LETTERS
Threonine phosphorylation post-translationally regulates protein secretion in Pseudomonas aeruginosa Joseph D. Mougous1, Casey A. Gifford1, Talia L. Ramsdell1 and John J. Mekalanos1,2 Secreted proteins are crucial to the arsenal of bacterial pathogens. Although optimal activity of these proteins is likely to require precise regulation of release, the signalling events that trigger secretion are poorly understood. Here, we identify a threonine phosphorylation event that post-translationally regulates the Hcp secretion island-I-encoded type VI secretion system of Pseudomonas aeruginosa (H-T6SS). We show that a serine–threonine kinase, PpkA, is required for assembly of the H-T6SS and for secretion of Hcp1. PpkA activity is antagonized by PppA, a Ser–Thr phosphatase. These proteins exhibit reciprocal effects on the H-T6SS by acting on an FHA domaincontaining protein, termed Fha1. Colocalization experiments with the T6S AAA+ family protein, ClpV1, indicate that Fha1 is a core scaffolding protein of the H-T6SS. Mutations affecting this H-T6S regulatory pathway provide a molecular explanation for the variation in Hcp1 secretion among clinical P. aeruginosa isolates. This mechanism of triggering secretion may be general, as many T6SSs contain orthologues of these proteins. Post-translational regulation of protein secretion by Thr phosphorylation is unprecedented in bacteria, and is likely to reflect the requirement for T6S to respond rapidly and reversibly to its environment. Type VI protein secretion systems are widely represented in proteobacteria and have been implicated in the virulence of several human pathogens1–8. The genome of P. aeruginosa PAO1 contains three T6S loci, termed Hcp secretion islands (HSIs). One of these loci, HSI-I, is coordinately regulated with critical virulence determinants such as type III secretion and exopolysaccharide production9–12. Additionally, HSI-I was shown to be important for chronic P. aeruginosa infection of the rat lung, and the H-T6SS has been found to be active in P. aeruginosa present in the lungs of cystic fibrosis patients6,13. Genes that encode proteins with homology to Ser–Thr protein kinases and phosphatases are components of many T6S loci, including HSI-I of P. aeruginosa3 (Fig. 1a and see Supplementary Information, Table S1). PpkA (PA0074), the kinase encoded by HSI-I, has been shown to
1 2
autophosphorylate and to phosphorylate artificial substrates in vitro14. The putative protein phosphatase encoded by HSI-I (PA0075) has not been studied previously and will henceforth be referred to as PppA. To determine whether PpkA and PppA are involved in the function of the H-T6SS, deletions of ppkA and pppA were constructed in wildtype (PAO1) and ΔretS genetic backgrounds and probed for changes in H-T6S activity. As secretion of an Hcp1 orthologue is a general feature of T6SSs, including the H-T6SS, Hcp1 export was measured as an initial gauge of H-T6SS activity4,5,8. The ΔretS background was included in these studies based on our prior observation that this mutant overexpresses the H-T6SS and abundantly secretes Hcp1, whereas the wildtype background produces the H-T6SS at low levels but fails to secrete Hcp1 (ref. 6). In the ΔretS background, deletion of ppkA abrogated Hcp1 secretion (Fig. 1b and see Supplementary Information, Fig. S1a). Conversely, inactivation of pppA in ΔretS decreased intracellular Hcp1 levels and increased the quantity of secreted Hcp1. The effects of ΔppkA and ΔpppA on Hcp1 secretion in wild-type cells mirrored those observed in the ΔretS background. Wild-type bacteria do not secrete Hcp1, thus deletion of ppkA had no apparent effect on Hcp1 localization (Fig. 1b and see Supplementary Information, Fig. S1b). Interestingly, deletion of pppA in wild-type bacteria resulted in the activation of Hcp1 secretion. Neither ppkA nor pppA deletions had an obvious effect on overall Hcp1 levels in the ΔretS or wild-type backgrounds. We also analysed periplasm-enriched fractions for the presence of Hcp1 in ppkA and pppA deletions prepared in wild-type and ΔretS backgrounds. Periplasmic Hcp1 was observed in all strains examined, suggesting that secretion is blocked at the stage of transport through the outer membrane in wild-type and ΔretS ΔppkA cells (see Supplementary Information, Fig. S1). Identification of a gene product required for periplasmic localization of Hcp1 will be required to conclusively demonstrate that periplasmic Hcp1 is a bona fide transport intermediate. Next, the effect of inactivating ppkA and pppA on H-T6SS assembly was determined. The assembly of the H-T6SS was assayed using a chromosomally encoded, stable and active ClpV1–GFP chimera that was previously shown to localize to discrete foci within the cell in a manner dependent on a functional H-T6SS6. ClpV1 is a AAA+ family
Department of Microbiology and Molecular Genetics, Harvard Medical School, 200 Longwood Avenue, Boston, MA 02115, USA. Correspondence should be addressed to J.J.M. (e-mail:
[email protected])
Received 30 November 2006; accepted 14 May 2007; published online 10 June 2007; DOI: 10.1038/ncb1605
NATURE CELL BIOLOGY VOLUME 9 | NUMBER 7 | JULY 2007
797
LETTERS
76
C
S
C
79
80
84
Δ retS Δ pppA
88
rG 1 vg
clp V1 86 87
Δ retS
c
pp
hc p1 82 83
89 Δ retS Δ ppkA
Δ re tS Δ
Δ re tS
Δ re tS Δ
pp
b
1 78
pA
73
kA
72
fh a
pA pp
kA pp 71
icm F1
a P. aeruginosa HSI-I
S
C
S
Mr(K) 60
TMA-DPH
RNAP
GFP
Hcp1-V
pA
kA
Δp p
Δp p
W T
Δh c
p1
17
Mr(K) 17
C 17
Hcp1
Merge
S
d
140
**
e
*
**
120
100 5 80 Foci per cell
Foci fluorescence intensity (arbitrary units)
6
60
3
40
2
20
1
0
0 ΔretS
ΔretS Δ pppA
Δ retS Δ ppkA
Figure 1 PpkA and pppA reciprocally regulate Hcp1 secretion and H-T6SS assembly. (a) Overview of the genetic organization of HSI-I. Open reading frames (ORFs) previously characterized or characterized herein are white, conserved hypothetical T6S ORFs are grey and HSI-I ORFs not conserved in other T6S loci are black. (b) PpkA and pppA promote and inhibit Hcp1 secretion, respectively. Immunoblot analysis of Hcp1 in the cell (C) and supernatant (S) of strains in the ΔretS background. RNA polymerase (RNAP) is used as a fractionation control. The blot was probed simultaneously with the anti-Hcp1 and anti-RNAP antibodies. Immunoblot analysis of Hcp1 secretion in the wild-type (WT) background is shown in the lower panel. (c) PpkA and pppA promote and inhibit H-T6SS assembly, respectively.
798
4
Δ retS
Δ retS Δ pppA
Δ retS ΔppkA
Fluorescence microscopy of strains containing clpV1–gfp in addition to the indicated genotypes. The GFP channel in images of ΔretS and ΔretS ΔpppA are scaled equally for direct visual comparison. The membrane dye TMA-DPH was used in this and subsequent microscopic analyses to outline the position of the cells. The scale bar represents 5 μm. (d, e) Quantitative analysis of fluorescence microscopy images indicates that deletion of pppA causes (d) increased ClpV1 recruitment (e) to fewer T6S foci. Uncropped images of the blots in b are shown in the Supplementary Information, Fig. S5. The error bars represent one s.d. from three microscopic fields in which 20 bacteria per field were randomly selected. A single asterisk indicates P
