Periplasmic de-acylase helps bacteria don their biofilm coat

One of the survival strategies used by planktonic bacteria when under stress is to encase their community within an extracellular matrix composed of biopolymers, such as polysaccharides, DNA, and proteins, thereby forming a biofilm that adheres to surfaces and interfaces (1–3). The chemical structure of the polysaccharides involved varies widely among different species of bacteria, but one structure that is common to many medically important biofilm-forming bacteria (both Gram-negative and Gram-positive) is that of partially de-N-acetylated poly-β-1,6-N-acetyl-d-glucosamine (dPNAG) (Fig. 1A) (4). This polymer differs from the more biologically familiar chitin in its 1,6 rather than 1,4 linkage pattern, but would be similarly insoluble were it not for the partial de-N-acetylation that exposes highly polar ammonium substituents on the chain. Similarly, chitin can be converted to a partially de-N-acetylated form known as chitosan. In PNAS, Little et al. (5) use structural, computational, and functional studies to explore the role of PgaB, a de-N-acetylase enzyme that is known to be critical in both the tailoring and extracellular export of the dPNAG component of biofilms produced by Escherichia coli.

Fig. 1.

(A) The structure of dPNAG. (B) The synthase-dependent pathway of dPNAG production and export encoded by pgaABCD in E. coli.

In E. coli dPNAG is produced by a synthase-dependent pathway, one of three main mechanisms by which bacteria produce and secrete polysaccharides (6) [the other two are the Wzx/Wzy-dependent and the ATP-binding cassette (ABC) transporter-dependent pathways (6, 7)]. Common components of synthase-dependent pathways are an inner membrane-spanning synthase protein, an inner membrane bis-(3′-5′)-cyclic dimeric guanosine monophosphate (c-di-GMP) receptor protein, an outer membrane-associated tetratricopeptide repeat (TPR)-containing protein, and an outer membrane-spanning β-barrel porin, with polysaccharide-tailoring enzymes also being present in many of the pathways (6). In the dPNAG pathway, these functions are apparently encoded on the pgaABCD genes. As shown in Fig. 1, the synthase, PgaC, has a transmembrane region with a glycosyltransferase domain on its cytosolic face and is responsible for both the synthesis of the PNAG precursor of dPNAG from UDP-N-acetylglucosamine and its translocation across the inner membrane into the periplasm (8). Regulation of its activity can be conferred by binding of the bacterial second messenger, c-di-GMP, to its receptor protein PgaD. Located within the outer membrane is the porin PgaA, which has a predicted C-terminal β-barrel domain and likely facilitates dPNAG export across the outer membrane. This protein has an N-terminal TPR-containing domain that is thought to protect the polysaccharide in the periplasm and may also be involved in PgaA–PgaB or PgaA–dPNAG interactions (4, 6, 9). Sitting in the periplasm between the two is PgaB itself, a metal-dependent, low-efficiency de-N-acetylase enzyme that modifies the PNAG precursor, partially (∼3–5%) de-N-acetylating it to form dPNAG before its export by PgaA.

PgaB activity appears to be necessary for polysaccharide export and biofilm formation because abrogation of its activity by gene deletion or active site mutation results in accumulation of PNAG in the periplasm (9). This observation prompted interest in its structure, revealing a unique two-domain arrangement wherein both domains adopt (β/α)_x barrel folds (4). The N-terminal domain has a (β/α)₇ barrel fold and belongs to the CAZy family four carbohydrate esterases (CE4) but distinguishes itself in having a unique circularly permutated arrangement and lacking a canonical aspartate residue typically involved in catalysis, thereby likely accounting for its low de-N-acetylase activity. The C-terminal domain has a (β/α)₈ barrel fold and is structurally related to the CAZy family GH18 and GH20 glycoside hydrolases, having a pronounced groove that could potentially accommodate several saccharide units.

Upon the basis of the PgaB structure, Little et al. (5) have performed a series of functional, structural, and computational studies to comprehensively determine the roles of the individual domains of PgaB within the functional context of the whole enzyme and its role in dPNAG production and export. Enzyme assays show that the two-domain protein must be intact to catalyze de-N-acetylation of β-1,6-GlcNAc oligomers, neither single domain being active, either individually or in combination. Docking of a modeled tetrahedral oxyanion intermediate in the de-N-acetylation of β-1,6-(GlcNAc)₅ suggests binding within the cleft formed between the N- and C-terminal domains. Important contacts are formed between the saccharide units and both domains, with catalytic residues and the essential metal ion located in the N-terminal domain.

Little et al. (5) further characterize the function of the glycoside hydrolase-like C-terminal domain of PgaB, which had previously been shown to lack catalytic activity. Such seemingly inactive glycoside hydrolase mutants in which catalytic residues have been replaced are an increasingly uncovered phenomenon, particularly in the case of chitinases, and generally thought to represent simple binding domains (10, 11). Indeed, through fluorescence perturbation studies, the authors (5) measure weak but progressively tighter binding of β-1,6-GlcNAc oligomers of increasing size, whereas chito-oligosaccharides are shown not to bind. Surprisingly, the binding of partially de-N-acetylated oligomers was not studied. Furthermore, Little et al. solve a crystal structure of the C-terminal domain in complex with a PNAG oligomer that clearly identifies the electronegative groove as the binding site, with clear density for the four reducing end-sugar residues. The authors also identify a previously unobserved, conformationally flexible β-hairpin loop that appears to be important in the formation of the binding-site.

Functional studies with long PNAG oligomers are challenging because of their limited solubility. Instead, Little et al. (5) used brute-force molecular dynamics simulations to predict how monomers of N-acetylglucosamine and glucosamine (modeled as glucosammonium) would bind to the PgaB structure. The resulting patterns of predicted binding density on the surface of PgaB provide a picture of how units of PNAG or dPNAG may associate with the enzyme. These densities overlap with the de-N-acetylation site and interdomain cleft and the binding groove of the C-terminal domain (predicted to prefer the de-N-acetylated glucosammonium). The authors also predict binding of saccharide units over a region spanning the two, suggesting that after de-N-acetylation, dPNAG/PNAG continually associates with PgaB and slides along its surface in a processive manner. Little et al. are confident in the power of these simulations as they demonstrate that this method correctly predicts the binding sites of N-acetylglucosamine and glucosamine in crystal structures of the C-terminal domain solved in complex with those monosaccharides.

In the past decade, during which bacterial polysaccharide secretion pathways have been studied, the way in which these polysaccharides are handled during their transport through the periplasm has remained an important question. ABC-transporter–dependent pathways for capsular polysaccharide export involve a multiprotein complex that spans the periplasm and funnels the polysaccharide out (12). In the case of synthase-dependent pathways, and dPNAG production and export in E. coli in particular, Little et al. (5) show that PgaB is important not only in modifying the polysaccharide to its partially de-N-acetylated form in the periplasm, but also in closely interacting with the polymer, which is apparently wound about the C-terminal domain following its modification, moving processively over the surface of the protein before it is threaded through the β-barrel porin of the PgaA protein for transport across the outer membrane. These results suggest that such protein–polysaccharide interactions may be generally important for protecting or directing the polysaccharide in the periplasm during export by the synthase-dependent pathways widely involved in biofilm formation. For example, a recent crystal structure of the BscA–BscB complex, which makes up the cellulose synthase in Rhodobacter sphaeroides (involved in biofilm formation in that organism), shows that the BscB extends 60 Å into the periplasm and includes two putative carbohydrate-binding domains that could interact closely with the nascent cellulose in that pathway and assist in directing its export (13). Little et al. (5) present data suggesting that PgaB plays this role in dPNAG production in E. coli, in addition to tailoring the PNAG precursor to form the mature dPNAG polysaccharide. Evidently, PgaB is closely intertwined with dPNAG production in biofilm formation.