Novel proteins that modulate type IV pilus retraction dynamics in Pseudomonas aeruginosa

Abstract

Pseudomonas aeruginosa uses type IV pili to colonize various materials and for surface-associated twitching motility. We previously identified five phylogenetically distinct alleles of pilA in P. aeruginosa, four of which occur in genetic cassettes with specific accessory genes (J. V. Kus, E. Tullis, D. G. Cvitkovitch, and L. L. Burrows, Microbiology 150:1315-1326, 2004). Each of the five pilin alleles, with and without its associated pilin accessory gene, was used to complement a group II PAO1 pilA mutant. Expression of group I or IV pilA genes restored twitching motility to the same extent as the PAO1 group II pilin. In contrast, poor twitching resulted from complementation with group III or group V pilA genes but increased significantly when the cognate tfpY or tfpZ accessory genes were cointroduced. The enhanced motility was linked to an increase in recoverable surface pili and not to alterations in total pilin pools. Expression of the group III or V pilins in a PAO1 pilA-pilT double mutant yielded large amounts of surface pili, regardless of the presence of the accessory genes. Therefore, poor piliation in the absence of the TfpY and TfpZ accessory proteins results from a net increase in PilT-mediated retraction. Similar phenotypes were observed for tfpY single and tfpY-pilT double knockout mutants of group III strain PA14. A PilAV-TfpY chimera produced few surface pili, showing that the accessory proteins are specific for their cognate pilin. The genetic linkage between specific pilin and accessory genes may be evolutionarily conserved because the accessory proteins increase pilus expression on the cell surface, thereby enhancing function.

FIG. 1.

Representative twitching motility of recombinant P. aeruginosa PAO1 NP strains. A group II strain PAO1 NP mutant was transformed with constructs expressing genes from group I strain 1244 (accessory gene tfpO) (boxed in red), group III strain PA14 (accessory gene tfpY) (boxed in yellow), group IV strain Pa5196 (accessory genes tfpW and tfpX) (boxed in black), and group V strain Pa281457 (accessory gene tfpZ) (boxed in white). The PAO1 wild type and the PAO1 NP mutant, both carrying the vector pBADGr, are positive and negative controls, respectively. Twitching motility was tested at two different arabinose concentrations, 0.01% and 0.2%; the cognate PAO1 pilin gene (pilA_II) restores wild-type motility at a 0.2% concentration of arabinose. The group I pilin gene complements the group II mutant to wild-type levels at 0.2% arabinose, but coexpression of the tfpO accessory gene at either arabinose concentration reduces motility. The group III and group V pilin genes do not complement well at either arabinose concentration tested, although motility is increased upon coexpression of their cognate accessory genes, tfpY and tfpZ, respectively, with 0.2% arabinose. The group IV pilin gene behaves aberrantly in that it complements the motility of PAO1 NP to the same extent as the cognate PAO1 pilin gene at 0.01% arabinose but not at 0.2%; this effect is independent of the tfpW and tfpX accessory genes. Below the twitching plates are growth controls showing that none of the observed differences in motility are due to growth defects.

FIG. 2.

Intact mass spectra of group III and group V pilins. The observed masses of 17,398 Da (group III pilin) and 17,696 Da (group V pilin) correspond to those predicted from the amino acid sequences (17,388 kDa and 17,705 kDa, respectively), showing that the pilins are not posttranslationally modified. The secondary peak is likely due to contaminating potassium or sodium adducts.

FIG. 3.

Comparison of twitching zones in recombinant strains. (A) Twitching zones on 0.2% arabinose plates of the PAO1 NP strain expressing the indicated genes from strains PA14 (group III) and Pa281457 (group V). Each gene marked with a pound sign (#) contains a premature stop codon introduced by site-directed mutagenesis as outlined in Materials and Methods. (B) Quantitation of twitching zone areas (mm²); average of 12 individual zones for each sample.

FIG. 4.

Loss of accessory proteins reduces surface piliation but not whole-cell pilin pools. Representative SDS-PAGE gels showing that the lack of the pilin accessory genes reduces surface piliation in PAO1 recombinant strains expressing group III PA14 (A) and group V Pa281457 (B) genes. Pilins are ∼15 kDa, and flagellins used as a loading control are ∼50 kDa. Pilin levels were normalized to flagellin levels by densitometry using ImageJ (Scion). The presence of the accessory protein resulted in a 3-fold increase (group III, TfpY) or a 10-fold increase (group V, TfpZ) in the amount of surface pili recovered. Western blots of whole-cell lysates from PAO1 recombinant strains obtained using α group III pilin (C) or α group V pilin (D) sera show that the lack of surface piliation is not due to decreased pilin levels in whole-cell pools. Each gene marked with a pound sign (#) contains a premature stop codon introduced by site-directed mutagenesis as outlined in Materials and Methods.

FIG. 5.

Recovery of surface pili in a pilT mutant background. Representative SDS-PAGE gels showing that the expression of the pilin gene cassettes from PA14 (group III [A]) and Pa281457 (group V [B]) strains in a PAO1 NP-pilT double mutant resulted in the expression of large amounts of surface pili regardless of the presence of the pilin accessory gene (pound sign indicates a premature stop codon). Pilin levels were normalized to flagellin levels in each lane by densitometry using ImageJ (Scion). Note that the samples prepared from the PAO1 NP-pilT background were diluted 10-fold compared to those from the PAO1 NP background due to the large amounts of pilin present in these samples. The change in the amount of surface pilin recovered from each strain, relative to what was seen for the NP strain complemented with both pilin and accessory genes, is indicated in each lane.

FIG. 6.

Surface piliation and twitching is reduced in a PA14 tfpY knockout. (A) Representative plate showing that twitching motility is reduced in a tfpY mutant compared with the wild type (wt) but can be complemented back to wild-type levels in the presence of pilA_III-tfpY. (B) Average area in mm² of a minimum of six twitching zones per strain. Asterisks indicate significant differences (P ≤ 0.05; Student's t test) between the motilities of wild-type and mutant strains. (C) SDS-PAGE gel of sheared surface proteins from the PA14 wild type and the PA14 tfpY::FRT mutant complemented with pBADGr, pilA_III, or pilA_III-tfpY. In the absence of TfpY surface piliation is reduced; complementation of the tfpY::FRT mutant with pilA_III alone results in increased surface piliation, although motility does not return to wild-type levels as shown in panels A and B. Pilin levels were normalized to flagellin levels by densitometry using ImageJ (Scion), and the change in the amount of surface pilin recovered relative to what was found for the wild-type control is shown in each lane. (D) SDS-PAGE gel of sheared surface proteins. Although surface piliation of a tfpY mutant is low, it is recovered in the tfpY-pilT double mutant, although not to the same level as a pilT single mutant. Pilin levels were normalized to flagellin levels by densitometry using ImageJ (Scion). Note that the pilT and tfpY-pilT mutant samples have been diluted fivefold relative to the other samples due to the large amount of surface pili expressed by these strains. The change in the amount of surface pilin recovered relative to the wild type is shown in each lane.

FIG. 7.

Complementation of PAO1 NP with a group V pilin-group III accessory gene chimera. The group III PA14 accessory protein TfpY cannot replace the function of the group V Pa281457 accessory protein TfpZ. (A) Twitching motility in the PAO1 NP recombinant strain is similar to that seen for the PAO1 NP recombinant strains lacking a functional accessory protein. (B) Representative SDS-PAGE gel showing that surface piliation is reduced 10-fold in the strain expressing the chimera (lane 3) compared to that expressing the original pilA_V-tfpZ cassette (lane 2) and similar to the level obtained by the expression of pilA_V alone (lane 1). Pilin levels were normalized to flagellin levels by densitometry using ImageJ (Scion). (C) Whole-cell lysates probed with α group V pilin antisera show that the reduction of surface piliation in the chimeric strain is not due to differences in whole-cell pilin levels. The flagellin (∼50 kDa) serves as a loading control.