An Extended Cyclic Di-GMP Network in the Predatory Bacterium Bdellovibrio bacteriovorus

Abstract

Over the course of the last 3 decades the role of the second messenger cyclic di-GMP (c-di-GMP) as a master regulator of bacterial physiology was determined. Although the control over c-di-GMP levels via synthesis and breakdown and the allosteric regulation of c-di-GMP over receptor proteins (effectors) and riboswitches have been extensively studied, relatively few effectors have been identified and most are of unknown functions. The obligate predatory bacterium Bdellovibrio bacteriovorus has a peculiar dimorphic life cycle, in which a phenotypic transition from a free-living attack phase (AP) to a sessile, intracellular predatory growth phase (GP) is tightly regulated by specific c-di-GMP diguanylate cyclases. B. bacteriovorus also bears one of the largest complement of defined effectors, almost none of known functions, suggesting that additional proteins may be involved in c-di-GMP signaling. In order to uncover novel c-di-GMP effectors, a c-di-GMP capture-compound mass-spectroscopy experiment was performed on wild-type AP and host-independent (HI) mutant cultures, the latter serving as a proxy for wild-type GP cells. Eighty-four proteins were identified as candidate c-di-GMP binders. Of these proteins, 65 did not include any recognized c-di-GMP binding site, and 3 carried known unorthodox binding sites. Putative functions could be assigned to 59 proteins. These proteins are included in metabolic pathways, regulatory circuits, cell transport, and motility, thereby creating a potentially large c-di-GMP network. False candidate effectors may include members of protein complexes, as well as proteins binding nucleotides or other cofactors that were, respectively, carried over or unspecifically interacted with the capture compound during the pulldown. Of the 84 candidates, 62 were found to specifically bind the c-di-GMP capture compound in AP or in HI cultures, suggesting c-di-GMP control over the whole-cell cycle of the bacterium. High affinity and specificity to c-di-GMP binding were confirmed using microscale thermophoresis with a hypothetical protein bearing a PilZ domain, an acyl coenzyme A dehydrogenase, and a two-component system response regulator, indicating that additional c-di-GMP binding candidates may be bona fide novel effectors.

Importance: In this study, 84 putative c-di-GMP binding proteins were identified in B. bacteriovorus, an obligate predatory bacterium whose lifestyle and reproduction are dependent on c-di-GMP signaling, using a c-di-GMP capture compound precipitation approach. This predicted complement covers metabolic, energy, transport, motility and regulatory pathways, and most of it is phase specific, i.e., 62 candidates bind the capture compound at defined modes of B. bacteriovorus lifestyle. Three of the putative binders further demonstrated specificity and high affinity to c-di-GMP via microscale thermophoresis, lending support for the presence of additional bona fide c-di-GMP effectors among the pulled-down protein repertoire.

FIG 1

Cyclic di-GMP control over the progress of B. bacteriovorus cell cycle: wild-type B. bacteriovorus maintains an obligate predatory biphasic life cycle. (A) Nonreplicating free-living and fast-swimming vibrioid AP cells search for prey. (B and C) Encountering of the prey cell is followed by irreversible attachment (B) and then by DgcB-dependent and CdgA-associated invasion into the prey periplasm (C). CdgA is a cyclic di-GMP effector that interacts with the invasion-essential type IVa pilus regulatory hub. (D) During entry and settlement, the bdelloplast is formed. (E) Within the bdelloplast, the predator enters GP, consumes the prey and grows as a multinucleoid filament. (F and G) Upon nutrient depletion, the filament divides to progeny AP cells that lyse and escape from the bdelloplast (F), the latter necessitating DgcA, which is essential for development of both flagellar and gliding motility (G). The nascent AP cells are ready for another predation cycle. (H and I) HI mutants are not restricted to the predatory lifestyle and, while remaining predatory, they can grow saprophytically via the DgcC pathway, retaining the dimorphic cell cycle.

FIG 2

Experimental scheme. Total protein harvested from fresh B. bacteriovorus HD100 AP cells and axenically replicating B. bacteriovorus HD100 M1.1 HI mutant cells was investigated by CCMS. Of 3,584 encoded proteins (23), 777 were detected by LC-MS/MS, but only 84 proteins presented competitive cyclic di-GMP specific binding. These binders demonstrated c-di-GMP binding exclusive to the AP (blue circle), unique to the HI mutants (green circle), or in both (red circle). AP-specific and common binders represent cyclic di-GMP effectors functional in the AP. HI-specific binders represent cyclic di-GMP effectors functional in the GP. The functionality of shared binders in the GP cannot be determined due to possible deregulated protein production in HI mutants (15).

FIG 3

Functional classification of the CCMS-identified cyclic di-GMP binders in B. bacteriovorus HD100 (84 binders) (A), P. aeruginosa (65 binders) (B), and S. Typhimurium SL1344 (36 binders) (C). Functional data for P. aeruginosa and S. Typhimurium were obtained from Nesper et al. (18). Only 5 of 20 functions (energy metabolism, motility, protein metabolism, amino acid catabolism, and nucleotide metabolism) were shared between the three bacteria. Of these categories, “energy metabolism” was enriched in all three. Other functions were differentially represented, possibly reflecting the adaptation of the c-di-GMP network to the different lifestyles of the three bacteria.

FIG 4

Validation of cyclic di-GMP binding by microscale thermophoresis (MST). The PilZ effector Bd2717, the two-component response regulator Bd2402, and the acyl-CoA dehydrogenase Bd2924 were His tagged, produced, purified, and fluorescently labeled before subjected to MST analysis with increasing concentrations of cyclic di-GMP or cyclic di-AMP. The thermophoretic movement of the three proteins was affected by c-di-GMP but not by c-di-AMP. (A) Bd2717 bound c-di-GMP with K_d = 0.176 μM. (B) The thermophoretic movement of Bd2717 unaffected by c-di-AMP, provided as an example. (C) Bd2402 bound c-di-GMP with K_d = 0.399 μM. (C) Bd2924 bound c-di-GMP with K_d = 7.67 μM. Values are averages of three triplicates. Error bars indicate the standard deviations. MST plots of Bd2402 and Bd2924 with c-di-AMP are presented in Fig. S4 in the supplemental material.

FIG 5

Phase-dependent cyclic di-GMP binding. A total of 84 c-di-GMP binders in B. bacteriovorus HD100, divided into 19 KOP and 65 putative novel, demonstrated three c-di-GMP binding patterns: AP specific (21/84), HI specific (41/84), and common AP and HI binding (22/84). It was thus concluded that at least 41 of 84 binders were GP related. The common binders are either AP-specific binders, abnormally produced and functional in the HI mutant (15), or binders shared between the AP and the GP. Since 8 of these were encoded by AP genes and 14 are GP genes, this group might be a composite group.