Cyclic Di-GMP-Regulated Periplasmic Proteolysis of a Pseudomonas aeruginosa Type Vb Secretion System Substrate

Abstract

We previously identified a second-messenger-regulated signaling system in the environmental bacterium Pseudomonas fluorescens which controls biofilm formation in response to levels of environmental inorganic phosphate. This system contains the transmembrane cyclic di-GMP (c-di-GMP) receptor LapD and the periplasmic protease LapG. LapD regulates LapG and controls the ability of this protease to process a large cell surface adhesin protein, LapA. While LapDG orthologs can be identified in diverse bacteria, predictions of LapG substrates are sparse. Notably, the opportunistic pathogen Pseudomonas aeruginosa harbors LapDG orthologs, but neither the substrate of LapG nor any associated secretion machinery has been identified to date. Here, we identified P. aeruginosa CdrA, a protein known to mediate cell-cell aggregation and biofilm maturation, as a substrate of LapG. We also demonstrated LapDG to be a minimal system sufficient to control CdrA localization in response to changes in the intracellular concentration of c-di-GMP. Our work establishes this biofilm signaling node as a regulator of a type Vb secretion system substrate in a clinically important pathogen.

Importance: Here, the biological relevance of a conserved yet orphan signaling system in the opportunistic pathogen Pseudomonas aeruginosa is revealed. In particular, we identified the adhesin CdrA, the cargo of a two-partner secretion system, as a substrate of a periplasmic protease whose activity is controlled by intracellular c-di-GMP levels and a corresponding transmembrane receptor via an inside-out signaling mechanism. The data indicate a posttranslational control mechanism of CdrA via c-di-GMP, in addition to its established transcriptional regulation via the same second messenger.

FIG 1

Mechanism of biofilm formation in P. fluorescens. (A) Overview of LapD- and LapG-mediated regulation of LapA localization to the outer membrane. Differential recruitment of LapG by LapD in response to cytoplasmic c-di-GMP levels controls LapA's surface attachment. When LapG is not sequestered to the inner membrane by LapD, LapG proteolytically cleaves the N terminus of LapA between two alanine residues of a TAAG motif, releasing the adhesin from the outer membrane, thus liberating the cell from the biofilm matrix. (B) Web Logo plot (24) of putative LapA-like substrates identified previously (4) reveals a series of conserved residues (green) upstream of the TAAG cleavage sequence (red).

FIG 2

LapG proteolysis of P. aeruginosa CdrA^Cterm. (A) Primary structure and domain organization of CdrA, depicting the location of the LapG cleavage site and the putative C-terminal “cysteine hook.” A second, N-terminal proteolytic site identified previously to be between residues 437 and 438 (9) is also depicted. (B) Cartoon model for the localization of CdrA at the outer membrane via its cognate TPS transporter CdrB, in which the C terminus of CdrA resides in the periplasm while the N terminus is excreted into the extracellular space. Such a topological orientation is the opposite of that of LapA-like proteins, in which the C terminus is extracellular (Fig. 1A). (C) Proteolysis of purified CdrA^Cterm by purified LapG-sfGFP produces 23-kDa and 5-kDa fragments (indicated by asterisks) in the presence of calcium but not EGTA. The sequence of the first 10 residues of the 5-kDa proteolytic fragment determined by Edman degradation is shown. A variant of CdrA^Cterm in which the double-alanine motif is mutated to arginine residues (TRRG) prevents proteolysis. (D) Proteolytic processing of CdrA^Cterm and LapA^Nterm by P. fluorescens (left) and P. aeruginosa (right) LapG-sfGFP demonstrates functional conservation across the two systems. LapA^Nterm and CdrA^Cterm proteolytic products are indicated by # and *, respectively.

FIG 3

Characterization of LapG specificity with the pCleevR assay. (A) Primary structure of pCleevR constructs in which various CdrA and LapA sequences containing their respective LapG cleavage sites are flanked N terminally by a His₆-tagged SUMO protein and C terminally by sfGFP A²²⁶R. (B) Length of the pCleevR sequence influences rate of cleavage by LapG. CdrA^short/LapA^short (as defined in panel A) sequences are cleaved more slowly than CdrA^long/LapA^long sequences, which are processed by LapG at approximately the same rate as CdrA^Cterm and LapA^Nterm. Cleavage products are indicated by asterisks (for CdrA) and number symbols (for LapA). Intact substrates are indicated with triangles. (C) Assessment of the relative rates of cleavage of CdrA^shortP′ and CdrA^shortP substrates compared with the CdrA^long substrate indicates that residues N terminal to the cleavage site are responsible for this increased rate of proteolysis. The fact that CdrA^randomP, which has the same length and number of residues flanking the cleavage site as CdrA^shortP′ but is of unrelated sequence, is cleaved more slowly than CdrA^shortP′ indicates that the enhanced rate is sequence dependent. The motif TAAG, in and of itself, is not sufficient for LapG targeting (right). Note that CdrA^long migrates with an aberrantly smaller size than CdrA^short in SDS-PAGE. (D) Single-amino-acid substitutions in CdrA^long identify residues D²⁰⁹⁷, P²⁰⁹⁸, and L²¹⁰¹ but not L²¹⁰⁰ as crucial for LapG-mediated cleavage. As with the TRRG variant of CdrA^Cterm (Fig. 2C), the analogous TRRG CdrA^long variant is not cleaved by LapG.

FIG 4

Regulation of CdrAB-dependent biofilm growth by LapG in P. aeruginosa. (A) Assessment of biofilm formation in response to CdrAB and LapG expression in static cultures of the P. aeruginosa PAO1 ΔcdrA strain (9) as measured by crystal violet staining (see Materials and Methods). Top, Western blot of CdrA released into the supernatant of these cultures using an anti-CdrA antibody developed previously (9). (B) Biofilm formation in P. aeruginosa expressing CdrAB and LapG (+Ara), including a plasmid in which the TAAG cleavage site of CdrA has been mutated to TRRG. Above the plots are Western blots of CdrA from whole-cell and supernatant fractions of these cultures, demonstrating that the TRRG mutation prevents LapG-dependent release of CdrA into the supernatant fraction. Asterisks indicate P values of <0.01 (*), <0.001 (**), <0.0001 (***), and <0.00001 (****) between the indicated samples.

FIG 5

Regulation of CdrA-dependent cell autoaggregation by the LapDG receptor system in a heterologous host. (A) Direct visualization of E. coli BL21 cultures expressing CdrAB and periplasmic P. aeruginosa LapG-sfGFP. The pBAD and pRSF expression constructs for each culture are as indicated (written above for the top row and below for the bottom row). (B) Similar visualizations of E. coli cell autoaggregation as in panel A but with full-length LapD coexpression alongside CdrAB and LapG-sfGFP. (C) Top, propensity for the cultures shown in panels A and B with robust phenotypes to sediment (as described in Materials and Methods). Higher values indicate greater autoaggregation. Red and green dashed lines indicate values obtained for native, dispersed E. coli cells and those aggregating due to expression of only CdrAB. Red and green asterisks above each sample indicate statistical comparisons to the cells grown with empty plasmids and cells expressing only CdrAB, respectively. The numbers of asterisks indicate P values as described in the legend for Fig. 4. Those that are not statistically different are indicated with “ns.” Bottom, for each of the cultures in the top panel, protein levels of CdrA in supernatant and whole-cell lysates were analyzed by Western blotting. Also analyzed in the whole-cell lysates were LapG-sfGFP (by in-gel fluorescence) and LapD (by Western blotting with an anti-His₅ antibody). Approximate molecular masses are shown to the right of each gel. (D) Cultures shown in the bottom row of panel A were imaged by differential interference contrast and GFP fluorescence at a magnification of ×40 to reveal microscopic features of the autoaggregated cells (left column). Cells expressing only the red fluorescent protein mKate did not aggregate with the CdrAB-expressing cells (right column). A small fraction of cells expressing CdrAB appear as long filaments, independent of the expression of active or inactive LapG. However, this appearance has no obvious impact on the overall autoaggregation phenotypes. Also, cells that appear unlabeled show a lower level of fluorescence that is detectable upon longer exposure times. (E) Similar overlays of differential interference contrast and GFP fluorescence images of cell aggregates from LapD-coexpressing strains displaying robust phenotypes shown in panel B. Bottom, mutations used to alter the functionality of LapD and LapG mapped onto their respective structural models.

FIG 6

Regulation of CdrA-dependent cell autoaggregation by c-di-GMP. (A) Direct visualization of E. coli BL21 cultures expressing variants of CdrAB, LapG-sfGFP, LapD, and WspR. The pBAD, pRSFDuet, and pACYCDuet expression constructs used in each culture are as indicated in panel B. (B) Propensity for the cultures shown in panel A to sediment. Higher values indicate greater autoaggregation. Red and green dashed lines indicate values obtained for native, dispersed E. coli cells and those aggregating due to expression of only CdrAB. Red and green asterisks above each sample indicate statistical comparisons to the cells grown with empty plasmids and cells expressing only CdrAB, respectively. The numbers of asterisks indicate P values as described in the legend for Fig. 4. Those that are not statistically different are indicated with “ns.”

FIG 7

Model for two-tier regulation of CdrA via c-di-GMP. Transcription of the cdrAB operon (bottom left) is initiated when c-di-GMP (yellow) binds the transcriptional regulator FleQ. Elevated c-di-GMP levels also result in second-messenger binding to the EAL domain of LapD, which in turn sequesters periplasmic LapG. Together these events allow CdrA to be secreted to and anchored within the outer membrane via its cognate outer membrane transporter CdrB. The N terminus of CdrA (purple) binds to the PORTA domain of CdrB and is processed by an unknown proteolytic activity, acting in the periplasm (as shown), within the transporter, or at the cell surface. When c-di-GMP levels are lowered by phosphodiesterases (PDEs), LapD adopts the autoinhibited state (bottom right), releasing LapG into the periplasm, where it cleaves the C-terminal tail of CdrA (red) at the TAAG cleavage site and releases this adhesin into the extracellular space, promoting cellular disaggregation.

FIG 8

Phylogenetic tree of LapG orthologs, displaying the diversity of the LapDG signaling system in bacteria. For this analysis, 5,220 bacterial genomes were downloaded from the NCBI database and initially searched for a putative LapG-like protease (identified by the conserved DX₉WX₁₂DCED[YF]X₃KX_20–40HX_12–17LD motif). These genomes were subsequently searched for a LapD-like protein (motif F[DE]X[GS]X[YF]X_20–40PXW[FLIVM]X_16–18GWX_47–51[RP]L) and a LapG-like substrate possessing a DPX_2–3LX₂[TA]AAG motif. LapG sequences were aligned by Clustal Omega (25) and analyzed phylogenetically by maximum likelihood with the JTT model for amino acid substitutions (26). Colored shadings highlight the class of bacteria to which individual species belong. Beside each shaded group, the corresponding number of species containing a LapD ortholog and a putative LapG substrate is indicated relative to the number of LapG orthologs within this group. Note that for many species, while LapD and LapG orthologs can be identified, no readily apparent LapA- or CdrA-like proteolysis motif was found. This observation suggests that other novel LapG targets may be present in these microbes.