Agromonas: a rapid disease assay for Pseudomonas syringae growth in agroinfiltrated leaves

Significance Statement This technical advance introduces a robust and reliable disease assay to monitor bacterial growth of Pseudomonas syringae in agroinfiltrated tissues using an antibiotics cocktail to select against Agrobacterium tumefaciens.


INTRODUCTION
Understanding the plant immune system and microbial pathogenicity is essential to improve plant biotechnologies and crop protection. To evaluate the level of resistance of a plant or the virulence of bacterial pathogens, the routine method relies on infection assays that quantify bacterial growth (i.e. colony count assays). Colony count assays are usually performed on stable transformant plants. However, generation of stable transgenic lines is time and resource consuming, and is limited to plant species that are amenable to genetic transformation. Therefore, there is a need for faster disease assays particularly in the studies of the model plant pathogen Pseudomonas syringae, which causes important economic damages in many plant species (Mansfield et al., 2012).
Rapid overexpression and transcript depletion of various exogenous and endogenous genes is facilitated by Agrobacterium tumefaciens-mediated transient expression (agroinfiltration). Agroinfiltration is used throughout plant science to study protein localisation and for their biochemical characterisation. Agroinfiltrated leaves are routinely used to study the interaction between the model plant Nicotiana benthamiana and the potato blight pathogen Phytophthora infestans (Chaparro-Garcia et al., 2011;Bozkurt et al., 2014;Dagdas et al., 2018) and other Phytophthora species. However, agroinfiltrated N. benthamiana leaves are not routinely used for disease assays with P. syringae. One problem is that selective isolation of P. syringae from agroinfiltrated tissue is challenging because there is an overlap of endogenous or introduced antibiotics resistance. For example, rifampicin is commonly used to select for antibiotic-resistance genes in the genome of A. tumefaciens and P. syringae strains, whereas resistance to kanamycin is frequently used to maintain plasmids in both bacteria and therefore these antibiotics are not useful for selective isolation in co-inoculated tissue.
A solution to this problem is to use selection for endogenous bacterial resistance to antibiotics. The combination of 10 µg ml À1 cetrimide, 10 µg ml À1 fucidin and 50 µg ml À1 cephaloridine (CFC; Figure 1a) permits the selection of Pseudomonas species (Mead and Adams, 1977). CFC is used for the selective isolation of Pseudomonas species during the microbiological examination of environmental, clinical, food and plant samples (Krueger and Sheikh, 1987;Hill et al., 2005;Pantazi et al., 2008;Fones et al., 2010;Straub et al., 2018), but has not yet been exploited in combination with agroinfiltrated leaves.
Here, we use CFC selection to establish a rapid and easy disease assay to quantify growth of P. syringae from agroinfiltrated leaves. We tested whether immunity can be studied in agroinfiltrated leaves even though these leaves contain A. tumefaciens. The method we developed is called 'agromonas' because it is based on agroinfiltration followed by inoculation of Pseudomonas syringae by both infiltration and spray inoculation. We demonstrate that the agromonas assay can be applied to different P. syringae strains, and demonstrate its adaptability to study the impact of immune components and bacterial effectors on P. syringae growth in planta.

CFC facilitates Pseudomonas syringae selection from agroinfiltrated tissues
To confirm that CFC facilitates P. syringae selection, we tested three different P. syringae strains that are pathogenic on N. benthamiana. We tested P. syringae pv. tomato DC3000, the causative agent of the bacterial speck disease of tomato lacking the type III effector gene hopQ1-1 [PtoDC3000(ΔhQ)]; P. syringae pv. tabaci 6605 (Pta6605), the causative agent for wildfire disease in tobacco; and P. syringae pv. syringae B728a (PsyB728a), the causative agent of bacterial brown spot of bean. We also included A. tumefaciens GV3101 (AtumGV3101), the non-oncogenic strain that is routinely used for agroinfiltration. These strains were infiltrated into N. benthamiana leaves and, at 3 days post-infiltration (3 dpi), leaf extracts were generated, diluted in water and plated out on LuriaÀBertani (LB) medium with or without CFC selection. All tested P. syringae pathovars grew equally well on LB medium supplemented with or without CFC (Figure 1b), demonstrating that CFC does not affect P. syringae growth. By contrast, AtumGV3101 did not grow at all on plates containing CFC ( Figure 1b). Likewise, CFC also blocks growth of the nononcogenic AtumC58C1 as well as all tested Xanthomonas strains (Table S1), confirming the selectivity of CFC.
To apply CFC selection to facilitate the selective isolation of P. syringae from agroinfiltrated leaves, N. benthamiana leaves were first infiltrated with AtumGV3101. Two days after agroinfiltration, each of the three different P. syringae strains were infiltrated into the agroinfiltrated regions. Three days later, leaf homogenates were plated onto LB plates containing CFC or gentamicin ( Figure 2a). While P. syringae strains were specifically isolated on CFC plates, AtumGV3101 was isolated on plates containing gentamicin (Figure 2b,c). No colonies with Agrobacterium morphology were detected on CFC plates (Figure 2b), demonstrating that A. tumefaciens cannot grow on CFC plates, even when P. syringae is growing. These data demonstrate that CFCcontaining medium facilitates the selection of living Pseudomonas spp. from agroinfiltrated leaves. Comparison with the samples taken immediately upon P. syringae (b) Pseudomonas syringae grows on CFC selection, Agrobacterium does not. Nicotiana benthamiana leaves were infiltrated with 1 9 10 6 CFU ml À1 P. syringae or 1 9 10 8 CFU ml À1 AtumGV3101, and bacterial populations were determined 3 days later using colony count method using LuriaÀBertani (LB) plates containing CFC or not. Error bars represent SE of n = 3 biological replicates. Student's t-test statistics (***P < 0.001). CFU, colony-forming units. (a) Experimental procedure for agromonas assay. Two days after agroinfiltration, agroinfiltrated leaves are infiltrated with Pseudomonas syringae bacteria. Bacterial growth is measured 3 days later (3 days post-infiltration, 3 dpi) by a classic colony count on LuriaÀBertani (LB) agar plates containing cephaloridine, fucidin and cetrimide (CFC). (b) CFC selects P. syringae from agroinfiltrated leaves. Agroinfiltrated leaves were infiltrated with 1 9 10 6 CFU ml À1 PtoDC3000(ΔhQ) and 3 days later leaf extracts were diluted, and each dilution was plated onto medium supplemented with or without CFC. Pictures were taken 48 h later. (c) Selective isolation of Pseudomonas spp. from agroinfiltrated leaves. Agroinfiltrated leaves were infiltrated with 1 9 10 6 CFU ml À1 P. syringae and, at 0 and 3 dpi, leaf extracts were plated on medium containing CFC or gentamicin to select P. syringae or Agrobacterium, respectively. Error bars represent SE of n = 3 biological replicates. infiltration (0 dpi) shows that P. syringae populations grow at least 10-fold in agroinfiltrated tissues ( Figure 2c). While comparing unmixed infection with mixed infection (Figures 1b and 2c), we observed that the presence of A. tumefaciens suppresses P. syringae growth, as previously reported . We compared the growth of PtoDC3000(ΔhQ) in the presence and absence of AtumGV3101. PtoDC3000(ΔhQ) grew ninefold less in N. benthamiana leaves infiltrated with AtumGV3101 as compared with a non-agroinfiltrated sample ( Figure S1), confirming that A. tumefaciens reduces P. syringae growth in planta. Therefore, it is essential to use agroinfiltrated leaves expressing the empty vector (EV) as control in agromonas assays.

PAMP receptors reduce bacterial growth in agromonas assay
To demonstrate that the agromonas assay can be used to study genes that confer immunity, we used two pattern recognition receptors (PRRs) that are absent in N. benthamiana. We tested tomato FLS3 (flagellin-sensing 3) and Arabidopsis EFR (EF-Tu receptor), which recognize the flgII-28 epitope of flagellin and the elf18 epitope of EF-Tu, respectively (Kunze et al., 2004;Zipfel et al., 2006;Cai et al., 2011;Hind et al., 2016).
To confirm the functionality of FLS3 and EFR upon agroinfiltration, we transiently expressed FLS3 and EFR in N. benthamiana leaves and measured the production of reactive oxygen species (ROS) upon treatment with flgII-28 and elf18. Leaves transiently expressing FLS3 were able to release a ROS burst upon flgII-28 treatment, whereas EFR and EV expressing leaves remained unresponsive to flgII-28 (Figure 3a). Likewise, leaves that transiently express EFR were able to release an oxidative burst upon elf18 treatment, whereas FLS3 and EV expressing leaves remained unresponsive to elf18 ( Figure 3a). These results show that FLS3 and EFR are functional in N. benthamiana, consistent with previous studies (Lacombe et al., 2010;Hind et al., 2016).
We next tested whether agroinfiltrated leaves expressing FLS3 have enhanced resistance to P. syringae upon infection. Agroinfiltrated leaves of N. benthamiana plants transiently expressing FLS3 showed reduced bacterial growth of both PtoDC3000(ΔhQ) and Pta6605 strains compared with leaves expressing the EV control ( Figure 3b). Similarly, EFR transient expression caused a strong reduction in the growth of PtoDC3000(ΔhQ), Pta6605 and PsyB728a (Figure 3b). Altogether these data demonstrate that agroinfiltration of FLS3 and EFR increases immunity to P. syringae in N. benthamiana.
We also measured bacterial growth of AtumGV3101 in the same extracts using gentamicin selection. While no effect on AtumGV3101 growth was detected in leaves transiently expressing EV and FLS3 (Figure 3c), transient expression of EFR in N. benthamiana reduced AtumGV3101 growth (Figure 3c), consistent with a previous study using EFR transgenic plants (Lacombe et al., 2010).

Depletion of host immunity gene increases bacterial growth in agromonas assay
To test if we could also promote bacterial growth by depleting a host immune component in agromonas assays, we depleted NbFLS2 by RNAi using hairpin (hp) constructs (Yan et al., 2012). We monitored ROS production after flg22 treatment to confirm NbFLS2 depletion. Agroinfiltration of hpFLS2 had no effect on flg22-induced ROS production 3 days after agroinfiltration, but suppressed the response 10 days after agroinfiltration in contrast to hpGFP (Figure 4a,b), confirming the selective depletion of NbFLS2.
We next inoculated these agroinfiltrated leaves depleted for NbFLS2 with Pta6605 to measure plant immunity to bacteria. NbFLS2 depletion using hpFLS2 resulted in significantly more P. syringae growth compared with the hpGFP Figure 3. PAMP receptors reduce bacterial growth in the agromonas assay. (a) Transient expression of tomato FLS3 and Arabidopsis EFR in Nicotiana benthamiana confers flgII-28 and elf18 responsiveness, respectively. Leaf discs from agroinfiltrated leaves expressing FLS3 (blue), EFR (red) or empty vector (EV; grey) were treated with 100 nM flgII-28 or elf18, and reactive oxygen species (ROS) was measured in relative light units (RLU). Error intervals (shaded regions) represent SE of n = 12 biological replicates. (b) Transient expression of FLS3 or EFR reduces Pseudomonas syringae growth. Two days after agroinfiltration, agroinfiltrated leaves expressing FLS3 (blue), EFR (red) or EV (grey) were spray-inoculated with the indicated strains of P. syringae (at 1 9 10 8 CFU ml À1 ), and bacterial growth was measured 3 days later using cephaloridine, fucidin and cetrimide (CFC) selection. Error bars represent SE of n = 3 biological replicates. Student's t-test statistics (*P < 0.05). (c) Transient expression of EFR, but not FLS3, affect Agrobacterium growth. Bacterial growth of AtumGV3101 was measured by plating the leaf extracts described in (b) on medium containing gentamicin. Error bars represent SE of n = 3 replicates. Student's t-test statistics (*P < 0.05).
© 2020 The Authors.  Figure 4c). These data are consistent with the reported role of NbFLS2 in immunity to P. syringae (Segonzac et al., 2011), demonstrating that depletion with hairpin constructs can be used in agromonas assays to study the role of endogenous immune components. By contrast, AtumGV3101 grew equally well in both hpGFP and hpFLS2 expressing leaves (Figure 4d), consistent with the absence of immunogenic sequences in flagellin of A. tumefaciens that are recognized by NbFLS2 .

Rapid functional analysis of immune components in agromonas assay
To illustrate that the agromonas assay can be used for fast functional analysis (Figure 5a), we generated the nonphosphorylatable mutant EFR Y836F known to be inactive in elf18-induced signalling (Macho et al., 2014). Indeed, agroinfiltrated leaves expressing EFR Y836F were unable to mount an elf18-induced ROS burst, unlike wild-type EFR (Figure 5b), confirming the non-functionality of this EFR Y836F mutant. Consequently, leaves expressing EFR Y836F were more susceptible to PtoDC3000(DhQ), in contrast to leaves expressing wild-type EFR (Figure 5c). Consistent with a role for EFR in conferring resistance to A. tumefaciens, growth of AtumGV3101 was also reduced by the functional EFR receptor, but not in the presence of EFR Y836F (Figure 5d). These data are consistent with the reported crucial role of Y836 of EFR in immunity to P. syringae (Macho et al., 2014), and illustrate that the agromonas assay can be used for quick and robust functional analysis of immune components.

T3 effector increases bacterial growth in agromonas assay
To demonstrate that bacterial growth can be increased in agromonas assays by transient expression of microbial effectors, we tested the type III (T3) effector AvrPto, which is a kinase inhibitor blocking PRRs (Xiang et al., 2008). As expected, expression of AvrPto blocked the ROS burst induced by flgII-28 when co-expressed with FLS3 (Figure 6a), consistent with an earlier study (Hind et al., 2016). In addition, AvrPto expression also blocked the ROS burst induced by flg22 (Figure 6b), consistent with an earlier study (Xing et al., 2007). Consequently, agroinfiltration of AvrPto increased growth of PtoDC3000(DhQ) in leaves transiently expressing FLS3 (Figure 6c), demonstrating that agroinfiltration of pathogenic microbial effector suppresses host defence. By contrast, AtumGV3101 grew equally well on leaves agroinfiltrated with AvrPto (Figure 6d), indicating that AvrPto does not affect A. tumefaciens growth by blocking PRRs. (a) Experimental procedure for studying the role of endogenous immune components in agromonas assay. Ten days after agroinfiltration, agroinfiltrated leaves are spray-inoculated with P. syringae bacteria. Bacterial growth is measured 3 days later by a classic colony count on LuriaÀBertani (LB) agar plates containing cephaloridine, fucidin and cetrimide (CFC). (b) FLS2 depletion reduces reactive oxygen species (ROS) production upon flg22 treatment. Leaves were agroinfiltrated (OD 600 = 0.2) with hpGFP (grey) or hpFLS2 (blue) and, at 3 and 10 dpi, leaf discs were treated with 100 nM flg22. Error intervals represent SE of n = 12 replicates. (c) FLS2 depletion increases P. syringae growth. Agroinfiltrated leaves expressing hpGFP (grey) or hpFLS2 (blue) were spray-inoculated at 10 dpi with 1 9 10 8 CFU ml À1 Pta6605 and bacterial growth was measured 3 days later using CFC selection. Error bars represent SE of n = 3 biological replicates. Student's t-test statistics (**P < 0.01). (d) FLS2 depletion does not affect Agrobacterium growth. Bacterial growth of AtumGV3101 was measured by plating the leaf extracts described in (c) on medium containing gentamicin. Error bars represent SE of n = 3 replicates. Student's t-test statistics. (b) Phosphomutant EFR Y836F is unable to trigger reactive oxygen species (ROS) burst upon elf18 treatment. Leaves were agroinfiltrated with EFR (red), EFR Y836F (light red) or empty vector (EV; grey), and the ROS burst was measured at 3 dpi in leaf discs treated with 100 nM elf18. Error intervals represent SE of n = 12 replicates. (c) Phosphomutant EFR Y836F is blocked in elf18-triggered immunity. Two days after agroinfiltration, agroinfiltrated leaves expressing EFR, EFR Y836F or EV were spray-inoculated with 1 9 10 8 CFU ml À1 PtoDC3000(ΔhQ) and bacterial growth was measured 3 days later using cephaloridine, fucidin and cetrimide (CFC) selection. Error bars represent SE of n = 3 biological replicates. Student's t-test statistics (*P < 0.05). (d) Agroinfiltration of EFR, but not EFR Y836F , reduces Agrobacterium growth. Bacterial growth of AtumGV3101 was measured by plating the leaf extracts described in (c) on medium containing gentamicin. Error bars represent SE of n = 3 biological replicates. Student's t-test statistics (**P < 0.01). Secreted immune hydrolases reduce bacterial growth in agromonas assay We recently discovered that the b-galactosidase NbBGAL1 contributes to FLS2-mediated immunity by initiating the hydrolytic release of flagellin elicitors, presumably by removing the terminal glycan from the flagellin polymer (Buscaill et al., 2019). To test whether immunity triggered by NbBGAL1 can be detected in the agromonas assay, we agroinfiltrated NbBGAL1. Consistent with previous studies (Buscaill et al., 2019;Kriechbaum et al., 2020), apoplastic fluids from leaves of N. benthamiana bgal1 mutant plants transiently overexpressing NbBGAL1 had strong b-galactosidase activity as NbBGAL1 can cleave galactose from FDG (fluorescein di-b-D-galactopyranoside) and no such activity was detected in the EV control (Figure 7a). These EV and NbBGAL1 expressing leaves were spray-inoculated with Pta6605, which carries BGAL1-sensitive glycans. Leaves overexpressing NbBGAL1 had reduced bacterial growth as compared with EV control leaves (Figure 7b), demonstrating that transient expression of NbBGAL1 in N.
benthamiana increases resistance to Pta6605. We also inoculated agroinfiltrated leaves with PsyB728a, which carries BGAL1-insensitive glycans. Bacterial growth of PsyB728a was not affected by NbBGAL1 when compared with EV expressing leaves (Figure 7c), consistent with the fact that NbBGAL1 acts in immunity only against strains carrying sensitive glycans. AtumGV3101 grew equally well on both plants agroinfiltrated with NbBGAL1 and EV (Figure 7d), indicating that NbBGAL1 does not affect Agrobacterium growth, consistent with the absence of immunogenic flagellin peptides triggering NbFLS2 .
Finally, we tested whether the NbBGAL1 orthologue in Arabidopsis, AtBGAL8, can provide immunity to strains Figure 6. T3 effector suppresses immunity in the agromonas assay. (a,b) Expression of AvrPto blocks reactive oxygen species (ROS) production upon flgII-28 and flg22 treatment. Leaf discs from agroinfiltrated leaves expressing FLS3 with empty vector (EV; blue), FLS3 with AvrPto (red) or EV alone (grey) were treated with 100 nM flgII-28 (a) or flg22 (b) and the ROS burst was measured in RLU. Error intervals represent SE of n = 12 replicates.
(c) Agroinfiltration of AvrPto increases Pseudomonas syringae growth in Nicotiana benthamiana and suppresses FLS3-mediated immunity. Two days after agroinfiltration, agroinfiltrated leaves expressing FLS3 in combination with either AvrPto (red) or EV (blue) were spray-inoculated with 1 9 10 8 CFU ml À1 PtoDC3000(ΔhQ) and bacterial growth was measured 3 days later using cephaloridine, fucidin and cetrimide (CFC) selection. Error bars represent SE of n = 3 biological replicates. Student's t-test statistics (**P < 0.01). (d) AvrPto does not affect growth of AtumGV3101. Bacterial growth of AtumGV3101 was measured by plating the leaf extracts described in (c) on medium containing gentamicin. Error bars represent SE of n = 3 biological replicates. Student's t-test statistics. (b) Agroinfiltration of NbBGAL1 and AtBGAL8 reduce Pta6605 growth. Two days after agroinfiltration, agroinfiltrated leaves expressing NbBGAL1 or AtBGAL8 were spray-inoculated with 1 9 10 8 CFU ml À1 Pta6605 and bacterial growth was measured 3 days later using cephaloridine, fucidin and cetrimide (CFC) selection. Error bars represent SE of n = 6 biological replicates; t-test P-values (*P < 0.05). (c) NbBGAL1 or AtBGAL8 do not reduce PsyB728a growth. Two days after agroinfiltration, agroinfiltrated leaves expressing NbBGAL1 or AtBGAL8 were spray-inoculated with 1 9 10 8 CFU ml À1 PsyB728a and bacterial growth was measured 3 days later using CFC selection. Error bars represent SE of n = 3 biological replicates; t-test P-values. carrying NbBGAL1-sensitive glycans. NbBGAL1 and AtB-GAL8 share 72% amino acid identity, and AtBGAL8 carries the catalytic residues ( Figure S2). Similarly to NbBGAL1, AtBGAL8 has b-galactosidase activity when produced by agroinfiltration in N. benthamiana bgal1 mutant plants (Figure 7a) and, as with NbBGAL1, agroinfiltration of AtB-GAL8 reduces growth of Pta6605 (Figure 7b) but did not affect the growth of PsyB728a (Figure 7c). Thus, similarly to NbBGAL1, AtBGAL8 can confer specific immunity to P. syringae strains carrying BGAL1-sensitive glycans. By contrast, AtumGV3101 grew equally well on agroinfiltrated leaves expressing AtBGAL8 (Figure 7d), indicating that AtBGAL8 does not affect Agrobacterium growth. These data demonstrate that the agromonas assay can be used to study diverse components of the immune system.

DISCUSSION
Agromonas is an easy, robust and simple assay to monitor P. syringae growth in agroinfiltrated leaf tissues both upon infiltration and spray inoculation of P. syringae. Using well-established PAMP receptors, a T3 effector and immune hydrolases, we demonstrated that the agromonas assay can be used to study components that enhance immunity (e.g. FLS3, EFR and BGAL1) and reduce immunity (e.g. hpFLS2 and AvrPto). This assay is now routinely used in our lab to study various components of plant immunity, perform structureÀfunction analysis, and study putative roles of effectors in immune suppression. This manuscript sets examples and parameters for this assay so it can be used widely by the research community.

Four practical considerations to take away
This manuscript establishes the methodology of the agromonas assays. As for all disease assays, experimental conditions are of fundamental importance. There are four essential parameters that should be considered for agromonas assays.
First, it is essential to use EV and hpGFP controls in overexpression and depletion assays, respectively. Agrobacterium suppresses P. syringae growth directly and/ or indirectly, so a mock control (i.e. buffer infiltration) is not very useful. The control should be based on the same Agrobacterium strain carrying the same vector and at the same final OD 600 .
Second, it is worth simultaneously monitoring growth of Agrobacterium simply using gentamicin selection, when the P. syringae strain being used is not gentamicin resistant. Occasionally, Agrobacterium growth is also affected by modulation of host immunity and this may affect protein expression. However, in the cases described here, Agrobacterium growth was reduced by EFR consistent with the literature (Kunze et al., 2004), but immunity to P. syringae was still enhanced, indicating that P. syringae growth was suppressed by EFR not by reduced Agrobacterium levels.
Third, it is crucial to allow sufficient protein accumulation in the agroinfiltrated leaves prior to P. syringae infection. This may differ between proteins. Likewise, depletion of endogenous proteins by hpRNAi may need time. For instance, 3 days upon agroinfiltration of hpFLS2 was not enough to deplete endogenous NbFLS2.
Fourth, we recommend analysing at least 3 (ideally 6) plants per condition per experiment, and repeating experiments at least three times. This is normal practice in P. syringae infections, and will undoubtedly display the high reproducibility of the agromonas assay.

Five limitations of agromonas assays
Despite the broad versatility of the agromonas assay, we would like to point out five limitations of the assay that should be considered for future experiments.
First, excessively high expression levels could cause artefacts, but this problem can be mitigated by using different promoters (Grefen et al., 2010). We like to note that many microscopy studies on fluorescent tagged proteins show the expected subcellular localisations (Bally et al., 2018), indicating that N. benthamiana usually delivers proteins at the intended site. Overexpression of host proteins may also mitigate the functions of specific effectors, for example, when these suppress this host protein.
Second, A. tumefaciens strains employed for agroinfiltration are non-oncogenic but they still trigger host immunity. Indeed, agroinfiltration into Nicotiana tabacum leaves elicits a low level of callose deposition , and the csp22 epitope of Cold Shock Protein (CSP) of A. tumefaciens is recognised by the receptor CORE (cold shock protein receptor) of N. benthamiana (Felix and Boller, 2003;Wang et al., 2016). CSP recognition increases in adult plants because CORE is expressed higher in 6-weekold plants (Wang et al., 2016), but we found that agromonas assays work in both young (3 weeks old) and adult plants (6 weeks old). Increased host immunity is visible in reduced P. syringae growth, and may be a limitation to study specific immune components.
Third, some immune responses are suppressed by agroinfiltration. For instance, agroinfiltration into N. tabacum leaves results in reduced abscisic acid levels and salicylic acid production . Likewise, A. tumefaciens strain GV3101(pMP90) used for our experiments is known to produce cytokinin through the transzeatin synthase encoded by the Ti plasmid, which induces stromules and changes the position of chloroplasts, a phenomenon that is reduced when using strain LBA4404 (Erickson et al., 2014).
Fourth, despite the fact that N. benthamiana is commonly used as a model for plantÀpathogen interactions (Goodin et al., 2008) and has the PRR co-receptors SOBIR1 and BAK1 (Heese et al., 2007;Liebrand et al., 2013), it may not possess all supporting components when testing immune components from other plant species. For example, N. benthamiana has ZAR1, but lacks ZED1 for the recognition of the bacterial effector HopZ1 (Baudin et al., 2017).
Fifth, quantification of bacterial growth by colony count assays remains a bottle neck for high-throughput screening. Alternative pathogen infection assays have been developed to increase throughput. For instance, bacterial DNA can be quantified by real-time polymerase chain reaction (PCR; Ross and Somssich, 2016), but this technique does not distinguish between living and dead bacteria, overestimating the titres of living bacteria (Rooney et al., 2020). An alternative approach for monitoring bacterial density is using bioluminescence (Fan et al., 2008), but this method requires the transformation of each bacterial strain with the luxCDABE operon (Fan et al., 2008). Also, bioluminescence reflects the metabolic state rather than bacterial viability, and cannot be used to detect low titres. However, bioluminescence might be a powerful approach to increase throughput in specific contexts.

Six opportunities of using agromonas assays
We believe that the agromonas assay can be applied to a wide range of applications. There are at least six main opportunities for a wider application of the agromonas assay.
First, the agromonas assay is a rapid and easy method to subject plant and pathogen proteins in immunity for structureÀfunction analysis, without the need of transgenic plants. For instance, we tested the non-phosphorylatable mutant EFR Y836F (Macho et al., 2014) to demonstrate that agromonas assay facilitates rapid structureÀfunction analysis of immune components.
Second, because CFC selects Pseudomonas spp., the agromonas assay works without genetic manipulation of P. syringae, allowing the study of the available repertoire of P. syringae mutants and strains. For instance, PtoDC3000 polymutants (Wei et al., 2015) and PsyB728a mutants (Vinatzer et al., 2006) can be tested in agromonas assays. However, phenotypes associated with effector functions may not always be visible when the effector target is overexpressed by agroinfiltration.
Fourth, for plant species that are not amenable to agroinfiltration, N. benthamiana remains an excellent heterologous expression system to study proteins from various organisms (plants, microbes and animals). We demonstrate this by studying EFR and AtBGAL8 from Arabidopsis and FLS3 from tomato. We anticipate that agromonas assays can be used to study more immune-related genes or quantitative trait locus (QTL) from various plants and microbes. For instance, PAMPs receptors LORE and LYM1/3 (Willmann et al., 2011;Ranf et al., 2015) are absent in N. benthamiana and should reduce bacterial growth in agromonas assays.
Fifth, P. syringae growth can be monitored in agromonas assay using both infiltration and spray inoculation, testing both post-and pre-invasive immunity. Here, we used the method of inoculation described in literature for each immune component tested (Xiang et al., 2008;Segonzac et al., 2011;Macho et al., 2014;Buscaill et al., 2019). We anticipate that leaf-dipping and vacuum inoculation should also be applicable in agromonas assays.
Sixth, agromonas assays can be used to study depletion of host immunity genes. We demonstrated this by depleting NbFLS2 using hpFLS2, which resulted in a significant enhancement of P. syringae growth. Similar results were observed with depletion of NbFLS2 by virus-induced gene silencing (VIGS; Segonzac et al., 2011). However, VIGS is at least 4 weeks slower than hpRNAi technique, and VIGS requires approval for work with modified tobacco rattle virus. We anticipate that agromonas assays can also be used to study the depletion of additional endogenous immune components, such as mitogen-activated protein kinases, calcium-dependent protein kinases, transcription factors, PAMP receptors, and nucleotide-binding/leucinerich repeats. However, depletion of these components with transient hpRNAi will depend on the stability of the protein.
In conclusion, the ability to characterise immune components from plants and microbes using agromonas assays will speed up our understanding of the plant immune system and microbial colonisation, and generate promising strategies for crop protection.

EXPERIMENTAL PROCEDURES Plants
Nicotiana benthamiana plants were grown in a growth chamber at 21°C and~60% relative humidity with a 16 h photoperiod and a light intensity of 2000 cd sr m À2 .

Silencing by RNA interference (RNAi)
An intron-containing hairpin RNA (ihpRNA) construct targeting a conserved region in FLS2a/b was designed to silence both FLS2 homologues (i.e. NbD013936.1 and NbD024362.1) detected in the NbDE database (Kourelis et al., 2019). The 300-bp fragment of GFP (SeqA ; Table S3) and the 300-bp fragment of FLS2a/b (SeqB ;  Table S3) used for RNAi were commercially synthesized (Invitrogen, Carlsbad, CA, USA). The fragments were cloned into the pRNAiGG vector (Yan et al., 2012) using BsaI restriction sites resulting in vector pPB070 and pPB072, respectively (Table S2). The binary constructs were transformed into AtumGV3101 (pMP90) by freeze-thawing, and transformants were selected by kanamycin resistance. Three-week-old N. benthamiana leaves were agroinfiltrated with the hairpin silencing construct at a final OD 600 = 0.2. Further experiments (ROS production and spray infection) were performed 10 days after agroinfiltration.

Agroinfiltration
For transient expression of proteins in N. benthamiana, overnight cultures of A. tumefaciens GV3101 (AtumGV3101) carrying binary vectors were harvested by centrifugation. Cells were resuspended in induction buffer (10 mM MgCl 2 , 10 mM MES pH5.0, and 150 µM acetosyringone) and mixed (1:1) with bacteria carrying silencing inhibitor P19 at OD 600 = 0.5. After 1 h at 21°C, cells were infiltrated with a needleless syringe into the abaxial side of three leaves of 4week-old N. benthamiana. Leaves were harvested and processed at the indicated days after agroinfiltration.

Bacterial strains
The bacterial strains used in this study are listed in Table S4. Pseudomonas and Xanthomonas strains were grown in LB medium at 28°C. For the infection assays, bacteria were cultured in LB medium containing 10 mM MgCl 2 at 28°C. Agrobacterium tumefaciens strains were grown in LB medium containing 50 lg ml À1 rifampicin, 10 lg ml À1 gentamicin and 50 lg ml À1 kanamycin at 28°C.

Oxidative burst assays
The generation of ROS was measured by a luminol-based assay on leaf discs adapted from Smith and Heese (2014). Luminol (Sigma-Aldrich, Saint Louis, MO, USA) was dissolved in dimethyl sulphoxide at a concentration of 10 mg ml À1 and kept in the dark. Horseradish peroxidase (HRP; Thermo Fisher Scientific, Waltham, MA, USA) was dissolved in water at a concentration of 10 mg ml À1 . Leaf discs (6 mm diameter), were incubated overnight in water in Petri dishes. Leaf discs from agroinfiltrated leaves were deposed in a 96-well plates, one leaf disc per well (Costar, Kennebunk, ME, USA); 100 µl of 25 ng µl À1 luminol, 25 ng µl À1 HRP and 100 nM elf18 (Kunze et al., 2004) or 100 nM flg22 (Felix et al., 1999) or 100 nM flgII-28 (Cai et al., 2011) was added and chemiluminescence was recorded immediately in relative light units (RLU) using an Infinite M200 plate reader (Tecan, Mannedorf, Switzerland). Measurements were taken every minute for 1 h. Standard errors were calculated at each time point and for each treatment. Experiments were repeated at least three times.

Bacterial growth upon inoculation
For syringe inoculations, an overnight culture was washed and resuspended in sterile water to a density of 1 9 10 6 CFU ml À1 and infiltrated into agroinfiltrated leaves using a blunt syringe via the abaxial side of the leaves. For spray inoculations, an overnight culture was washed and resuspended in sterile water to a density of 1 9 10 8 CFU ml À1 and sprayed onto adaxial surfaces of agroinfiltrated leaves. Before inoculation, plants were covered with a humidified dome for 1 day. After infection, plants were re-covered with the dome and kept for 3 days in a growth cabinet at 21°C. For determination of in planta bacterial growth, three leaf discs (1 cm diameter) were excised 3 days after inoculation from inoculated leaves. Each leaf disc was soaked in 15% H 2 O 2 for 2 min to sterilise the leaf surface. Leaf discs were washed twice in sterile water and dried under sterile conditions for 30 min. Leaf discs were then ground in sterile water for 5 min using the tissue-lyser and metal beads (Biospec Products, Bartelsville, OK, USA). Serial dilutions of the homogenate were plated onto LB agar supplemented with either gentamicin (10 µg ml À1 ) for selection of AtumGV3101 or CFC (Oxoid TM C-F-C Supplement) at 19 concentration, prepared according to the manufacturer's instructions for selection of P. syringae strains and incubated at 28°C. Colonies were counted after 36 h incubation at 28°C. The P-value was calculated using the two-tailed Student's t-test to binary compare bacterial growth between agroinfiltrated plants.

Statistics
All values shown are mean values, and the error intervals shown represent standard error of the mean (SE), unless otherwise indicated. P-values were calculated using the two-tailed Student's ttest. All experiments have been reproduced and representative datasets are shown.