Ligand-induced dimerization of Drosophila peptidoglycan recognition proteins in vitro

Abstract

Drosophila knockout mutants have placed peptidoglycan recognition proteins (PGRPs) in the two major pathways controlling immune gene expression. We now examine PGRP affinities for peptidoglycan. PGRP-SA and PGRP-LCx are bona fide pattern recognition receptors, and PGRP-SA, the peptidoglycan receptor of the Toll/Dif pathway, has selective affinity for different peptidoglycans. PGRP-LCx, the default peptidoglycan receptor of the Imd/Relish pathway, has strong affinity for all polymeric peptidoglycans tested and for monomeric peptidoglycan. PGRP-LCa does not have affinity for polymeric or monomeric peptidoglycan. Instead, PGRP-LCa can form heterodimers with LCx when the latter is bound to monomeric peptidoglycan. Hence, PGRP-LCa can be said to function as an adaptor, thus adding a new function to a member of the PGRP family.

Fig. 1.

CD spectra of PGRP-LCa and PGRP-LCx with and without His tags in 10 mM phosphate buffer, pH 7.2/100 mM Na₂SO₄. The color key is as follows: light blue, PGRP-LCx; dark blue, PGRP-LCx(His)₆; red, PGRP-LCa; yellow, PGRP-LCa(His)₆.

Fig. 2.

Peptidoglycan structures. Peptidoglycan is composed of alternating N-acetylglucoseamine (GlcNAc) and N-acetylmuramic acid (MurNAc) units forming unbranched glycan strands. The strands are crosslinked by two tetrapeptides connected from the third residue of one tetrapeptide to the fourth residue of another tetrapeptide. Often, the connection is by means of an interpeptide bridge of variable but strain-specific length and composition. The general structure is shown for the E. coli peptidoglycan type, and the variable part (shown in boxes) is shown for other species. The structure of TCT is shown in a gray box. In this monomeric subunit, the N-acetylmuramic acid residue is in its 1,6-anhydro form, the result of enzymatic cleavage from intact peptidoglycan.

Fig. 3.

Binding of PGRPs to isolated polymeric peptidoglycan. PGRP-SA, sLCa, and sLCx were incubated with polymeric peptidoglycan from M. luteus (M.l.), S. aureus (S.a.), B. megaterium (B.m.), E. coli (E.c.), L. casei C (L.c.), L. plantarum (L.p.), or L. fermentum (L.f.). Bound protein (B) was separated from free protein (F) by using centrifugation, and proteins were visualized by using Coomassie staining.

Fig. 4.

PGRP-LCx is localized in the cytoplasmic membrane with the PGRP domain facing outwards. Shown are the results of the immunostaining of S2 cells (Left) and cells overexpressing PGRP-LCx(His)₆ (Right) with a primary antibody against the His epitope and a secondary antibody with an Alexa Fluor 488 conjugate. Cells were fixed but not permeabilized. The images are maximum through-focus projections of 0.5-μm-thick optical sections. Lateral sections are shown as insets.

Fig. 5.

PGRP-LCx binds monomeric peptidoglycan. Monomeric peptidoglycan (TCT) was incubated with His-tagged PGRP proteins, which subsequently were immobilized on a metal chelate matrix. After washing, TCT was eluted from the matrix-bound proteins with 0.1% trifluoroacetic acid. RP-HPLC chromatograms from the eluate after incubating TCT with sLCx(His)₆ (line a), sLCa(His)₆ (line b), SA(His)₆ (line c), and Hepes buffer (line d). The absorbance at 215 nm (solid lines) refers to the left y axis, and the acetonitrile gradient (dashed line) refers to the right y axis.

Fig. 6.

PGRP-LCx/PGRP-LCa heterodimers are induced by monomeric peptidoglycan. (A) Dimerization was assayed on a solid-phase metal-chelate affinity matrix. Proteins were preincubated in the presence or absence of monomeric peptidoglycan (TCT) and then applied to the matrix. Proteins that did not adhere to the matrix [free (F)] were collected, and bound (B) proteins were eluted from the matrix with a strip buffer. Fractions were applied to SDS/PAGE followed by Coomassie staining. (B) Dimerization in the presence of insoluble, polymeric, peptidoglycan (PGN) was assayed by incubating proteins and peptidoglycan for 45 min followed by separation of free protein from bound by centrifugation and analysis of the fractions on SDS/PAGE/Coomassie staining.

Fig. 7.

A model for elicitor-induced PGRP-LC dimerization and signal transduction. The model rationalizes the findings in the present paper and in earlier RNA interference studies (4, 14) and incorporates findings from Stenbak et al. (19). (A) PRGP-LCx (x) is a membrane-bound receptor with the PGRP domain on the outside of the cell. The cross indicates that LCa (a) cannot bind peptidoglycan ligands. (B) Polymeric peptidoglycan (PGN), consisting of alternating units of N-acetylmuramic acid (blue) and N-acetylglucosamine (purple), binds to LCx. Only when LCx is bound to an ultimate anhydrous form (open circle) of N-acetylmuramic acid is a conformational change (in red) induced. This change leads to LCx dimer formation and the activation of the Imd pathway. However, gross PGRP-LCx binding can be to all muramic acid units. Activation by polymeric peptidoglycan relies on the homophilic interaction of PGRP-LCx and by binding to peptidoglycan. LCa is excluded from this type of dimerization because it cannot bind peptidoglycan. (C) TCT (or another peptidoglycan monomer in anhydrous form) binds to LCx, causing the conformational change. (D) The conformational change leads to the heterodimer formation between LCx and LCa needed for peptidoglycan monomer-induced signal transduction.