Delivery of CdiA nuclease toxins into target cells during contact-dependent growth inhibition

Abstract

Bacterial contact-dependent growth inhibition (CDI) is mediated by the CdiB/CdiA family of two-partner secretion proteins. CDI systems deploy a variety of distinct toxins, which are contained within the polymorphic C-terminal region (CdiA-CT) of CdiA proteins. Several CdiA-CTs are nucleases, suggesting that the toxins are transported into the target cell cytoplasm to interact with their substrates. To analyze CdiA transfer to target bacteria, we used the CDI system of uropathogenic Escherichia coli 536 (UPEC536) as a model. Antibodies recognizing the amino- and carboxyl-termini of CdiA(UPEC536) were used to visualize transfer of CdiA from CDI(UPEC536+) inhibitor cells to target cells using fluorescence microscopy. The results indicate that the entire CdiA(UPEC536) protein is deposited onto the surface of target bacteria. CdiA(UPEC536) transfer to bamA101 mutants is reduced, consistent with low expression of the CDI receptor BamA on these cells. Notably, our results indicate that the C-terminal CdiA-CT toxin region of CdiA(UPEC536) is translocated into target cells, but the N-terminal region remains at the cell surface based on protease sensitivity. These results suggest that the CdiA-CT toxin domain is cleaved from CdiA(UPEC536) prior to translocation. Delivery of a heterologous Dickeya dadantii CdiA-CT toxin, which has DNase activity, was also visualized. Following incubation with CDI(+) inhibitor cells targets became anucleate, showing that the D.dadantii CdiA-CT was delivered intracellularly. Together, these results demonstrate that diverse CDI toxins are efficiently translocated across target cell envelopes.

Figure 1. CdiA^UPEC536 and HA-CdiA^UPEC536 are exported to the cell surface.

A) Schematic of the CdiA^UPEC536 exoprotein depicting the locations of the inserted N-terminal hemagglutinin (HA) epitope tag and the CdiA-CT region (residues Val3016– Ile3242) used to generate anti-CdiA-CT^UPEC536 polyclonal antibodies. Regions corresponding to the secretion signal sequence and the toxic tRNase domain are indicated. The vertical VENN sequence demarcates the N-terminal margin of the variable CdiA-CT sequence. Residue numbers are shown by the scale bar below. B) Whole-cell immunoblot analysis of E. coli cells expressing CdiA^UPEC536 and HA-CdiA^UPEC536. Cells expressing HA-CdiA^UPEC536 (HA-CdiA), CdiA^UPEC536 (CdiA) or no effector protein (CDI^-) were fixed without permeabilization and stained with anti-HA or anti-CdiA-CT^UPEC536 (anti-CdiA-CT) antibodies. Where indicated, cells were treated with proteinase K (proK) prior to fixation. Stained cells were spotted onto nitrocellulose membrane and analyzed with an Odyssey infrared imager.

Figure 2. CdiA^UPEC536 is transferred to the surface of target cells.

A) Immunofluorescence microscopy of CDI^UPEC536 co-cultures. Non-fluorescent inhibitor cells (expressing either CdiA^UPEC536 or HA-CdiA^UPEC536) were mixed with red fluorescent target cells (2∶1 inhibitor-to-target ratio) for 1 h, then analyzed by fluorescence microscopy using anti-CdiA-CT^UPEC536 (anti-CdiA-CT) or anti-HA antibodies as described in Methods. Green fluorescence is indicative of anti-CdiA-CT and anti-HA staining of cell surfaces. Inhibitor cells in the CDI^- panel carry an empty vector without a CDI system. Where indicated (+), samples were treated with proteinase K prior to fixation, but cells were not permeabilized. B) Transfer of CdiA-CT^UPEC536 antigen to target cells. The percentage of bamA ⁺ and bamA101 target cells with CdiA-CT^UPEC536 antigen was quantified after one and two hours of co-culture with CdiA^UPEC536 inhibitor cells. The bamA101 allele was complemented with plasmid pBamA (corresponding to pDAL950 in Table 1). C) Transfer of HA antigen to target cells. Quantifications in panels B and C were determined by analysis of 150 target cells from two independent experiments. The reported values represent the average ± SEM.

Figure 3. CdiA-CT^UPEC536 is delivered into target cells.

A) Anti-CdiA-CT^UPEC536 immunofluorescence microscopy of CDI^UPEC536 co-cultures. Non-fluorescent inhibitor cells expressing either CdiA^UPEC536 (CdiA) or HA-CdiA^UPEC536 (HA-CdiA) were mixed with red fluorescent target cells (2∶1 inhibitor-to-target ratio) for 1 h. Cells were then fixed and permeabilized for fluorescence microscopy using anti-CdiA-CT^UPEC536 (anti-CdiA-CT) antibodies as described in Methods. Green fluorescence is indicative of anti-CdiA-CT immunostaining. Where indicated (+), cells were treated with proteinase K prior to fixation. The histogram quantifies the percentage of target cells (average ± SEM) with surface and internal CdiA-CT^UPEC536 antigen staining. At least 150 target cells from two independent experiments were scored for the quantification of CdiA delivery. B) Anti-HA epitope immunofluorescence microscopy of CDI^UPEC536 co-cultures. Co-culture conditions and sample preparation was as described in panel A except that anti-HA antibodies were used for immunofluorescence. Green fluorescence is indicative of anti-HA immunostaining. Where indicated (+), cells were treated with proteinase K prior to fixation.

Figure 4. The CdiA-CT^{3937 −2} DNase is delivered into E. coli cells.

A) The chimeric EC93-Dd3937 CDI system is functional. Mock inhibitors (CDI^-) or CDI³⁹³⁷ cells were co-cultured at a 1∶1 ratio with target bacteria that carry plasmid pBR322 (cdiI ^-) or pDAL852 (cdiI ³⁹³⁷⁻²). Co-cultures were sampled at the indicated times and viable target cells quantified as colony forming units (CFU) per mL. Reported values are the average ± SEM for three independent experiments. For panels B, C and D, inhibitor cells were labeled with GFP (green) and targets with DsRed (red) to differentiate the two cell populations by fluorescence microscopy. Competitions were conducted as described for panel A, and cells were stained with DAPI to visualize DNA. B) Inhibitor cells expressing CdiA-CT³⁹³⁷⁻² were co-cultured with cdiI^- targets, C) Mock inhibitors (CDI^-) were co-cultured with cdiI^- targets. D) Inhibitors expressing CDI³⁹³⁷ were co-cultured with targets expressing CdiI³⁹³⁷⁻² immunity.

Figure 5. DNase delivery is dependent upon the BamA receptor.

A) bamA101 target cells are resistant to the chimeric EC93-Dd3937 CDI system. Mock inhibitors (CDI^-) or CDI³⁹³⁷ cells were co-cultured with bamA ⁺ cells or bamA101 cells at a 1∶100 inhibitor to target ratio. Cultures were sampled at the indicated times and viable target cells quantified as colony forming units (CFU) per mL. Reported values are the average ± SEM for three independent experiments. B) Visualization of DNase delivery. Competitions were conducted as described for panel A, but inhibitors were labeled with GFP (green) and targets with DsRed (red) to differentiate the two cell populations by fluorescence microscopy. Cells were stained with DAPI to visualize DNA. C) Quantification of anucleate target cells. The percentage of anucleate target cells (defined as cells lacking DAPI staining) was determined by inspection of random microscopy fields. Two independent co-cultures were analyzed and between 186–260 target cells were scored for each competition. Values represent the average ± SEM.