Control of biofilm formation by an Agrobacterium tumefaciens pterin-binding periplasmic protein conserved among diverse Proteobacteria

Abstract

Biofilm formation and surface attachment in multiple Alphaproteobacteria is driven by unipolar polysaccharide (UPP) adhesins. The pathogen Agrobacterium tumefaciens produces a UPP adhesin, which is regulated by the intracellular second messenger cyclic diguanylate monophosphate (c-di-GMP). Prior studies revealed that DcpA, a diguanylate cyclase-phosphodiesterase, is crucial in control of UPP production and surface attachment. DcpA is regulated by PruR, a protein with distant similarity to enzymatic domains known to coordinate the molybdopterin cofactor (MoCo). Pterins are bicyclic nitrogen-rich compounds, several of which are produced via a nonessential branch of the folate biosynthesis pathway, distinct from MoCo. The pterin-binding protein PruR controls DcpA activity, fostering c-di-GMP breakdown and dampening its synthesis. Pterins are excreted, and we report here that PruR associates with these metabolites in the periplasm, promoting interaction with the DcpA periplasmic domain. The pteridine reductase PruA, which reduces specific dihydro-pterin molecules to their tetrahydro forms, imparts control over DcpA activity through PruR. Tetrahydromonapterin preferentially associates with PruR relative to other related pterins, and the PruR-DcpA interaction is decreased in a pruA mutant. PruR and DcpA are encoded in an operon with wide conservation among diverse Proteobacteria including mammalian pathogens. Crystal structures reveal that PruR and several orthologs adopt a conserved fold, with a pterin-specific binding cleft that coordinates the bicyclic pterin ring. These findings define a pterin-responsive regulatory mechanism that controls biofilm formation and related c-di-GMP-dependent phenotypes in A. tumefaciens and potentially acts more widely in multiple proteobacterial lineages.

Keywords: biofilm; cyclic diguanylate monophosphate; protein structure; pterin; regulation.

Fig. 1.

PruR binds pterins and is required to control both the DGC and PDE activity of DcpA. (A) PruA reaction and relevant pterin molecule structures. i): PruA uses NADPH as a cofactor and catalyzes the reduction of a 7,8-dihydropterin substrate to a 5,6,7.8-tetrahydropterin. Atoms are numbered in the dihydropterin. R indicates side groups as shown in (ii)–(iv): ii): R1—monapterin; iii): R2—neopterin; and iv): R3—pABA-glutamate (folate). (B) In vitro pterin binding assays were performed as described in supplemental methods with purified Δ_SS-PruR (50 µM), NADPH, and with or without His₆-PruA. HPLC fractioned reactions were examined by fluorescence for the oxidized pterins (excitation: 356 nm; emission: 450 nm). UV absorbance at 283 nm was used to measure folate relative to standards. PruR was incubated with the following pterin or folate species: H₂MPt, dihydromonapterin; H₄MPt, PruA-generated-tetrahydromonapterin; H₂NPt, dihydroneopterin; H₄NPt, PruA-generated tetrahydroneopterin; H₂F, dihydrofolate; and H₄F, tetrahydrofolate. Bars are averages of triplicate assays with error bars as SD and analyzed by standard one-way ANOVA and post hoc Tukey analysis (P values relative to H₄MPt, *, <0.05, ns, not significant). (C) Biofilm assays of the A. tumefaciens C58 wild type (WT, gray bar), a ΔdcpA mutant (black bars), or a ΔdcpAΔpruR double mutant (white bars) containing the vector control or a plasmid-borne P_lac-dcpA fusion induced with 500 µM isopropyl β D 1-thiogalactopyranoside (IPTG) expressing either the wild-type dcpA or catalytic site mutants (DGC⁻, E308A; PDE⁻, E431A; DGC⁻PDE⁻, both). The ratio of acetic acid–solubilized CV absorbance (A₆₀₀) from 48 h biofilm assays normalized to the OD₆₀₀ planktonic turbidity from the same culture. Assays were performed in triplicate and error bars are SD; P values calculated comparing complementation of dcpA in the ΔdcpA strain compared to the ΔdcpAΔpruR strain by standard two-tailed t test. (P values, * <0.05, **** <0.0001)

Fig. 2.

PruR is a periplasmic protein. (A) Signal P prediction of the PruR N-terminal signal sequence. Black arrow, predicted signal peptidase cleavage site; gray arrow, putative cleavage and lipidation site; gray text, predicted lipidated Cys19. (B) Western blot of Sodium dodecylsulfate polyacrylamide gel electrophoresis (SDS-PAGE) using α-PruR polyclonal antisera to probe extracts of wild-type A. tumefaciens C58 WT and the ΔpruR mutant on its own or expressing either the P_lac-pruR plasmid or the P_lac-Δ_SSpruR plasmid (both plasmids also expressing dcpA). Cultures were grown to similar densities with or without induction with 400 μM IPTG and fractionated to separate the cytoplasmic/membrane fraction (C lanes) from the periplasmic fraction (P lanes). Antibody binding was detected with goat anti-rabbit secondary antibody conjugated to horseradish peroxidase (GAR-HRP) and chemiluminescentsubstrate exposed on a BioRad ChemiDoc. (C) Biofilm assays of WT or a ΔpruR mutant derivative with the empty vector plasmid (−) or harboring a plasmid-borne P_lac fusion expressing either pruR or Δ_SSpruR (both plasmids also express dcpA). Ratio of acetic acid-solubilized CV A₆₀₀ from 48 h biofilm assays normalized to the OD₆₀₀ planktonic turbidity of the same culture. Assays performed in triplicate and error bars are SD; P values calculated by standard two-tailed t test (P values, *** <0.01).

Fig. 3.

PruR forms a complex with the periplasmic region of DcpA. Western blots with whole cell suspensions that were either untreated or incubated with DSS cross-linker (0.75 mM) and separated on 10% SDS-PAGE gels. Antibody binding was detected with GAR-HRP secondary antibody and chemiluminescent substrate exposed on a BioRad ChemiDoc. Nonspecific bands serve as protein loading controls. (A and B) Cell suspensions were prepared from A. tumefaciens mutants deleted for the cellulose operon to reduce clumping (Δcel); with ΔcelΔpruR and ΔcelΔdcpA mutants (C) Wild-type A. tumefaciens or ΔdcpA ectopically expressing the cytoplasmic (DcpA_cyt) or periplasmic (DcpA_peri) domains of DcpA from P_lac. Antibodies were α-PruR polyclonal antibody (1:40,000 dilution) and α-DcpA_Peri (1:20,000 dilution). 400 µM IPTG was added to induce P_lac. Complexes are indicated as labeled, PruR-DcpA, 87 kDa; DcpA, 71 kDa; PruR; 16 kDa. Red, blue, and black triangles in panel C indicate the full-length DcpA-PruR complex, the expected size of the PruR-DcpA_cyt complex (~66.5 kDa), and the PruR-DcpA_peri complex (~37.8 kDa), respectively.

Fig. 4.

Deletion of pruA diminishes the interaction between DcpA and PruR. Western blot probing for the PruR-DcpA_peri complex following in vivo DSS cross-linking in the A. tumefaciens wild type or a ΔpruA mutant expressing the P_lac-pruR-dcpA_peri plasmid (400 μM IPTG). Probed with polyclonal antibody preparations (A) α-PruR, 1:40,000; (B) α-DcpA_peri, 1:20,000. Antibody binding was detected with GAR-HRP secondary antibody and a chemiluminescent substrate on a BioRad ChemiDoc. Nonspecific bands serve as protein loading controls.

Fig. 5.

The pruR-dcpA operon is conserved across multiple Proteobacteria. (A) Domain structure for PruR and DcpA; transmembrane domains and signal sequence are black lines; the DcpA periplasmic domain is purple; the DGC domain is green; the PDE domain is lime green; Pt indicates the pterin-binding activity of the protein. The arrow diagrams indicate presumptive operon structure for homologs of the pruR-dcpA operon from multiple pathogens. A subset of DcpA homologs are truncated relative to DcpA and contain only the DGC domain, and degenerate DGC domains predicted to be catalytically inactive are marked with an asterisk. Klebsiella pneumoniae pneumoniae ATCC 700721; Vibrio cholerae, O1 biovar El Tor str. N16961; Vibrio vulnificus CMCP6; Pseudomonas aeruginosa PA01. Gene and domain sizes are proportional. (B) Radial phylogram of APB, BBP, DBP, EBP, and GMP representing bacterial families. Bolded family names and red stars indicate taxa with pruR genes linked with a dcpA-periplasmic domain. Green and blue dots indicate a PruR linked to a DcpA-periplasmic domain with a GGDEF/EAL (PruR-DcpA) or a TorS-BaeS HK (PruR-DTB) cytoplasmic domain, respectively. Tree was generated using Interactive Tree of Life (24).

Fig. 6.

The structure of PruR is conserved in multiple Alpha- and Gammaproteobacteria. (A) The overall structure of PruR from A. tumefaciens (7kos) is depicted. The secondary structure elements are labeled and shown in red (α helices), yellow (β-strands), and green (loops). (B) Superpositions are presented of PruR from A. tumefaciens (orange, 7kos; violet, 7kou), K. pneumoniae (pink, 7rkb) V. cholerae (wheat, 7kp2), V. vulnificus (light green, 7kom). The peptide main chains of all structures are depicted as ribbons. (C) Multiple sequence alignments are presented for PruR proteins from A. tumefaciens (At 7kos), K. pneumoniae (Kp 7rkb) V. cholerae (Vc 7kp2), V. vulnificus (Vv 7kom) with secondary structure elements from At PruR mapped above. The lime green circles mark pterin-binding site residues conserved across all proteins.

Fig. 7.

The PruR structure is a degenerate SUOX fold and PruR binds pterins. Superposition of A. tumefaciens PruR (orange, 7kou) with the MoCo-binding domain of (A) chicken liver SUOX (teal, PDB 1sox) and (B) E. coli YedY (gray, PDB 1xdq). The structures are depicted as transparent cartoons, except for key determinants of the MoCo-binding region, which are opaque (see SI Appendix, Fig. S10A for full structure comparisons). The MoCo and conserved Cys185 of SUOX and Cys102 of YedY are shown as balls-and-sticks (carbon, teal or gray; oxygen, red; nitrogen, dark blue; sulfur, yellow), the molybdenum is shown as a mauve or olive sphere for SUOX and YedY, respectively, water molecules as small cyan spheres, and hydrogen bonds as navy, dashed lines. W70, conserved in PruR, is also shown as sticks. Structures in images on the left were rotated 90° around the y-axis toward the viewer to obtain images on the right (as indicated by dashed arrows) where the extended β-hairpins in SUOX and YedY are marked. (C) A zoomed-in view of superposition of A. tumefaciens (orange, 7kou) and K. pneumoniae (pink, 7rkb) PruR. The neopterin and conserved residues of the pterin-binding pocket are shown as sticks. Residue sidechains are numbered according to A. tumefaciens. (D) Superposition of all observed neopterin-binding modes from crystal structures of K. pneumoniae (7rkb) and V. cholerae (7kp2). The binding pocket is represented as an electrostatic surface potential (blue-to-red, positive-to-negative charge) and neopterin as stick models. Carbons are in yellow (alternative conformation A, V. cholerae), gray (alternative conformation B, V. cholerae, 7kp2), and green (K. pnuemoniae, 7rkb), with oxygen in red and nitrogen in blue. (E) Congo Red staining by inoculating 3 µL spots of the A. tumefaciens wild type and ΔpruR harboring plasmid-borne P_lac-pruR alleles on a single ATGN-CR plate supplemented with 75 µg/mL Congo Red and 400 μM IPTG and incubated at 30 °C for 48 h. (F) DSS in vivo cross-linking of the A. tumefaciens wild type or a ΔpruA mutant expressing the P_lac-pruR wild-type and mutant alleles (400 μM IPTG). The blot was probed with α-PruR polyclonal antibody (1:40,000), and GAR-HRP secondary antibody, and antibody binding was detected with chemiluminescent substrate using a BioRad ChemiDoc. Nonspecific bands serve as protein loading controls.