Stable pH Suppresses Defense Signaling and is the Key to Enhance Agrobacterium-Mediated Transient Expression in Arabidopsis Seedlings

Agrobacterium-mediated transient expression is a powerful analysis platform for diverse plant gene functional studies, but the mechanisms regulating the expression or transformation levels are poorly studied. Previously, we developed a highly efficient and robust Agrobacterium-mediated transient expression system, named AGROBEST, for Arabidopsis seedlings. In this study, we found that AGROBEST could promote the growth of agrobacteria as well as inhibit the host immunity response. When the factor of agrobacterial growth is minimized, maintaining pH at 5.5 with MES buffer was the key to achieving optimal transient expression efficiency. The expression of plant immunity marker genes, FRK1 and NHL10, was suppressed in the pH-buffered medium as compared with non-buffered conditions in Col-0 and an efr-1 mutant lacking the immunity receptor EFR recognizing EF-Tu, a potent pathogen- or microbe-associated molecular pattern (PAMP or MAMP) of A. tumefaciens. Notably, such immune suppression could also occur in Arabidopsis seedlings without Agrobacterium infection. Furthermore, the PAMP-triggered influx of calcium ions was compromised in the pH-buffered medium. We propose that the enhanced transient expression efficiency by stable pH was due to inhibiting calcium ion uptake and subsequently led to suppressing immunity against Agrobacterium.

L-77 and detergent Triton X-100 are also used to improve the cuticular penetration of fluid on the leaf surface and can significantly enhance the efficiency 17 . The optimization of Agrobacterium-mediated transformation is mainly based on trial-and-error approaches, but the theoretical explanation is seldom addressed. A major breakthrough was the identification of the immune receptor in A. thaliana, which revealed the key molecular interaction between A. tumefaciens and A. thaliana. A. thaliana triggers its immune responses via recognizing elongation factor-thermo unstable (EF-Tu) of A. tumefaciens as a pathogen-or microbe-associated molecular pattern (PAMP or MAMP) via EF-Tu receptor (EFR); however, the elevated immune responses also hinder the transformation efficiency 18 . A mutant with loss of EFR function (i.e., efr-1) shows a significant increase in transient transformation efficiency and hence is used as a tool for rapid analysis of gene function or promoter activity 19 . However, the mutant background of efr-1 may not be suitable for generating new mutants or for defence-related studies. This limitation was overcome in part by the development of a highly efficient transient transformation method known as Agrobacterium-mediated enhanced seedling transformation (AGROBEST), which optimizes the pH-buffered medium by supplementation with AB salts and glucose 19 .
Use of the Agrobacterium-mediated transiently expressed β-glucuronidase (GUS) reporter as an assay system significantly increased the transient transformation efficiency of wild-type Col-0 and efr-1 mutant seedlings. The advantage of the method is to enhance the transient expression efficiency regardless of the genetic background of A. thaliana. In our previous study 19 , we showed an obvious enhancement in transient expression efficiency of efr-1 seedlings over two other reported methods (FAST 20 and Marion et al. 21 ). Thus, the conditions optimized in the AGROBEST method may be able to suppress the residual defence mechanism on top of EFR-mediated resistance in A. thaliana. This residual defence mechanism could be rather primitive and essential for A. thaliana to protect against pathogens when a specific receptor for pathogen recognition (e.g., EFR for EF-Tu and FLS2 for flagellin) is not available in the plant species. However, the AGROBEST condition may also make A. tumefaciens more effective to infect A. thaliana independent of the host defence mechanism.
Here, we aimed to unveil the rationale underlying why AGROBEST can significantly enhance the transient expression/transformation efficiency in A. thaliana seedlings. We found that although the AGROBEST condition can increase the growth of A. tumefaciens during co-inoculation with A. thaliana, the main reason for the enhanced transient expression efficiency is stable acidic pH at 5.5. Further analysis showed that the stable pH was able to suppress the A. thaliana defence signalling pathway with or without Agrobacterium infection. Also, we provide evidence that the suppression of immunity could be due to ineffective calcium ion influx in the stable pH environment. The mechanistic insights gained from this study advance our understanding of Agrobacteriumplant interactions and provide invaluable information when using AGROBEST for dissecting gene function and regulatory pathways.

Results
AGROBEST promotes Agrobacterium growth and attachment to both Col-0 and efr-1 seedlings.
AGROBEST enhancing transient expression efficiency may be due to increased virulence or growth of Agrobacterium cells or promoting a plant's susceptibility to Agrobacterium infection. Thus, we first investigated whether the AGROBEST condition could enhance the growth of Agrobacterium cells during inoculation with A. thaliana seedlings. We compared the viable Agrobacterium cell number in the medium and that associated in seedlings under two conditions, regular MS medium and AGROBEST. Transient GUS expression was 10-fold higher with AGROBEST than MS medium 19 . The MS medium contains ½ MS salts and 0.5% sucrose, and pH was adjusted to 5.5 with KOH in the presence of 200 μM AS. The AGROBEST medium includes an equal amount of the above-mentioned MS medium and AB-MES medium supplemented with 200 μM AS. The key difference between these two media is the addition of AB salts, glucose, and 2-(N-morpholino)ethanesulfonic (MES) to maintain pH at 5.5. In Col-0 seedlings, Agrobacterium cell number in medium was greater with AGROBEST than MS medium as early as 1 day post-infection (dpi) and the growth and difference continued up to 3 dpi. In efr-1 seedlings, Agrobacterium cell number in medium was also greater with AGROBEST than MS medium although the difference at 1 dpi is not statistically significant. Both Col-0 and efr-1 seedlings had more in planta Agrobacterium cells with AGROBEST than MS medium at 2 dpi and 3 dpi (Fig. 1a). Therefore, the AGROBEST medium could promote the growth of Agrobacterium cells during co-inoculation in both Col-0 and efr-1 seedlings. To our surprise, in general, transient GUS expression was 5-to 10-fold higher in the efr-1 mutant than Col-0 under the AGROBEST or MS condition 19 , but we could not detect a conclusive difference in number of Agrobacterium viable cells from either medium or in planta between co-culture of Col-0 and efr-1 seedlings (Fig. 1a). Similar results were obtained from at least two independent experiments with multiple biological replicates except the in planta Agrobacterium viable cells in efr-1 mutant were sometimes a bit higher than Col-0 ( Fig. 1a) or no statistically different.
Besides counting the in planta viable cells, we used scanning electron microscopy (SEM) to examine Agrobacterium cells attached to Arabidopsis seedlings. In general, for both Col-0 and efr-1 seedlings, density of attached agrobacterial cells was lower in roots than cotyledons; more cells were consistently attached to the abaxial than adaxial side of cotyledons. Hence, we compared the MS and AGROBEST conditions for agrobacterial cells attached to the abaxial side of cotyledons of Col-0 and efr-1 seedlings (Fig. 1b). For both Col-0 and efr-1 seedlings, density of agrobacterial cells on seedlings was higher with AGROBEST than MS medium. Similar to viable cell counting results, Agrobacterium cell density on the cotyledon surface did not differ between efr-1 and Col-0 seedlings despite some variations in different areas of the same cotyledon.
In summary, the AGROBEST condition promoted cell growth in the medium and increased the overall cell number associated with seedlings and the attached cells on the cotyledon surface. Despite the cell number available for transformation being significantly higher with AGROBEST than MS medium, the marked increase in number of attached cells on the cotyledon with AGROBEST medium is nonetheless an interesting phenomenon to investigate whether the medium can make the plant tissue more welcoming to the bacteria. Thus, we further investigated whether the AGROBEST medium could have any additional impact on enhancing transient expression or transformation efficiency in addition to increased cell number in the inoculum.
Stable pH at 5.5 is optimal for Col-0 and efr-1 seedling transformation. In our previous study 19 , we demonstrated that the AGROBEST condition is an optimal condition for A. thaliana seedling transient expression. As well, stable acidic pH at 5.5 buffered with MES or phosphate and AB salts in the co-culture medium are critical criteria in the optimization procedure. Thus, we next excluded the influence of medium composition such as AB salts on transformation efficiency and focused on the pH values to determine whether pH 5.5 is indeed optimal for transformation. The induction of vir genes by the phenolic compound AS can be activated at acidic pH 5.5-6.0 22 , so we used the defined AGROBEST medium buffered with MES to pH 5.0, 5.5, and 6.0 to test the effect of different acidic pH on transient expression efficiency. The highest transient GUS expression for both Col-0 and efr-1 seedlings was at pH 5.5 (Fig. 2). The result for the best efficiency at pH 5.5 was expected, but we were surprised by the very low transient GUS expression at pH 6.0, the pH capable of efficient induction of vir gene expression in vitro 22 . Furthermore, we noticed greater enhancement for efr-1 seedlings as compared with Col-0 at pH 5.5 versus pH 5.0 and 6.0. These results suggest that the massive increase in transient expression efficiency at pH 5.5 may not simply be explained by cell number in the inoculum nor activation of vir genes. Therefore, we designed the following experiments to study the role of stable pH 5.5 in seedlings under constant (e.g., inoculum cell density) and variable (e.g., with or without MES buffer) conditions.

Stable acidic pH suppresses the defence response in Arabidopsis seedlings.
To study the effect of stable pH in Arabidopsis seedlings, we used a high cell density (OD 600 = 2.0) of agrobacteria and shorter time (1 day) to minimize the growth difference in bacteria during inoculation. Because pH 5.5 is the optimal condition for transient transformation, we used the defined MS medium of adjusted pH 5.5 with or without MES buffer to test whether the transient expression efficiency was affected when the buffering system was absent. At fixed Agrobacterium cell density and inoculation conditions, transient GUS activity was strikingly increased in Col-0 seedlings grown in MES-buffered versus non-buffered medium (Fig. 3a). The data suggest that maintaining a stable pH to 5.5 is a key factor in achieving high transient expression/transformation efficiency. In addition, the GUS activity in Col-0 seedlings with MES medium was comparable to that in efr-1 seedlings without MES, with still an effect for efr-1 seedlings in MES on top of the -MES condition (Fig. 3a). This enhancement could be due to the suppression of the defence response by stable pH. Therefore, we detected the expression of some well-known defence marker genes: FLG22-induced receptor-like kinase 1 (FRK-1) and NDR1/HIN1-like 10 (NHL10). Both FRK1 and NHL10 are highly inducible upon contact with bacterial PAMPs 23 . The expression of FRK1 and NHL10 was highly induced in both Col-0 and efr-1 seedlings after Agrobacterium infection (Fig. 3b). In both Col-0 and efr-1 seedlings, FRK1 expression at 1 dpi was significantly suppressed with MES medium, with suppression levels about 1 log 10 (10-fold) lower with than without MES (Fig. 3b). Similarly, in both Col-0 and efr-1 seedlings, the expression of NHL10 at 1 dpi was suppressed more than 1.5 log 10 (50-fold) with than without MES (Fig. 3b).
Despite the significantly higher Agrobacterium-induced FRK1 and NHL10 expression in Col-0 than efr-1, the levels with MES suppressing the marker gene expression were approximately similar between Col-0 and efr-1 seedlings. The ability of MES to suppress the defence gene expression in efr-1 seedlings indicates that the suppression could be EFR-independent. We also explored the possibility that the effect of suppression by MES was also observable when bacterial cells were absent. Although the expression was low in mock controls (4-day-old Arabidopsis seedlings grown in co-culture medium for 1 day without Agrobacterium infection), FRK1 and NHL10 expression was also suppressed by MES (Fig. 3b). Similar suppression effects were observed in Arabidopsis seedlings grown in the original AGROBEST versus MS co-culture media for 3 days without Agrobacterium infection (Fig. 3c). Although not as significant as with Agrobacterium infection, the MES-mediated suppression effect on FRK1 and NHL10 expression was still observable, with about a 40% and 20% decrease, respectively, in seedlings without agrobacteria. The results confirm that maintaining stable acidic pH 5.5 is the key factor (adjustment) suppressing the defence response and leading to enhanced efficiency of Agrobacterium-mediated transient transformation of seedlings.

Stable acidic pH inhibits the influx of calcium ion triggered by PAMPs.
We speculated that maintaining a stable pH could lead to an imbalance in ion movement between the medium (or apoplasts) and cytosol, which undermines the normal defence signalling pathway. Calcium (Ca 2+ ) ion plays important roles in defence signalling pathways [23][24][25] . Influx of Ca 2+ ions from apoplasts into cytosol upon pathogen challenge is considered an important event to trigger downstream signalling pathways 26 . An increase in intracellular Ca 2+ ion is sensed by some calcium-dependent protein kinases (CDPKs) and subsequently triggers the downstream defence pathways 23 . PAMPs such as flg22 and elf18 can induce Ca 2+ influx from apoplasts into the intracellular space 27 . Thus, we compared cytosolic calcium levels upon PAMP induction in the medium with or without MES by using an aequorin-based method 28 . Aequorin is a 22-kDa protein that can bind calcium ions and a luciferin molecule, coelenterazine. Upon binding of calcium ions, aequorin converts coelenterazine to coelenteramide by its oxygenase activity. During the process, blue light at λ = 469 nm is emitted and detectable with a luminometer 29 . We used aequorin-expressing A. thaliana seedlings to detect the influx of calcium ions upon PAMP challenge. The main purpose of this experiment was to investigate any impact of calcium ion influx when pH was maintained at 5.5 with MES. When no PAMPs were added to the seedlings, cytosolic calcium ion level showed no increase (Fig. 4a,b), so seedlings were in a resting condition. In contrast, flg22 (Fig. 4a) or elf18 (Fig. 4b) could induce a prominent influx of calcium ions into the cytosol of Arabidopsis seedlings, which was consistent with previous observations 27 . Although an influx of calcium ions was observed in the sample with MES, the magnitude and the influx rate in MES-added samples were significantly lower than in the samples without MES. Calcium ion levels showed a steeper and higher increase in flg22-treated samples without MES (Fig. 4a). A maximum level, at about 0.8 µM, of cytosolic calcium ions was detected and it also appeared earlier, at about 3 min, as compared with about 4 min (0.6 µM) for samples with MES. A higher level was maintained until the end of the detection time (10 min) when MES was absent. Similarly, elf18 treatment (Fig. 4b) triggered a higher (~0.8 µM) and faster (~3:30 at maximum) influx of calcium ions without MES. A peak shape of the influx was observed only in samples without MES, whereas the levels decreased to similar levels for samples with MES at the end of the detection time (10 min). In addition, without PAMP treatment, the calcium ion levels were generally lower in seedlings with than without MES (Supplementary Figure S1). This observation may echo the reduced expression of defence marker genes in the presence of MES (Fig. 3b).

Discussion
Our previous study showed that the highly enhanced transient expression efficiency of efr-1 seedlings was achieved by two key factors, stable pH 5.5 and AB salts present in the AGROBEST medium 19 . Our current study found that the AGROBEST condition promoted Agrobacterium growth and attachment onto the seedling surface and suppressed the defence response. This situation may lead to enhanced transient GUS expression, an assay generally used to reflect transient transformation efficiency. We also extend the findings that buffered pH 5.5 is the key to this enhancement. The stable pH 5.5 suppressed the expression of the immune marker genes FRK1 and NHL10 independent of EFR, which could occur without Agrobacterium infection. Without MES to maintain pH at 5.5 during Agrobacterium infection, the pH value of the MS medium tended to fluctuate, often with up to 0.2 of pH increase after Agrobacterium infection. This finding is consistent with an alkalization effect upon triggering an immune response 18,30 although we sometimes also observed decreased pH after Agrobacterium infection in non-buffered medium. Thus, we propose that the stable acidic pH could lead to a physical effect or a signal mimicking the absence of pathogens to trick plants to staying in a resting state with compromised immune response. Although the number of Agrobacterium cells attached to the cotyledon surface of both Col-0 and efr-1 seedlings was greater with AGROBEST than MS medium, transient transformation efficiency is not necessarily associated with a higher number of bound Agrobacterium cells. We previously reported a 5-to 10-fold higher transient GUS expression in the efr-1 mutant than Col-0 under the AGROBEST or MS condition at 3 dpi with infection; the initial cell density was OD 600 = 0.02 19 . With the same experimental setup in this study, both viable cell count and microscopy observations suggested no significant increase in number of plant-associated agrobacterial cells on efr-1 seedlings versus Col-0 seedlings under the same condition (Fig. 1). This result is indeed consistent with previous findings of agroinfiltration in adult leaves revealing no higher Agrobacterium viable cells in leaves of the efr-1 mutant, which is more susceptible to Agrobacterium-mediated transient GUS expression than Col-0 18 . Thus, EFR does not seem to play a major role in Agrobacterium attachment and rather acts on sensing the EF-Tu to trigger downstream immune responses in resistance to Agrobacterium-mediated transformation.
The suppression of the defence-response marker genes by MES indicates that the enhanced transient expression efficiency could be due to the seedlings in stable acidic pH being less capable of defence against A. tumefaciens and hence T-DNA more effectively transferred to the host. To this end, T-DNA-mediated transient GUS expression is a hallmark and standard assay to reflect Agrobacterium-mediated transient transformation efficiency [19][20][21]31 . However, we do not exclude that the increased GUS activity promoted by AGROBEST could also be due to increased T-DNA-encoded GUS gene expression besides increased copy number of T-DNA translocated from agrobacteria into the plant cells. Future work to develop a technically feasible tool quantifying the copy number of T-DNA in planta will shed light on the association of T-DNA copy number and transient gene expression levels inside the plant cells. Uptake of calcium ions mainly relies on plasma membrane Ca 2+ /H + exchangers, and export of H + /proton is a key event for the exchanger to function. Zhai et al. 32 reported that the proton pump inhibitor vanadate could supress the activity of the Ca 2+ /H + exchanger in the Arabidopsis membrane. In addition, optimal pH for the exchanger was 7.0, and acidic pH would reduce the efficiency of the exchanger 32 . We speculated that maintaining the pH at 5.5 with MES could reduce the activity of such Ca 2+ /H + exchangers by inhibiting the proton movement across the plasma membrane. As the calcium ion influx is reduced, the upregulation of defence-response genes would be inhibited and thus the defence response would be weakened. The weakening of the defence response could make the Agrobacterium cells more effective for cell surface attachment as well as T-DNA transfer, which would lead to reduced transient expression or transformation efficiency. Because the effect of stable pH mainly targets the physical exchange of protons and calcium ions across the plasma membrane, the effect would be independent of many evolved defence mechanisms requiring molecular interactions. The enhanced transformation efficiency in efr-1 seedlings supports that the immunity suppressed by stable pH is an extra mechanism that could be independent of the EFR signalling pathway. The activation of residual defence responses in efr-1 seedlings may be via recognition of other PAMPs that rely on calcium signalling cascades such as CDPKs 23 . Thus, inhibition of Ca 2+ influx is able to further suppress the defence response. Besides EF-Tu 30 , peptidoglycan and cold-shock protein but not flagellin of Agrobacterium can elicit plant immunity [33][34][35] . Practically, although knocking out EFR in A. thaliana can significantly enhance transient expression, the presence of other Agrobacterium PAMPs can still trigger the defence response to impede efficient Agrobacterium-mediated transformation, especially for some Arabidopsis ecotypes with unknown genetic background. Suppression of basal immunity via stable pH could be an alternative way to achieve satisfactory transient expression for various ecotypes. The mechanism of Ca 2+ influx should be highly conserved across the plant kingdom. Thus, the findings in this study may be applied to increase both stable and transient transformation in other plant species recalcitrant to Agrobacterium-mediated transformation.

Methods
Materials and growth condition. Arabidopsis thaliana seeds of ecotype Columbia-0 (Col-0) and T-DNA insertion mutant efr-1 (SALK_044334) were obtained from the Arabidopsis Biological Resource Center (Columbus, Ohio, USA). A. thaliana seedlings were grown in a growth room at 22 °C under a 16-hr/8-hr lightdark cycle (75 μmol m −2 s −1 ). Agrobacterium tumefaciens strain C58C1 (pTiB6S3ΔT) H36 harbouring pBISN1 37 was used in the study for transient transformation assay. A. tumefaciens cells were routinely grown in 523 medium 38 containing appropriate antibiotics with shaking (220 rpm) at 25 °C.
Agrobacterium infection in Arabidopsis seedlings. The procedures for the seedling transient transformation assay were adapted from our previous report 19 with modifications. The 523 grown A. tumefaciens cells were harvested and pre-induced with 200 μM acetosyringone in AB-MES (17 mM K 2 HPO 4 , 8 mM NaH 2 PO 4 , 18 mM NH 4 Cl, 2 mM KCl, 1.25 mM MgSO 4 , 100 µM CaCl 2 , 10 µM FeSO 4 , 50 mM MES, 2% glucose (w/v), pH 5.5) at 25 °C for 16 hr and resuspended in appropriate buffer with cell density OD 600 = 0.02 or 2.0. A. thaliana seeds were germinated in regular MS medium (1/2 MS, 0.5% sucrose [w/v], pH 5.5) (~10 seeds in each well) and 4-day-old seedlings were then incubated in 1 ml respective co-culture medium in one well of 6-well plates for Agrobacterium infection up to 3 days (for infection at OD 600 = 0.02) or 1 day (for infection at OD 600 = 2.0) followed by the co-culture medium being replaced by fresh MS medium (1/2 MS, 0.5% sucrose [w/v], pH 5.5) with timentin (100 µg/ml) for another 3 days. The AGROBEST method used was described previously and the co-culture medium used was MS medium and AB-MES at a 1:1 ratio. The seedlings were then removed for GUS staining immediately or stored at −80 °C before RNA extraction or quantitative GUS activity assay.
Plant RNA extraction and quantitative RT-PCR. Total RNA was extracted from A. thaliana seedlings by using the RNeasy Plant Mini Kit (Qiagen). First-strand cDNA was synthesized from 4 μg total RNA with SuperScript III Reverse Transcriptase (Invitrogen) and oligo (dT) primers. Quantitative PCR involved using the QuantStudio 12 K Flex Real Time PCR system (Applied Biosystems) with the Power SYBRR Green PCR Master Mix (Invitrogen). Primers for A. thaliana FRK1 (At2g19190) and NHL10 (At2g35980) 23 were synthesized according to the sequences of a previous study. ACTIN 2 (At3g18780) was an internal control. Data were compared by Student t test from three technical replicates for each independent experiment. GUS staining and activity assay. Seedlings were stained with 5-bromo-4-chloro-3-indolyl glucuronide (X-Gluc) at 37 °C for 6 hr as described 39 . GUS activity assay was determined by the conversion of 4-methylumbelliferyl-β-D-glucuronide (4-MUG) to 4-methylumbelliferone (4-MU). 4-MU fluorescence (ex. 356 nm, em. 455 nm) was measured by using a 96 microtiter-plate reader (Bio-Tek Synergy Mx) and the specific GUS enzyme activity was calculated based on 4-MU standards. Data were compared by Student t test from three biological replicates for each independent experiment.

Agrobacterium viable cell count and electron microscopy. Seedlings infected with A. tumefaciens
were rinsed briefly with 0.9% saline. Seedling-associated A. tumefaciens cells were extracted by grinding the seedlings in 0.9% saline (1 ml for 10 seedlings from each well). The extracts or inoculation medium were serially diluted and spread on low-salt LB (0.5% NaCl) agar plates for incubation at 25°C for 2 days before counting colony-forming units (cfu). Agrobacterium cells on the seedling surface were observed by scanning electron microscopy (SEM, FEI Quanta 200). The samples were first mixed with fixation solution (2.5% glutaraldehyde, 4% paraformaldehyde in 0.1 M sodium phosphate buffer, pH 7.0) at room temperature overnight. The samples were rinsed with 0.1 M sodium phosphate buffer, pH 7.0 three times and post-fixed in 1% osmium tetroxide (OsO 4 ) for 4 h then rinsed with 0.1 M sodium phosphate buffer, pH 7.0 three times 40 . Before viewing, samples were dehydrated in an ethanol series and dried with use of a Hitachi HCP-2 critical point dryer. SEM at 20KV was used for viewing.
Cytosolic calcium ion measurement. Aequorin-expressing Arabidopsis seeds were a gift of Prof. Marc Knight (Durham University, Durham, UK). Measurement of cytosolic calcium ions was based on a published report 28 with modifications. Ten-day old seedlings were first incubated in MS medium with or without MES for 1 day, transferred into 1 well of 12-well plates with 500 μl reacting solution (MS medium at pH 5.5 with 10 mM KCl, 10 mM CaCl 2 , 100 mM MgCl 2 and 10 μM coelenterazine, + /− MES) and vacuum-infiltrated for 5 min, then incubated in the dark for 3 h. The same reacting solution (500 μl, with or without flg22 (0.2 μM)/elf18 (0.2 μM)) was loaded into the solution automatically in the reader (Bio-Tek Synergy Mx), and luminescence at λ = 469 nm was measured immediately at every 10 s for 10 min. Data were compared by Student's t test from three biological replicates for each independent experiment.