The Legionella IcmSW complex directly interacts with DotL to mediate translocation of adaptor-dependent substrates

Abstract

Legionella pneumophila is a Gram-negative bacterium that replicates within human alveolar macrophages by evasion of the host endocytic pathway through the formation of a replicative vacuole. Generation of this vacuole is dependent upon the secretion of over 275 effector proteins into the host cell via the Dot/Icm type IVB secretion system (T4SS). The type IV coupling protein (T4CP) subcomplex, consisting of DotL, DotM, DotN, IcmS and IcmW, was recently defined. DotL is proposed to be the T4CP of the L. pneumophila T4SS based on its homology to known T4CPs, which function as inner-membrane receptors for substrates. As a result, DotL is hypothesized to play an integral role(s) in the L. pneumophila T4SS for the engagement and translocation of substrates. To elucidate this role, a genetic approach was taken to screen for dotL mutants that were unable to survive inside host cells. One mutant, dotLY725Stop, did not interact with the type IV adaptor proteins IcmS/IcmW (IcmSW) leading to the identification of an IcmSW-binding domain on DotL. Interestingly, the dotLY725Stop mutant was competent for export of one class of secreted effectors, the IcmSW-independent substrates, but exhibited a specific defect in secretion of IcmSW-dependent substrates. This differential secretion illustrates that DotL requires a direct interaction with the type IV adaptor proteins for the secretion of a major class of substrates. Thus, by identifying a new target for IcmSW, we have discovered that the type IV adaptors perform an additional role in the export of substrates by the L. pneumophila Dot/Icm T4SS.

Figure 1. dotL mutants are defective for intracellular growth.

(A) Schematic of the dotL mutants. Twelve dotL mutants, each containing a single amino acid change, were identified. A black line indicates their individual locations within the 783 amino acid protein DotL and the residue changed is shown. The DotL transmembrane domains are shown by a hatched box and the Walker A box motif (GSTGSGKT) is shown by a gray box. A. castellanii (B) and A/J mouse bone marrow derived macrophages (C) were infected with a wild-type strain (Lp02) (filled squares), a T4SS-deficient strain (Lp03) (filled triangles), a ΔdotL strain containing a dotL⁺ complementing clone (filled inverted triangles) and two representative dotL mutants (Q222R (open diamonds) and Y725Stop (open circles)). Fold growth was calculated by dividing the average colony forming units (CFUs) of triplicate wells at a given time point by the average CFUs at time zero. Error bars represent the standard deviation from the mean and the results are representative of three independent experiments.

Figure 2. Recruitment of IcmSW to the membrane by the dotL mutants.

Subcellular localization of IcmS was assayed using L. pneumophila lysates separated into total (T), soluble (S) and membrane (M) fractions. ΔA ΔL is the ΔdotA ΔdotL double mutant and all of the dotL mutants were expressed in a ΔdotL background. Protein fractions were analyzed by western blot using an IcmS-specific antibody. Fractions were loaded proportionally and are representative of three independent experiments.

Figure 3. Identification of the IcmSW-binding domain on DotL.

(A) E. coli lysates co-expressing His:SW and various GST:DotL fragments truncated at their amino terminus were bound to a Ni-NTA column. The presence of the indicated GST:DotL fusions was assessed in the total (T), flow-through (F), and bound fractions (B) in a GST western. A schematic of the C-terminal amino acids 631–783 of DotL illustrates the indicated deletions. (B) L. pneumophila lysates expressing C-terminal deletions of DotL were separated into total (T), soluble (S) and membrane (M) fractions. Membrane recruitment of IcmS was determined by western blot using an IcmS-specific antibody. (C) IcmS recruitment was assayed for a dotL mutant containing a deletion of the putative IcmSW-binding domain on DotL (amino acids 671–753). Results are representative of three independent experiments.

Figure 4. Identification of the domain(s) of DotL that is necessary and sufficient to bind IcmSW.

DotL domains necessary for binding IcmSW were identified by assaying L. pneumophila strains expressing 10 amino acid (A) and 20 amino acid (B) internal deletions of DotL. Membrane recruitment of IcmS was determined by western blot using an IcmS-specific antibody against total (T), soluble (S) and membrane (M) fractions. (C) The DotL domain sufficient to bind IcmSW was identified by binding lysates co-expressing His:SW and various GST:DotL fragments to a Ni-NTA column. The presence of GST:DotL in the total (T), flow-through (F), and bound fractions (B) was determined by western blotting using a GST-specific antibody. Results are representative of three independent experiments. (D) A schematic of DotL containing amino acids 631–783 is shown to illustrate the properties of the DotL IcmSW-binding domain. The sufficient domain is indicated by a gray box and the necessary domains are denoted by striped boxes.

Figure 5. DotL directly interacts with IcmSW.

(A) Schematic of four DotL fragments that were assayed for interaction with His:SW. DotL amino acids are indicated on the right and the location of the IcmSW-binding domain is shown at the top. (B) E. coli lysates containing His:SW and GST:DotL fragments were sequentially exposed to Ni-NTA and glutathione sepharose resins. Samples were separated on a denaturing SDS-PAGE gel and stained with Coomassie Blue to view isolated proteins. Arrows indicate GST:DotL, IcmW, and His:IcmS proteins, a bracket identify several GST:DotL degradation products, and molecular markers are shown on each side of the gel.

Figure 6. Growth phase dependent instability of DotL in the absence of icmS.

DotL levels were analyzed by western blotting using a DotL-specific antibody on samples harvested at various stages of growth. Samples include Lp02 or ΔicmS strains expressing wild-type DotL, the DotLY725Stop mutant, DotL1–773, or a DotL mutant lacking a domain necessary for IcmSW-binding (DotL1–694/715–783). Samples ranged from mid-exponential phase (lane 1, OD₆₀₀ ∼2.5) to late stationary phase (lane 8, OD₆₀₀ ∼3.5). Results are representative of three independent experiments.

Figure 7. The dotLY725Stop mutant has a specific secretion defect for IcmSW-dependent substrates.

Export of CyaA fusions to the following Dot/Icm substrates was assayed: A. SdeA, B. SidD, C. VipA, D. RalF, E. LnaB, and F. LidA. The CyaA:substrate fusions were expressed in a wild-type strain (black bars), a strain lacking icmS (white bars), a strain with the dotLY725Stop mutant integrated onto the chromosome (striped bars), and a dotLY725Stop ΔicmS double mutant (gray bars). cAMP was measured by ELISA from triplicate wells and error bars represent the standard deviation from the mean.

Figure 8. Model for secretion of IcmS/IcmW-dependent effectors by DotL.

The transmembrane-spanning L. pneumophila T4SS complex is represented by a purple cylinder, components of the T4CP subcomplex (DotL, DotM, IcmS, and IcmW) are labeled and the inner and outer membranes are indicated as IM and OM, respectively. (A) Export of IcmSW-dependent substrates in a wild type strain. (B) Rejection of IcmSW-dependent substrates in a dotL mutant strain that lacks the IcmSW-dependent binding domain (dotL ^−S/W). (C) Jamming of IcmSW-dependent substrates in a ΔicmS mutant, leads to DotL degradation by ClpAP.