Physiological and Molecular Understanding of Bacterial Polysaccharide Monooxygenases

SUMMARY

Bacteria have long been known to secrete enzymes that degrade cellulose and chitin. The degradation of these two polymers predominantly involves two enzyme families that work synergistically with one another: glycoside hydrolases (GHs) and polysaccharide monooxygenases (PMOs). Although bacterial PMOs are a relatively recent addition to the known biopolymer degradation machinery, there is an extensive amount of literature implicating PMO in numerous physiological roles. This review focuses on these diverse and physiological aspects of bacterial PMOs, including facilitating endosymbiosis, conferring a nutritional advantage, and enhancing virulence in pathogenic organisms. We also discuss the correlation between the presence of PMOs and bacterial lifestyle and speculate on the advantages conferred by PMOs under these conditions. In addition, the molecular aspects of bacterial PMOs, as well as the mechanisms regulating PMO expression and the function of additional domains associated with PMOs, are described. We anticipate that increasing research efforts in this field will continue to expand our understanding of the molecular and physiological roles of bacterial PMOs.

INTRODUCTION

Cellulose (C₆H₁₀O₅)_n and chitin (C₈H₁₃O₅N)_n are the two most abundant biopolymers on Earth (1, 2) and are composed of β(1-4)-linked d-glucose (Glc) and β(1-4)-linked N-acetyl-d-glucosamine (GlcNAc), respectively. These biopolymers are widely distributed due primarily to their high resiliency and versatility in combining with proteins and other compounds to form hybrid complexes (3, 4). Cellulose is the primary component of plant cell walls, but it is also found in some bacteria, fungi, and protozoa (5). Cellulose has twice the tensile strength of chitin, conferring an advantage as a structural component for land plants over chitin (6). On the other hand, chitin has played a widespread role in numerous living organisms since the Cambrian life explosion (3, 7). In crustaceans and insects, chitin supports the exoskeleton and the attachment of muscles to joints and defends against pathogens and predators (8–10). In yeasts, amoebas, fungi, and sponges, chitin provides rigidity and strength to the cell wall (11–14). Also, chitin confers buoyancy in some marine photosynthetic microorganisms by increasing the cell surface area (15).

Bacteria have developed specialized strategies to efficiently degrade chitin and cellulose. Cellulose-degrading bacteria are prevalent in terrestrial and submarine soil (16), whereas chitin-degrading bacteria are prevalent in both terrestrial and aquatic environments (17). The degradation of cellulose and chitin predominantly involves two secreted-enzyme families that work synergistically with one another: glycoside hydrolases (GHs) and polysaccharide monooxygenases (PMOs) (18–23). GHs, which are common in bacteria (17, 24, 25), have been known for a relatively long period of time and have been the subject of recent reviews (26, 27). Conversely, PMOs are a recent addition to the known biopolymer degradation machinery; their oxidative chemistry was first reported in 2010 (19), and since then, they have been identified in eukaryotic, bacterial, and viral genomes (28). PMOs are commonly associated with fungal metabolism and biomass degradation, and the potential for these proteins in the biorefinery industry has already been extensively reviewed (20, 29, 30). Bacterial PMOs have also been implicated in having numerous physiological roles. Thus, here we focus on PMO-containing bacteria and their diverse physiological and functional roles.

MOLECULAR ASPECTS OF POLYSACCHARIDE MONOOXYGENASES

The crystalline structures of cellulose and chitin make these biopolymers resistant to hydrolysis. PMOs use an oxidative process leading to the cleavage of β-1,4 glycosidic linkages, creating new chain ends for glucanase action (19, 31, 32). PMOs are intriguing enzymes, not only as they oxidize C-H bonds but also because this chemistry does not require separation of the polysaccharide chain from the crystalline matrix backbone for bond cleavage. Since their discovery, significant contributions to the biochemical characterization of PMOs have been made, and other reviews have thoroughly covered PMO structure and the copper active site relating to catalysis (28, 33–35); therefore, only a brief summary is given here.

Classification of Polysaccharide Monooxygenases

The Carbohydrate-Active enZYmes (CAZy) database (36), classifies PMOs as auxiliary activity (AA) enzymes and places all PMOs into four families based on sequence homology: AA9 (formerly GH61), AA10 (formerly CBM33), AA11, and AA13. The AA9, AA11, and AA13 families are found in fungi and oxidize cellulose, chitin, and starch, respectively (37–40). All bacterial PMOs belong to the AA10 family and are active on either cellulose or chitin (19, 32, 41). To date, only approximately 20 out of the over 2,400 putative bacterial PMOs present in the CAZy database have been biochemically characterized, leaving many unanswered questions about this large group.

Evolution of Bacterial PMOs

Bacterial PMOs most likely share a distant common ancestor with fungal PMOs belonging to the AA9 PMO families (42); however, no AA9 or AA11 PMOs are found in bacterial genomes. Phylogenetic clustering of bacterial PMO sequences separates AA10 PMOs into two major clades: clade I and clade II. Clade I consists of chitin-active PMOs, whereas clade II consists of cellulose-active PMOs (23, 40, 42). Further selection for more favorable binding interactions with substrates and other proteins likely influenced the origin of additional subclades (23, 42). An interesting question arises as to whether bacterial PMOs initially evolved to utilize chitin or cellulose substrates. One could speculate that since chitin was present before cellulose, PMOs could have originated to degrade chitin. The first chitin fossil record dates to 505 million years ago (middle Cambrian) and was isolated from a skeletal component of early sponges (43), although chitin probably originated in protozoans 1,400 million years ago (44). Cellulose originated in land-adapted plants later, around 450 to 470 million years ago (45). Indeed, recent phylogenetic analyses suggested that bacterial PMOs first originated to degrade chitin and then evolved for cellulose degradation (42).

Mechanism

PMOs are mononuclear copper-dependent monooxygenases that hydroxylate either the C-1 or C-4 position of the glycosidic bond, forming an unstable intermediate, which decomposes with concomitant bond cleavage. Following bond cleavage, C-1 hydroxylation ultimately produces aldonic acids through a hemiacetal intermediate, whereas C-4 hydroxylation ultimately leads to ketoaldoses through a hemiketal intermediate (Fig. 1). Loop flexibility, particularly the “loop 2” region, has been reported to have a role in conveying oxidative regioselectivity, although the molecular basis for regioselectivity remains unanswered (46). The chemical mechanism of PMOs, including the catalytic steps and the active-site oxidant, still remains to be determined (28, 47). PMOs are also called lytic PMOs (LPMOs) since C-1/C-4 hydroxylation ultimately results in the cleavage of the glycosidic bond. We have commented on this nomenclature previously (28).

FIG 1.

The reaction catalyzed by polysaccharide monooxygenases. Oxidation at the C-1 or C-4 position of the glycosidic linkage produces aldonolactones and 4-ketoaldoses, respectively (19, 152). Aldonolactone products are most commonly observed as aldonic acids, and 4-ketoaldoses are most commonly observed as gemdiols. For cellulose, R indicates Glc and R′ indicates OH; for chitin, R indicates GlcNAc, and R′ indicates NHCH₃CO.

Two one-electron reductions and two proton transfers are necessary to activate O₂ for the oxidation of the glycosidic bond (Fig. 1). Small-molecule reductants and photosynthetic pigments (48) have been shown to directly serve as electron donors for bacterial PMOs. Recently, it was shown that phenols coupled with members of the glucose-methanol-choline (GMC) family of oxidoreductases were effective electron donors for fungal PMOs (49). Also, cellobiose dehydrogenase (CDH), a known electron donor for fungal PMOs (50, 51), has also been shown to serve as an electron donor for bacterial PMOs (52). Thus, although not yet identified, extracellular oxidoreductases or redox-active enzymes likely serve as biological redox partners for bacterial PMOs. Due to the range of environments in which bacterial PMOs are found, as well as their diverse functions (see below), the biologically relevant electron donor may vary based on the bacterial environment.

Structure of Bacterial PMOs

The first crystal structure of a bacterial PMO, that of the Gram-negative bacterium Serratia marcescens, was reported in 2005 (53). Since then, multiple crystal structures and one nuclear magnetic resonance (NMR) structure (54) have been reported for various organisms, including Serratia marcescens, Enterococcus faecalis (55, 56), Vibrio cholerae (57), Streptomyces coelicolor (41), Jonesia denitrificans (58), and Bacillus amyloliquefaciens (34). These structures show that the catalytic domains of bacterial PMOs have a β-sandwich fold similar to that of the other PMO families. The active site coordinates a type II mononuclear copper with two histidines and the N-terminal amine in a T-shaped histidine brace (Fig. 2). The active site is found on a flat substrate binding surface ideal for interactions with crystalline substrates (19, 37–40, 54–56, 59, 60), which allows PMOs to cleave the glycosidic bond without the need to separate the polysaccharide chain from the crystalline matrix. Chitin-active bacterial PMOs have a small pocket near the copper active site that might serve to accommodate the acetyl group found on the glucosamine moiety (41).

FIG 2.

Crystal structure of ScLPMO10B from Streptomyces coelicolor (PDB accession number 4OY6) (153). The histidine brace motif, consisting of the two copper-coordinating histidine residues, is shown in the inset. The surface of the protein is shown in light gray and illustrates the flat substrate binding surface.

The first crystal structures of a PMO with bound cello-oligosaccharides were recently reported for the fungus Lentinus similis (61). These structures provide insight to the molecular aspects involved in substrate binding for bacterial PMOs. Hydrogen bond contacts, a lone pair-aromatic interaction involving His1, and a CH-π interaction help to bind the cello-oligosaccharide (61). The CH-π interaction forms between the pyranose ring of the carbohydrate and a highly conserved Tyr on the surface of the PMO. The hydrogen-bonding network that forms between the oligosaccharide, a water molecule, and N-terminal His1 may serve as a proton transfer pathway needed to stabilize a reactive Cu-O₂ intermediate (61, 62).

Domain Structure

PMOs are active as single-domain proteins; however, in some cases, the full-length polypeptide has been found to contain additional domains. These additional domains include carbohydrate binding modules (CBMs), fibronectin type III-like domains (FnIIIs), hydrolase domains, and polycystic kidney disease domains (PKDs) (20). Domains with unknown function have also been found fused to PMOs in the AA11 family (39). CBMs are classified into numerous families, based on amino acid sequence. These domains contain aromatic residues proposed to increase binding to various polysaccharides (63). Likewise, CBMs fused to PMOs have been proposed to increase polysaccharide specificity and binding (64, 65) and to modulate activity (18, 41). CBMs have also been implicated in aiding in substrate processivity by GHs (66), but to date, processivity has yet to be shown for PMOs. FnIIIs and PKDs are typically associated with mediating cell adhesion and protein-protein interactions (67–73). Therefore, these domains may facilitate substrate binding; however, exactly how they do so is not known, especially since the PMO catalytic domain is capable of binding the substrate. Thus, a physiological perspective is necessary to achieve a deeper understanding of the role of multidomain PMOs.

Studies on the V. cholerae PMO GbpA (VCA0811), a PMO with four domains, provide some insight into the physiological role of PMO domain architecture. GbpA is composed of four domains: a catalytic PMO domain, a chitin binding domain, and two bacterial flagellin-like domains (57). Binding of V. cholerae to mucin in the intestinal epithelium of the host is mediated by the GbpA catalytic domain and the two surface flagellin-like domains (57); it is unclear if the catalytic domain has an oxidative role in this process. On the other hand, in an aquatic habitat, the catalytic domain and the chitin binding domain of GbpA attach to the exoskeleton of crustaceans and cleave chitin (74). PMO-associated domains can contribute to our understanding of the physiological roles of these enzymes. Thus, a detailed understanding of PMO-associated domains found in other bacteria as well as the substrates with which these PMOs interact is needed not only to advance the field but also to have a better understanding of the physiological roles of these domains.

PHYSIOLOGICAL ASPECTS OF POLYSACCHARIDE MONOOXYGENASES

Genomic Comparison of Bacterial PMOs

The Pfam hidden Markov model for bacterial PMOs can be found under accession number PF03067 (75, 76), and it can also be accessed under InterPro database accession number IPR004302. This database identifies PMO families based on sequence homology. However, this database is not comprehensive, and putative PMOs are still being discovered, creating new Pfams (77, 78). Through the PF03067 database, putative PMOs seem to be extensively present in the phyla Proteobacteria, Actinobacteria, and Firmicutes. The phylum Chloroflexi possesses only three species with PMOs, the phylum Bacteroidetes possesses two species, and the phyla Chlamydiae and Verrucomicrobia possess one species each, indicating that these phyla have a low number of PMOs. The Armatimonadetes, Cyanobacteria, Fibrobacteres, Planctomycetes, Spirochaetae, and Thermotogae phyla did not contain any putative PMOs. These results are in agreement with data from two previous reports (17, 40).

The presence of PMOs in bacterial genomes appears to be influenced by numerous variables, including habitat and lifestyle. Since PMOs are oxygen dependent, it is not surprising that anaerobic microbial communities lack PMOs (17). Likewise, it is not surprising that many bacteria isolated from soil or decomposing biomass possess PMOs. Xylanimonas cellulosilytica, a Gram-positive bacterium of the actinobacterial family possessing two PMOs (40), was isolated from a decaying tree and shown to degrade both cellulose and xylan (79). Similarly, Saccharophagus degradans, a Gram-negative proteobacterium possessing one PMO (40), is a carbohydrate-degrading marine bacterium that can degrade chitin and cellulose, among other polymers (80). However, there are examples of aerobic cellulolytic bacteria that do not possess PMOs. Both Acidothermus cellulolyticus, a Gram-positive bacterium of the actinobacterial family isolated from an acidic hot spring (81), and the saprophytic Gram-negative proteobacterium Agrobacterium radiobacter, which grows on decayed plant matter (82), lack PMOs. These examples suggest that other variables beyond lifestyle could dictate the presence of PMOs in bacterial genomes.

Environmental Factors: Regulator Elements of Cellulose and Chitin Metabolism

Some bacteria use cellulose and chitin as signaling molecules (Fig. 3) (83) and as markers for nutrient availability and developmental control (84). In polymer-enriched habitats, bacteria secrete PMOs as well as cellulases and chitinases (18–23, 85) to degrade cellulose and chitin into smaller soluble oligosaccharides; this could possibly provide the bacterium with simpler carbon and nitrogen sources (Fig. 3). PMOs can be upregulated when bacteria are grown in the presence of cellulose or chitin (86, 87), indicating that bacteria have developed mechanisms to regulate the expression of these PMOs in response to these potential substrates (86–88) (Fig. 3 and see below). For example, the thermophilic actinobacterium Thermobifida fusca secretes the PMOs E7 and E8 when grown on cellulolytic polymers; these PMOs work synergistically with T. fusca GHs to form cellobiose, a sugar utilized for growth (89). Genes that are in the same operon are coexpressed and can provide valuable insight into gene function and protein interactions. However, PMOs do not appear to be organized into operons but instead are distributed throughout the whole genome close to regulatory systems involved in sensing cellulose or chitin (88, 90).

FIG 3.

Bacterial sensing of cellulose and chitin. Bacteria can sense cellulose and chitin through CebR and DasR, respectively. In turn, CebR and DasR regulate the expression of PMOs and other GHs. PMOs cleave cellulose and chitin chains, generating new chain ends more readily accessible by GHs for further degradation.

The genus Streptomyces illustrates very well the complex network of the transcriptional control of both cellulose- and chitin-active PMOs. Streptomyces sp. SirexAA-E has six predicted PMOs. During growth on cellulose, Streptomyces sp. SirexAA-E secretes three putative PMOs, SACTE_3159, SACTE_6428, and SACTE_2313, whereas the putative PMOs SACTE_2313, SACTE_0080, and SACTE_6493 are secreted when grown with chitin (88). SACTE_2313 was the only putative PMO secreted under conditions of both cellulose and chitin growth, suggesting substrate-specific responses for the other four PMOs.

CebR, the master regulator of cellulose/cello-oligosaccharide catabolism, is involved in sensing cellulose and cellooligosaccharide following their uptake in Streptomyces (86). In the presence of cello-oligosaccharides, including cellobiose, binding of CebR to DNA is weakened, allowing downstream gene transcription to be enhanced (91) (Fig. 3). Genes regulated by CebR include cellobiose/cellotriose ABC transporters, cellulose-active PMOs, and other cellulases (86). In Streptomyces, these genes are found close to a palindromic CebR binding element, which regulates their expression under cellulose growth (86, 88).

Chitin and chito-oligosaccharide catabolism, on the other hand, is regulated by the GntR/HutC family regulator DasR, a global transcriptional regulator in Streptomyces (84). In response to the uptake of the chitin monomer N-acetylglucosamine (GlcNAc), DasR regulates numerous genes, including those involved in the phosphotransferase system, the chitobiose ABC transporter system, as well as chitin-active PMOs and other chitinases. Interesting, the DasR-DNA complex does not dissociate in the presence of GlcNAc; rather, glucosamine-6-phosphate, a intermediate in GlcNAc metabolism, is the effector molecule (84) (Fig. 3). In Streptomyces coelicolor, the SCO0481, SCO2833, SCO6345, and SCO7225 genes were upregulated by DasR (87). Based on sequence alignments, these genes encode putative chitin-active PMOs (see Fig. 5). When Streptomyces coelicolor A3 was grown in soil, SCO7225 (chiM) was the only putative PMO that was upregulated (92). Microarray analyses confirmed that there were 21- and 3-fold upregulations of the putative PMOs SCO7225 and SCO2833, respectively, in soil with and without chitin; there was no change in SCO6345 under these conditions (93). These changes in PMO transcripts are consistent with a role for chitin in gene regulation.

FIG 5.

Sequence alignment of conserved regions of bacterial PMOs. Green highlighting shows copper-coordinating residues, gray highlighting shows conserved residues on the putative substrate binding surface, and purple highlighting shows a conserved aromatic residue by the active site. Numbers indicate the residue of the reference sequence (italics). Shown are bacterial PMOs predicted to oxidize C-1 of cellulose (A), C-1/C-4 of cellulose (B), and chitin (C).

Besides cellulose- and chitin-rich environments, other environmental factors may influence the expression levels of PMOs. The ability of the Gram-negative bacterium Vibrio cholerae to survive under diverse environmental conditions is enhanced through attachment to the chitinous exoskeletons of zooplankton (94, 95), and the PMO GbpA (VCA0811) plays an important role in chitin degradation for growth (96). Ecological studies have determined that GbpA and the ability to bind to chitin are influenced by environmental factors such as increases in temperature (97) and quorum sensing (98). The detailed functions of GbpA are described below in the section on V. cholerae.

FUNCTIONS OF BACTERIAL PMOs

Bacterial PMOs have been implicated in various functions, including nutrition, endosymbiosis, and virulence in pathogenic organisms (88, 96, 99–101) (Fig. 4). It is still too early to definitively know the function of bacterial PMOs, and it is highly likely that the physiological role of PMOs varies from organism to organism. In this section, we discuss the advantages that PMOs confer when organisms rely on carbon sources other than glucose, the gene regulatory network involved in sensing cellulose and chitin, the primary role of PMOs in facilitating host-microbe interactions, and, finally, the properties of PMOs as antifungal agents and virulence factors.

FIG 4.

Putative functions of bacterial PMOs. Following secretion, bacterial PMOs can oxidize a range of polysaccharide substrates, resulting in PMOs being implicated as having various functions, including degrading biomass, being involved in endosymbiotic relationships, and serving as virulence factors of pathogenic bacteria and as antifungal agents, in addition to providing a nutritional source for bacteria.

Host-Microbe Interactions

Endosymbiosis.

Symbiosis between host organisms and cellulolytic bacteria led to the evolution of new feeding strategies. Host organisms have taken advantage of bacterially secreted enzymes, including PMOs, GHs, and other cellulases, to deconstruct cell walls and degrade plant biomass. These symbiotic relationships modulate the interactions of the host organisms with their environment. The first documented acquisition of symbiotic cellulolytic microorganisms was by lower termites originating around 150 million years ago, leading to enormous evolutionary and ecological success (102, 103).

Some strains of Streptomyces sp. are symbiotic with bark beetles and wood wasps, which rely on the degradation of plant biomass (104). These insects take advantage of secreted biomass-degrading enzymes to facilitate energy accumulation inside plant cell walls for larva development (104). Similarly, the giant snail Achatina fulica possesses high cellulolytic activity, mainly due to the abundance of resident microbiota in the gastrointestinal tract (105). Streptomyces sp. strain I1.2 was isolated from this herbivore invertebrate and displayed high growth rates on cellulose as the sole carbon source (105). Among genes that putatively encode enzymes involved in plant cell wall deconstruction, there are four genes encoding putative PMOs (105). The authors of that study concluded that Streptomyces is involved in symbiotic associations with A. fulica and that this association is important for lignocellulose degradation (105). In the marine environment, host-proteobacterium symbiotic relationships potentially contributed to the evolutionary success of bivalves in a cellulose-enriched habitat (106). Unlike terrestrial symbiotic relationships, where cellulolytic bacteria are found in the gut, these microbiotas are found in the gills of the bivalves (107–110). The gammaproteobacterium Teredinibacter turnerae, isolated from the gills of bivalves, contains numerous GHs and other carbohydrate-active enzymes as well as one predicted PMO (TERTU_0046, which contains the two histidines in the active site critical for PMO activity but lacks the conserved regions of the bacterial PMOs depicted in Fig. 5). Additional studies are needed to know when the genes encoding these putative PMOs are induced and to biochemically characterize the enzyme as a PMO.

Antifungal properties.

As the cell wall of fungi is composed primarily of chitin (13), it is not surprising that some bacterial PMOs possess antifungal activity (100, 111–113). Cbp50 (annotated CBM33), from the Gram-positive soil-dwelling bacterium Bacillus thuringiensis, was the first PMO definitively shown to have antifungal properties (100). Cpb50 strongly binds β-chitin but can also bind colloidal chitin and cellulose (100). By binding β-chitin present in the fungal cell wall, Cpb50 may prevent chitin biosynthesis during cell division (100). Other studies have suggested that other PMOs, such as CHB1 and CHB2, may interact with the fungal cell wall. CHB1 and CHB2 are two chitin binding proteins with sequence homology to known PMOs (Fig. 5), are secreted by Streptomyces in the presence of α-chitin, and could bind to mycelia from fungi (111, 112). Another PMO in Streptomyces, CHB3, was later shown to interact with α-chitin, β-chitin, and chitosan, possibly allowing Streptomyces to bind various polysaccharide substrates found in fungal cell walls (114). Subsequently, PMOs similar to CHB1 and CHB2 from Bacillus amyloliquefaciens were shown to bind chitin from fungi (113). However, those studies did not conclude whether CHB1 and CHB2 have antifungal proprieties. It is also important to note that not all PMOs enhance antifungal proprieties of bacteria. It has been suggested that CBP21 in Serratia marcescens, BtCBP in B. thuringiensis, and BliCBP in Bacillus licheniformis are able to bind and degrade chitin but do not have any effect on inhibiting fungal growth (115). However, these PMOs were not suspended in a solution containing reducing agents, complicating interpretations of the results.

Bacterial PMOs as virulence factors.

The biological roles of bacterial chitinases and PMOs and their interactions with insects and fungi can be understood from an environmental perspective where these interactions often result in pathogenesis. Chitin is present in the midgut of most invertebrates, where it serves as a major structural component (116), functioning as a mechanical barrier for protection and compartmentalization of digestive processes (117). Pathogenic bacteria must breach this layer to invade the host organism, and there is growing evidence that PMOs are involved in this process.

Paenibacillus larvae, a Gram-positive bacterium pathogenic to honeybee larvae, is able to degrade the chitin-containing gut during infection (118, 119). Once this layer is breached, P. larvae can initiate the invasive growth phase (99). PlCBP49, a chitin-active PMO (Fig. 5), was shown to be key for honeybee larva colonization, as its absence abolished P. larvae virulence (99).

Role of PMOs in Human Infection

In recent years, attention has been drawn to the role that bacterial chitinases and PMOs play in humans, despite the fact that chitin is not an endogenous component of mammals (120, 121). The bacterial pathogens that show a correlation between PMO activity and infectibility are V. cholerae, Pseudomonas aeruginosa, and Listeria monocytogenes. Unfortunately, too little is currently known to review the pathogenic role that PMOs may play in Enterococcus faecalis infections, where the putative PMO EF0362 is upregulated in the presence of blood and urine (122, 123), and in other pathogenic bacteria with predicted PMOs, such as Serratia marcescens, Bacillus anthracis, and Legionella pneumophila.

Vibrio cholerae.

As mentioned above, V. cholerae can survive under low-nutrient environmental conditions by attaching to the chitinous exoskeletons of zooplankton (94). Chitin is a substrate for the PMO GbpA (124), which is induced along with chitinases when grown in the presence of chitin and chito-oligosaccharides (125). Not only is V. cholerae able to use free chitin particles or the chitinous exoskeleton of copepods as the sole carbon and nitrogen source for growth (96), but also chitin is used for the production of ammonia to increase toxicity to heterotrophic grazers (126).

Kirn et al. hypothesized that the colonization factors promoting adhesion in the natural environment of V. cholerae could be the same for intestinal mucosal surfaces and identified GbpA as the protein responsible for attachment to epithelial cells (101). GbpA may have originated to function in adhesion with environmental biotic substrates and then evolved to be a virulence factor in pathogens (74). V. cholerae colonization of the human intestine is mediated, in part, by GbpA; GbpA interacts with mouse intestinal mucus, with the principal component being the glycoprotein mucin (127). GbpA and mucin mutually enhance expression: GbpA stimulates mucin secretion in a concentration-dependent manner, and mucus secretion increases the level of GbpA (127).

Quorum sensing and cyclic di-GMP (c-di-GMP) signaling are connected pathways that sense and integrate environmental cues, controlling a broad variety of physiological functions in V. cholerae (128). Recently, both of these pathways have been shown to regulate GbpA. At a high cell density, GbpA expression is repressed by HapR, a central regulator of cell density-dependent quorum sensing in V. cholerae (129). Furthermore, GbpA is degraded by two known quorum sensing proteases, HapA and PrtV (98), strongly indicating that GbpA is regulated through quorum sensing. Genome neighborhood analysis revealed that gbpA is near two c-di-GMP binding riboswitches called Vc1 and Vc2 (130). A recent study confirmed that the second messengers c-di-GMP and cyclic AMP (cAMP) regulate gbpA transcription (131). The levels of c-di-GMP and cAMP are regulated by numerous abiotic signals, confirming that GbpA expression is fine-tuned in response to environmental signals (131).

Interestingly, GbpA is not the only PMO predicted to be present in V. cholerae. Although uncharacterized, VCA0140 (which has the two histidines in the active site critical for PMO activity but does not possess the conserved regions of the bacterial PMOs present in Fig. 5) was also strongly expressed when V. cholerae was associated with living copepods (125). Additional studies are needed to characterize and help understand the function of VCA0140.

Pseudomonas aeruginosa.

Pseudomonas aeruginosa is a pathogenic Gram-negative bacterium that inhabits soil and freshwater ecosystems. P. aeruginosa and the pathogenic yeast Candida albicans often form polymicrobial biofilms, especially in the lungs of patients with cystic fibrosis (132). The interaction of these two pathogenic organisms is mainly antagonistic (132). P. aeruginosa can grow on hyphae and kill fungal C. albicans but cannot interact with or kill yeast-form C. albicans (133), likely due to differences in the proteins and carbohydrates present in the cell walls (134).

Although P. aeruginosa is unable to grow on chitin as a sole carbon source (135), the putative PMO CbpD (PSPA7_4667) is secreted when grown with colloidal chitin-enriched medium. Research into the antagonistic interaction between P. aeruginosa and C. albicans provided evidence that CbpD mediates binding and adhesion to the chitin present in C. albicans hyphae (136). Thus, P. aeruginosa cannot kill the yeast form of C. albicans, as CbpD probably cannot interact with the outermost glycoprotein layer of C. albicans yeast cells (136).

There is growing evidence that CbpD has a virulence role in this opportunistic pathogen. First, there is a strong association between CbpD and cystic fibrosis. Cystic fibrosis patients have been found to have high levels of CbpD in their lungs (137), and microarray and protein analyses showed that CbpD was upregulated in strains associated with cystic fibrosis compared to other laboratory strains (137). In addition, in a strain isolated from an acute transmissible cystic fibrosis case, CbpD was abundant in the secretome compared with the laboratory-adapted strain when grown with artificial sputum medium containing mucin (138). Notably, CbpD has been found only in clinical isolates of P. aeruginosa but not in nonpathogenic strains isolated from soil (139).

Second, microarray analysis showed that CbpD is important during the early formation of biofilm in cystic fibrosis lung rather than maintaining the biofilm (140). The transcription of cbpD and its secretion through the type II system are probably under the control of the quorum sensing system (141). Taken together, these results suggest that CbpD plays a potentially pathogenic role in distinct and specific areas of inflammation and disease and is important for the attachment of the surface leading to biofilm formation (137, 138).

It is important to note that P. aeruginosa also possesses the putative PMOs CbpA and CbpE (142, 143). Continued studies will shed light on functional homologies and differences between CbpD, CbpA, and CbpE and their function in P. aeruginosa.

Listeria monocytogenes.

L. monocytogenes is a saprophytic Gram-positive bacterium found in water, soil, and decaying plants (144) that causes foodborne infection leading to septicemia, meningoencephalitis, gastroenteritis, and perinatal infections (145). The gene lmo2467 encoding a PMO (LmPMO10) in L. monocytogenes is not essential for either chitinolytic activity (146) or growth in media with different carbon sources as alternatives to glucose (147). Another study confirmed that LmPMO10 was not present in the secretome, nor was it upregulated when cells were grown in the presence of different types of chitin (148). Since LmPMO10 is not essential for growth in basic chitin medium, it was suggested that it could have a role during pathogenesis. To date, the only transcriptional study of LmPMO10 under different growth conditions revealed that this gene was upregulated when subjected to low temperature or hypoxia, or during stationary-phase growth, but not in blood from healthy human donors or in the presence of the intestinal lumen of mice (149). In addition, LmPMO10 was upregulated at the initial stage of biofilm formation (150), similar to the PMO in P. aeruginosa. A mutant lacking LmPMO10 did not influence bacterial invasion or replication in tissue culture cell lines; however, this mutant exhibited a defect in bacterial colonization in the spleen and liver of infected mice, and LmPMO10 contributed to virulence in the bloodstream of infected mice (151). These studies unfortunately are not conclusive with regard to the possible role of LmPMO10 in virulence, and the exact mechanism of action must still be clarified.

PERSPECTIVES

PMOs in pathogens have been shown to be active on numerous polysaccharides, including cellulose and chitin, yet definitive physiological roles remain unclear. These proteins have been implicated in diverse functions, including facilitating endosymbiosis, providing a source of nutrition, enhancing virulence, and serving as virulence factors in pathogenic organisms (Fig. 4). With the assumption that the putative PMO active sites are functional, the key question is, what are the substrates? A starting point for any putative PMO would be a carbohydrate moiety; however, this opens the door to a great many possibilities. The diversity of carbohydrates beginning with the monosaccharide building blocks generates a long list of potential structures. These building blocks can then be linked into oligo- and then further to polysaccharides with potential branch points leading to diverse structures. Glycan attachment to proteins creates a further diverse array of structure and, correspondingly, potential PMO substrates. It is also possible that saccharide-containing glycans found on mammalian cells and other β(1-4)-linked polysaccharides encountered in pathogenic bacterial biofilms may be the targets of bacterial PMOs. Strategies to screen these extensive substrate classes are clearly needed. An illustrative example is GbpA in V. cholerae, which has been shown to degrade chitin. It was also shown to interact with the glycoprotein mucin, and thus, a role in adhesion was proposed, but mucin as a substrate was not investigated. The physiological roles of PMOs will certainly be connected to the environmental habitat of the bacteria, and clues to function will no doubt be found in these habitats as well as in interactions with other organisms in these natural habitats or with hosts. Mechanisms of host-microbe interactions and pathogenesis will certainly emerge from these studies. Ultimately, a better understanding of PMO expression and function could lead to a deeper understanding of the molecular and physiological mechanisms of bacteria for biomass conversion, feeding strategies, as well as symbiotic relationships.

ACKNOWLEGMENTS

We thank Tyler Detomasi, Yirui Guo, Ben Horst, Christopher Lemon, Elizabeth Ndontsa, Minxi Rao, and Elise Span for critical reading of the manuscript.

This review was prepared by M.A., J.A.H., and M.A.M. M.A. and J.A.H. conducted the primary literature search and contributed to the development, organization, and writing of the manuscript. M.A.M. contributed to the organization, writing, and editing of the manuscript. All authors read and approved the final manuscript.

We declare no conflict of interest.

Biographies

Marco Agostoni obtained his B.Sc. in marine biology with an emphasis on microalgal physiology from the Marche Polytechnic University in Italy. He then moved to the Marine Science Institute at the University of Texas at Austin, where he received his M.Sc. in Marine Science, focusing on harmful algal blooms. During his Ph.D. in the Molecular and Cell Biology Program and the MSU/DOE Plant Research Lab at Michigan State University, he worked on identifying the regulatory roles of the second messenger cyclic di-GMP in cyanobacteria. After he obtained his Ph.D. in May 2015, he began his postdoctoral fellowship at the University of California, Berkeley, where he is unraveling the physiological role of PMOs during host-pathogen interactions and in the environment. The knowledge gained will help expand the understanding of the possible functional roles of bacterial PMOs. His broader interests lie in leveraging microorganisms as effective tools for practical applications.

John A. Hangasky obtained bachelor's of science degrees in chemistry and forensic science from the University of New Haven in New Haven, CT. He then pursued graduate studies at the University of Massachusetts, Amherst, where he received his Ph.D. in chemistry in 2014, based on his work characterizing alpha-ketoglutarate-dependent oxygenases responsible for mediating the hypoxic response of the transcription factor HIF (hypoxia inducible factor). Following his Ph.D., he joined the laboratory of Michael A. Marletta at the University of California, Berkeley, where he is currently a postdoctoral research associate. His current research interests range from identifying and characterizing novel PMOs and the polysaccharides that they degrade to studying the molecular mechanisms of PMOs.

Michael A. Marletta earned an A.B. in chemistry and biology from Fredonia, State University of New York. After completion of a Ph.D. at UCSF with George Kenyon and a postdoctoral fellowship at MIT with Chris Walsh, Dr. Marletta joined the faculty at MIT, where investigations into the mammalian biosynthesis of nitrate led to the laboratory's findings on nitric oxide (NO). He continued structure-function studies involving NO at the University of Michigan and now at the University of California, Berkeley. The emergence of the Energy Biosciences Institute at Berkeley led the Marletta group toward studies into the enzymology of cellulose degradation, subsequently leading to the joint discovery of the polysaccharide monooxygenases. Those studies continue and have been expanded to include PMOs in pathogens.