulating Arabidopsis clock opens a new and exciting research direction to elucidate the molecular mechanisms underlying clock-defense interplay during host-pathogen interactions. Our data also illustrate the circadian clock being decentralized, which likely allows organisms to adapt to the changing environment in the presence of pathogens and other biotic and abiotic stresses. RnA-seq analysis. Two sets of high-resolution RNA-seq data from 100 µM MJ or 5 µM COR treated samples were used for gene expression analysis 18,19 . The log 2 transformed fold changes of expression of circadian genes (Table S1) and defense genes (Table S2) were used to generate the heatmap using the heatmap.2 function in R package gplots. The circadian genes were annotated to be related to rhythmic processes according to Arabidopsis Information Resources (Table S1) and the defense genes were reported previously 21 (Table S2).

Methods
qRt-pcR analysis. RNA extraction and qRT-PCR were performed as previously described 39,40 . Primers used in qRT-PCR are listed in Table S3.  15,16,18,19,27,29,[41][42][43] ) and our preliminary experiments to test the concentrations for each chemical that induced changes of clock activity in a dosage dependent manner but did not cause overstress in plants. MJ and JA-Ile treatments with 10 µM and 100 µM demonstrated dosage-dependent phenotypes, including clock activity and seeding growth. But 100 µM coronatine drastically stunted plant growth and induced high anthocyanin production, suggesting plants under extreme stress. Thus, we used 1 and 10 µM for COR in this report.
Immediately after the treatments, the plants were measured for luminescence with an Omega Luminescence Reader (BMG LABTECH, Inc.) in LL with 90 µmol m −2 s −1 photon flux density. LUC activity was measured at 1-h intervals for 5 days and analyzed for amplitude, period, and phase with the R package MetaCycle 44 . All luciferase assay experiments were repeated three times with similar results.