Loss-of-function mutation of REDUCED WALL ACETYLATION2 in Arabidopsis leads to reduced cell wall acetylation and increased resistance to Botrytis cinerea.

Nearly all polysaccharides in plant cell walls are O-acetylated, including the various pectic polysaccharides and the hemicelluloses xylan, mannan, and xyloglucan. However, the enzymes involved in the polysaccharide acetylation have not been identified. While the role of polysaccharide acetylation in vivo is unclear, it is known to reduce biofuel yield from lignocellulosic biomass by the inhibition of microorganisms used for fermentation. We have analyzed four Arabidopsis (Arabidopsis thaliana) homologs of the protein Cas1p known to be involved in polysaccharide O-acetylation in Cryptococcus neoformans. Loss-of-function mutants in one of the genes, designated REDUCED WALL ACETYLATION2 (RWA2), had decreased levels of acetylated cell wall polymers. Cell wall material isolated from mutant leaves and treated with alkali released about 20% lower amounts of acetic acid when compared with the wild type. The same level of acetate deficiency was found in several pectic polymers and in xyloglucan. Thus, the rwa2 mutations affect different polymers to the same extent. There were no obvious morphological or growth differences observed between the wild type and rwa2 mutants. However, both alleles of rwa2 displayed increased tolerance toward the necrotrophic fungal pathogen Botrytis cinerea.

Plant cells have semirigid cell walls composed of polysaccharides and proteins. Some cells form a secondary wall after cessation of extension, and such walls usually contain lignin as well. In both primary and secondary walls, polysaccharides including cellu-lose, hemicelluloses, and pectins constitute the major components. Hemicelluloses are a diverse group of polysaccharides that include xylans, xyloglucans, (gluco) mannans, and mixed-linkage glucans (Scheller and Ulvskov, 2010). Pectic polysaccharides are acidic and have complex structures that are characterized by a high content of GalUA residues (Mohnen, 2008).
Many cell wall polysaccharides are esterified with O-acetyl substituents to varying extents. Acetylation is commonly found in the backbone of mannans and xylans and in the side chains of xyloglucans (Scheller and Ulvskov, 2010). Acetylation of the GalUA residues of pectin is found in both homogalacturonan and rhamnogalacturonan I (RGI; Ishii, 1997;Mohnen, 2008), and acetylation of rhamnosyl residues of RGI has been reported (Sengkhamparn et al., 2009). Extraction and separation of cell wall polymers often involves alkaline treatment, which removes acetate esters. Therefore, it is conceivable that other polymers also contain acetyl esters in their native state. The biological significance of polysaccharide O-acetylation is not known, but it is known that acetylation of pectin affects its physicochemical properties in vitro. Highly acetylated pectins have poor gelling properties and limited value as functional food ingredients, unlike, for example, less acetylated pectins from citrus fruit (Ralet et al., 2003). Xylans in secondary cell walls, particularly in woody species, can be highly acetylated, and this affects the physicochemical properties of the xylans in the wall (Ebringerová et al., 2005). Acetylation of hemicelluloses, particularly xylans, has implications for the use of biomass as feedstocks for biofuel production and other industrial processes. Saccharification of biomass with enzymes is negatively affected by the presence of acetyl esters (Selig et al., 2009). However, even more importantly, the released acetate and conversion products of acetate are inhibitory to the organisms used to ferment sugars into ethanol and other products (Helle et al., 2003;Carroll and Somerville, 2009).
The mechanism of O-acetylation of polysaccharides is not well understood. Acetylation takes place in the Golgi apparatus, and a donor substrate for pectin acetylation was demonstrated to be acetyl-CoA (Pauly and Scheller, 2000). Presumably, acetylation of other polysaccharides takes place in the same place and uses the same substrate, but this remains to be determined. To date, the only protein known to be involved in acetylation of a cell wall polysaccharide is a protein known as Cas1p that was identified in the pathogenic fungus Cryptococcus neoformans (Janbon et al., 2001). A loss-of-function mutant of the CAS1 gene produces capsules with nonacetylated glucuronoxylomannan and displays hyperpathogenicity. The biochemical mechanism of Cas1p has not been shown. Homologs of Cas1p exist in many other eukaryotes, including plants. In public databases, many of these proteins are annotated as O-acetyltransferase, but there is no direct evidence for this assignment. In this study, we investigated the four CAS1 homologs in Arabidopsis (Arabidopsis thaliana) and show that inactivation of one of the homologs, REDUCED WALL ACETYLATION2 (RWA2), results in cell walls with reduced acetyl-ester content.

Phylogenetic Analysis of Cas1p Homologs in Plants and Isolation of T-DNA Insertional Mutants
BLAST searches with the C. neoformans Cas1p protein sequence identified Cas1p homologs in a wide range of plant species, including all species with fully sequenced genomes. Representatives of basal plants Selaginella moellendorfii (lycophyte) and Physcomitrella patens (moss) have one and three genes, respectively, encoding Cas1p homologs (Fig. 1). A complete compilation of Cas1p homologs in plants can be found at the Phytozome database (www.phytozome.net), which currently lists 49 members of the gene family. The identified sequences fall into a well-defined clade, implying that they are all orthologous. We identified four genes in Arabidopsis encoding putative Cas1p-like proteins and named them RWA1 (At5g46340), RWA2 (At3g06550), RWA3 (At2g34410), and RWA4 (At1g29890) based on the biochemical phenotype of the rwa2 mutants.
A T-DNA insertion line for each gene was identified and designated rwa1-1 (SAIL_205_F09), rwa2-1 (GK-571F07), rwa3-1 (SALK_133630), and rwa4-1 (SAIL_ 205_D08; Fig. 2A). All four homozygous mutants appeared to be indistinguishable in terms of growth and morphology from the wild type. Reverse transcription (RT)-PCR was conducted to confirm that the T-DNAs disrupt the transcription of target genes (Fig. 2B). The total wall acetylation level and monosaccharide composition were determined for each of the four rwa mutants and compared with the ecotype Columbia-0 (Col-0) wild type and the qrt1 mutant (the background genotype for SAIL lines rwa1-1 and rwa4-1). Alcoholinsoluble residue (AIR) was extracted from stems and leaves of mature plants and analyzed for content of alkali labile acetyl esters. A significant difference in cell wall acetate was found between wild-type and rwa2-1 leaves, whereas the other rwa mutants did not show any significant changes (Fig. 3). AIR from rwa2-1 leaves released 17.0% 6 5.2% (P , 0.05, Student's t test) less acetic acid compared with the wild type upon saponification with 0.09 M NaOH. In repeated experiments with leaves of 4-to 8-week-old plants, cell wall preparations from rwa2-1 consistently released 15% to 30% less acetic acid compared with the wild type. No significant difference between mutants and the wild type was found in the stem samples (data not shown).
One possible explanation for the reduced cell wall acetyl content could be a reduced presence of a particular acetylated polysaccharide. However, no significant difference in wall monosaccharide composition was found, indicating that the change in acetylation is likely to be independent of polysaccharide composition (Supplemental Fig. S1).

Expression Patterns of RWA Genes
The lack of observable biochemical phenotypes in the rwa1, rwa3, and rwa4 mutants and the moderate change in acetylation in rwa2 mutants suggested redundancy between the RWA proteins. Unfortunately, only the RWA1 gene is represented on publicly available array data; therefore, we used real-time RT-PCR to investigate expression patterns (Fig. 4). All four RWA genes are ubiquitously expressed, but the RWA1, RWA3, and RWA4 genes have a predominant expression in the more mature parts of inflorescence stems, whereas RWA2 is relatively more highly expressed in leaves and inflorescence stem tops. The high expression of RWA1 in tissue with secondary wall formation is in agreement with array data (Oikawa et al., 2010). The expression patterns of the four RWA genes are consistent with the decrease in acetylation being observed only in rwa2 mutants and only in leaves.

Identification of a Second Mutant Allele of rwa2 with Reduced Wall Acetylation
In the subsequent work, we focused on RWA2. The rwa2-1 line was crossed with the Col-0 wild type, and cell wall acetylation of F1 plants was studied to determine if the trait is recessive or dominant. The F1 generation plants displayed equivalent levels of acetylation compared with the wild type, indicating that the reduced acetylation phenotype in rwa2-1 is recessive (data not shown). Two additional T-DNA insertional mutants of RWA2, rwa2-2 (SALK_002742) and rwa2-3 (SALK_013562), were obtained from the SALK collection (Fig. 2). rwa2-2 has an insertion 213 bp upstream of the putative start codon, while rwa2-1 and rwa2-3 have insertions within the seventh and 13th introns, respectively. Whereas semiquantitative RT-PCR displayed a significant expression level for rwa2-2, no full-length transcript was detected for rwa2-3 (Fig. 2B). Thus, we chose rwa2-3 to be used as the second mutant allele for further analyses. Cell wall preparations from rwa2-3 and F1 plants of a cross between rwa2-1 and rwa2-3 mutants showed a similar level of reduction in wall acetate as seen in the rwa2-1 mutant (Fig. 5). These results confirm that the reduced acetate in the rwa2 mutants is linked to disruption of the RWA2 gene.

Nonacetylated Xyl-Containing Epitope Is Detected in Epidermal Primary Cell Walls of rwa2 Mutants
Polysaccharide-specific antibodies are powerful tools to investigate cell wall composition and structure in situ. LM23 is a novel rat monoclonal antibody raised against xylogalacturonan that has dual specificity to-  . Cell wall acetylation levels in rwa mutants. Acetic acid was determined after saponification of AIR derived from mature leaves. Rosette leaves of eight 6-week-old short-day-grown plants were pooled for AIR preparation, and three pools per genotype were analyzed. The qrt1 mutant was added to the analysis as a parental line (rwa1-1 and rwa4-1 mutants are homozygous for the qrt1 mutation; Preuss et al., 1994). rwa2-1 displayed a significant reduction of about 20% in acetic acid release compared with the Col-0 wild type (WT; P , 0.05, t test), whereas none of the other samples were significantly different from each other. Values shown are means 6 SE (n = 3).
ward xylogalacturonan and xylan polymers. The binding of LM23 to xylan and pectic components requires the absence of acetyl groups and is generally increased by saponification treatments (J.P. Knox and S.E. Marcus, unpublished data). Alkali pretreatments have been shown to efficiently remove acetyl groups (Verhertbruggen et al., 2009a;Marcus et al., 2010). To further investigate the acetate-deficient phenotype of the rwa2 mutants, equivalent transverse stem sections of the wild type, rwa2-1, and rwa2-3 were immunolabeled with LM23. In the epidermis of wild-type stems, the epitope was detected at cell corners at a very low level (Fig. 6A). In contrast, the LM23 epitope strongly occurred in the outer epidermal cell corners of rwa2 (Fig. 6B). As shown in Figure 6, the LM23 epitope was unmasked at the epidermal cell walls of both the wild type and rwa2 following alkali pretreatment. LM23 binding was more pronounced in rwa2 epidermal cell walls after the alkaline treatment when compared with a pH 7.0 treatment, suggesting that the RWA2 disruption did not induce a complete deacetylation of the polysaccharide recognized by LM23 in the epidermal cell walls. The inactivation of RWA2 did not induce any increase in LM23 binding to the xylan-rich secondary cell walls of stem interfascicular fibers, whereas pretreatments with alkali did (see Supplemental Fig.  S2, where the binding of LM10 xylan and LM20 homogalacturonan are also shown as controls). This indicated that the rwa2 mutation did not impact upon secondary cell wall xylan polysaccharides.
To determine the nature of the polysaccharide recognized by LM23 in the epidermis, equivalent sections were labeled with the LM10 xylan antibody. No LM10 binding was detected irrespective of alkaline pretreatment in either the wild type or rwa2. Thus, the LM23 epitope detected in the epidermal cell corners of rwa2 is concluded to correspond to unacetylated xylogalacturonan. Similar observations were made at cell cor-ners of cortical and pith parenchyma (data not shown), which indicates that the action of RWA2 upon the LM23 epitope was not specific to the epidermal cells but rather to the primary cell walls.

RWA2 Is Involved in Acetylation of Both Pectic and Nonpectic Polysaccharides
To investigate whether pectin is the only polysaccharide for which acetylation is affected by RWA2 and if a difference in pectin O-acetylation affects the efficiency of pectinase digestion, we prepared pectin extracts by treating cell wall preparations with endopolygalacturonase (EPG) and pectin methylesterase (PME). For these experiments, the cell wall material was washed first with buffer containing SDS and then with chloroform:methanol (1:1, v/v) to remove proteins from the cell wall preparations. This washing method was chosen over a previously described method with phenol:acetic acid:water (2:1:1, v/v; Jensen et al., 2008) since the residual phenol interfered with measurement of the released acetic acid. Subsequent treatment with EPG and PME released similar amounts of pectic material from rwa2-1 and the wild type, and we did not detect any significant difference in monosaccharide composition (Supplemental Fig.  S3). The acetyl ester content both in the pectin extracts and in the residue, which was substantially depleted in pectin, was 15% to 20% lower in rwa2-1 than in the wild type (i.e. the same level of decrease observed in total cell wall preparations; Figs. 3, 5, and 7). In two independent experiments, the reduction of acetate in the supernatant from rwa2-1 was 16% 6 3%, which is significantly different from the wild type (P , 0.05, ANOVA) and not significantly different from the 19% 6 3% reduction in acetate in the starting material . Expression profiles of the four RWA genes. RNA was isolated from different tissues and quantified by real-time PCR. The expression levels are shown relative to expression in leaves. Three reference genes were used for the analysis, ACT2, UBC, and PP2AA3, and DC T values were calculated using the average C T values for the three reference genes.
from rwa2-1. The similar decrease in acetylation in starting material, pectinase-released material, and residue after pectinase treatment led to the conclusion that acetylation of both pectic and nonpectic polysaccharides is affected by the rwa2 mutation.

RWA2 Is Involved in O-Acetylation of Xyloglucan
The O-acetylation of the hemicellulose xyloglucan was assessed by oligosaccharide mass profiling (OLIMP; Lerouxel et al., 2002;Obel et al., 2009). The analysis of leaf material demonstrated that the abundance of the O-acetylated fucosylated xyloglucan oligosaccharides was significantly reduced in rwa2-1 and rwa2-3 compared with the wild type (P , 0.01, t test; Fig. 8). A significant reduction of O-acetylated nonfucosylated oligosaccharides could not be detected by OLIMP. The total abundance of the fucosylated oligosaccharides (nonacetylated plus O-acetylated) did not change, indicating that the overall carbohydrate side chain substitution pattern was not affected by the reduction in O-acetylation. The reduction in acetate level of xyloglucan was comparable to that found in the previous experiments for total cell walls and pectins (Figs. 3, 5, and 7). Taken together, these observations indicate that RWA2 has broad specificity for a range of primary cell wall polysaccharides.

Subcellular Localization of RWA2
To determine its subcellular localization, RWA2 was transiently expressed as a translational GFP fusion protein in Nicotiana benthamiana. Since placement of a GFP tag in one end of a protein may cause mistargeting , we made constructs with GFP placed either at the N-terminal or the C-terminal end of the RWA2 protein. In both cases, the expressed RWA2 fusion protein was detected in the endoplasmic reticulum (ER; Fig. 9). Since polysaccharides are synthesized in Golgi and acetylation of pectin has been detected in this compartment (Pauly and Scheller, 2000), Golgi localization could be expected for a polysaccharide acetyltransferase. However, the unspecific acetylation related to RWA2 suggests that its function is at a biochemical step upstream to the actual O-acetyl transfer onto the polysaccharide.
Polysaccharide O-acetylation inhibits hydrolase action; hence, a reduced acetylation could lead to a more easily deconstructed feedstock for the production of biofuels or other biomass-derived products (Selig et al., 2009). However, a potential drawback could be that pathogenic organisms would also have easier access to the cell wall polysaccharides and the mutant plants would be more susceptible to disease. Therefore, we tested the response of the plants to infections with B. cinerea, a common necrotrophic fungal pathogen. Excised leaves of the wild type, rwa2-1, and rwa2-3 were inoculated with B. cinerea and observed for 3 d after inoculation (Fig. 10). Both rwa2-1 and rwa2-3 mutants showed a substantially increased resistance to the fungus, as seen from the size of lesions on the leaves. The experiment was repeated three times with the same results. In contrast, no discernible difference in susceptibility was observed for the mutants as compared with the wild type when they were subjected to infection by powdery mildew (Golovinomyces Figure 6. Indirect immunofluorescence detection of cell wall epitopes at the epidermis of equivalent transverse sections of wild-type (WT) and rwa2 stems. For the right panels, the LM23 epitope was rarely detected at the epidermal cell corners of stem sections pretreated with phosphate buffer (pH 7.0; A) but the epitope was unmasked after pretreatment with sodium carbonate (pH 12.4; C and D) and to a lower extent when the RWA2 gene was disrupted (B). For the left panels, regardless of the pH pretreatment, the LM10 xylan epitope was not detected at the epidermis of wild-type or rwa2 plants, suggesting that the polysaccharide recognized by LM23 at the epidermis is xylogalacturonan and not xylan. Double-headed arrows show the occurrence of the LM23 epitope at the epidermis. Arrowheads indicate the outer surface of the epidermis. Bars = 20 mm. Figure 7. RWA2 plays a key role in acetylation of both pectic and nonpectic polysaccharides. Cell wall extracts were treated with EPG and PME for 24 h. Acetic acid release upon saponification was determined for supernatant and residue as well as for the starting material. The sugar content in cell wall material (CWM) and supernatant was calculated per g dry weight of cell wall material. For the residue, the sugars are calculated per g dry weight of residue. Values shown are means 6 SE (n = 5). WT, Wild type. cichoracearum), a biotrophic fungal pathogen (Supplemental Fig. S4).
The reason for increased resistance to B. cinerea is not clear, but it might be either a direct effect of the altered polysaccharide acetylation or an indirect effect via the induction of defense responses. The expression of PDF1.2 and PAD3 has been linked to B. cinerea resistance (Penninckx et al., 1998;Berrocal-Lobo et al., 2002;Kliebenstein et al., 2005) as markers for jasmonate/ ethylene-dependent mechanisms and camalexin accumulation, respectively. We tested if these genes were constitutively activated in the mutants, but we did not detect any differences in the expression levels of these genes (Supplemental Fig. S5).

DISCUSSION
To our knowledge, this is the first report identifying a protein involved in cell wall polysaccharide O-acetylation in plants. The data clearly demonstrate that the RWA2 protein is involved in polysaccharide O-acetylation in Arabidopsis primary walls. To examine the function of RWA proteins, we analyzed mutant lines for all Arabidopsis RWA gene family members regarding polysaccharide acetylation. Loss of function of RWA2 resulted in a significant decrease of about 20% in degree of acetylation. Presumably, the homologs RWA1, RWA3, and RWA4 are also involved in polysaccharide acetylation, even though the single loss-offunction mutants did not reveal any significant decrease in overall wall acetylation. Largely overlapping functions and expression patterns among these proteins might explain that only rwa2 mutants showed a decrease in acetylation and only of about 20%. RWA1, RWA3, and RWA4 proteins are more closely related to each other based on sequence comparison with RWA2 ( Fig. 1; Supplemental Fig. S6), and the three genes are predominantly expressed in stem tissue, whereas RWA2 has a relatively high expression in leaves (Fig. 4). These expression patterns are consistent with the observation of changes in acetylation levels in leaves and in the epidermis of stems of rwa2 mutants.
Determining which specific polysaccharides are affected in the rwa2 mutants is challenging. Standard extraction of wall polymers with alkali could not be used, and enzymatic degradation of both pectin and Figure 8. O-Acetylation of xyloglucan is affected in rwa2. Destarched walls were treated with a xyloglucan-specific endoglucanase and their xyloglucan oligosaccharide composition was profiled by matrix-assisted laser-desorption ionization time of flight mass spectrometry (OLIMP). For one-letter xyloglucan nomenclature, see Fry et al. (1993). Values shown are means 6 SE (n = 3 biological replicates, each with 8-10 technical replicates). WT, Wild type. hemicelluloses may be affected by the content of acetyl esters in the polysaccharide backbones. Nevertheless, the data presented here indicated that numerous wall polymers are affected to a similar extent. In situ antibody labeling indicated that acetylation of a xylogalacturonan differs between the wild type and rwa2 mutants, and analysis of solubilized pectic polysaccharides confirmed this conclusion. However, the amount of acetate present in the pectic fraction corresponds to about one acetic acid residue per four GalUA residues. Considering the small abundance of xylogalacturonan present in Arabidopsis pectin (Zandleven et al., 2007), other pectic polysaccharides (i.e. homogalacturonan and/or RGI) must also have reduced O-acetylation in the mutant. Interestingly, the remaining residue after pectinase extraction, which is highly depleted in pectin, showed the same relative decrease in acetylation in the rwa2 mutant compared with the wild type. Furthermore, specific analysis of xyloglucan showed that the oligosaccharides released with endoxyloglucanase had the same relative decrease in acetylation as the pectic polysaccharides in the rwa2 mutants compared with the wild type. Immunofluorescence labeling indicated that the rwa2 mutations had no effect on acetylation of xylans in the secondary walls. RWA1, RWA3, and RWA4 are highly expressed in tissue with secondary walls (Oikawa et al., 2010; Fig. 4), but the corresponding mutants did not show changes in acetylation, presumably due to redundancy. Future studies of double, triple, and quadruple mutants will be required to reveal the relative importance of the four genes in different tissues.
Taken together, these observations indicate that RWA2 is involved in acetylation, but not in the specific acetylation of a particular polymer. Therefore, it would appear that RWA proteins are active at an upstream biochemical event prior to the actual transfer of an acetyl group to a polysaccharide substrate.
The RWA proteins have sequence similarity to Cas1p isolated from C. neoformans. However, as noted in a recent paper (Anantharaman and Aravind, 2010), C. neoformans Cas1p and the homologs in metazoans are larger proteins than RWA proteins, and the sequence similarity is only with the C-terminal domain of Cas1p, which is predicted to contain more than 10 transmembrane helices (Supplemental Fig. S6). The N-terminal domain of Cas1p and the metazoan homologs have some similarity to the DUF231 family of proteins, a large family of proteins in plants with 45 members in Arabidopsis (Anantharaman and Aravind, 2010). Some DUF231 proteins have been implicated in the formation of esters in plant cell walls but not specifically linked to acetyl esterification, and a role in methylesterification has been proposed (Bischoff et al., 2010). Some DUF231 proteins have been localized to the Golgi vesicles (Oikawa et al., 2010). The DUF231 proteins appear to be typical type II membrane proteins with a single N-terminal membrane-spanning helix and a globular domain inside the Golgi, and the predicted topology is similar to the N-terminal part of Cas1p. The bioinformatics study of Anantharaman and Aravind (2010) indicated that the C-terminal domain of Cas1p and the plant RWA proteins belong to a larger class of acyltransferases with 10 transmembrane helices, whereas the N-terminal domain is an esterase domain. Another recent study showed that overexpression of the human Cas1 homolog in mammalian cells led to increased O-acetylation of sialic acid (Arming et al., 2010), suggesting that metazoan Cas1 homologs are sialic acid O-acetyltransferases. Sialyltransferase and overexpressed human Cas1 were shown to be colocalized in Golgi vesicles (Arming et al., 2010). Interestingly, it is known that mammalian O-acetylation of sialic acid takes place inside the Golgi vesicles by an integral membrane protein that utilizes acetyl-CoA from the cytoplasm to form an acetylated intermediate and releases CoA back into the cytoplasm (Higa et al., 1989). The potential role of an N-terminal esterase domain in this process is not understood, but we suggest that it could function as a transesterase. In this study, we found RWA2 to be located in the ER rather than in the Golgi, which may be consistent with the protein acting at an earlier step in the acetylation process, for example involving transport and an acetylated intermediate. The broad specificity of acetylation associated with RWA2 suggests that other proteins, likely DUF231 proteins, are involved in specific acetylation, perhaps as transesterases using the intermediate formed by RWA2. Obviously, since the biochemical function of none of these proteins has been demonstrated, there are other possible explanations, and there may well be additional proteins necessary for polysaccharide O-acetylation both in plants and in the fungi/metazoa lineage. The possibility also remains that DUF231 and RWA proteins form complexes in plants and that the observed ER localization of RWA2 is an artifact of overexpression. Future studies will be required to address these questions.
No visible growth or morphological differences were observed between the rwa2 mutants and the wild type. Therefore, reduction in acetylation by 20% does not compromise the plant performance under the standard experimental conditions. In spite of the normal growth phenotype, the rwa2 mutants showed a remarkably increased resistance toward B. cinerea. Increased resistance against this pathogen has previously been reported for other cell wall-related mutants. Arabidopsis plants expressing PME inhibitors (AtPMEI-1 or AtPMEI-2) are more resistant against B. cinerea as compared with wild-type plants due to the fact that highly methylesterified pectin serves as a poorer carbon source for the growth of this fungus (Lionetti et al., 2007). Arabidopsis plants expressing an attenuated version of a polygalacturonase, PGII, from Aspergillus niger are also more resistant against this pathogen and the bacterial pathogen Pseudomonas syringae due to a constitutive activation of defense responses (Ferrari et al., 2008). Other cell wall mutants have also been shown to have increased resistance to various pathogens through constitutive activation of defense signaling pathways involving salicylic acid (Nishimura et al., 2003), jasmonic acids/ethylene (Ellis et al., 2002), or abscisic acid (Hernández-Blanco et al., 2007). Underlying mechanism by which the reduction in the acetylation content in the cell wall increases the defense capacity against B. cinerea are unknown but could involve one or more of the mechanisms outlined above. However, transcript analysis of untreated plants did not indicate a constitutive activation of the jasmonic acid-regulated pathway or increased camalexin accumulation (Supplemental Fig. S5). The analysis does not exclude the possibility that rwa2 plants have a faster and/or stronger activation of these defense mechanisms in response to B. cinerea compared with the wild type. Alternatively, other defense mechanisms may be activated in rwa2 plants that are independent of jasmonate/ethylene and camalexin.
Acetylation of polysaccharides has attracted attention as an important property that affects downstream processing of biomass in biofuel production (Helle et al., 2003;Carroll and Somerville, 2009;Selig et al., 2009). We have shown here that a 20% reduction in acetylation can be achieved without any detectable effect on plant growth. Since acetylation and RWA proteins are conserved, they are likely to have an important biological role, but at least a 20% reduction appears to be tolerated in Arabidopsis. Future studies will be needed to show how large a reduction can be achieved without adverse growth effects. Nevertheless, one may ask if a 20% reduction in acetate is significant from an economical viewpoint. Technoeco-nomic modeling of a biofuel production facility indicates that a 20% reduction in O-acetylation would lead to a 10% reduction in production costs of ethanol (Klein-Marcuschamer et al., 2010). Such a reduction would be highly significant from an economic perspective, but the assumptions need to be tested in a pilot scale. From a practical viewpoint, the identification of RWA proteins as key proteins involved in acetylation also suggests that it may be possible to develop crops such as sugar beet (Beta vulgaris) and potato (Solanum tuberosum) with less acetylated pectins that can be more readily used as structural food ingredients. From a scientific viewpoint, identification of the RWA2 protein has given us a tool to determine the mechanism of polysaccharide acetylation and the biological role of this type of cell wall modification.

Plant Materials and Mutant Identification
Arabidopsis (Arabidopsis thaliana) plants for cell wall extraction were grown in Percival Arabidopsis growth chambers under short-day conditions (10-h photoperiod) at 22°C and 70% relative humidity. Plants for general use were grown with a 16-h photoperiod at 22°C. Plants for pathoassay were grown with a 12-h photoperiod and 50% to 60% relative humidity. For all plants, the light intensity was about 125 mmol photons m 22 s 21 . T-DNA insertion lines were identified in the SAIL (Sessions et al., 2002), GABI-Kat (Rosso et al., 2003), and SIGnAL (Alonso et al., 2003) collections and obtained from the Arabidopsis Biological Resource Center: rwa1-1 (SAIL_205_F09), rwa2-1 (GK-571F07), rwa2-3 (SALK_013562), rwa3-1 (SALK_133630), and rwa4-1 (SAIL_205_D08). The presence of the T-DNA insertion was confirmed and homozygous mutant lines identified by PCR. Primer sequences are listed in Supplemental Table S1.

Phylogenetic Analysis
BLAST searches were conducted with Cryptococcus neoformans Cas1p amino acid sequence against genomes of selected organisms, namely Arabidopsis, poplar (Populus trichocarpa), rice (Oryza sativa), Selaginella moellendorfii, Physcomitrella patens, and Homo sapiens. Additional BLAST searches were conducted with the identified plant sequences to ensure that all orthologs had been identified. A multiple alignment was generated using ClustalX (2.0) using full-length sequences and manually adjusted. The phylogenetic tree was generated with PHYLIP (version 3.69; Felsenstein, 2009) and visualized using TreeView (Page, 1996).

RNA Extraction and RT-PCR
Entire aboveground parts of 4-week-old plants were used for semiquantitative RT-PCR, and various parts of 6-week-old plants were used for quantitative RT-PCR. For the stem samples, the main inflorescence stem was cut into bottom, middle, and top parts of equal length. Plant materials were frozen in liquid nitrogen and ground in a bead mill (Qiagen). RNA was extracted using the RNeasy Plant Mini Kit (Qiagen) and the RNase-Free DNase Set (Qiagen) according to the manufacturer's protocol. cDNA was synthesized with SuperScriptIII reverse transcriptase (Invitrogen) with 12-to 18-bp oligo(dT) primers (Invitrogen) according to the manufacturer's protocol. Real-time quantitative PCR was performed with Fast SYBR Green Master Mix (Applied Biosystems) on the StepOnePlus Real-Time PCR System (Applied Biosystems) with StepOne software version 2.0. Transcript levels were calculated using the DDC T method (DDC T = DC T,sample 2 DC T,reference ) for comparing relative expression. Three reference genes with a stable expression level among different tissue types (Czechowski et al., 2005) were used for the analysis, Actin2 (ACT2; At3g18780), polyubiquitin gene (UBC; At5g25760), and regulatory subunit A3 of protein phosphatase 2A (PP2AA3; At1g13320), and DC T values were calculated using the average C T values for the three reference genes.

Cell Wall Preparation and Extraction of Pectin
For the measurement of acetic acid release and the analysis of monosaccharides, AIR was prepared and enzymatically destarched according to the previously described method (Harholt et al., 2006). The destarched AIR was ground with a bead mill (Qiagen) with two 3-mm beads at 30 Hz for 3 min. Uniform and small particle size is essential for obtaining consistent results for acetic acid release measurement.
For pectin extraction, cell walls were prepared as follows. Mature leaves were frozen in liquid nitrogen and ground with a mortar and pestle. The ground tissue was suspended in 100 mM sodium phosphate buffer (pH 7.0) and centrifuged at 3,220g for 10 min. Resuspension and centrifugation were repeated twice with grinding buffer and then once with sodium phosphate buffer supplemented with 0.5% SDS. The resulting pellet was incubated overnight in sodium phosphate buffer supplemented with 0.5% SDS at 4°C. After the overnight incubation, the cell wall extract was washed three times with sodium phosphate buffer to remove the SDS. The pellet was resuspended with chloroform:methanol (1:1, v/v) and incubated for 1 h at 4°C. The suspension was filtered through a nylon mesh (pore size, 48 mm; Pall Life Sciences). Cell wall material remaining on the filter was rinsed thoroughly with acetone and destarched according to Harholt et al. (2006). Pectin was extracted from 50 mg of destarched cell wall material by adding 1 mL of enzyme mix (50 mM cyclohexane diamine tetraacetic acid [CDTA], 50 mM ammonium formate, 0.05% sodium azide, 1 unit mL 21 Aspergillus aculeatus EPG [Megazyme], and 1 unit mL 21 Aspergillus niger PME [a kind gift from Danisco], pH 4.5) and incubating for 24 h. After the incubation, the samples were heat inactivated at 105°C for 15 min, cooled to ambient temperature, and centrifuged at 21,000g for 10 min. The supernatants were transferred to new tubes and stored at 4°C. The remaining pellets were washed twice with digestion buffer and then twice with 70% (v/v) ethanol.

Determination of Acetyl Esters
AIR (10 mg) was saponified in 550 mL of 0.09 M NaOH at ambient temperature overnight. The sample was neutralized by adding 50 mL of 1 M HCl and 900 mL of 1 M Tris-HCl (pH 8.4). The suspension was centrifuged at 21,000g for 10 min immediately before the measurement of acetic acid. Acetic acid was determined using the Acetic Acid Detection Kit (R-Biopharm) adapted into a 96-well plate format. Absorbances were measured at 340 nm with a plate reader (SpectraMax M2; Molecular Devices). First, 100 mL of Sol I, 20 mL of Sol II, and 150 mL of saponified sample were added to the wells, mixed thoroughly, and read at 340 nm to give A 0 . Subsequently, 10 mL of 10fold-diluted Sol III was added to the wells, mixed thoroughly, incubated for 2 min at ambient temperature, and read at 340 nm to give A 1 . Finally, 20 mL of 10-fold-diluted Sol IV was added to the wells, mixed thoroughly, and incubated for 10 min at ambient temperature, and the absorbance was read every 2 min until a stable reading was obtained to give A 2 . A 0 , A 1 , and A 2 values were applied into a formula to calculate DA according to the manufacturer's instructions. The amount of acetic acid released was calculated using a standard curve using an acetic acid solution (Sol V) standard.

OLIMP Analysis
Leaves from 8-week-old plants were freeze dried, ground, extracted with 70% ethanol, and centrifuged. The pellet was washed with 1:1 chloroform: methanol, and the cell wall material was subsequently digested with xyloglucanspecific endoglucanase from A. aculeatus (Pauly et al., 1999) and analyzed by OLIMP according to the previously described method (Obel et al., 2009;Gü nl et al., 2010).

Cloning and Subcellular Localization of RWA2
RWA2 was PCR amplified from cDNA derived from RNA extracted from Arabidopsis inflorescence stems. The PCR product was first introduced into a Gateway donor vector (Invitrogen) using BP reaction, then to destination vectors pMDC43 and pMDC83 (Curtis and Grossniklaus, 2003) using LR reaction according to the manufacturer's protocol (Invitrogen). The resulting plasmids expressing RWA2 protein with an N-terminal GFP fusion (pDMC43-GFP-RWA2) or C-terminal GFP fusion (pDMC83-RWA2-GFP) were transformed into Agrobacterium tumefaciens (strain C53) and transiently expressed in 3-to 4-week-old leaves of Nicotiana benthamiana as described previously . The A. tumefaciens strains transformed with the RWA2 fusion constructs were coinfiltrated with A. tumefaciens transformed with constructs for known marker proteins tagged with mCherry, red florescent protein (RFP; Nelson et al., 2007). The Golgi marker is the cytoplasmic tail and transmembrane domain soybean (Glycine max) a-1,2-mannosidase I, and the ER marker is signal peptide of AtWAK2 (for Arabidopsis wall-associated kinase 2) with a C-terminal ER retention signal HDEL (Nelson et al., 2007). Marker protein constructs were obtained from the Arabidopsis Biological Resource Center. Confocal laser scanning microscopy was performed using a Zeiss LSM 710 microscope equipped with an argon laser (488 nm for GFP excitation) and an In Tune laser (536 nm for RFP excitation). Emission was collected at 510 to 545 nm (GFP) and 570 to 625 nm (RFP). The pinhole diameter was set at one Airy unit. Images were processed in Photoshop (Adobe Systems).

Immunolabeling Procedures
Inflorescence stems of the wild type, rwa2-1, and rwa2-3 with equivalent growth (same age, similar stem height [approximately 20 cm]) were selected for sectioning. Following a fixation overnight in 4% paraformaldehyde in 50 mM PIPES, 5 mM EGTA, and 5 mM MgSO 4 (pH 6.9) at 4°C, 1.5-cm sections from the middle part of the stems (10 cm distal from the rosette) were embedded on 7% agarose gels. Transverse stem sections (60 mm thick) were made using a Leica VT1000S vibratome. At least 20 stems were analyzed with each antibody and for each genotype.
The immunolabeling procedure was carried out as described by Verhertbruggen et al. (2009b) using the following rat monoclonal antibodies: LM10 xylan (McCartney et al., 2005), LM20 pectic homogalacturonan (Verhertbruggen et al., 2009a), and LM23 (J.P. Knox and S.E. Marcus, unpublished data). Each antibody was in the form of hybridoma supernatant and used at 10-fold dilution. An anti-rat monoclonal antibody coupled with fluorescein isothiocyanate was used as a secondary antibody. The sections were mounted in Citifluor AF1 with a glass coverslip and observed with a Leica D4000B epifluorescence microscope coupled with a Leica DC500 camera. Antibody binding was analyzed through a GFP filter. To confirm that the LM23 antibody recognizes an esterified epitope, sections were pretreated with 0.1 M sodium carbonate (pH 12.4) or phosphate-buffered saline (pH 7.0) for 60 min and then washed extensively in phosphate-buffered saline prior to indirect immunofluorescence labeling.

Pathogen Response Assays
Botrytis cinerea IK2018 isolated from strawberry (Fragaria spp.) fruit was obtained from Dr. Birgit Jensen (University of Copenhagen) and maintained on potato dextrose agar (Difco). Spores were collected in 3 mL of 12 g L 21 potato dextrose broth (PDB; Difco) by gentle rubbing and filtered through Miracloth (EMD Chemicals) to remove mycelium. The number of spores was counted using a hemocytometer, and the suspension was adjusted to 5 3 10 5 conidiospores mL 21 in PDB for infection of leaves. Rosette leaves from 4-week-old soil-grown Arabidopsis plants were placed in petri dishes containing 0.6% agar, with the petiole embedded in the medium. Inoculation was performed by placing 5 mL of a suspension of 5 3 10 5 conidiospores mL 21 in 12 g L 21 PDB on each side of the middle vein. The plates were incubated at +22°C with a 12-h photoperiod. High humidity was maintained by covering the plates with a clear plastic lid. Lesion diameter was determined by analysis of high-resolution digital images of infected leaves using ImageJ (Abramoff et al., 2004). Included scale objects allowed standardization of measurements across images.
Golovinomyces cichoracearum UCSC1 was maintained on cucumber (Cucumis sativus) var Bush Champion (Burpee). Plants were grown for 3 weeks in growth chambers at 22°C with a 16-h photoperiod. Plants were inoculated by placing a 1.3-m-tall settling tower over the flat and using compressed air to disperse the spores from two to three cucumber leaves infected 10 d prior to inoculation. After 5 min, the plants were returned to a separate growth chamber with 80% humidity. Representative rosette leaves were photographed 12 d post inoculation.

Supplemental Data
The following materials are available in the online version of this article.
Supplemental Figure S2. Indirect immunofluorescence detection of the LM10, LM20, and LM23 epitopes in stem sections of the wild type and rwa2-1.
Supplemental Figure S3. Monosaccharide composition analysis of wildtype and rwa2-1 cell wall before and after pectinase treatment.
Supplemental Figure S4. Infection of wild-type and rwa2 mutant leaves with G. cichoracearum.
Supplemental Figure S5. Expression analysis of pathogen response genes.
Supplemental Figure S6. Multiple sequence alignment of Arabidopsis RWA proteins and C. neoformans Cas1p.
Supplemental Table S1. Primers used for PCR.