Defeated by the nines: nine extracellular strategies to avoid microbe-associated molecular patterns recognition in plants

This review discusses and illustrates nine extracellular strategies that plant pathogens and symbionts use to avoid the recognition by cell surface receptors.


Introduction
Plants are a rich source of nutrients for many organisms (Cardinale et al., 2011). Their root and aerial systems are exposed to a wide range of microbes including bacteria, fungi, and oomycetes. Some of these microbes are plant pathogens that cause detrimental effects on plant fitness (Dangl and Jones, 2001;Dodds and Rathjen, 2010), while symbiotic microbes can promote plant growth, such as rhizobacteria and arbuscular mycorrhizal fungi (Pérez-de-Luque et al., 2017).
When microbes enter the extracellular space within plant tissues (the apoplast), they interact with host cells that carry surface receptors that recognize conserved molecules, conventionally called microbe-associated molecular patterns (MAMPs, Jones and Dangl, 2006). MAMPs are fragments of proteins or glycans that are essential for the biology of microbes and are absent in the host plant. Examples are fragments of flagellin and peptidoglycans (PGN) from bacteria and fragments of chitin and b-glucans from fungi and oomycetes. These MAMPs are recognized on the host cell surface by pattern recognition receptors (PRRs). PRRs are transmembrane receptor-like kinases (RLKs) or receptor-like proteins (RLPs) that often carry extracellular leucine-rich repeat (LRR) or lysine motif (LysM) domains to confer MAMP recognition (Couto and Zipfel 2016;Tang et al., 2017;Boutrot and Zipfel, 2017;Schellenberger et al., 2019). PRR activation triggers a series of immune responses (Bigeard et al., 2015), resulting in pattern-triggered immunity (PTI) and preventing colonization by nonadapted microbes (Bigeard et al., 2015;Andersen et al., 2018). Importantly, many MAMPs require hydrolytic release from their precursors before they can be perceived by PRRs. For instance, hydrolytic release from precursors is required for the recognition of flagellin and elongation factor Tu (EF-Tu), and for the recognition of chitin and b-glucan.
To colonize plants, adapted pathogens and symbionts have adopted strategies to avoid recognition by the immune system. Many effector proteins that act inside the host cell interfere in signaling downstream of PRRs (Toruño et al., 2016). In this review, however, we describe nine extracellular strategies that plant pathogens and symbionts use to avoid recognition by PRRs (Table 1). These strategies occur at three levels: MAMP production ( Figure 1A), MAMP release ( Figure 1B), and MAMP perception ( Figure 1C).

Level 1: Three strategies Preventing MAMP production
We distinguish three different strategies that microbes use to prevent the accumulation of MAMP precursors (Level 1, Figure 1A). The first strategy is to accumulate mutations in protein-based MAMPs to avoid recognition of the MAMP fragment. The second strategy is a similar genetic adaptation to alter glycan-based MAMPs beyond recognition. The third strategy is to downregulate the accumulation of MAMP precursors upon infection. The latter can involve transcription factors, epigenetic regulation, and posttranscriptional control. Specific examples of the three strategies are illustrated in Figure 2.

Strategy 1: Polymorphisms in protein MAMPs
Polymorphisms in MAMP protein sequences are a frequently used strategy to avoid detection. Sequence polymorphisms have been described for MAMPs from bacterial flagellin, EF-Tu, and RaxX.
In most angiosperms, recognition of bacterial flagellin is mediated by Flagellin Sensing 2 (FLS2), a PRR with extracellular LRRs (Gómez-Gómez and Boller, 2000). FLS2 recognizes peptides from a highly conserved 22-amino acids region in the N-terminal domain of flagellin, called flg22 (Felix et al., 1999;Zipfel et al., 2004). However, some flagellated bacteria carry flg22 sequences that are not recognized by FLS2. flg22 peptides from most e-, d-, and a-proteobacteria induce moderate, weak, or no response, respectively, in contrast to flg22 peptides from the majority of cand b-proteobacteria, which trigger strong responses (Cheng et al., 2020). For instance, flagellin of the crown gall disease pathogen Agrobacterium (Agrobacterium tumefaciens, an a-proteobacterium) possesses a different flg22 sequence that does not trigger FLS2-mediated immune responses in most plant species including Arabidopsis (Arabidopsis thaliana; Felix et al., 1999). However, some cand b-proteobacteria also evade FLS2 recognition with sequence polymorphism. For example, the bacterial wilt pathogen Ralstonia solanacearum (a b-proteobacterium) and some strains of Xanthomonas campestris pathovar (pv.) campestris (Xcc; a c-proteobacterium) have nonrecognizable versions of the flg22 sequence that evade FLS2-mediated defenses in their respective hosts (Pfund et al., 2004;Sun et al., 2006;Wei et al., 2018). Specifically, an aspartate (D) to valine (V) substitution at amino acid position 14 within the flg22 sequence of flagellin from Xcc results in reduced immune responses associated with increased virulence of Xcc on Arabidopsis (Figure 2A; Sun et al., 2006). Similarly, X. oryzae pv. oryzae (Xoo) and pv. oryzicola (Xoc) evade rice (Oryza sativa) FLS2-mediated recognition with substitutions in the flg22 sequence (Wang et al., 2015). Consistent with selection for immune evasion, nonpathogenic strains of X. arboricola pv. juglandis carry the conserved flg22 sequence whereas pathogenic strains carry polymorphisms within the flg22 sequence that evades recognition by FLS2 (Cesbron et al., 2015). Evasion of flagellin recognition by flg22 polymorphisms is also observed with symbiotic bacteria. For instance, the plant beneficial endophytic bacterium Burkholderia phytofirmans (a b-proteobacterium) and the plant beneficial bacterium Sinorhizobium meliloti (an a-proteobacterium) carry flg22 sequences with weak elicitor activity in grapevine (Vitis vinifera) and with no elicitor activity, respectively (Felix et al., 1999;Trdá et al., 2014).
Flagellin is also recognized by FLAGELLIN-SENSING 3 (FLS3), which is present only in some solanaceous plant species including some cultivars of tomato (Solanum lycopersicum), potato (S. tuberosum), and pepper (Capsicum annuum; Hind et al., 2016). FLS3 recognizes flgII-28, a 28amino acid peptide from the central region of the flagellin protein. Interestingly, polymorphisms within flgII-28 sequences are observed between strains of Pseudomonas syringae pv. tomato (Pto; a c-proteobacterium). Flagellin of the Col338 strain of Pto contains a valine (V) residue at position 13 in the flgII-28 sequence, and this peptide triggers a weaker immune response in tomato than the flgII-28 peptide from the PtoT1 strain, which carries an alanine (A) residue at this position (Cai et al., 2011).
Evasion of recognition caused by flagellin polymorphisms has also been described for animal pathogens. For instance, the human pathogenic bacteria Campylobacter jejuni, Helicobacter pylori, and Bartonella bacilliformis evade flagellin recognition by Toll-Like Receptor 5 (TLR5) with mutations within the key recognition sites of the flagellin N-terminal D1 domain (Andersen-Nissen et al., 2005).
Besides flagellin, evasion with protein polymorphism has also been reported for peptides containing the first 18 amino acids of bacterial Elongation Factor Thermal unstable (EF-Tu), called elf18, which is recognized in Arabidopsis by the EF-Tu Receptor (EFR; Kunze et al., 2004;Zipfel et al., 2006). Polymorphism within the elf18 sequence correlates with different elicitation activity. For instance, elf18 peptides from Xcc and PtoDC3000 trigger only 0.8%-3.2% of the immune activity as compared to peptides from Agrobacterium, Ralstonia, and other Xanthomonas and Pseudomonas strains (Lacombe et al., 2010).
In conclusion, protein MAMP polymorphism is an efficient and frequently used strategy to evade host immunity, employed by both pathogenic and symbiotic bacteria. The polymorphisms in MAMPs highlight the exceptional genetic plasticity associated with host adaptation of bacteria.
However, amino acid sequence polymorphism is of course restricted by protein function. Flagellin, for instance, cannot accept many substitutions in the flg22 sequence without affecting flagellin function because this region acts as a hinge that facilitates important conformational changes in the flagellin structure when reversing the spin of flagellar rotation (Fliegmann and Felix, 2016;Wang et al., 2017). And obviously, while evasion of immunity by MAMP polymorphism  is relevant for protein-based MAMPs, different strategies are needed for nonproteinaceous MAMPs.

Strategy 2: Polymorphisms in glycan MAMPs
Glycan polymorphism is the second strategy used by microorganisms to evade host immunity. This strategy is illustrated here by the deacetylation of chitin by fungi.
Many fungi evade host immunity by deacetylating chitin into chitosan. Chitin is a structural element of fungal cell walls and chitin fragments are conserved MAMPs that are almost universally recognized in the plant kingdom by Chitin Elicitor Receptor Kinase 1 (CERK1) as a signature of fungal invasion (Pusztahelyi, 2018). Chitosan, however, induces a weaker immune response than chitin, so the deacetylation of chitin is frequently used by plant-associated fungi to avoid recognition (Vander et al., 1998). The soil-borne pathogenic fungus Verticillium dahliae, for example, secretes Polysaccharide Deacetylase 1 (VdPDA1) to deacetylate chitin oligomers and prevent chitin-triggered immunity in cotton (Gossypium hirsutum) plants ( Figure 2B; Gao et al., 2019). Fusarium oxysporum f. sp. vasinfectum PDA1 is also required for virulence during wilt disease in cotton (Gao et al., 2019). Similarly, the wheat stripe rust fungus Puccinia striiformis f. sp. tritici suppresses chitin-induced plant defense by secreting the Polysaccharide Deacetylase Pst13661 (Xu et al., 2020). The wheat (Triticum aestivum) stem rust fungus P. graminis f. sp. tritici, the broad bean (Vicia faba) rust fungus Uromyces fabae, and the maize (Zea mays) anthracnose fungus Colletotrichum graminicola also use chitosan instead of chitin in their hyphae (El Gueddari et al., 2002). Likewise, the endophytic fungus Pestalotiopsis sp. avoids recognition by secreting chitin deacetylases (PesCDA; Cord-Landwehr et al., 2016). Chitosan is also produced by human pathogens to evade immunity. The human fungal pathogen Cryptococcus neoformans, for instance, secretes chitin deacetylase CnCda4 to further deacetylate chitosans that are already partially deacetylated by other chitin deacetylases to evade host immunity (Hembach et al., 2020).
In conclusion, the evolution of glycoforms on the exposed portion of microbes is an efficient strategy to dampen the host immune response. More examples of microbial glycan polymorphisms that evade immunity remain to be discovered. For instance, bacterial pathogens of animals modify the structure of their peptidoglycans (PGNs) through deacetylation and thus evade the antibacterial activity of lysozyme and delay pro-inflammatory immune responses (Boneca et al., 2007;Shimada et al., 2010;Wolf et al., 2011;Grifoll-Romero et al., 2019). However, PGN modification remains to be described for plant pathogens.

Strategy 3: Downregulating MAMP production
Another microbial strategy to evade host detection is to reduce the amount of MAMPs by downregulating their biosynthesis during infection. This strategy has been described for both bacteria and fungi.
Pathogenic, opportunistic, and commensal bacteria downregulate flagellin biosynthesis during infection. Biosynthesis of flagella in Pseudomonas is downregulated by the second messenger cyclic-di-GMP (cdG; Hickman and Harwood, 2008). Elevated cdG levels in the plant pathogen P. syringae, the plant opportunist P. aeruginosa and the plant commensal P. fluorescens reduce flagellin levels, and thus contribute to the evasion of FLS2-mediated immune response in Nicotiana benthamiana and Arabidopsis (Pfeilmeier et al., 2016). Flagellin protein levels are also downregulated in PtoDC3000 via reduction in flagellin expression by the gene expression regulator AlgU to avoid host immune responses (Bao et al., 2020).
Downregulation of flagellin genes is also observed upon entry of P. syringae pv. syringae B728a (PsyB728a) into the leaf of bean (Phaseolus vulgaris) plants (Yu et al., 2013). Comparison of the global transcriptome profiling of PsyB728a in epiphytic and apoplastic sites reveals that the mean induction level of genes related to flagellar biosynthesis and motility was >4.5-fold greater on leaf surface than in planta (Yu et al., 2013). Regarding the underlying sensory mechanism to downregulate flagellin expression, Escherichia coli downregulates flagellar genes in response to immobilization with anti-flagellar antibodies (Cullender et al., 2013), indicating that flagellin downregulation occurs when bacteria are immobilized.
In fungal pathogens, downregulation of b-glucan biosynthesis reduces exposure to glycan MAMPs. For instance, during the biotrophic phase of infection, the fungal maize pathogen C. graminicola downregulates the expression of genes encoding Killer toxin resistant 5 (KRE5) and KRE6, which are key enzymes for the biosynthesis of b-glucan ( Figure 2C; Oliveira-Garcia and Deising, 2016). However, KRE5 and KRE6 expressions are indispensable for the formation of appressoria and necrotrophic hyphae. Consistent with a need for b-glucan, KRE genes are also required for full virulence of fungal human pathogens C. neoformans and Candida albicans (Herrero et al., 2004;Gilbert et al., 2010).
In conclusion, downregulating MAMP precursor levels is a common strategy used by both fungi and bacteria to avoid host detection. Obviously, this strategy is only beneficial for the microbe when production of the MAMP precursor is not required for full virulence. For instance, bacterial proliferation and spread do not rely on flagellin after host entry and the altered fungal cell wall composition may no longer need b-glucans upon host entry.

Level 2: Preventing MAMP release
MAMPs discussed in this section are released from microbes by host-secreted hydrolases, such as chitinases and proteases. We distinguish four strategies to block MAMP release from microbes ( Figure 1B). MAMP precursors are protected against host hydrolases by microbial proteins and glycans, and host hydrolases are also inhibited and disintegrated. Specific examples of the four strategies are illustrated in Figure 3.

Strategy 4: Hiding MAMP precursors with proteins
Microbial organisms can evade host recognition by secreting proteins that cover MAMP precursors to prevent hydrolytic The Plant Cell, 2021 Vol. 33, No. 7 release of MAMPs. There are seven examples of this strategy, involving structurally unrelated secreted proteins used by both pathogenic and symbiotic fungi.
The tomato leaf mold fungus (Cladosporium fulvum syn. Passalora fulva) produces Avr4, a member of Carbohydratebinding module family 14 (CBM14). Avr4 specifically binds to chitin in the fungal cell wall to protect it against plant chitinases, which release chitin elicitors (Van den Burg et al., 2006; Figure 3A). Homologs of Avr4 are present in other pathogenic fungi, indicating a similar protection of chitin cell walls by other fungi (Stergiopoulos et al., 2010).
Likewise, the hemibiotrophic xylem-invading fungus V. nonalfalfae prevents chitin hydrolysis by secreting VnaChtBP, a CBM18 protein that binds chitin and suppresses chitin-triggered host immunity (Volk et al., 2019). VnaChtBP is present in 28 V. nonalfalfae isolates, suggesting a high evolutionary stability and testifying its importance for the fungal lifestyle.
Plant beneficial fungi that are parasites of pathogenic fungi also avoid plant immune response by covering MAMPs with proteins. For example, the mycoparasite Clonostachys rosea (sin. Gliocladium roseum) produces CBM50 members LysM1 and LysM2 to protect hyphae against chitinases to prevent MAMP-induced defenses during wheat root colonization by its host F. graminearum (Dubey et al., 2020).
In conclusion, covering MAMP precursors with proteins to prevent their hydrolysis is a mechanism used by many fungal pathogens and symbionts. Remarkably, several structurally unrelated carbohydrate-binding proteins (Avr4, ChtBP, Vd2LysM, ChELP1 and ChELP2, SnTox1, LysM2, and FGB1) have convergently evolved to protect fungal hyphae from hydrolysis by plant chitinases and glucanases, which would otherwise release MAMP from their precursors. In addition to preventing MAMP release, these proteins can also promote virulence by strengthening the cell wall.

Strategy 5: Shielding MAMP precursors with glycans
Glycosylation of MAMP precursors to shield them from hydrolytic release of MAMPs is the fifth strategy to evade host O-glycosylation of flagellin is very common in bacteria, including important plant pathogens from the genera Xanthomonas, Pseudomonas, Burkholderia, Dickeya, Erwinia, Pantoea, and Pectobacterium (Taguchi et al., 2010;Ichinose et al., 2013;De Maayer and Cowan, 2016). This flagellin glycosylation is currently thought to prevent the hydrolytic release of the flagellin MAMP ( Figure 3B). For instance, glycosylation mutants of P. syringae pv. tabaci 6605, P. syringae pv. glycinea race 4 and X. campestris pv. campestris XcA lacking the flagellin glycosyltransferase (FGT1) are less virulent on tobacco, soybean (Glycine max), and cabbage (Brassica oleracea), respectively (Takeuchi et al., 2003;Taguchi et al., 2006;Ichinose et al., 2013). Flagellin glycosylation is also important for Acidovorax avenae, a Gram-negative bacterial pathogen causing leaf blight in rice. Flagellin isolated from the avirulent N1141 strain induces immune responses, whereas flagellin from the virulent K1 strain does not. These flagellin protein sequences are identical but their glycosylation pattern is different: strain K1 carries a 2,150-Da O-glucan while strain N1141 harbors a 1,600-Da O-glycan (Hirai et al., 2011). Thus, glycosylation avoids flagellin recognition, presumably by preventing the hydrolytic release of immunogenic flagellin fragments.
Consistent with shielding glycans, plant-secreted b-galactosidase 1 (BGAL1) acts in immunity by promoting the release of immunogenic peptides from glycosylated flagellin of PtoDC3000 and P. syringae pv. tabaci 6605 (Pta6605), which both carry a terminal-modified viosamine (mVio) sugar on flagellin O-glycans (Buscaill et al., 2019). BGAL1 is not required to release the flagellin MAMP from the Dfgt1 mutant of Pta6605, which produces nonglycosylated flagellin. Interestingly, pv. syringae B728a (PsyB728a) evades host immunity by having O-glycans on flagellin that are resistant to hydrolysis by BGAL1 (Buscaill et al., 2019), even though PsyB728a has an flg22 sequence recognized by FLS2 (Segonzac et al., 2011). Importantly, mVio biosynthesis genes are absent from PsyB728a, which instead carries a putative (1,2)-linked terminal GlcNac (N-acetylglucosamine) on its Oglycan (Yamamoto et al., 2011;Chiku et al., 2013). This suggests that different glycoforms on flagellin are required for the colonization of different hosts and that hosts may use different glycosidases to release flagellin MAMPs.
Bacteria also use polymorphic lipopolysaccharides (LPSs) to evade immunity. LPSs are the major component of the outer membrane of Gram-negative bacteria and consist of lipid A, a di-glucosamine carrying four to seven fatty acids, and an oligosaccharide core region that carries an O-polysaccharide (OPS) comprising a variable number of oligosaccharide repeats. OPS composition is highly diverse among bacterial species and strains. Glycans covering bacterial LPS may alter host responses. For instance, the plant pathogenic bacterium Xylella fastidiosa possesses a long chain O-antigen that delays recognition by the host plant (Rapicavoli et al., 2018). Mutant X. fastidiosa lacking these O-antigens induces faster immune responses (Rapicavoli et al., 2018). In addition, genes that encode glycosyltransferase domains and cause strong virulence phenotypes when disrupted in PsyB728a are suspected to be involved in the biosynthesis of O-antigens that decorates LPS (Helmann et al., 2019).
Glycosylation of fungal cell walls is also used to prevent MAMP release. For instance, the rice blast fungus Magnaporthe oryzae, the rice brown spot fungus Cochlioborus miyabeanus, and the rice sheath blight fungus Rhizoctonia solani accumulate a-1,3-glucan on the surface of infectious hyphae (Fujikawa et al., 2009(Fujikawa et al., , 2012. Fungal mutants with reduced a-1,3-glucan levels have reduced virulence, indicating that a-1,3-glucan may mask chitin and bglucans in the fungal cell wall to shield it against hydrolytic MAMP release (Fujikawa et al., 2012). The large phylogenetic distance between these ascomycete and basidiomycete rice pathogens indicates that this strategy is a widespread mechanism that may be used by fungal pathogens to evade host innate immunity.
Glycan shielding of MAMP precursors has also been described for human pathogens, including viruses, fungi, and bacteria, as a strategy to evade recognition by the host immune system (Walls et al., 2016;Hernández-Chávez et al., 2017;Poole et al., 2018;PradHan et al., 2019). For instance, flagellin glycosylation reduces recognition of human bacterial pathogens, providing an evasive strategy for P. aeruginosa (Arora et al., 2005) and B. cenocepacia (Hanuszkiewicz et al., 2014). In conclusion, glycan shielding of MAMP precursors is a common strategy used by many pathogens to enhance infection. These findings predict that many more examples of glycan shielding will be discovered for plant pathogens and symbionts.

Strategy 6: Blocking MAMP release by inhibiting the activity of host hydrolases
The inhibition of MAMP-releasing hydrolases is another strategy used by plant pathogens. Examples of this strategy have been identified in oomycetes and bacterial pathogens.
Phytophthora species secrete glucanase inhibitor proteins (GIPs) during invasion of their hosts, which themselves inhibit MAMP release through extracellular Endo-b-1,3-Glucanases (EGases). For example, P. sojae secretes GIP1 to inhibit soybean EGaseA, thereby preventing EGaseA-mediated release of elicitor-active glucan oligosaccharides ( Figure 3C; Rose et al., 2002). Analysis of tomato leaves inoculated with P. infestans showed that P. infestans GIPs and tomato EGases form stable complexes in the apoplast (Damasceno et al., 2008), indicating that GIPs-mediated inhibition of EGases to prevent MAMP release is a common strategy used by Phytophthora in different hosts.
Thus, EPI1 might suppress PC2-elicited host immunity by inhibiting the protease that releases the elicitor.
Bacterial pathogens also produce hydrolase inhibitors to prevent MAMP release. For instance, PtoDC3000 produces the small molecule BGAL1 inhibitor galactosyrin (Buscaill et al., 2019). BGAL1 promotes the release of MAMPs from glycosylated flagellin carrying mVio on their O-glycan (see Strategy 5). BGAL1 is suppressed by galactosyrin during infection to prevent the release of immunogenic peptide from flagellin (Buscaill et al., 2019). GIP1, EPI1, and galactosyrin are just the first examples of pathogen-derived inhibitors targeting MAMP-releasing host hydrolases. Further studies on widespread pathogen-derived inhibitors will probably uncover many more examples. However, in addition to preventing MAMP release, these inhibitors also protect the physical integrity of flagellin and the microbial cell wall by preventing their degradation.

Strategy 7: Disintegrating host-derived hydrolases
Destruction of MAMP-releasing host hydrolases is the seventh strategy used by invading microorganisms to evade immunity. We currently know three unrelated classes of proteases from fungal pathogens that disintegrate host chitinases to prevent MAMP release.
In conclusion, SSEP1, FoSep1, FoMep1, and other fungalysins represent different protease families that cleave plant chitinases to prevent MAMP release. The phylogenetic distance between these fungi and proteases indicates that the inactivation of chitinases evolved convergently. Besides preventing MAMP release, these proteases also protect the physical integrity of the cell wall by preserving chitin.

Level 3: Preventing MAMP perception
Once MAMPs are released, we know two effective strategies that prevent the perception of MAMPs by PRRs ( Figure 1C). One strategy degrades MAMPs before they reach their receptors, the other strategy sequesters MAMPs before they are perceived. Specific examples of the two strategies are illustrated in Figure 4.

Strategy 8: Degrading MAMPs
The eighth strategy used by pathogens is based on the targeted degradation of released MAMPs by pathogen-derived proteases. Examples are the bacterial effectors AprA and LasB and fungal effectors with chitinase activity (EWCAs), explained below.
Pseudomonas species secrete alkaline protease ArpA, a 50-kD zinc metalloprotease (MEROPS family M10 of clan MA), which prevents flagellin-triggered immune responses by degrading flagellin monomers and flg22 ( Figure 4A; Bardoel et al., 2011). Flagellin polymers resist degradation by AprA and this preserves the integrity of the flagellum (Bardoel et al., 2011). In Arabidopsis, AprA prevents flg22-and flagellin-induced immune responses and delays stomatal closure. AprA of PtoDC3000 is required for its full virulence on both Arabidopsis and tomato. Interestingly, AprA is widespread among human-and plant-pathogenic bacteria, including the bacterial plant pathogen P. syringae and human pathogen P. aeruginosa, suggesting a conserved infection mechanism among bacteria.
In addition to AprA, P. aeruginosa secretes a second zinc metalloprotease, the 33-kD pseudolysin LasB (MEROPS family M4 of clan MA), which also degrades flagellin and acts in concert with AprA to prevent flagellin-mediated immune recognition (Casilag et al., 2016). The production of two different proteases targeting flagellin monomers likely provides P. aeruginosa with robust immune suppression mechanisms.
Fungal pathogens also degrade MAMPs during infection. The cucurbit powdery mildew fungus Podosphaera xanthii releases effectors with chitinase activity (EWCAs) when penetrating melon (Cucumis melo) plant cells to degrade immunogenic chitin oligomers and thereby prevents the activation of chitin-triggered immunity (Martínez-Cruz et al., 2021). Remarkably, EWCA homologs are also widely distributed in plant fungal pathogens but also in fungi that are pathogens of insects, nematodes, and animals, suggesting a conserved infection mechanism among fungi (Martínez-Cruz et al., 2021).
In conclusion, the degradation of MAMPs is an efficient mechanism to avoid recognition but only a few of these proteases have been identified. Additional pathogen-produced proteases that promote virulence (Chandrasekaran et al., 2016;Figaj et al., 2019) may also act by degrading MAMPs. Likewise, pathogen-secreted glycosidases may degrade glycan-based MAMPs. For instance, the human pathogen Histoplasma capsulatum secretes endo-b-1,3-glucanase Eng1 to evade host detection (Garfoot et al., 2016) and many plant pathogens having glycan-based MAMPs secrete glycosidases (Ospina-Giraldo et al., 2010;Murphy et al., 2011;Vermassen et al., 2019;McGowan et al., 2020). Notably, MAMP degradation must require fine regulation of these hydrolases to avoid unspecific or premature degradation of MAMP precursors or inadvertent MAMP release.

Strategy 9: Sequestering released MAMPs
Elicitor sequestration is the ninth strategy used by plant pathogens to evade recognition. In this strategy, pathogensecreted proteins bind released elicitors to prevent them from binding host receptors. Many fungi secrete proteins to sequester chitin elicitors.
During infection, C. fulvum secretes Extracellular protein 6 (Ecp6), a LysM-containing protein of the CBM50 family that binds chitin fragments. Ecp6 binds these elicitors with greater affinity than the chitin receptor, so C. fulvum evades chitin recognition by using Ecp6 to sequester chitin fragments, (Figure 4B;De Jonge et al., 2010). The widespread presence of Ecp6 orthologs suggests that this is a common strategy of many pathogenic fungi to avoid host recognition (De Jonge and Thomma, 2009). Indeed, the fungal wheat pathogen Mycosphaerella graminicola produces Mg3LysM, an Ecp6 homolog, that plays a major role in pathogen virulence on wheat plants by preventing the elicitation of chitin-induced immunity (Marshall et al., 2011). Similarly, the rice blast fungus M. oryzae produces Secreted LysM protein-1 (Slp1), which binds chitin fragments and prevents chitintriggered immunity (Mentlak et al., 2012). Interestingly, Vd2LysM from V. dahlia and ChELP1 and ChELP2 from C. higginsianum, also sequester chitin oligomers and additionally protect chitin polymers against chitinases (Strategy 4; Takahara et al., 2016;Kombrink et al., 2017).
Surprisingly, even necrotrophic fungi use LysM proteins to evade immunity. For instance, the necrotrophic fungus R. solani, which kills seedlings and causes root rot in a broad range of plant species, secretes RsLysM, which associates with chitin oligomers to prevent early chitin perception during sugar beet (Beta vulgaris) colonization (Dölfors et al., 2019).
LysM proteins are also used by animal fungal pathogens and contribute to fungal virulence during host invasion. The insecticidal fungus Beauveria bassiana secretes the two LysM effectors Blys2 and Blys5 that bind chitin and prevent the activation of immunity in insects (Cen et al., 2017).
Bacteria also use the sequestration strategy by targeting MAMP precursors. For instance, the endophyte bacterium Bacillus subtilis BSn5 enhances its colonization of Arabidopsis and voodoo lily (Amorphophallus konjac) through minimizing the stimulation of flg22-induced defense by producing the antimicrobial peptide (lantibiotic) subtilomycin, which binds to its own flagellin (Deng et al., 2019). The presence of subtilomycin biosynthesis genes in genomes of other bacteria suggests that flagellin sequestration is a common strategy of endophytic bacteria to adapt to endosphere niches (Deng et al., 2019).
MAMP sequestration is common to fungal and bacterial microbes, but more details remain to be discovered in other invading microorganisms. The success of this strategy depends on the competition between the microbial-derived sequestering protein and the host PRR.

Concluding remarks and perspectives
Successful plant-associated microbes evade extracellular detection by the host immune system. While certain immune evasion mechanisms are used by microbes from diverse kingdoms, other mechanisms have so far only been described for certain microbes. However, it seems unlikely that these strategies are pathogen-specific, prompting us to expect that this is likely to change with further development of the field.
For most of the nine strategies described above, suppression MAMP perception also results in increased stability of the MAMP precursor. It has therefore not always been robustly demonstrated that the observed enhanced virulence associated with the strategy is caused by evading MAMP recognition or by increased stability of the MAMP precursor. A good way to investigate this experimentally would be to test for enhanced virulence in the absence of the PRR, as this would indicate an important role in the stabilization of the MAMP precursor.
We can think of at least three additional extracellular strategies that would prevent MAMP recognition ( Figure 1D). First, MAMP recognition can be blocked by receptor shedding. This has been described for animal pathogens, but not for plant pathogens. For example, the fungal respiratory pathogen Coccidioides posadasii secretes the Metalloproteinase Mep1 during endospore differentiation. Mep1 digests the Spherule Outer Wall glycoprotein (SOWgp), resulting in the prevention of host recognition mediated by this receptor (Hung et al., 2005). Similarly, the opportunistic pathogen of human lungs P. aeruginosa secretes the metalloproteinase LasB, which cleaves the human urokinase-type Plasminogen Activator Receptor (uPAR) through domain-specific endoproteolysis (Leduc et al., 2007). There have been no reports of PRR inactivation by shedding in plant-microbe interactions yet. By contrast, ectodomain shedding of the legume Symbiosis Receptor Kinase (SYMRK) causes the formation of a signaling complex with Nod Factor Receptor 5 (NFR5, Antolín-Llovera et al., 2014).
A second possible strategy is the use of antagonists. MAMP antagonists would bind PRRs and prevent the binding of MAMPs to their respective receptors and therefore inhibit PRR function. For instance, C-terminal truncations of the flagellin flg22 elicitor can block flg22 perception by FLS2 in tomato (Meindl et al., 2000), but the existence and use of MAMP antagonists during infection remains to be reported.
A third possible strategy is to alter the apoplastic conditions such that MAMP release and/or perception is inhibited. Although this mechanism has not yet been demonstrated, the regulation of hydrolases by pH and redox status would offer opportunities for pathogens to interfere in MAMP release. Interestingly, several plant pathogens secrete homologs of plant regulatory peptides to suppress host immunity. For instance, the fungal wilt pathogen F. oxysporum secretes a functional homolog of rapid alkalinization factors (RALFs), peptide hormones that trigger an increase of apoplastic pH and enhances fungal colonization (Masachis et al., 2016;Thynne et al., 2017).
In conclusion, the extracellular detection of MAMPs by plants is an active and exciting research field. The presence of many MAMPs, hydrolytic enzymes, and hydrolase inhibitors implicate a large and mostly unexplored area of research, still holding most of its secrets. Increased understanding of this extracellular battlefield of both animal and plant pathogens will ultimately translate into new strategies for the prevention of infectious diseases.

Supplemental data
The following materials are available in the online version of this article.
Supplemental Data Set S1. Overview of strategies employed by microbes to evade MAMP recognition in plants.