pH Gradient Mitigation in the Leaf Cell Secretory Pathway Alters the Defense Response of Nicotiana benthamiana to Agroinfiltration

Partial neutralization of the Golgi lumen pH by ectopic expression of influenza virus M2 proton channel stabilizes acid-labile and protease-susceptible recombinant proteins in the plant cell secretory pathway. Here, we assessed the impact of M2 channel expression on the proteome of Nicotiana benthamiana leaf tissue infiltrated with the bacterial gene vector Agrobacterium tumefaciens, keeping in mind the key role of pH homeostasis on secreted protein processing and the involvement of protein secretion processes in plant cells upon microbial challenge. The proteomes of leaves agroinfiltrated with an empty vector or with an M2 channel-encoding vector were compared with the proteome of non-infiltrated leaves using a iTRAQ quantitative proteomics procedure. Leaves infiltrated with the empty vector had a low soluble protein content compared to non-infiltrated leaves, associated with a strong decrease of photosynthesis-associated proteins (including Rubisco) and a parallel increase of stress-related secreted proteins (including pathogenesis-related proteins, protease inhibitors and molecular chaperones). M2 expression partly compromised these alterations of the proteome to restore original soluble protein and Rubisco contents, associated with higher levels of translation-associated (ribosomal) proteins and reduced levels of stress-related proteins in the apoplast. Proteome changes in M2-expressing leaves were determined both transcriptionally and post-transcriptionally, to alter the steady-state levels of proteins not only along the secretory pathway but also in other cellular compartments including the chloroplast, the cytoplasm, the nucleus and the mitochondrion. These data illustrate the cell-wide influence of Golgi lumen pH homeostasis on the leaf proteome of N. benthamiana plants responding to microbial challenge. They underline in practice the relevance of carefully considering the eventual off-target effects of accessory proteins used to modulate specific cellular or metabolic functions in plant protein biofactories.

108 expression on leaf protein content at the cell-wide scale (Fig. 1). A significant adjustment of the 109 protein complement was suggested by a low total soluble protein content of 5.7 mg.g -1 fresh 110 weight in EV-infiltrated leaves compared to 9.4 mg.g -1 in non-infiltrated leaves 6 d post-infiltration 111 (post-ANOVA LSD; P<0.05) (Fig. 1A). Protein content reduction in agroinfiltrated leaves was 112 associated with a downregulation of ribulose-1,5-bis-phosphate carboxylase/oxygenase (Rubisco) 113 large and small subunits and a concomitant upregulation of protein bands in the 20-35-kDa range 114 ( Fig. 1B). By contrast, protein content in M2 vector-infiltrated leaves was estimated at 8.4 mg.g -1 115 fresh weight, statistically similar to non-infiltrated leaves (LSD; P>0.05) (Fig. 1A). Rubisco subunit 116 band intensities were also comparable in M2 vector-infiltrated and non-infiltrated leaves, two to 117 three times more intense than the corresponding bands in EV-infiltrated leaves (Fig. 1B).

124
We conducted a iTRAQ analysis of non-infiltrated and infiltrated leaf protein extracts to 125 estimate the overall impacts of agroinfiltration and M2 expression on the leaf soluble protein 126 complement (Fig. 1C,D) and to characterize the specific effects of these treatments at the cell-127 wide scale (Figs. 2-5). A total of 5,928 unique peptides were detected by MS/MS, allowing for 128 the identification of 2,388 proteins at a confidence level of 95% (P<0,05). A little more than 50% 129 (i.e. 1,255) of these proteins were identified based on at least two unique peptides and used for 130 further comparative assessments (Supplemental Table S1). On a leaf fresh weight (FW) basis, 131 425 proteins were downregulated, and 50 proteins upregulated, by at least twofold in EV-infiltrated 132 leaves compared to 214 proteins downregulated and 54 upregulated in M2 vector-infiltrated 133 leaves (Fig. 1C). On a total soluble protein (TSP) basis, 202 proteins were downregulated, and protein numbers flagged as down-(162) or upregulated (79) upon M2 expression (Fig. 1C). A 136 pairwise comparison of MS/MS data produced for EV-and M2 vector-infiltrated leaves was 137 performed to estimate the overall impact of M2 proton channel activity on agroinfiltration-induced 138 proteome changes (Fig. 1D). On a TSP basis, only 10 proteins were upregulated, and 19 139 downregulated, by at least twofold in M2 vector-infiltrated leaves compared to EV-infiltrated 140 leaves, out of 1,255 proteins monitored. These data suggested overall a qualitative impact of M2 141 expression on the host proteome limited, on protein-specific basis, to a relatively small number of 142 up-or downregulated proteins. They confirmed, by contrast, the strong impact of agroinfiltration 143 on the leaf proteome, attenuated to some extent by Golgi pH alteration in M2-expressing cells.  (Fig. 2, Supplemental Fig. S2). Agroinfiltration was shown previously to  158 peroxidases), molecular chaperones (e.g. heat shock proteins), PR proteins (e.g. ß-glucanases, 159 chitinases, osmotins) and protease inhibitors (e.g. Kunitz proteins) (Supplemental Table S2).

166
Unlike stress-related proteins, several proteins involved in mRNA translation, photosynthesis 167 and ATP biosynthesis were strongly downregulated in EV-infiltrated leaves (Fig. 2B, upper panel).
I and II, chlorophyll-binding proteins, ATP synthase subunits and translation-associated ribosomal 170 proteins (Supplemental Table S3) accounted for more than 75% of the downregulated proteins 171 (Fig. 2B, lower panel). A small proportion (13%) of these proteins were chloroplast genome-172 encoded but the vast majority were encoded by the nuclear genome as observed for the 173 upregulated proteins (Fig. 2B,   proteins downregulated in EV-infiltrated leaves were found at higher levels in M2-expressing 204 leaves ( Fig. 3A) including, along with Rubisco ( Fig. 1), a range of chloroplastic proteins involved 205 in photosynthesis, ATP biosynthesis and mRNA translation (Supplemental Table S4). Likewise, 206 nuclear genome-encoded proteins upregulated in leaves upon EV infiltration including PR 207 proteins, Kunitz protease inhibitors and stress-related oxidoreductases (Supplemental Table S5), 208 were found at lower levels in M2-expressing leaves (Fig. 3B).

209
Venn diagrams were produced, and a principal component analysis (PCA) performed, to 210 visually compare the proteomes of non-infiltrated and infiltrated leaves (Fig. 4, Fig. 5 Table S1) revealed strongly divergent proteomes in non-infiltrated 222 and EV-infiltrated leaves, compared to M2-expressing leaves exhibiting a hybrid, intermediate

233
Immunoblotting and reverse transcriptase (RT)-qPCR analyses were conducted to statistically confirm the validity of our proteomic inferences and to determine whether proteome changes in 235 leaves upon EV infiltration or M2 ectopic expression were transcriptionally or posttranscriptionally 236 regulated (Fig. 6, Fig. 7). In line with the Coomassie blue-stained gels above ( Fig. 1), immunoblot 237 signals for the large and small subunits of Rubisco were the most intense in non-infiltrated leaf 238 samples and the least intense in EV-infiltrated leaf samples (post-ANOVA LSD, P<0.05) (Fig. 6A).

239
Likewise, PR-2 and PR-3 proteins were readily detected in both EV-and M2 vector-infiltrated leaf

248
Gene expression trends for up-and downregulated proteins followed a similar path in 249 agroinfiltrated leaves compared to non-infiltrated leaves but could not explain the distinct 250 accumulation patterns observed for some of these proteins in EV-and M2 vector-infiltrated leaves.

251
For instance, mRNA transcript numbers for the two subunits of Rubisco decreased to similar 252 levels in EV-and M2 vector-infiltrated leaves (Fig. 7A) but higher levels of both subunits were 253 found in M2-expressing leaves (Fig. 1, Fig. 6A). Similarly, transcript numbers for the prominent  were only ~20-70% the levels measured in EV-infiltrated leaves ( Table 1)

289
We characterized soluble protein profiles in the apoplast of non-infiltrated, EV-infiltrated and 290 M2-expressing leaves to document the eventual impact of M2 on protein secretion (Fig. 8).

291
Agroinfiltration was shown previously to trigger a strong upregulation of defense protein secretion protein antibodies (Fig. 6B). In line with iTRAQ and immunodetection data (Fig. 3B, Fig. 6B), the 300 major band at 34 kDa was found in M2-expressing leaves at levels about half the corresponding PAGE 11 levels in EV-infiltrated leaves (Fig. 8B) despite similar numbers of mRNA transcripts for 302 endochitinase A in leaves under either treatment (Fig. 7). A post-transcriptional mitigating effect 303 of M2 on protein release in the apoplast was further substantiated with GFP variant pHluorin as 304 a recombinant protein model (Fig. 8C). We recently reported a positive effect of M2 on pHluorin            Table S1). The PCA was performed on log-normalized data using the R software,  Rubisco large (RbcL) and small (RbcS) subunits, PR-2 proteins and PR-3 proteins were detected 461 by immunoblotting on nitrocellulose sheets following 12% (w/v) SDS-PAGE in reducing conditions.

462
Rubisco subunits were detected with polyclonal IgG raised in rabbits against RbcL (Agrisera, Prod.

465
Nonspecific binding on nitrocellulose sheets was prevented by incubation in blocking solution (5%     Table S1 Complete list of confidently identified proteins following iTRAQ analysis

835
Each bar is the mean of three biological (plant) replicates ± SE.