Proteomics and transcriptomics of the BABA-induced resistance response in potato using a novel functional annotation approach

Background Induced resistance (IR) can be part of a sustainable plant protection strategy against important plant diseases. β-aminobutyric acid (BABA) can induce resistance in a wide range of plants against several types of pathogens, including potato infected with Phytophthora infestans. However, the molecular mechanisms behind this are unclear and seem to be dependent on the system studied. To elucidate the defence responses activated by BABA in potato, a genome-wide transcript microarray analysis in combination with label-free quantitative proteomics analysis of the apoplast secretome were performed two days after treatment of the leaf canopy with BABA at two concentrations, 1 and 10 mM. Results Over 5000 transcripts were differentially expressed and over 90 secretome proteins changed in abundance indicating a massive activation of defence mechanisms with 10 mM BABA, the concentration effective against late blight disease. To aid analysis, we present a more comprehensive functional annotation of the microarray probes and gene models by retrieving information from orthologous gene families across 26 sequenced plant genomes. The new annotation provided GO terms to 8616 previously un-annotated probes. Conclusions BABA at 10 mM affected several processes related to plant hormones and amino acid metabolism. A major accumulation of PR proteins was also evident, and in the mevalonate pathway, genes involved in sterol biosynthesis were down-regulated, whereas several enzymes involved in the sesquiterpene phytoalexin biosynthesis were up-regulated. Interestingly, abscisic acid (ABA) responsive genes were not as clearly regulated by BABA in potato as previously reported in Arabidopsis. Together these findings provide candidates and markers for improved resistance in potato, one of the most important crops in the world.


Background
Potato is today the third largest food crop in the world. One of the largest threats to potato production is late blight disease, caused by Phytophthora infestans, which is currently controlled by frequent use of fungicides. Extensive efforts to find more efficient weapons against late blight are ongoing worldwide. One alternative could be integration of induced resistance (IR) in current management strategies. IR, i.e. exposing plants to abiotic or biotic stress leading to improved resistance to subsequent pathogen attack both locally and systemically [1], is thought to function by activating or enhancing defence signalling [2]. This signalling is suggested to depend either on salicylic acid (SA) or jasmonic acid/ethylene (JA/ET) for the two classical forms of IR: systemic acquired resistance (SAR) and induced systemic resistance (ISR), respectively [3]. Resistance can be induced by treatment of plants with the xenobiotic amino-acid derivative dL-β-aminobutyric acid (BABA) [4,5]. BABA can enhance resistance in a wide range of plants against several types of pathogens, e.g. treatment increases the resistance of potato against P. infestans [4].
We have earlier shown by microscopy that a direct activation of defence responses associated to the formation of HR-like lesions takes place at efficient BABA-IR in the potato Phytophthora system [6]. This is in contrast to a "priming effect" proposed in other plant-pathogen systems [7]. The molecular mechanisms behind BABA-IR remains unclear and a better understanding on the molecular level could lead to the identification of biomarkers which in the future can be implemented in plant protection or breeding strategies.
Two microarray studies investigating BABA-IR have been conducted in Arabidopsis. Tsai et al., 2011 [8], reported that SA-responsive genes were both directly up-regulated by BABA treatment and up-regulated after priming by BABA and subsequent Pseudomonas syringae pv. tomato DC3000 infection. They also showed that BABA inhibits the Arabidopsis response to the bacterial phytotoxin coronatine, thus suppressing the coronatineassociated JA responses. The second microarray study of BABA-IR in Arabidopsis conducted by Zimmerli et al., 2008 [9] showed that BABA priming for increased thermotolerance involved ABA-associated transcription factors. They also showed that neither ET nor SA played a part in the acquired thermotolerance. These two microarray studies in Arabidopsis demonstrate the complexity of BABA-IR mechanisms. Furthermore, BABA has been shown to reduce vegetative growth and to cause major alterations in plant amino acid balance in Arabidopsis, effects that were reversed by L-glutamine treatment [10,11]. Based on these results, it was suggested that BABA provokes a stress-induced morphogenic response (SIMR) in Arabidopsis by activation of a general amino acid inhibition response. Another report indicates that the lectin-receptor kinase VI.2 is required for BABA priming of the PTI response and for BABA-IR against Pseudomonas syringae DC3000 in Arabidopsis [12].
In potato, a very important host for many pathogens such as P. infestans, the only previous RNA study relating to BABA was a time-series study using cDNA-AFLP [13]. Around 60 potato transcripts were found to be differentially expressed after BABA treatment and a clear overlap of genes affected by BABA and Phytophthora inoculation was found. Studies in transgenic NahG expressing potato plants, with an impaired SA metabolism, indicate that BABA-IR against P. infestans is dependent on SA [14]. BABA has also been shown to cause a direct activation of basal defence responses such as reactive oxygen species, hypersensitive response-like spots, phenolics and PR-1 expression in potato [6]. To elucidate molecular defence responses activated by BABA in potato, a genome-wide microarray analysis based on the recently released Solanum tuberosum group phureja genome [15] was performed in combination with labelfree quantitative proteomic analysis of the apoplast secretome. Earlier reports showed that mRNA regulation in mammalian cells only explains about 40% of the changes in protein abundance [16] and the apoplast is a major interaction arena between plants and pathogens.
Genome-wide analysis is often hampered by incomplete and outdated functional annotations of genes and derived microarray probes. To this end, we present a network-based approach to retrieve information from orthologous gene families across 26 sequenced plant genomes, which increased the number of annotated probes by close to 9000.

Plant material and growing conditions
Potato plants (cv. Desiree) were grown under controlled conditions with 16 h/8 h day and night regime with a fluorescent lamp light intensity of 200 μmol m −2 s −1 . The temperature was set to 20°C and the relative humidity was 65%. Pots were circulated every week to avoid positional effects. All sampling was done before flowering and only fully expanded leaves from the middle part of the plants were used. Plants were treated with BABA or water 2 days before harvest. For the treatments plants were randomly chosen, labeled and then placed back in the growth chamber evenly dispersed between treatments. For the detached leaf assays we used 29-day-old plants and for the molecular analysis we used 35-day-old plants.

BABA treatment
Plants were treated with BABA (DL-3-aminobutyric acid, Sigma) dissolved in water at a concentration of 1.0 or 10 mM. BABA was sprayed on the plants until saturation (ca 40 mL) and control plants were sprayed with water as earlier described [6].

Phytophthora inoculation and measurement
Four detached leaflets from each of three Desiree plants were infected with P. infestans two days after treatment with either 10 mM BABA or water (control). P. infestans strain SE03058 isolated in mid-Sweden (mating type A1, virulence 1, 2, 3, 4, 7, 10, 11) was used with a sporangial concentration of 36,000 sporangia ml −1 . The inoculated leaflets were kept in the dark at 15°C during the first day. On the second day the temperature was changed to 20°C with 16 h light/8 h dark. The size of lesion diameter was measured 7 days post inoculation (dpi). Data from two separate experiments were combined for a total of 18 biological samples. The difference in lesion sizes between the treatments were analysed by Student's t-test with Minitab v.16.

Gene ontology annotation and enrichment analysis
Genome annotation is commonly done by orthologous inference, the transfer of annotation across genomes based on orthologous relationships between genes in different genomes. There are a number of different algorithms available to determine orthologous relationships [17]. OrthoMCL is an orthology detection method which clusters genes into orthologous gene families based on sequence similarity [18]. However, the OrthoMCL software can currently not be applied to large datasets. Thus a parallel version of OrthoMCL (Parallel-OrthoMCL) was developed in house in order to allow the OrthoMCL algorithm to be applied to large datasets. Orthologous gene families across 26 plant species were constructed using Parallel-OrthoMCL. 25 plant species available from PLAZA [19] were included in the analysis and the protein translations of the longest representative transcript from the Solanum tuberosum group phureja genome was downloaded from the Solanaceae Genomics Resource [ftp://ftp.plantbiology.msu.edu/pub/data/SGR/GO_ annotations/]. The OrthoMCL percent match parameter was set to 50% and the MCL inflation value was set to 2.0. These are values commonly used by other genome annotation pipelines such as PLAZA [19]. Once gene families had been constructed, potato genes were annotated using a network-based annotation as follows: (i) all available GO annotations in PLAZA were mapped onto the appropriate genes in each of the 25 species contained in PLAZA; (ii) each potato gene was then assigned the GO terms of all the other genes from the 25 other species that were present in its specific gene family: (iii) the probes from the potato microarray (Agilent JHI Solanum tuberosum 60 k v1) were then annotated by assigning to each probe the GO terms of the genes corresponding to that probe as determined by BLASTing the probe sequences against the nucleotide transcript sequences from the potato genome. The probes of the potato microarray were also annotated using the potato gene GO annotation file downloaded from the Solanaceae Genomics Resource (SGR annotation). As described above, probes were annotated by assigning each probe the GO terms of the genes corresponding to that probe. The two microarray probe annotations were compared to identify the number of probes annotated in each annotation and also to identify the distribution of the number of GO terms assigned to each probe.

RNA sample preparation and microarray analysis
In total one leaflet from each of three plants per treatment was used for RNA analysis. Five leaf discs of 10 mm diameter were punched out of each leaflet, pooled, and immediately frozen in liquid nitrogen. The frozen samples were homogenized for 30 s at 30 rpm in a Retsch Mixer Mill MM 200 (Retsch GmbH, Haan, Germany) followed by extraction of RNA, DNase treatment and purification using RNeasy Mini kit (Qiagen GmbH, Hilden, Germany). RNA concentration and purity (260/ 280 nm >1.8) was checked by a ND-1000 NanoDrop (Wilmington, USA) and integrity of the samples were analyzed with an Experion™ Automated Electrophoresis System (Bio-Rad Laboratories, Hercules, USA). Each sample was diluted in RNase-free water to a concentration of 200 ng/μl and mRNA expression analysis was done with a custom-made Agilent expression array based on the predicted transcripts of the potato genome (version 3.4). The complete microarray design is available in ArrayExpress (E-MTAB-1655). The arrays were run according to the instructions and recommendations of the supplier (Agilent). The probe intensities were background corrected and normalised using the quantile method in the Limma R-package [20]. Fold changes and standard errors were obtained by fitting a linear model to each gene expression values and standard errors were smoothed based on the empirical Bayes method. Genes with p-values below 0.05 after adjustment (Benjamini-Hochberg) were regarded as significant. The microarray data was deposited in ArrayExpress, accession number: E-MTAB-1545. Significantly differentially expressed genes were analyzed for Enrichment of Gene Ontology Terms in GOEast using default settings on the basis of a custom made GO enrichment list for the microarray. To give an overview of all significantly enriched GO terms ReviGO was applied to cluster terms based on semantics [21]. Based on the potato genome (version 3.4) and manual curation, genes and proteins were functionally characterized by the MapMan mapping file.

Secretome extraction and visualisation
Two days after BABA treatment, three biological replicates in the form of separate plants were subjected to analysis. Four leaflets per plant were harvested on ice. Secretome proteins were extracted by vacuum infiltration with a phosphate buffer directly after harvest followed by centrifugation as described by Ali et al., 2012 [22] and Alexandersson et al., 2013 [23]. For visualisation of the apoplastic proteins, 30 μl of the samples were separated on 14% SDS polyacrylamide gels. The samples were mixed with 6 × SDS sample buffer containing DTT and denatured at 93°C for 3 min prior to electrophoresis. The gels were run for 20 min at 80 V followed by 1.5 h at 100 V and staining overnight in Coomassie Brilliant Blue (10% (v/v) acetic acid, 45% (v/v) methanol, and 0.25% (w/v) Coomassie R-250). Finally, the gels were briefly washed in dH 2 O before 3 h in destaining solution (10% (v/v) acetic acid and 40% (v/v) methanol).

Identification and quantification of proteins by LC-MS/MS
For identification of the apoplastic proteins, 30 μl of the secretome samples were separated with SDS-PAGE as described for the visualisation of proteins, but the gel was only run 10 mm. The gels were stained with Coomassie Brilliant Blue and the stained portions cut into 1-2 mm 2 pieces. Tryptic digestion of the gel pieces and peptide elution was carried out as described by Ali et al., 2012 [22]. The eluted peptides were concentrated to 50 μl by rotary evaporation and stored at −80°C. Six μl of sample was subjected to HPLC-MSMS analysis using an Eksigent nanoLC2D HPLC system coupled to an Orbitrap XL. The peptides were loaded onto a pre-column (Agilent Zorbax 300SB-C18, 0.3 mm ID, 5 mm, 5 μm particle size) connected to an analytical column (Agilent Zorbax 300SB C18, 75 μm ID, 150 mm, 3.5 μm particle size). The analytical column was pre-equilibrated for 10 min using buffer consisting of 0.1% formic acid (FA), 5% acetonitrile (ACN) at a flow rate of 10 μl/min, and the peptides were separated in an 0.1 FA buffer using a 55 min linear gradient from 5% to 40% ACN followed by a 5 min linear gradient from 40% to 80% ACN at a flow rate of 350 nl/min. The eluted peptides were analysed using an LTQ Orbitrap, operated in data-dependent mode to automatically perform Orbitrap-MS and LTQ-MSMS analysis. Survey scan spectra (400-2000 Da) were acquired using the Orbitrap mass analyzer with the resolution R = 60000. Automatic gain control was enabled. The seven most intense ions were selected for fragmentation in the LTQ, using a mass window of 2 Da for precursor ion selection. The precursor ions were fragmented with normalised collision energy of 35 (with activation Q set to 0.25 and an activation time of 30 ms). Dynamic exclusion was enabled with a repeat count of 2, a repeat duration of 20 seconds, exclusion duration of 120 seconds, an exclusion list size of 499 and a 10 ppm exclusion mass width relative to both low and high. This raw data was then converted to mgf files using ProteoWizard [24]. The Proteios software environment [25] was used to search the files with Mascot against a database consisting of the protein sequences from the potato genome and all Solanaceae sequences from UniProt extended with an equal size random part, with conserved protein length and amino acid distribution for the random part. Search tolerances were set to 5 ppm for MS and 0.5 Da for MS/MS. One missed cleavage was allowed. Carbamidomethylation of cysteine residues was selected as a fixed modification and oxidation of methionine residues was selected as a variable modification. Proteios was used to identify hits at an estimated peptide spectrum match false discovery rate (FDR) of 0.01 using the random database mentioned above. For quantitative analysis, a label-free approach based on precursor intensities was used [26], with msInspect feature detection [27], and Proteios alignment as described previously [28]. Quantitative data from different conditions was normalised using a linear transformation procedure where the intensity of each feature was divided by the sum of all features in the same run and then multiplied by the average of the summed intensities of all runs for that replicate. Missing values were set to 1. This data was then log-transformed and analysed using an ANOVA based approach in the Qlucore software (www.qlucore.se). Peptides that were differentially abundant were identified by performing a twogroup comparison Qlucore with a FDR < 0.1 according to the Benjamini-Hochberg procedure for determining adjusted p-value, q. Only peptides uniquely identifying one protein in the potato genome (version 3.4) were used in order to avoid ambiguous peptides identifying several similar proteins.

P. infestans growth after BABA treatment
To find out if BABA at a concentration of 10 mM could restrict P. infestans growth in potato cv. Desiree, a detached leaf assay was performed. BABA led to significantly reduced lesion sizes of P. infestans in Desiree at 7 dpi (p = 0.012; Figure 1). Small necrotic lesions were visible on the leaves after treatment with BABA at 10 mM but not 1 mM (not shown). These were not observed in the controls, where leaves were treated with water. BABA at 1 mM was not tested since we earlier demonstrated a threshold level of 2.5 mM BABA for successful restriction of P. infestans growth in potato [29].

New network-based annotation of potato genes
Network-based annotation of the potato microarray probes and PGSC gene models applying parallelised OrthoMCL across 26 sequenced plant genomes was compared to the annotation of the potato microarray probes provided by Solanaceae Genomics Resource (SGR). Of the available ortholog detection algorithms, OrthoMCL has previously been shown to have the best performance (and the lowest false positive rate) in ortholog detection across multiple genomes [17]. As can be seen from Figure 2, 8616 additional probes out of a total of 45126 receive an annotation based on the 26 genomes with the SGR annotation. The number of ontology terms assigned to a probe is one of the indicators of the amount of information linked to the probe. In comparison to the earlier annotation, the number of GO terms assigned per probe was much larger with a median of 34 GO terms assigned per probe for the new annotation versus 6 for the SGR annotation (Additional file 1: Figure S1). Thus, the new annotation using gene families constructed from 26 plant genomes provides a more complete annotation of the potato microarray than was possible with the annotation file provided by the SGR.

Overall transcript and protein changes
In the microarray analysis, 5378 transcripts were significantly (adj. p-value <0.05) differentially expressed in potato leaf 48 h after treatment with 10 mM BABA (Table 1). In the corresponding apoplast secretome samples, 91 proteins (50 up-and 41 down-regulated) were found to change in abundance (Table 1). BABA applied at the lower concentration (1 mM) had almost no influence on gene expression. Only six differentially expressed transcripts were identified (1 up-and 5 down-regulated), in contrast to 24 apoplastic proteins (11 up-and 13 down-regulated) with a significantly changed abundance (Tables 1 and 2). The only transcript up-regulated by 1 mM BABA, a MutT domain protein (G400017400), was also up-regulated by 10 mM BABA.

Gene ontology analysis and gene expression after 10 mM BABA treatment
With more than 5000 transcripts affected by 10 mM BABA, it is evident that many functional mechanisms were altered. In order to identify processes where gene expressional changes occur, a gene ontology (GO) enrichment  analysis was done based on the extended functional annotation described above. The 15 most enriched GO terms are described in Table 3. In all, a large number of GO terms (>500) were significantly enriched (p <0.1; Additional file 2: Table S1). To give an overview of all GO terms we used ReviGO, which clusters GO terms based on semantic similarities [21]. This analysis shows two major clusters of terms associated to stress response and metabolism as well as two minor clusters related to development and metabolism ( Figure 3).
In Table 4, probes corresponding to the 50 most upand down-regulated microarray transcripts were divided into functional categories. A full list of significant differentially expressed genes after 10 mM BABA treatment is included in Additional file 3: Table S2. Enriched GO terms after BABA treatment included processes associated with induction of plant defence such as plant-type hypersensitive response, necrotic spotted lesions, cell surface receptor signalling pathways, innate-immune response, defence response to fungus, and systemic acquired resistance (SAR). Examples of genes linked to these processes were two NDR1 homologs (G400032148, G400032231), two EDS1 splice variants (G400033029), two phytoalexin-deficient 4-2 protein splice variants (G400019873), several ethylene receptor splice variants (e.g. G400031819) and PR1 homologs (e.g. G400005112). When categorising genes to cellular compartments one of the most enriched GO terms is cytoplasmic membranebound vesicle (Table 3). This category included KiTH-2 (G400008097-G400008100) and Wound-induced protein WIN1 (G400019435) similar to AtPR4, which were among the most up-regulated transcripts after BABA treatment (Table 4). Among these genes was also the potato PEN1 (G400021331) homolog. WIN1 together with an Nb cell death marker (G400010131), osmotin (PR-5; G400003057), and Nectarin 5 (G400031325) were among the four most up-regulated transcripts involved in stress response. One Kunitz proteinase inhibitor and one serine protease inhibitor are among the most down-regulated transcripts related to stress response (G400009512, G400010128). There were no differentially expressed genes among the 17 genes annotated as R-genes on the array. However, the expression of some other LRR transcripts changed significantly. A TIR/NBS/LRR protein (G400001756) is up-regulated 8 times and three NBS-LRR type proteins (G400015872, G400027797, G400043650) are up-or down-regulated about two times. Only subtle expression differences were seen among other members in this large gene family.
Significant enrichments were found for genes associated with the hormones ethylene, ABA, SA, JA and cyclopentenone, but of the processes related to hormone metabolism, only gibberellin biosynthesis was enriched. In this pathway, several splice variants of copalyl diphosphate synthases, which act early in biosynthesis of gibberellins, were sharply up-regulated (Additional file 2: Table S1 and Additional file 3: Table S2). Furthermore, several gibberellin 20-oxidases (G400036251, G400039968, G400040776, G400036003), which catalyze the conversion of different gibberellin forms, as well as gibberellin 2-oxidase 1, which is able to de-activate gibberellins [30], were down-regulated by BABA, except for one transcript (G400015796) that was slightly up-regulated. In addition, several gibberellic acid insensitive dwarf (GID) proteins (G400022493, G400022492, G401008665, G400018153, G400015446) which are GAreceptors regulating the stability of DELLA proteins, were also affected by BABA treatment, and two splice variants of DELLA proteins (GAI) (G400019212) were up-regulated.  Total number differentially expressed transcripts (adj. p-value <0.05) and proteins (FDR < 0.1) with altered abundance in leaves (cv. Desiree) treated with BABA (1 and 10 mM). Samples were analysed two days after treatment of the leaf canopy with BABA. Figure 3 ReviGO analysis of enriched GO terms. Significantly enriched GO terms were clustered by semantic similarities and four overall clusters pointed out.   In JA-regulation, two Jaz proteins were oppositely regulated, with transcripts representing Jaz1 (G400002930) up-regulated and Jaz3 (G400032119) down-regulated. Several splice variants of MYC1 (G401010822) and MYC2 (G404010822 G400005525, G402010822) transcription factors were down-regulated, whereas two splice variants of a JA-induced WRKY (G400019824) were up-regulated. JAZs act as negative regulators of the JA signalling pathway by binding to MYC transcription factors [31]. In JA biosynthesis (although not significant in the GO analysis), allene oxide cyclase (AOC) (G401012679), was down-regulated, whereas splice variants of lipoxygenase (LOX) (G400010859, G400007793) and Acyl-coenzyme A oxidase (G400010548, G400020620) were up-regulated.
In the brassinosteroid (BR) pathway, several brassinosteroid insensitive 1-associated receptor kinase 1 s (BAKs) were up-regulated. Among the 50 most down-regulated are transcripts encoding two forms of the DWARF1/ DIMINUTO protein (G400021142) which are involved in early brassinosteroid biosynthesis. The Arabidopsis DWARF1 knock-out mutant has stunted growth and an altered secondary cell wall [33].
Treatment with 10 mM BABA also affected the regulation of several transcripts involved with amino acid metabolism e.g. related to proline, arginine and glutamine biosynthesis in the glutamate pathway. The most profound up-regulation was found in transcripts for biosynthesis of aromatic amino acids, e.g. chorismate mutase 1 (G400001438), arogenate dehydratase (G400007122) and arogenate dehydrogenase 1 (G400030683, G400020334) involved in phenylalanine and tyrosine biosynthesis [34]. Several transcripts of tryptophan decarboxylase (TDC) (G400018359, G400018358, G400018357, G400005444, G400035067) and one transcript of strictosidine synthase (G402017288) were highly up-regulated.
BABA also seemed to influence genes involved in cell wall structure, e.g., several cellulose synthase genes were slightly up-regulated whereas one was found among the 50 most down-regulated transcripts (Table 4).

Secretome SDS-PAGE and LC-MS/MS analysis
The quality of the apoplast samples were checked by separation on SDS-polyacrylamide gels prior to protein identification by mass spectrometry (Figure 5). A thick band previously identified by us as PR-1 [35] was found after application of 10 mM BABA. A faint PR-1 band was seen after 1 mM BABA whereas no PR-1 band was detected in the water control.

Novel functional annotation and overall changes
In this paper we provide the potato research community with a more comprehensive functional annotation of the potato PGSC gene models and potato microarray probes (Agilent JHI Solanum tuberosum 60 k v1; Additional files 4 and 5). This new functional annotation was used for the interpretation of the BABA-IR response. BABA at a concentration of 10 mM alters the expression of over 5000 transcripts and secretion of over 90 proteins to the apoplast. This is in contrast to the finding that only six transcripts (1 up-and 5 down-regulated) and 24 proteins (11 up-and 13 down-regulated) were affected by 1 mM BABA. This may explain the results from an earlier study by Liljeroth et al., 2010 [29], where 5 mM BABA was required to affect P. infestans infection rate, whereas BABA applied at concentrations below 2.5 mM failed to restrict P. infestans growth in the three tested potato cultivars. In agreement with this, we found that treatment of potato cv. Desiree with 10 mM BABA has a clear inhibitory effect of P. infestans infection. Whether the differences seen after treatment with 1 mM BABA, especially in the apoplast, lead to an increased resistance against other pathogens still has to be tested.
Furthermore, we can confirm the up-regulation of family members of proteinase inhibitor type-2:s, ubiquitinconjugating proteins, glutathione S-transferases, lipoxygenases, histone H3.2 and a sesquiterpene synthase 2 previously identified by Li et al., 2009 [13] who reported  The only up-regulated transcript by 1 mM BABA, a MutT/nudix domain protein, was also up-regulated by 10 mM BABA. Several MutT/nudix were also found to be up-regulated after BABA treatment in Arabidopsis [9]. This family is not well characterised functionally, but some members use their hydrolase activity to maintain cell homeostasis and have been implied to play a role in plant defence. For example, over-expression of AtNUDX2 resulted in increased pyrophosphatase activity toward ADP-ribose and increased tolerance to oxidative stress [36] and AtNUDX7 has been identified as an SA-independent factor in Enhanced Disease Susceptibil-ity1 (EDS1)-controlled plant defence and programmed cell death [37]. In line with these observations, we found an ADP-ribose polymerase 2 and several other ADPribosylation factors as well as EDS1 to be transcriptionally up-regulated by 10 mM BABA in potato.
When comparing the transcripts and apoplastic protein abundance after 10 mM BABA treatment, regulation of 13 of the 70 proteins with peptides matching the predicted potato genome models overlapped. The samples are taken two days after treatment with BABA in both the transcript and protein analysis, but the overlap might have been greater if the RNA samples were taken earlier since there is an expected time lag between transcript and protein synthesis. There are also differences in transcript and protein stability. Another reason for these differences might be that protein abundance in the apoplast is at least partially regulated by protein transport and vesicle trafficking, and thus not directly dependent on transcript levels. Cytoplasmic membrane-bound vesicle was the most enriched GO term and several transcript encoding proteins in vesicle trafficking, such as clathrin assembly protein, Epsin-2 and PEN1 were all found to be between 2.5-and 11-fold up-regulated. The low overlap between transcript regulation and apoplast protein abundance confirms the value of combining the two methods.

Immune responses and sterols
Several typical stress-response genes up-regulated by BABA in potato (e.g. PR-1, PR-2, PAD4, thaumatin, aspartyl protease family protein) have previously been found to be up-regulated in an Arabidopsis microarray study, where 46 genes in total were up-regulated about 3 days after treatment with 0.25 mM BABA [8]. In the present study, no known annotated R-genes and only a few other NBS-LRR proteins were up-regulated by BABA, which might indicate that BABA-IR is not mediated by increased expression of these genes through R-gene recognition and effector-triggered immunity (ETI). Instead most of the upregulated genes and proteins affected by BABA were PRproteins and components involved in hormone signalling and defence signalling such as PAMP triggered immunity, PTI. The up-regulated potato PEN1 homolog, which encodes membrane associated syntaxin, is involved in PTI, where it plays a part in the vesicle-based secretion system [38]. EDS1, PAD4 and SAG101, genes implicated to play a part in non-host resistance [39], were also up-regulated by 10 mM BABA. However, none of them have been shown to be required for the defence against P. infestans in Arabidopsis [40][41][42]. Another gene among the four most upregulated transcripts involved in stress response encodes osmotin (PR-5). This protein might play a role in BABA-IR to drought stress, as its abundance was increased in the leaf proteome of crabapple seedlings that were treated with BABA prior to drought stress exposure [43].
In a recent study, Kopischke et al. [44] reported that sterol balance plays a role in Arabidopsis non-host resistance to P. infestans. In their study, the erp1 mutant lacking sterol ester accumulation experienced increased cell death. Oomycetes lack the ability to synthesize sterols, which are essential for their reproduction, and therefore depend on their hosts for sterol compounds [45]. The sterol biosynthesis within the mevalonate pathway was clearly down-regulated in the present study, whereas the biosynthesis of another branch of the pathway, namely sesquiterpene phytoalexin, was up-regulated ( Figure 3). These results are in line with earlier findings in potato, where wound-induced sterol and steroid glycoalkaloid synthesis are suppressed in favour of sesquiterpenoid phytoalexin biosynthesis when potato tubers are exposed to elicitors or pathogens that induce HR [46,47]. This redirection of the isoprenoid biosynthetic pathway has been linked to changes in expression of 3-hydroxy-3 methylglutaryl-coenzyme A reductase (HMGR) genes, where hmg2 and hmg3 are up-regulated whereas hmg1 is suppressed [48,49]. In the present study two HMGRs were up-regulated by 10 mM BABA. In a study conducted by Ros et al. [50], sesquiterpene synthase and vetispiradiene synthase were up-regulated after infection with P. infestans in two field-resistant cultivars, Bettina and Indira. It had earlier been shown that arachidonic acid, a fatty acid present in the lipids of oomycetes, can suppress steroidglycoalkaloids [51] and act as an elicitor of sesquiterpene phytoalexins [52] with the involvement of LOX [53,54] in potato tubers. LOX is also among the 50 most upregulated transcripts by 10 mM BABA in this study. However, BABA-IR in potato has been shown to require SA but not 9-LOX-derived oxylipins [14].
When comparing the present secretome data with that obtained from three genotypes, SW93-1015, Sarpo Mira (P. infestans incompatible) and Desiree (P. infestans compatible), 24 and 72 hpi with P. infestans (Ashfaq Ali, personal communication), patatin 3 (P400017707) increased in abundance after BABA treatment as well as infected SW93-1015 and Sarpo Mira, but not in susceptible Desiree. This is in line with previous findings that tobacco NTPAT1-3 is involved in cell death during HR [55], since HR are formed in SW93-1015 and Sarpo Mira but not Desiree. A role for patatin-like proteins in cotton HR has also been proposed [56].

Amino acid metabolism
Earlier studies in Arabidopsis have revealed that BABA treatment alters amino acid composition and can be linked to an amino acid associated alternation of growth [10,11]. The BABA-mediated growth reduction and IR to Pseudomonas syringae pv. tomato DC3000 was compromised by treatment with L-glutamine [11]. The amino acid L-serine and the stress-responsive amino acids proline and asparagine were reported to be induced by BABA in Arabidopsis, whereas the aspartic acid content decreased . Transcripts involved in the proline biosynthesis were also observed after treatment with 10 mM BABA in the present study. Recently, a study in Arabidopsis revealed decreased abundance of several amino acids, including serine and glutamic acid, after treatment with potassium phosphite [57], and of aspartic acid after BABA treatment. However, whether a BABA-induced altered amino acid balance leads to changed leaf morphology in potato similar to those seen in Arabidopsis, remains to be investigated.
Another up-regulated transcript was arginine decarboxylase, which is involved in the biosynthesis of the polyamine putrescine by catalyzing the synthesis of agmatine from arginine. Polyamines are involved in a wide range of stress responses, partly as reactive oxygen scavengers [58]. Arginine decarboxylase has previously been reported to increase systemically in tobacco leaves after BABA treatment [59], which suggests a role for polyamines in BABA-IR.

Hormones
SA has previously been shown to be important for basal defence against P. infestans and required for BABA-IR in the same potato pathosystem [14,60]. In the SA signalling pathway, transcripts of a phenlyalanine ammonia lyase (PAL) homolog were two fold up-regulated, whereas isochorismate synthase (ICS) did not change significantly. ICS, in contrast to PAL, is directly involved in the SA biosynthesis and has been shown to be upregulated after BABA-treatment of Arabidopsis [8,9]. In contrast to SA, genes involved in JA, ET, GA and ABA biosynthesis were affected by BABA, and GO term analysis pointed to an enrichment of genes in GA biosynthesis.
In the present study, two splice variants of DELLA proteins, involved in GA signalling, were up-regulated. DELLA proteins play a role in resistance to necrotrophs and susceptibility to biotrophs by modulating the SA/JA signalling [61]. This modulation is accomplished by competitive binding of DELLAs to JAZ proteins, which prevents interaction between JAZ and MYC2, thus promoting MYC2-induced responses such as LOX [61]. Navarro et al. (2008) further suggested that GA might change this balance by degradation of DELLAs, thus promoting resistance to biotrophs and susceptibility to necrotrophs. Thus, it would be interesting to further investigate the role of DELLAs in BABA-IR in potato.
9-cis-epoxy-carotenoid dioxygenase 2 (NCED2), which is involved in the biosynthesis of, for example, ABA, was up-regulated. In a previous study in Arabidopsis, BABA treatment was shown to lead to the accumulation of several ABA responsive elements and it was suggested that crosstalk occurs between BABA and ABA signalling [9]. The ABA responsive transcription factors, identified as BABA-induced in Arabidopsis e.g. ABI1, 2 and 3, were not differentially expressed in BABA-treated potato. Instead, the only ABA-related transcription factor we could identify as differentially regulated, a bZIP, was repressed after BABA treatment.
Several genes related to BRs and auxins, which can affect plant development and cell wall formation, changed expression after treatment with BABA e.g. TIR1 and several auxin-response factors (ARFs), but not ARF5, were down-regulated. Repression of auxin-related processes has repeatedly been observed in defence signalling. In Arabidopsis, Wang et al. [62] found that a number of auxin responsive genes were down-regulated after treatment with the SA analog benzothiadiazole S-methyl ester (BTH). With some exceptions, e.g. auxin-induced SAUR and auxin responsive factor 5 (ARF5), this is also true in our data. Furthermore, the over-expression of a rice IAA-amino synthetase, GH3, led to an enhanced disease resistance to the rice pathogen Xanthomonas oryzae, possibly due to down-regulation of expansins, leading to decreased loosening of the cell walls [62,63]. However, we see a clear up-regulation of several GH3 isoforms, but no significant change in expression of expansin transcripts in potato.

Secreted proteins
Data obtained from the analysis of the apoplastic proteins revealed a high number of proteins related to defence and metabolism in the secretome of BABA-treated leaves. Several pathogenesis-related (PR) proteins were found to be highly abundant following BABA treatment. These proteins are used as markers for pathogen response and are induced by different biotic and abiotic stresses [64]. Among these were an acidic class II 1, 3-beta glucanase (PR-2) and thaumatin. PR-2, but not thaumatin, also showed increased abundance after BABA treatment of drought-stressed crabapple seedlings [43]. Another highly induced protein of less known biological function was the alginate-lyase motif containing citrate binding protein not present in the Arabidopsis genome [65]. The transcript encoding the citrate binding protein was also among the most up-regulated and has been described to be induced in potato after inoculation with P. infestans [50].
A cell wall peroxidase was highly induced after BABA treatment. These are known to be involved in cell wall strengthening following stress challenge and function by combatting degradation enzymes produced by pathogens [66]. Furthermore, peroxidases are involved in a number of defence-related responses such as HR, cross-linking of phenolics, and phytoalexin production. Peroxidases can be induced by different stresses and some plant defence activators including BABA [67]. A Snakin (Sn-1) protein was also highly up-regulated. Snakin proteins are known for their antimicrobial role in plant defence [68,69]. In addition, Sn-1 was recently found to play a role in leaf primary metabolism and cell wall composition in potato, suggesting a possible role in plant growth and development [70]. Samples were analysed two days after treatment of the leaf canopy with BABA. Proteins with significantly changed abundance (FDR < 0.1) are included in the table.