Fig 1.
AfuA is a sugar-phosphate specific periplasmic binding protein.
(A) Ribbon diagrams of A.pleuropneumoniae AfuA complexed with either no ligand (right) or β-G6P (left). (B) Cleft region of AfuA depicting 2F0-Fc map contoured at 2σ around β-G6P. Residues used for mutagenic studies are indicated with text labels. Predicted hydrogen bonds are drawn as yellow dashed lines with distances indicated in angstroms (Å). (C) Representative ITC curve for a positive AfuA-ligand interaction. The curve shown is from a titration of A.pleuropneumoniae AfuA with a 10:1 ratio of G6P. (D) Representative ITC curve illustrating lack of an AfuA-ligand interaction. The curve shown is from a titration of A.pleuropneumoniae AfuA with a 10:1 ratio of glucose.
Table 1.
Binding constants between A.pleuropneumoniae AfuA and G6P.
Fig 2.
Electrostatic surface mapping of AfuA and additional ligand-bound structures.
(A) Electrostatic surface potential of apo (right) and G6P-bound (left) AfuA. Potential values range from -10 (red) to +10 kJ/mol·e (blue). (B) Magnified view of G6P-AfuA binding cleft with electrostatic surface map overlay. Residues proximal to the G6P molecule are shown as sticks. Water molecules are shown as red spheres. (C) Chemical structure of β-G6P. * denotes the anomeric carbon. (D, E) Ribbon diagrams showing A.pleuropneumoniae AfuA in complex with (D) F6P or (E) S7P. Inset boxes below each ribbon structure show the 2F0-Fc map contoured at 2σ around each ligand.
Table 2.
Binding constants between WT A.pleuropneumoniae AfuA and screened molecules.
Fig 3.
afuABC rescues a nutrient-limited E.coli strain.
(A) Solid media complementation of ΔuhpT in M9 minimal media (M9 MM) supplemented with 10mM glucose or G6P. 5μL drops were plated of the indicated dilutions and grown at 37°C for 30 hours before imaging. Plates are representative of n = 3 transformations. (B): OD600 readings of E.coli growth over 24 hours at 37°C in M9 minimal medium supplemented with 10mM glucose, G6P, fructose or F6P. Readings were taken every 15 minutes–data shown is parsed to hourly readings for clarity. Curve legend is the same in all panels and is indicated in the first panel. Error bars represent SEM of cell growth from n = 3 transformations.
Fig 4.
(A) The complete tree of all 268 putative AfuA homologues. Based on the tree, homologues were divided into: gammaproteobacteria (83 hits, purple), alpha and betaproteobacteria (121 hits, cyan) and other bacteria (64 hits, yellow). (B) The phylogenetic tree of AfuA in gammaproteobacteria. The 83 gammaproteobacteria AfuA homologues were reduced to 51 by removing sequences that shared >95% identity. These hits were divided into: 21 in the Enterobacteriaceae (red), 10 present in the Vibrionaceae (green), and 20 in the Pasteurellaceae (grey). The two AfuA homologues used in this study are highlighted in red.
Fig 5.
C. rodentium requires AfuA during host colonization.
(A) Gene organization, annotation and predicted functions for genomic region corresponding to OI-20 in C. rodentium. Percent identity to EHEC O157:H7 was calculated with ClustalW2 using the entire predicted sequence of each protein. (B) Growth curves in LB broth for WT and ΔafuA C. rodentium. Cultures were tracked with OD600 readings for the indicated amount of time (n = 2). (C) Actin pedestal formation by C. rodentium WT or ΔafuA. HeLa cells were infected for 8hrs with C. rodentium then stained with phalloidin (green), anti-C. rodentium Tir (red) and DAPI to detect DNA (blue). (D) Adherence of C. rodentium WT and ΔafuA to HeLa cells after 8hrs infection. Counts shown represent adherent bacteria on a monolayer of cells (see Methods) and are mean values ± SEM (n = 3 infections). (E) SDS-PAGE of type-3 secretion effectors by C. rodentium WT, ΔafuA and ΔescN after growth in DMEM. ΔescN is a negative control strain that is secretion-deficient. (F) C57BL/6 mice were orally infected with C. rodentium WT or ΔafuA and bacterial colonic burdens determined at 10 dpi. Data were combined from two independent experiments. Each symbol represents one animal (n = 8 for each infection). The median is indicated. (G) Competitive index (CI) of simultaneous C. rodentium WT and ΔafuA infection in C57BL/6 mice in stool 6 dpi, and colonic tissue 10 dpi. Data were combined from two independent experiments (n = 8). A CI < 1 indicates the WT strain outcompeted the knockout (6 dpi median CI = 0.49, p = 0.0313; 10 dpi median CI = 0.38, p = 0.0156).
Fig 6.
AfuABC substrates are present in the intestine during C. rodentium infection.
(A) Representative LC-MS/MS chromatograms with indicated peaks for 100ng G6P/F6P/M6P mix (top), uninfected stool (middle) and C. rodentium WT infected stool (bottom). (B) S7P peaks. G6P/F6P/M6P were identified with a 259→97 ion transition and S7P was identified with a 289→97 ion transition.
Table 3.
LC-MS/MS quantification of AfuA substrates across two sets of C. rodentium-infected mice.
Fig 7.
C. rodentium requires AfuA for efficient host-to-host transmission.
(A) Infection schematic for quantification of C. rodentium host-to-host transmission. C57BL/6 index mice were orally infected with C. rodentium WT or ΔafuA. At 6 dpi, index mice were added to a cage containing two naïve mice and 48h post-exposure, bacterial luminal and colonic burdens determined. (B, C) Ability of C. rodentium WT or ΔafuA to transmit between index and co-housed C57BL/6 mice. Bacterial colonic (B) and luminal (C) burdens were determined for both index and co-housed mice. Data were combined from two independent experiments. Each symbol represents one animal (n = 4 (index), 8 (co-housed) for each infection). The median is indicated.* = p<0.05.