The Interaction of a Bacterial Toxin (Aggregatibacter actinomycetemcomitans leukotoxin) with Its Receptor Depends On Both Lipid-Protein and Protein-Protein Interactions

The Interaction of a Bacterial Toxin (Aggregatibacter actinomycetemcomitans leukotoxin) With Its Receptor Depends on Both Lipid-Protein and Protein-Protein Interactions

Angela C. Brown, Patrik Nygren, Kathleen Boesze-Battaglia, Edward T. Lally

The gram negative bacterium Aggregatibacter actinomycetemcomitans is associated with human periodontal disease as well as systemic diseases such as infective endocarditis.  As part of its toxic arsenal, A. actinomycetemcomitans secretes a 114-kDa protein called leukotoxin (LtxA).  This toxin kills human white blood cells, thereby allowing the bacterium to flourish in the host.  The ability of LtxA to kill cells has been shown to require both cholesterol and the expression of lymphocyte-function associated antigen 1 (LFA-1), an alpha-L(CD11a)/beta-2(CD18) integrin.  Binding of LtxA to LFA-1 initiates a multi-step process that results in the cleavage of talin from the cytoskeleton and clustering of the LtxA/LFA-1 complex in large, cholesterol-rich rafts.  In this work, we sought to identify the role of both cholesterol and LFA-1 in this process, using confocal microscopy, surface plasmon resonance (SPR) and fluorescence resonance energy transfer (FRET) microscopy.

We first visualized LtxA internalization using confocal microscopy and demonstrated that internalization occurs only in the presence of LFA-1, and the toxin is located close to the membrane after internalization.  The functional significance of the interaction of LtxA and LFA-was explored using a K562 cell line (Science 301:1720, 2003) which expresses a CFP-tagged cytosolic alpha-L domain and a YFP-tagged beta-2 domain.  In these FRET microscopy experiments, we found that activation of LFA-1 with PMA causes transient cytosolic domain separation (measured by a decrease in FRET), as expected.  However, binding of LtxA to LFA-1 results in an increase in FRET, which is consistent with a mechanism in which LtxA binds to and brings the cytosolic domains closer together than they are in the inactive state.  Unlike activation, this ligating effect is not transient, lasting for at least 30 mins. 

The binding of LtxA to cytoplasmic domains of alpha-L (residues 1112-1170) and beta-2 (residues 724-769) was measured using SPR.  A peptide corresponding to the cytoplasmic domain of the beta-3 cytoplasmic domain (residues 746-786) served as a control for the experiments.  We determined that LtxA has a strong affinity for the cytosolic domains of both the alpha-L (K_D = 1.5 x 10^-8 M) and beta-2 (K_D = 4.2 x 10^-9 M) subunits and a lower affinity for the beta-3 cytoplasmic domain (K_D = 2.3 x 10^-7 M).  A series of shorter peptides were then synthesized to locate the regions of the cytoplasmic domains that interact with LtxA.  These results demonstrated that LtxA bound to the membrane-proximal region of the alpha-L domain and an intermediate region of the beta-2 domain.  Overall, these SPR results indicate that LtxA binds the cytoplasmic domains of LFA-1 strongly, suggesting that LtxA may displace cytoskeletal-linking proteins, such as talin, from LFA-1.  This is consistent with our previous observation that the LtxA/LFA-1 clustering requires cleavage of LFA-1 from its cytoskeletal tether.

In addition, the LtxA/LFA-1 clustering requires the presence of cholesterol, suggesting that either LtxA or LFA-1 binds to cholesterol, in addition to binding one another.  Because LFA-1 is not known to have an affinity for cholesterol, we measured the affinity of LtxA for membranes containing varying amounts of cholesterol.  We found that the affinity of LtxA for membranes in the absence of cholesterol (1.9 x 10^-8 M) is comparable to the affinity of LtxA for the alpha-L cytosolic domain of LFA-1.  In the presence of cholesterol, however, the affinity of LtxA for membranes increases substantially (4.6 x 10^-12 M), due to a drastic difference in the dissociation rate.  These results suggest that LtxA may bind to the membrane and LFA-1 simultaneously and that the binding of LtxA to cholesterol may act to hold the toxin and LFA-1 in a cholesterol-rich environment, thereby driving the cholesterol-dependent clustering of LtxA and LFA-1.

This work provides additional insight into the complex interaction of LtxA with its receptor and demonstrates that the process is driven by both protein-protein and lipid-protein interactions.

This work was supported by the NIH: R01DE09517 and F32DE020950.