Morgan, J.
(2016)
Biophysical studies of the ligand-affinity & binding of leukocyte integrin Mac-1 I-domain: an open & closed case?
PhD thesis, University of Nottingham.
Abstract
Leukocyte integrin Mac-1 plays an integral role in the innate immune system through its function as a cell adhesion receptor. In this role Mac-1 mediates leukocyte migration during the inflammatory response, an essential element of host defense. To date Mac-1 has over 30 identified ligands, with the major ligand-binding region recognised to be the I-domain of the alpha-chain. Within the I-domain interactions are localised around the MIDAS, these are typically characterised by an acidic residue from the ligand co-ordinating through the MIDAS bound Mg2+ ion.
This work has focused on the characterisation of the Mac-1 I-domain interaction with several key ligands including: GPIbα, kininogen, HKD5 and fibrinogen. The GPIbα interaction governs the firm adhesion of leukocytes to aggregated platelets to enable their migration into the surrounding tissue to combat infection. This is a classic example of the role of Mac-1 within inflammation. In the case of kininogen binding to Mac-1 I-domain, domain five (HKD5) of cleaved high molecular weight kininogen (HKa) is the proposed ligand, whilst domain three (HKD3) is a known ligand for GPIbα. As such this raises the prospect of kininogen acting as a potential mediator of the Mac-1 interaction with GPIbα and characterising the binding interfaces for both ligands on the I-domain surface is a key goal of this work to provide more understanding of how this could work mechanistically. Interestingly, the interaction of HKa with Mac-1 has been shown to inhibit the binding of fibrinogen, another ligand of interest, suggesting a shared binding interface on the I-domain. There is much still to understand about the mechanism and physiological function of the fibrinogen interaction with Mac-1; in vivo studies provide an early indication of a role for this interaction in governing cross-talk between the haemostatic and inflammatory systems. Collectively all three ligands alongside Mac-1 have been identified as key in immunothrombosis, which demonstrates an alternative beneficial aspect to the usually pathologically occurrence of thrombosis. The role of Mac-1 and its ligands in both these critical physiological events, the characterisation of their interactions for the future development of therapeutics to combat thrombosis and the improved understanding of the role of immunothrombosis in host defense is of the utmost importance.
This project has undertaken extensive studies of the Mac-1 I-domain by NMR and SPR spectroscopy for the purposes of understanding the structural basis of the I-domain ligand-affinity and characterising binding interfaces. The NMR studies have yielded the 80 % amide backbone assignment through the characterisation of the 15N-TROSY spectrum of the wild-type Mac-1 I-domain alongside two mutant constructs, T209A and F302W, through the analysis of 2D and 3D NMR experiments. This encompasses the assignment of the five key MIDAS residues, as well as the full characterisation of the C-terminal alpha helix, which was of particular importance for understanding the structural divergence observed between the 'open' and 'closed' X-ray crystal structures. Good coverage of assignment was obtained across the remaining I-domain structure with exception of some of the loop containing regions, where the more dynamic character of the structure hindered the resolution of these peaks. Overall, this provided a good foundation for the further characterisation of the wild-type Mac-1 I-domain ligand-binding interfaces. This characterisation was achieved for the ligands: GPIbα N-terminal domain, HKD5 and fibrinogen gamma-chain peptide P2-C in the presence of Mg2+. The interfaces defined by this work were noted to be focused around the MIDAS, as has been previously suggested by literature evidence of antibody binding studies to block these respective interactions. While each of the ligand interfaces were unique, there was overlap of the residues involved. This new characterisation of the interfaces adds weight to the proposal that with regards to Mac-1 binding HKa acts to bridge the interaction with GPIbα. This led to the further proposal that in this role HKa facilitates a 'capture' and 'hand-over' mechanism through its binding of GPIbα via domain 3 and Mac-1 via domain 5. In this manner, HKa bridges the interaction between platelets and leukocytes. HKa acts to bring the two cell-surface receptors into close proximity, thereby enabling the subsequent competitive displacement of HKD5 by GPIbα at the Mac-1 I-domain MIDAS. Furthermore, the similar interface of the fibrinogen binding site, also characterised by these studies, is in line with established evidence that HKa and fibrinogen exert inhibitory effects with regards to each others interaction with Mac-1. A possible hypothesis for the physiological role of a shared interface is that fibrinogen acts as an endogenous control to the HKa promoted binding of Mac-1 to GPIbα in order to prevent excessive vascular inflammation, which could occur if the interaction was allowed to proceed unchecked. In response to vascular injury fibrinogen is localised to the site by adhesion to the exposed subendothelial matrix. Therefore fibrinogen will be present at high concentrations in the vicinity of the activated platelets as the leukocyte recruitment process begins and would be positioned to provide a regulatory role.
The SPR studies of the binding profile for the wild-type and mutant I-domain constructs provided clear evidence on the essential role played by the metal cation in the Mac-1 I-domain interactions. The role of the cation was clearly demonstrated by the loss of binding observed for the wild-type I-domain following the addition of EDTA. Furthermore, the abrogation of binding for the T209A construct highlighted the particular importance of the MIDAS co-ordination of the cation to facilitate ligand-binding. The metal ion co-ordination was shown to be essential for ligand recognition in the cases of both GPIbα and kininogen with the T209A construct showing a total knock-down of binding, while significant but not total loss of binding was also seen for the fibrinogen interaction. The GPIbα N-terminal domain was demonstrated to be the result of a 1:1 interaction as was the kininogen interaction, although with an additional element of non-specific binding. While the fibrinogen interaction binding profile exhibited the character of a multi-site ligand, although the exact number could not precisely defined here. With regards to affinity GPIbα N-terminal domain and HKD5 interaction were found to be of a similar magnitude (K_D = 121 ± 9.7 µM and 77 ± 55 µM respectively) while the fibrinogen interaction was found to several fold stronger (K_D = 30 ± 3.3 µM). In comparison with other characterised integrin interactions these results fit best with those obtained from the non-activated I-domain conformation and as an extension from this it must be concluded, that the F302W mutant does not represent the higher ligand-affinity conformation reproduced elsewhere using engineered disulphide bonds.
The characterisation of the identified MIDAS localised interfaces and metal-ion dependence brings the Mac-1 I-domain mode of binding with both GPIbalpha and kininogen into the same territory as the characterised X-ray crystal structure complexes of Mac-1 with ic3b and related beta2-integrin LFA-1 with ICAM-1. Thus, further strengthening the proposed model of binding for these two ligands, as via the co-ordination of an acidic residue at the MIDAS. Altogether this work has helped to build a narrative of the physiological interplay of the Mac-1, GPIbα, kininogen and fibrinogen interactions.
Actions (Archive Staff Only)
 |
Edit View |
Tools
Tools
![[thumbnail of Juliet_Morgan.pdf]](https://eprints.nottingham.ac.uk/style/images/fileicons/application_pdf.png "Juliet_Morgan.pdf")