The most abundant constituents of oomycete cell walls are glucans, polysaccharides that consist of linked glucose units (Aronson et al., 1967; Sietsma et al., 1969). β-1,3 and β-1,6-glucan are the major components of oomycete cell walls, whereas cellulose, a β-1,4-glucan, forms a relatively small fraction (Aronson et al., 1967). β-1,6-Glucan is only found in oomycetes and fungi, whereas cellulose and β-1,3-glucan are present in plant cell walls too (Fesel and Zuccaro, 2015).

A β-glucan-triggered response, i.e., the accumulation of the phytoalexin glyceollin, was first observed when soybean (Glycine max) was treated with glucans isolated from cell walls of Phytophthora sojae (previously P. megasperma f. sp. glycinea and P. megasperma var. sojae) (Ayers et al., 1976). β-Glucans also trigger phytoalexin production in several other fabaceous species, and in potato (Solanum tuberosum), although this is a weaker response (Cline et al., 1978; Cosio et al., 1996). A purified β-1,3/1,6-glucan heptaglucoside was found to be one of the active molecules in eliciting production of phytoalexins in soybean (Sharp et al., 1984a,b). Laminarin, an oligomeric β-1,3-glucan with β-1,6-glucan branches isolated from the marine brown alga Laminaria digitata, is another pattern that can induce a plethora of defense-associated responses in Nicotiana tabacum, grapevine (Vitis vinifera), and the monocots rice (Oryza sativa) and wheat (Triticum aestivum; Inui et al., 1997; Klarzynski et al., 2000; Aziz et al., 2003). Furthermore, Arabidopsis thaliana is responsive to the β-glucan laminarin, although it does not respond to the oomycete-derived heptaglucoside elicitor. Arabidopsis responses to laminarin are mediated by the plant hormone ethylene, and do not seem to involve the well-known defense hormone salicylic acid (SA). In contrast, when Arabidopsis or tobacco plants are treated with a sulfated form of laminarin the expression of the SA-induced marker gene PR1 is induced (Ménard et al., 2004). Taken together, responses to β-glucans vary greatly depending on the specific β-glucan and plant species. Therefore, different plant species might have different receptors involved in the recognition of different β-glucan patterns.

Phytophthora-derived β-1,3-glucan was shown to bind soybean membranes (Yoshikawa et al., 1983). The glucan-binding protein (GBP) from soybean was identified and it was demonstrated that, when expressed in tobacco and Escherichia coli, GBP conferred β-glucan-binding activity. Furthermore, an antibody raised against GBP inhibited β-glucan-binding activity in soybean and reduced phytoalexin accumulation (Umemoto et al., 1997). Interestingly, GBP also shows β-glucanase activity and might release β-glucans from the pathogen’s cell wall (Fliegmann et al., 2004). After heterologous expression of soybean GBP in tomato, high affinity binding of the β-1,3/1,6-glucan heptaglucoside was observed. However, this did not result in activation of downstream defense responses in tomato (Mithöfer et al., 2000; Fliegmann et al., 2004). These data suggest that additional, probably membrane-bound, proteins are required to recognize the β-glucan patterns (Mithöfer et al., 2000).

Recently, glucan-chitosaccharides were isolated from the cell wall of the root oomycete Aphanomyces euteiches and were found as novel patterns that triggered calcium oscillations in the nucleus of root cells and induced defense genes in Medicago truncatula (Nars et al., 2013). How these molecules are perceived is not yet known, but there is a role for the nod factor perception (NFP) protein, a lysin motif (LysM)-RLK. NFP is involved in the recognition of microbial N-acetylglucosamine patterns and is required for nodule formation in interaction with Rhizobium bacteria. An nfp mutant was more susceptible to A. euteiches, whereas overexpression of NFP led to increased resistance, demonstrating its involvement in the perception of A. euteiches by M. truncatula (Rey et al., 2013). However, NFP was not required for the glucan-chitosaccharide-induced calcium oscillations, suggesting a regulatory function in defense for NFP rather than direct recognition (Nars et al., 2013).

Transglutaminases (TGases) are a widespread family of enzymes, found in prokaryotes and eukaryotes, that facilitate cross-linking between glutamine and lysine residues in proteins, thereby strengthening structures, e.g., cell walls (Lorand and Graham, 2003; Martins et al., 2014). The formation of a covalent bond between amino acid residues confers high resistance to proteolysis (Reiss et al., 2011). In oomycetes, TGases could protect cell walls from hydrolytic host enzymes. A 42-kDa TGase cell wall glycoprotein (GP42) of P. sojae functions as a potent elicitor of phytoalexin synthesis in the non-host parsley (Petroselinum crispum) (Parker et al., 1991). A 13-amino acid peptide fragment (Pep-13) derived from GP42 was found responsible for triggering immunity and was shown to bind to purified plasma membranes of parsley. Furthermore, Pep-13 elicits a multitude of defense responses, e.g., expression of defense-related genes and phytoalexin production (Nürnberger et al., 1994, 1995; Hahlbrock et al., 1995). Interestingly, Pep-13 treatment of potato resulted in a similar defense activation, with the distinct difference that it induced a hypersensitive response (HR; Halim et al., 2004).

GP42 homologs are only found in oomycetes and some marine bacteria belonging to the genus Vibrio that are pathogenic on fish and several marine invertebrates (Reiss et al., 2011). It is thought that an ancestral oomycete, from which species of Phytophthora, Pythium and downy mildews have evolved, acquired GP42 from Vibrio bacteria through horizontal gene transfer, giving an selective advantage over oomycetes that lack this TGase (Reiss et al., 2011). A 100 kDa monomeric plasma membrane protein from parsley was shown to bind to the Pep-13 ligand and thus may be part of the putative receptor complex (Nennstiel et al., 1998).

Application of mycelial extracts from P. infestans to potato tubers led to necrosis and accumulation of phytoalexins, predominantly rishitin, and lubimin. The molecules responsible for triggering this response were identified as the eicosapolyenoic acids (EPs), arachidonic acid (AA), and eicosapentaenoic acid (EPA; Bostock et al., 1981). Treating potato tuber slices with AA greatly reduced or even arrested growth of P. infestans (Bostock et al., 1982). EPs are components of Phytophthora cells and are seemingly not present in other microbial classes nor are they produced by higher plants (Robinson and Bostock, 2015). Interestingly, the downy mildew Hyaloperonospora arabidopsidis has lost the genes required for AA synthesis (Baxter et al., 2010). It is tempting to speculate that H. arabidopsidis has lost this ability through evolution as a way to avoid recognition.

Eicosapolyenoic acids induce the accumulation of antimicrobial compounds in many plant species, ranging from many solanaceous species, e.g., potato and tomato, to bean (Phaseolus vulgaris) and avocado (Persea americana) (Longland et al., 1987; Romero-Correa et al., 2014; Robinson and Bostock, 2015). Furthermore, in potato application of AA induced accumulation of reactive oxygen species (ROS), that could be involved in mediating synthesis of the phytoalexin rishitin from lubimin (Yoshioka et al., 2001).

EPs are able to trigger systemic acquired resistance in several plants species to different pathogens. The hormonal regulation of these responses seems to differ among plant species; in some the SA pathway is elicited, whereas in other species responses seem to rely on jasmonic acid (JA) or ethylene. It is postulated that this may be due to the concentration of EPs in the treatment (Robinson and Bostock, 2015). For example, Arabidopsis plants made to produce low levels of EP showed increased resistance to Botrytis cinerea, P. capsici and aphid feeding, but higher susceptibility to Pseudomonas syringae pv. tomato (DC3000). This was associated with higher levels of JA and enhanced expression of JA-related genes, but decreased SA levels and reduced expression of SA-related genes. Furthermore, low levels of AA administered to tomato leaves resulted in increased JA levels and decreased SA levels and higher resistance against Botrytis cinerea (Savchenko et al., 2010).

How, exactly, EPs are perceived remains to be resolved. EPs could be recognized directly by a membrane-bound receptor, leading to the activation of plant immunity. Another possibility is that plant membranes that readily incorporate AA (Ricker and Bostock, 1992), are perturbed leading to the release of endogenous patterns from the host cell cytoplasm. Or alternatively, AA can be used as a substrate for lipoxygenases, e.g., the potato LOX1, thereby producing oxylipin signals that trigger plant immunity. In the latter two scenarios recognition would be independent of plant PRRs (Robinson and Bostock, 2015).

Interestingly, treating potato with a combination of AA and β-1,3-glucans strongly increased the response to AA. β-glucans alone, however, did not trigger a response in potato (Preisig and Kuć, 1988).

Recently, the XEG1 (xyloglucanspecific endo-β-1,4-glucanase) protein was isolated from P. sojae culture filtrates (Ma et al., 2015). This secreted protein elicits cell death in N. benthamiana, N. tabacum, pepper (Capsicum annuum), tomato (S. lycopersicon) and soybean but not in maize (Zea mays) and cotton (Gossypium hirsutum). Analysis of the XEG1 protein sequence revealed that it belongs to the glycoside hydrolase GH12 family that is widespread amongst prokaryotic and eukaryotic microbes, especially in plant-associated microorganisms. Within the Phytophthora genus many GH12 proteins are found of which half trigger cell death in N. benthamiana. The downy mildew H. arabidopsidis also has three GH12 genes, however, none of them encode a protein that elicits cell death (Ma et al., 2015). Previously, it was demonstrated that fungal GH12 proteins are able to degrade β-glucan (Karlsson et al., 2002) and xyloglucan, a hemicellulose found in the plant cell wall (Master et al., 2008). Recombinant XEG1 protein partially released reducing sugars from both glucans, but was most active with a xyloglucan substrate. Mutations in the catalytic site of XEG1 strongly decreased xyloglucanase activity and abolished β-glucanase activity. In contrast, XEG1 enzyme activity was not required for the induction of cell death in N. benthamiana and soybean. Moreover, active and inactive recombinant XEG1 were able to induce resistance against P. sojae and Phytophthora parasitica var. nicotianae to a similar extent in soybean and N. benthamiana, respectively. Silencing as well as overexpression of XEG1 in P. sojae both led to reduced virulence on soybean through distinct mechanisms. Silenced P. sojae lines showed reduced virulence, but did not activate a stronger defense response in soybean, suggesting that XEG1 has a role in virulence, possibly through breakdown of cell wall components. XEG1 overexpression transformants induced more ROS accumulation and callose deposition compared to wildtype P. sojae, confirming the idea that XEG1 acts as a MAMP. A XEG1 PRR has not been identified but XEG1 requires the co-receptor SOMATIC EMBRYOGENESIS RECEPTOR-LIKE KINASE 3/BRI1-ASSOCIATED RECEPTOR KINASE (SERK3/BAK1) for triggering cell death, suggesting that a SERK3/BAK1-associated RLK or RLP recognizes XEG1 (Ma et al., 2015).

Necrosis and ethylene-inducing peptide 1 (Nep1)-like proteins (NLPs) form a family of secreted proteins mainly found in plant-associated microorganisms, and cytotoxic members are well known to induce necrosis and ethylene production in dicot plants (Bailey, 1995; Oome and Van den Ackerveken, 2014). Three types of NLPs have been identified: type 1 NLPs are found in bacteria, oomycetes and fungi, type 2 NLPs are found in fungi and bacteria and the newly identified type 3 NLPs are only present in fungi (Oome and Van den Ackerveken, 2014). Although many members of the NLP family are cytotoxic to plants, in recent years many non-cytoxic NLPs have been identified in fungal and oomycete species with a (hemi-)biotrophic lifestyle (Cabral et al., 2012; Dong et al., 2012; Zhou et al., 2012). In search of the function of 10 non-cytotoxic NLPs of the obligate biotrophic downy mildew H. arabidopsidis (HaNLPs) it was found that NLPs activate plant immunity in Arabidopsis (Oome et al., 2014). Expression of HaNLPs in Arabidopsis led to a severe growth reduction and increased resistance to H. arabidopsidis for 7 out of 10 HaNLPs. Only a small fragment of the tested HaNLP3 protein was sufficient to activate plant defense responses and immunity to downy mildew. This 20–24 amino acid fragment (nlp20/nlp24) contains two conserved regions. The second region is the heptapeptide motif GHRHDWE which is highly conserved in all NLPs (Oome and Van den Ackerveken, 2014). The first motif that starts with the AIMY amino acid sequence is highly conserved in type 1 NLPs (Oome et al., 2014). Treatment of Arabidopsis plants with synthetic nlp24 peptides corresponding to an oomycete, fungal and bacterial type 1 NLP resulted in the increased production of the defense-related phytohormone ethylene and high resistance to downy mildew. Conversely, a synthetic peptide of a type 2 NLP from the bacterial pathogen Pectobacterium carotovorum that lacks the AIMY motif was unable to elicit a response in Arabidopsis. Taken together, this demonstrated that the first motif contains the immunogenic part of nlp24 (Oome et al., 2014). Furthermore, nlp20, a peptide based on PpNLP, a cytotoxic P. parasitica type 1 NLP, was sufficient for MAPK activation, production of ROS, and increased callose deposition in Arabidopsis, but did not have any cytotoxic effect (Böhm et al., 2014). Other plant species were tested for their ability to respond to nlp peptides, revealing that nlp-triggered ethylene production was observed in several closely related Brassicaceae species, and also in more distantly related lettuce plants (Lactuca sativa), but not in solanaceous species such as tomato, potato, and N. benthamiana (Böhm et al., 2014).

In a screen for nlp20 sensitivity, a collection of T-DNA insertion mutants corresponding to 29 RLKs and 44 RLPs were tested for loss of nlp20-induced ethylene production. Furthermore, 135 natural accessions of Arabidopsis were also tested for the loss of nlp20 sensitivity. Two T-DNA insertion alleles of RLP23, rlp23-1, and rlp23-2 that were unable to express the receptor-like protein as well as three Arabidopsis accessions that carried a frameshift mutation resulting in a premature stop codon in RLP23 coding sequence were insensitive to nlp20. It was shown that the RLP23 LRR domain physically interacts with nlp20 in vitro and in planta (Albert et al., 2015). RLP23 lacks a cytoplasmic signaling domain but was shown to require the RLK SUPPRESSOR OF BIR1 1 (SOBIR1) for signaling. RLP23 and SOBIR1 interact in the absence of nlp peptides (Bi et al., 2014; Albert et al., 2015), whereas a second RLK, BAK1, was recruited only in presence of the ligand (Albert et al., 2015). Arabidopsis sobir1 and bak1-5/bkk1 mutants lost nlp20-responsiveness, indicating that SOBIR1 and BAK1 are required for RLP23 to function. Moreover, it was demonstrated that RLP23 is required for nlp peptide-induced resistance. Unlike wildtype Arabidopsis, nlp24-treatment of rlp23 mutants did not result in an increased resistance to H. arabidopsidis (Albert et al., 2015).

Many oomycete pathogens secrete small 10 kDa proteins called elicitins. The first proteins from this family identified were cryptogein and capsicein from Phytophthora cryptogea and Phytophthora capsici, respectively. These proteins were found to elicit necrosis, induce resistance, and cause increased production of ethylene as well as the phytoalexin capsidiol in tobacco plants (Ricci et al., 1989; Milat et al., 1991). Elicitin responses were observed in all tested Nicotiana spp., but not in other solanaceous species, such as tomato and eggplant. Furthermore, some Brassicaceae species also respond to elicitin; most radish cultivars (Raphanus sativus) and one turnip cultivar (Brassica campestris), but not Arabidopsis, showed necrosis after elicitin treatment (Kamoun et al., 1993). The gene encoding for P. infestans elicitin INF1 was found to be downregulated during early infection of potato. However, in the necrotrophic phase of infection inf1 expression was upregulated (Kamoun et al., 1997). Interestingly, N. benthamiana, a nonhost of P. infestans, gained susceptibility after silencing of inf1, demonstrating that the recognition of INF1 contributes to resistance (Kamoun et al., 1998).

Members of the Peronosporales, e.g., Phytophthora spp. and downy mildews are unable to synthesize sterols and must, therefore, acquire them during pathogenesis. Elicitin and elicitin-like sequences are also found in downy mildew pathogens, but no functional analysis has been performed on these proteins (Baxter et al., 2010; Cabral et al., 2011; Stassen et al., 2012; Sharma et al., 2015). Dehydroergosterol binding activity was shown for several elicitins in vitro. Furthermore, elicitins are able to catalyze sterol transfer between liposomes (Mikes et al., 1998). However, in vivo sterol-binding activity of elicitins has not been demonstrated. Interestingly, the oomycete pathogen A. euteiches is able to synthesize sterols and seems to lack elicitin genes (Gaulin et al., 2008, 2010).

The putative elicitin receptor was recently cloned from a wild potato (Solanum) that responds to the P. infestans elicitin INF1. A S. microdontum ecotype showed a clear cell-death response when inf1 was transiently expressed. Crosses with an unresponsive S. microdontum subspecies and further screening and genetic mapping resulted in the identification of the RLP ELR (elicitin response). Stable expression of ELR in S. tuberosum cv. Désirée conferred the cell death response after expression of inf1. Furthermore, ELR mediated a broad-spectrum response to elicitins of oomycetes: most tested elicitins induced a cell-death response in transgenic ELR potato, even though there is often low sequence similarity between elicitins (Du et al., 2015). Recognition might therefore be based on structural similarity rather than a small conserved peptide. ELR was shown to bind to SERK3/BAK1, but binding of the putative receptor to the RLP adaptor protein SOBIR1 or the elicitin ligand was not tested (Du et al., 2015). Intracellular perception, however, cannot be ruled out as elicitins have, anecdotally, been reported to be detected inside plant cells, e.g., the immunocytochemical localization of the elicitin quercinin in oak (Quercus robur) root cells infected with P. quercina (Brummer et al., 2002). ELR is thought to mediate extracellular recognition of elicitins, but direct binding to confirm the receptor function of ELR still needs to be demonstrated (Du et al., 2015). Previously, studies in tobacco suggested that INF1 binds to the cytoplasmic domain of a lectin RLK from N. benthamiana, NbLRK1 (Kanzaki et al., 2008). Silencing of NbLRK1 resulted in reduced INF1 responsiveness suggesting the RLK contributes to defense signaling. Although no ELR has been identified in tobacco yet, SERK3/BAK1 and SOBIR1 were found to be required for elicitin-triggered cell death in N. benthamiana (Chaparro-Garcia et al., 2011; Peng et al., 2015). It is, therefore, likely that ELR acts similar to RLP23 (Albert et al., 2015) and tomato Cf-4 (Postma et al., 2016), in that it requires both a BAK1-like RLK and SOBIR1-like RLK for pattern-triggered immunity.

A 34 kDa glycoprotein was isolated from P. parasitica var. nicotianae mycelium that triggered enhanced lipoxygenase activity as well as accumulation the defense-related cell wall hydroxyproline-rich glycoproteins in tobacco. This protein was localized to the internal and external layers of the hyphal cell wall (Séjalon-Delmas et al., 1997). The protein sequence revealed two cellulose-binding domains belonging to the carbohydrate binding module 1 (CBM1) family similar to that of fungal glycanases (Mateos et al., 1997; Gaulin et al., 2006). This putative function was corroborated by demonstrating protein binding to fibrous cellulose and plant cell walls. Furthermore, the protein was shown to have lectin-like activities; human red blood cells were readily agglutinated by this protein. Therefore, it was designated cellulose-binding elicitor lectin (CBEL). Moreover, CBEL was able to elicit necrosis, activate defense gene expression, and trigger immunity to P. parasitica var. nicotianae. No enzymatic activities for CBEL were observed, suggesting it acts as a pattern (Mateos et al., 1997).

Silencing of CBEL resulted in a severe reduction of adhesive abilities of P. parasitica var. nicotianae to cellulosic surfaces, but did not affect pathogenicity. Interestingly, knockdown mutants showed dispersed abnormal cell wall thickenings, indicating that CBEL might be involved in cell wall deposition in the pathogen (Gaulin et al., 2002). CBEL activity as a pattern is not limited to tobacco, as infiltration of CBEL in Arabidopsis leaves resulted in defense responses differentially dependent on the phytohormones SA, JA, and ethylene (Khatib et al., 2004). CBEL-induced necrosis was lost in JA-insensitive coi1 and ethylene-insensitive ein2 mutant plants, whereas PR1 and WAK1 expression, accumulation hydroxyproline-rich glycoproteins, and peroxidase activity was greatly reduced or abolished in an Arabidopsis NahG mutant that metabolizes SA (Khatib et al., 2004). Transient expression of CBEL as well as infiltration of recombinant CBEL in tobacco leaves resulted in rapid development of necrotic lesions. Immunocytochemistry revealed that the delivered CBEL was bound to the plant cell wall. Substitution of aromatic residues in CBEL that are possibly involved in cellulose binding reduced the necrosis-inducing activity. Necrosis-induction in tobacco was lost for three recombinant CBEL proteins (Y52A, Y188A, and Y52A_Y188A), that were also unable to induce defense-related genes at similar concentrations as native CBEL. Recently, it was shown that CBM1-1 is the main determinant in the interaction with cellulose; a mutation in CBM1-2 (Y188A) only showed a slight decrease in cellulose binding compared to wild type CBEL, whereas a mutation in CBM1-1 (Y52A) strongly decreased the binding capacity of CBEL and the double mutant (Y52A_Y188A) entirely lost the ability to bind cellulose (Martinez et al., 2015). Taken together, these data show amino acids in the two CBM1s, that were predicted to be important for cellulose binding, are important for elicitor activity.

To define the minimum CBEL pattern that triggers immunity, synthetic peptides of CBM1-1 and CBM1-2 were generated. CBM1-1synt and CBM1-2synt were sufficient to activate plant defense in tobacco and Arabidopsis, respectively. Intriguingly, recombinant CBEL but not recombinant CBEL_Y52A_Y188A, induced calcium fluxes in tobacco cells but not in protoplasts. This demonstrates that the plant cell wall and unmodified CBM1s are important for CBEL perception (Gaulin et al., 2006).

CBM1s are probably not essential for pathogens with an obligate biotrophic lifestyle; only one was detected in the Albugo laibachii genome and no clear CBM1-encoding genes were found in H. arabidopsidis, whereas Pythium ultimum and Phytophthora spp. contain multiple CBM1-encoding genes (Larroque et al., 2012). It has been proposed that adhesion of CBEL or its CBM1s perturb the cellulose status, and the perception of this disturbance leads to defense activation, but this remains to be proven (Dumas et al., 2008). The fact that BAK1 and RESPIRATORY BURST OXIDASE HOMOLOGUE (RBOH) D and F proteins are required for some of the CBEL-induced defense responses suggests that a PRR might be involved (Larroque et al., 2013). The oxidative burst triggered by pattern recognition is mediated by the NADPH oxidases RBOH D and F (Suzuki et al., 2011). Necrosis-induction by CBEL in bak1-4 and the rbohD/F double mutant was similar to the Col-0 Arabidopsis wildtype. However, no ROS production was detected in bak1-4 and rbohD/F and activation of MAP kinases was reduced in bak1-4 and delayed in rbohD/F compared to Col-0. The expression of JA-responsive genes WRKY11 and PDF1.2, but not the expression of the SA-responsive gene PR1, was also reduced in these mutant lines (Larroque et al., 2013). The dependence of some CBEL-induced responses on BAK1 suggests a role for an RLK or RLP in the perception of CBEL. Three Arabidopsis accessions were found that are unresponsive to CBEL, and may therefore offer a way to decipher CBEL-triggered immunity (Larroque et al., 2013).

A secreted apoplastic protein from P. parasitica called OPEL was recently discovered to trigger a plant immune response (Chang et al., 2015). OPEL contains a thaumatin-like domain, a glycine-rich domain, and a glycosyl hydrolase (GH) domain that has a putative laminarinase active site. OPEL seems to be oomycete specific; homologues were only found in Phytophthora spp. and other oomycetes such as H. arabidopsidis, Py. ultimum and A. laibachii. OPEL is expressed during early infection stages of P. parasitica, rapidly increasing transcript levels within 12 hours after inoculation on N. benthamiana. Furthermore, infiltration of N. tabacum with recombinant OPEL protein resulted in cell death, increased callose deposition, ROS accumulation, induction of defense-related genes and systemic acquired resistance against several pathogens. Moreover, transient expression of OPEL in N. benthamiana enhanced resistance to P. parasitica. It was shown that the GH domain was essential for the increased callose deposition and increased accumulation of ROS in N. tabacum. Although the OPEL GH domain contains a laminarinase signature active site motif, no laminarin or β-1,3-glucan enzymatic activity was detected in OPEL recombinant protein. Mutation of the putative laminarinase active site motif in the predicted GH domain abolished elicitor activity of OPEL, which suggests enzymatic activity of OPEL is required for triggering the defense response (Chang et al., 2015). The OPEL substrate has not been identified but is likely a polysaccharide in the plant cell wall. OPEL-released degradation products might, therefore, be perceived by plants as DAMPs.