Subcellular Localization of Arabidopsis Pathogenesis-Related 1 (PR1) Protein

The Arabidopsis thaliana pathogenesis-related 1 (PR1) is an important defense protein, so far it has only been detected in extracellular space and its subcellular sorting and transport remain unexplained. Using a green fluorescent protein (GFP) tagged full length, as well as a C-terminus truncated version of PR1, we observed that when expressed ectopically in Nicotiana benthamiana leaves, PR1 co-localizes only partially with Golgi markers, and much more prominently with the late endosome (LE)/multivesicular body (MVB) FYVE marker. The C-truncated version PR1ΔC predominantly localized to the endoplasmic reticulum (ER). The same localizations were found for stable Arabidopsis transformants with expression of PR1 and PR1ΔC driven by the native promoter. We conclude that the A. thaliana PR1 (AtPR1) undergoes an unconventional secretion pathway, starting from the C-terminus-dependent sorting from the ER, and utilizing further transportation via phosphatidyl-inositol-3-phosphate (PI(3)P) positive LE/MVB-like vesicles. The homology model of the PR1 structure shows that the cluster of positively charged amino acid residues (arginines 60, 67, 137, and lysine 135) could indeed interact with negatively charged phospholipids of cellular membranes. It remains to be resolved whether Golgi and LE/MVB localization reflects an alternative sorting or trafficking succession, and what the role of lipid interactions in it will be.


Introduction
In plants, the synthesis of various pathogenesis-related (PR) proteins is among the first of the biotic stress-induced responses. There are 17 evolutionarily conserved families of PR proteins. Some of them have direct antimicrobial function while others play a role as immunity signaling molecules [1,2].
PR1s are the most abundantly produced PR proteins upon pathogen attack (e.g., 2%-10% of the total tobacco leaf protein; [3,4]). Members of the PR1 family are highly conserved among plants; homologues have also been found in fungi, insects, and vertebrates including human. Together they form a superfamily of secreted proteins named CAP (from cystein rich secretory protein (CRISP), antigen 5, and PR1 proteins; [5]). Even though PR1s have been extensively studied, their exact function remains enigmatic. They have been suggested to have antifungal activity, or to play a role in host salicylic acid (SA)-mediated defense signaling and hypersensitive response (HR)-related cell death [6].
PR1s are present in plant genomes as highly multiplied gene families. In the genome of Arabidopsis thaliana, there are 22 PR1-type genes [7], mostly arranged in clusters of genes [6]. Only one of them, A. thaliana PR1 (AtPR1, At2g14610), is activated by pathogens, insects, or chemical treatments, whereas other PR1-type genes are constitutively expressed in roots and pollen [6].

Localization of AtPR1 after the Transient Expression in N. benthamiana
As the AtPR1 protein fused at the C-terminus to GFP was transiently expressed in N. benthamiana, we observed the GFP signal in compartments localized throughout the cytoplasm ( Figure 1A-C). We have also designed a truncated version of PR1 with deleted 50 amino acids from the C-terminus and fused it in frame again C-terminally to GFP (AtPR1∆C; Figure 1A,D,F). We expected that this protein will be predominantly secreted outside the cell, since it has no other known sorting signal. Surprisingly, the majority of the protein seemed to be retained in the ER and AtPR1∆C-GFP positive compartments are not observed in contrast to AtPR1-GFP. This is supported by the observation of a strong nuclear membrane signal in AtPR1∆C-GFP possibly coming from an over-loaded ER ( Figure 1D). In order to confirm the presence of the GFP signal, we have checked the emission spectra by a lambda scan mode ( Figure 1E,F), allowing us to distinguish between the real GFP signal and chloroplast autofluorescence.
The PR1-positive compartments were found to be highly mobile and of variable shape, with a diameter of approximately a few hundred nanometer up to one micrometer big (Video S1). Video S1 also indicates a possible occurrence of dilated subdomains of ER in transiently expressing N. benthamiana cells. As PR1 was previously reported to be localized extracellularly [8,12], in order to better visualize cell wall GFP signal, we subjected leaves to increased pH, as well as to plasmolysis by a high concentration of mannitol or NaCl. We observed a weak cell wall and a plasmamembrane signal after the mannitol induced plasmolysis, for both full length and C-truncated forms of PR1 ( Figure S1).
Based on these observations, we concluded that the AtPR1 protein is localized to a vesicle-like intracellular compartments and that localization into these compartments requires an intact C-terminus in transient N. benthamiana cell expression. PR1-GFP is also localized adjacent to the plasma membrane and weakly into the apoplastic cell wall space. sorting signal. Surprisingly, the majority of the protein seemed to be retained in the ER and AtPR1ΔC-GFP positive compartments are not observed in contrast to AtPR1-GFP. This is supported by the observation of a strong nuclear membrane signal in AtPR1ΔC-GFP possibly coming from an over-loaded ER ( Figure 1D). In order to confirm the presence of the GFP signal, we have checked the emission spectra by a lambda scan mode ( Figure 1E,F), allowing us to distinguish between the real GFP signal and chloroplast autofluorescence. The PR1-positive compartments were found to be highly mobile and of variable shape, with a diameter of approximately a few hundred nanometer up to one micrometer big (Video S1). Video S1 also indicates a possible occurrence of dilated subdomains of ER in transiently expressing N. benthamiana cells. As PR1 was previously reported to be localized extracellularly [8,12], in order to better visualize cell wall GFP signal, we subjected leaves to increased pH, as well as to plasmolysis by a high concentration of mannitol or NaCl. We observed a weak cell wall and a plasmamembrane signal after the mannitol induced plasmolysis, for both full length and C-truncated forms of PR1 ( Figure S1).
Based on these observations, we concluded that the AtPR1 protein is localized to a vesicle-like intracellular compartments and that localization into these compartments requires an intact C-terminus in transient N. benthamiana cell expression. PR1-GFP is also localized adjacent to the plasma membrane and weakly into the apoplastic cell wall space.

Co-Localization of PR1 with Compartmental Markers in N. benthamiana Transiently Transformed Cells
In order to uncover the identity of the AtPR1-positive compartment, we have performed co-localization studies of full-length PR1 using red fluorescent protein (RFP)-tagged markers for ER (dsRed-HDEL; [17]), cis-Golgi marker Glycine max Golgi mannosidase-RFP (GM-RFP; [18]), trans-Golgi marker sialyl transferase-RFP (ST-RFP; donated by Ian Moore, Oxford), extracellular marker pectin methyl esterase 1-RFP (PME1-RFP; [19]), autophagy marker ATG8-RFP [20], and PI(3)P-enriched early/late endosomal compartmental marker dsRed-FYVE [21] (Figure 2A-F). PR1-GFP only weakly colocalized with the ER (Figure 2). The overlap with the Golgi marker was partial and the majority of the signal overlapped with FYVE-positive compartment ( Figure 2). In some cases, these compartments had obvious structure, typical for MVBs-the presence of particles/vesicles within the vesicle ( Figure 3). Unfortunately, our time series scanning did not reveal any events resembling exosome secretion, fusion with vacuolar membrane, nor endocytotic events (Video S2). We performed the object-based quantification of colocalization of PR1-GFP with GM-RFP, ST-RFP and dsRed-FYVE. The highest proportion of the green signal overlapping with the red one was found for the combination of PR1-GFP with the dsRed-FYVE (Figure 3). The PR1 obviously does not share the secretion route with PME1 ( Figure 2D). No distinct co-localization with the ATG8 autophagy marker was observed ( Figure 2F). Even though transiently expressed proteins are found in cells affected by pathogen attack (Agrobacterium) and have increased expression of stress-related PR1 protein, this situation does not seem to lead to a distinct increase of autophagy. In the cells observed, the ATG8 signal is mostly cytoplasmic with only a few ATG8-RFP positive autophagosomal bodies that had only partial co-localization with the PR1-GFP-positive vesicles ( Figure 2F). In order to uncover the identity of the AtPR1-positive compartment, we have performed co-localization studies of full-length PR1 using red fluorescent protein (RFP)-tagged markers for ER (dsRed-HDEL; [17]), cis-Golgi marker Glycine max Golgi mannosidase-RFP (GM-RFP; [18]), trans-Golgi marker sialyl transferase-RFP (ST-RFP; donated by Ian Moore, Oxford), extracellular marker pectin methyl esterase 1-RFP (PME1-RFP; [19]), autophagy marker ATG8-RFP [20], and PI(3)P-enriched early/late endosomal compartmental marker dsRed-FYVE [21] (Figure 2A-F). PR1-GFP only weakly colocalized with the ER (Figure 2). The overlap with the Golgi marker was partial and the majority of the signal overlapped with FYVE-positive compartment ( Figure 2). In some cases, these compartments had obvious structure, typical for MVBs-the presence of particles/vesicles within the vesicle ( Figure 3). Unfortunately, our time series scanning did not reveal any events resembling exosome secretion, fusion with vacuolar membrane, nor endocytotic events (Video S2). We performed the object-based quantification of colocalization of PR1-GFP with GM-RFP, ST-RFP and dsRed-FYVE. The highest proportion of the green signal overlapping with the red one was found for the combination of PR1-GFP with the dsRed-FYVE (Figure 3). The PR1 obviously does not share the secretion route with PME1 ( Figure 2D). No distinct co-localization with the ATG8 autophagy marker was observed ( Figure 2F). Even though transiently expressed proteins are found in cells affected by pathogen attack (Agrobacterium) and have increased expression of stress-related PR1 protein, this situation does not seem to lead to a distinct increase of autophagy. In the cells observed, the ATG8 signal is mostly cytoplasmic with only a few ATG8-RFP positive autophagosomal bodies that had only partial co-localization with the PR1-GFP-positive vesicles ( Figure 2F).
The subset of markers was used for co-localization with the truncated form of PR1, PR1ΔC, which confirmed the previous observation that this version without the sorting signal contained in the C-terminus is localized on the ER (Figure 4).   The subset of markers was used for co-localization with the truncated form of PR1, PR1∆C, which confirmed the previous observation that this version without the sorting signal contained in the C-terminus is localized on the ER (Figure 4).

Localization of AtPR1 Expressed under the Native and 35S Promoter in Arabidopsis Cells
In order to describe PR1 in situ in Arabidopsis we have designed PR1-GFP and PR1ΔC-GFP constructs placed under the control of the native 2 kb promoter and used it for the stable transformation of Arabidopsis. Without induction, in seven days old seedlings, almost no GFP signal was detected, and after the SA induction of PR1 expression, in cotyledons, we could see only a very weak signal in the case of the full length and slightly stronger for the truncated version of PR1 ( Figure 5A,B). In order to check for the proteolysis/degradation of PR1-GFP, we have performed Western blot analysis using antibodies raised against PR1 and GFP ( Figure 6). We have detected high amount of endogenous PR1 (15 kDa), and significantly less of GFP tagged constructs (approximately 40 and 35 kDa for the full length and truncated version, respectively). Since the transgenic construct expression is driven by the same promoter, the fact that the expression levels differ indicates that either some distant positive transcriptional signals are missing (out of our 2 kb promoter fragment) in the transgenic PR1s, or the fused protein product is less stable than endogenous PR1. Along some unspecific bands in higher protein mass positions, there were several weaker bands of smaller proteins found with both a PR1-specific antibody that recognizes N-terminal part of PR1, as well as with an anti-GFP body which recognized the intact C-terminal part. These bands might come either from the antibody unspecificity, but also could be signs of the presence of partially proteolytically processed PR1-GFP proteins. In any case, both antibodies recognize well the full length vs. the Cterminally truncated PR1-GFP fusion versions.

Localization of AtPR1 Expressed under the Native and 35S Promoter in Arabidopsis Cells
In order to describe PR1 in situ in Arabidopsis we have designed PR1-GFP and PR1∆C-GFP constructs placed under the control of the native 2 kb promoter and used it for the stable transformation of Arabidopsis. Without induction, in seven days old seedlings, almost no GFP signal was detected, and after the SA induction of PR1 expression, in cotyledons, we could see only a very weak signal in the case of the full length and slightly stronger for the truncated version of PR1 ( Figure 5A,B). In order to check for the proteolysis/degradation of PR1-GFP, we have performed Western blot analysis using antibodies raised against PR1 and GFP ( Figure 6). We have detected high amount of endogenous PR1 (15 kDa), and significantly less of GFP tagged constructs (approximately 40 and 35 kDa for the full length and truncated version, respectively). Since the transgenic construct expression is driven by the same promoter, the fact that the expression levels differ indicates that either some distant positive transcriptional signals are missing (out of our 2 kb promoter fragment) in the transgenic PR1s, or the fused protein product is less stable than endogenous PR1. Along some unspecific bands in higher protein mass positions, there were several weaker bands of smaller proteins found with both a PR1-specific antibody that recognizes N-terminal part of PR1, as well as with an anti-GFP body which recognized the intact C-terminal part. These bands might come either from the antibody unspecificity, but also could be signs of the presence of partially proteolytically processed PR1-GFP proteins. In any case, both antibodies recognize well the full length vs. the C-terminally truncated PR1-GFP fusion versions.
Using prolonged (24 h) dark incubation in combination with SA treatment of seedlings, we managed to improve the intensity of GFP constructs signal. This way we could detect the localizations, which were in agreement with previous observations-PR1∆C-GFP on the ER and ER bodies, and PR1-GFP in a vesicles-like compartment present in the cortical cytoplasm, and also in the vacuole ( Figure 5C).   Using prolonged (24 h) dark incubation in combination with SA treatment of seedlings, we managed to improve the intensity of GFP constructs signal. This way we could detect the localizations, which were in agreement with previous observations-PR1ΔC-GFP on the ER and ER bodies, and PR1-GFP in a vesicles-like compartment present in the cortical cytoplasm, and also in the vacuole ( Figure 5C).
In order to further address the identity of PR1-positive intracellular compartment(s), we decided to use stable Arabidopsis transformation of PR1-GFP constructs driven by the 35S promoter. This allowed us to obtain a stronger signal without the need to boost the expression by SA, which in high concentrations and prolonged incubations, according to our bright field imaging, significantly affected the appearance and physiology of cells. Under the 35S promoter, in young seedlings, the AtPR1-GFP signal was well detected mainly in ER-bodies like the compartment in cotyledons, probably as an effect of overexpression, but the presence of PR1 compartments similar to those observed in N. benthamiana transient expression was also detected ( Figure 7A). The three-dimensional (3D) reconstruction performed for cotyledon epidermal cells allowed the side view/optical cross section which showed that the PR1 containing compartments are accumulated just beneath the outer surface (arrowhead, Figure 7B). Bigger bodies observable in inner parts of the cells are autofluorescent plastids. In order to further address the identity of PR1-positive intracellular compartment(s), we decided to use stable Arabidopsis transformation of PR1-GFP constructs driven by the 35S promoter. This allowed us to obtain a stronger signal without the need to boost the expression by SA, which in high concentrations and prolonged incubations, according to our bright field imaging, significantly affected the appearance and physiology of cells. Under the 35S promoter, in young seedlings, the AtPR1-GFP signal was well detected mainly in ER-bodies like the compartment in cotyledons, probably as an effect of overexpression, but the presence of PR1 compartments similar to those observed in N. benthamiana transient expression was also detected ( Figure 7A). The three-dimensional (3D) reconstruction performed for cotyledon epidermal cells allowed the side view/optical cross section which showed that the PR1 containing compartments are accumulated just beneath the outer surface (arrowhead, Figure 7B). Bigger bodies observable in inner parts of the cells are autofluorescent plastids.

Discussion
Here we analyze how the PR1 protein is transported through the cell to the extracellular space. We assayed the localization using the heterologous and transient Arabidopsis gene expression in Nicotiana. It was known before that in N. tabacum there are three acidic (1a, 1b, and 1c) and one basic (1g) PR1 isoforms that are induced upon tobacco mosaic virus (TMV) infection [2]. It is a common knowledge that the basic PR proteins are considered as vacuolar, while acidic as extracellular cargo [22]. However, the situation with PR1s charge and destination may not be that simple, especially in the case of basic AtPR1 that was reported to be secreted [6,12]. In addition, for instance in tomato, homologues of acidic tobacco PR1a, -b, and -c are basic proteins [2]. This shows that individual PR1 orthologs may have different trafficking routes and functions and at this moment, it is not clear how this might be related to their charge. Moreover, from of our results and previously published localization of PR1-RFP [8], it is obvious that one PR1 may have both vacuolar and extracellular localization. In our experiments, a small fraction of PR1 compartments co-localized with Golgi, but the signal overlap with the PI(3)P enriched compartment was clearly dominant. According to Berson et al. [23], Kim et al. [24], and Kolb et al. [25], FYVE marked compartments could be early or late endosomes. Size and morphology of the PR1-associated compartments clearly points to a late endosome identity on the way to maturation to MVBs. The PR1 protein in MVBs could be further targeted to a vacuole, but can be also secreted via exosomes. Targeting to the plasma membrane to release PR1 content into extracellular space, rather than to the vacuole, would be more consistent with the presumed extracellular function of PR1. In our experiments, however, only a weak signal is

Discussion
Here we analyze how the PR1 protein is transported through the cell to the extracellular space. We assayed the localization using the heterologous and transient Arabidopsis gene expression in Nicotiana. It was known before that in N. tabacum there are three acidic (1a, 1b, and 1c) and one basic (1g) PR1 isoforms that are induced upon tobacco mosaic virus (TMV) infection [2]. It is a common knowledge that the basic PR proteins are considered as vacuolar, while acidic as extracellular cargo [22]. However, the situation with PR1s charge and destination may not be that simple, especially in the case of basic AtPR1 that was reported to be secreted [6,12]. In addition, for instance in tomato, homologues of acidic tobacco PR1a, -b, and -c are basic proteins [2]. This shows that individual PR1 orthologs may have different trafficking routes and functions and at this moment, it is not clear how this might be related to their charge. Moreover, from of our results and previously published localization of PR1-RFP [8], it is obvious that one PR1 may have both vacuolar and extracellular localization. In our experiments, a small fraction of PR1 compartments co-localized with Golgi, but the signal overlap with the PI(3)P enriched compartment was clearly dominant. According to Berson et al. [23], Kim et al. [24], and Kolb et al. [25], FYVE marked compartments could be early or late endosomes. Size and morphology of the PR1-associated compartments clearly points to a late endosome identity on the way to maturation to MVBs. The PR1 protein in MVBs could be further targeted to a vacuole, but can be also secreted via exosomes. Targeting to the plasma membrane to release PR1 content into extracellular space, rather than to the vacuole, would be more consistent with the presumed extracellular function of PR1. In our experiments, however, only a weak signal is observed in extracellular space, which can be explained by the fact, that GFP, unlike RFP, is quenched in the acidic cell wall environment. In addition, in stable transformants, after SA induction and prolonged dark treatment, for the full-length construct we see a portion of PR1-compartments in the cortical cytoplasm, and also in intravacuolar bodies. Surprisingly, besides intracellular compartments, we observed for both PR1-and PR1∆C-GFP association with the plasma membrane, a situation so far unreported for plant PR1s, but well described for the human homolog Golgi-associated PR1 protein (GLIPR2/GAPR1) which was found to interact with the negatively charged membrane phospholipids [26]. We hypothesize that AtPR1 could also directly interact with cellular membranes. To this end, we constructed a homology model of the Arabidopsis PR1 protein (see Material and Methods, and Figure 8). Similarly to GLIPR2/GAPR1, the PR1 protein is positively charged and we found one cluster of positively charged amino acid residues (arginines 60, 67, 137 and lysine 135). These amino acid residues could interact with negatively charged phospholipids of the cellular membranes and they would orient the protein towards the lipid bilayer in such a manner that corresponds to the proposed interaction site of GLIPR2/GAPR1 [27]. Moreover, the highly positive surface of Arabidopsis PR1 and its localization into the lumen of the MVB, which is surrounded by the FYVE marker, could be explained by a direct interaction of PR1 with PI(3)P. However, this hypothesis needs to be verified by additional experiments, but our data does suggest amino acid residues could be directly involved and in the future tested by site-directed mutagenesis.
A very interesting and unexpected observation which might be related to the predicted PR1-membrane interaction is a retention of the AtPR1∆C in the ER when compared to the full length AtPR1. Export or retention from/in the ER are regulated by specific amino acid signals which are best studied in transmembrane secreted proteins (e.g., dilysine motif; [28]); much less is known about soluble or peripheral membrane proteins. It is also possible, that already in this step, the ER membrane enrichment by PI(3)P marks the ER exit sites, similarly to what was found for ER-derived autophagosome formation (PI(3)P is on the cytosolic side of the membrane) [29], or for parasitic Plasmodium protein secretion (PI(3)P is on the ER-lumen side) [30]. Interestingly, PR1∆C lacks several amino acid residues creating the potential PI(3)P-binding site (two β sheets in the predicted PR1 structure are completely missing; Figure S2) and moreover, a model of the mutated protein structure suggests that PR1∆C is more negatively charged than wild type PR1. The precise topology of PR1-lipid interactions is an important question for future research. observed in extracellular space, which can be explained by the fact, that GFP, unlike RFP, is quenched in the acidic cell wall environment. In addition, in stable transformants, after SA induction and prolonged dark treatment, for the full-length construct we see a portion of PR1-compartments in the cortical cytoplasm, and also in intravacuolar bodies. Surprisingly, besides intracellular compartments, we observed for both PR1-and PR1ΔC-GFP association with the plasma membrane, a situation so far unreported for plant PR1s, but well described for the human homolog Golgiassociated PR1 protein (GLIPR2/GAPR1) which was found to interact with the negatively charged membrane phospholipids [26]. We hypothesize that AtPR1 could also directly interact with cellular membranes. To this end, we constructed a homology model of the Arabidopsis PR1 protein (see Material and Methods, and Figure 8). Similarly to GLIPR2/GAPR1, the PR1 protein is positively charged and we found one cluster of positively charged amino acid residues (arginines 60, 67, 137 and lysine 135). These amino acid residues could interact with negatively charged phospholipids of the cellular membranes and they would orient the protein towards the lipid bilayer in such a manner that corresponds to the proposed interaction site of GLIPR2/GAPR1 [27]. Moreover, the highly positive surface of Arabidopsis PR1 and its localization into the lumen of the MVB, which is surrounded by the FYVE marker, could be explained by a direct interaction of PR1 with PI(3)P. However, this hypothesis needs to be verified by additional experiments, but our data does suggest amino acid residues could be directly involved and in the future tested by site-directed mutagenesis. A very interesting and unexpected observation which might be related to the predicted PR1-membrane interaction is a retention of the AtPR1ΔC in the ER when compared to the full length AtPR1. Export or retention from/in the ER are regulated by specific amino acid signals which are best studied in transmembrane secreted proteins (e.g., dilysine motif; [28]); much less is known about soluble or peripheral membrane proteins. It is also possible, that already in this step, the ER membrane enrichment by PI(3)P marks the ER exit sites, similarly to what was found for ER-derived autophagosome formation (PI(3)P is on the cytosolic side of the membrane) [29], or for parasitic Plasmodium protein secretion (PI(3)P is on the ER-lumen side) [30]. Interestingly, PR1ΔC lacks several On one side, our PR1 construct is present mainly in a compartment that corresponds to LE/MVB/ prevacuolar endosomal compartment (PVC), on the other side, it also colocalizes with both cisand trans-Golgi markers. It remains unsolved whether this ambivalence is an alternative sorting or a successive localization, both possibilities could be also distorted due to overexpressing PR1 conditions which prevented correct protein sorting. We believe that future work using more natural conditions for this protein, such as a pathogen presence, could bring more data on modes of trafficking and secretion of PR1.

Plant Material
The seeds of A. thaliana Col-0 (Nottingham Arabidopsis Stock Centre, University of Nottingham, UK) were surface-sterilized and plated onto a half-strength Murashige and Skoog medium (1/2 MS). The plants were propagated in vitro for 7 days (23 • C, 16/8 h), and used for treatments and/or microscopical observation, or were transferred into Jiffy tablets and cultivated in a growth chamber under the same cultivation conditions (23 • C, 16/8 h). N. benthamiana seeds were surface-sterilized and plated into Jiffy tablets and placed into a growth chamber under the same cultivation conditions (23 • C, 16/8 h).

Cloning Procedure
For preparation and cloning of At2g14610 PR1 constructs, due to the fact that it is an intronless gene, genomic DNA from wild type Col-0 Arabidopsis was used as a template for polymerase chain reaction (PCR). Primers enabling cloning into pENTR3C vectors (Thermo Fisher Scientific Inc., Waltham, MA, USA) were EcoORFPR1: 5 -AAGAATTCGATGAATTTTACTGGCTATTCTCG-3 and PR1Xho: 5 -AACTCGAGTATGGCTTCTCGTTCACATAATTCC-3 for the full length open reading frame (ORF) and the same forward but different reverse primer for the C-truncated version PR1∆Xho: 5 -AACTCGAGTTCGCAGCGTAGTTGTAGTTAGCC-3 . For cloning of PR1 with and endogenous promoter, the same two reverse primers were used, with EcoPR1: 5 -AAGAATTCGATGTTTGAGGTTGAGTACGATGG-3 as forward. Reverse primers were designed with an omitted STOP codon to allow in-frame fusion with GFP. Cloned constructs were transferred from an entry vector by a recombinant LR clonase of Gateway cloning system (Invitrogen, Thermo Fisher Scientific Inc., Waltham, MA, USA) into pGWB5 (for 35S driven expression) and pGWB4 (for endogenous promoter driven expression) vectors [31]. The constructs were used for the transformation of Agrobacterium tumefaciens GV3103 (Nottingham Arabidopsis Stock Centre, University of Nottingham, UK).

Transient Expression in N. benthamiana
Overnight Agrobacterium cultures were diluted in infiltration buffer (10 mM MgSO 4 ·7H 2 O with 100 µM acetosyringon; both from Sigma (Sigma GmbH, Erlangen, Germany) until optical density (OD) 0.1 and used for infiltration by syringe on the abaxial N. benthamiana leaf surface. The fluorescence was observed 48 h later.

Stable Transformation of A. thaliana
Arabidopsis wild type Col-0 was transformed by Agrobacterium mediated floral dip method [32].

Confocal Microscopy
For live-cell imaging, either infiltrated parts of N. benthamiana leaves or 5-7 day-old seedlings were used for observation under the ZeissLSM880 (Carl Zeiss GmbH, Jena, Germany) confocal microscope using ×63 oil immersion objective. Excitation wavelengths used were 488 nm for GFP and 561 nm for RFP and dsRed. The images were analyzed using Zen 2.1 Software (Carl Zeiss GmbH) and Fiji/ImageJ [33,34].

Treatments of Seedlings
In order to boost the expression of PR1 protein from the construct placed under the control of endogenous promoter, SA was used (250-500 µM) for at least 24 h in liquid 1/2 MS in sterile 6-multiwell plates. Seedlings/plants were further used for confocal microscopy or protein extraction. Plasmolysis was induced with 0.8 M mannitol or 1 M NaCl.

Sodium Dodecyl Sulfate-Polyacrylamide Gel Electrophoresis and Western Blot
Total protein extracts were performed from two-week old plants transformed with promoterPR1::PR1-GFP and promoterPR1::PR1∆C-GFP (after SA induction of PR1 expression) using protein extraction buffer described in Hala et al. [35]. In order to visualize 15 kDa PR1 protein, 15% acrylamide gel was used. Proteins were blotted onto a nitrocellulose membrane (quality of transfer was checked by Ponceau staining) and blocked overnight at 4 • C with 5% nonfat dry milk in PBS (137 mM NaCl, 2.7 mM KCl, 10 mM Na 2 HPO 4 , and 2 mM KH 2 PO 4 , pH 7.4) supplemented with 0.2% Tween 20. Primary antibodies anti-PR1 (Agrisera AB, Vanas, Sweeden) and anti-GFP (Roche, Basel, Switzerland) were diluted 1:1000 and incubated with the membranes for 3 h at room temperature in the blocking solution. Horseradish peroxidase-conjugated antibody (Promega, Madison, WI, USA) was applied followed by enhanced chemiluminescent ECL detection (Amersham, GE Healthcare, Chicago, IL, USA).

Homology Modelling of the PR1 Structure and the Electrostatics Calculation
The homology model of Arabidopsis PR1 and PR1∆C was built using the SWISS-MODEL portal [36] and evaluated by the ProSa-Web server [37]. Electrostatic potential of the protein was calculated by numerically solving the nonlinear Poisson-Boltzmann equation in the APBS program [38]. The VMD program [39] was used to prepare Figure 8 and Figure S2.