Elucidating structure-activity relationships for protein-hydrocarbon interactions: Using human CD1d protein as a model system

Lipids consist of water-insoluble hydrocarbon chains (HC). To fulfil their diverse and often vital roles in aqueous environments lipids partner with lipid-binding proteins. These proteins accommodate the water insoluble HC of lipids in deep hydrophobic channels, thereby shielding the HC from water. This type of interaction of HC with lipid-binding proteins is of crucial importance for diverse biological processes. New insights into the rules which govern protein-HC interactions will enhance our understanding of these processes and guide the development of new therapies.
A prototypical example for a HC-binding protein is human CD1d, which exerts key roles in the body's immune system. The proposed research will use this protein as a model system to provide a systematic study into protein-HC interactions. In particular, it will examine how selective chemical modifications in a given model HC affect binding to CD1d. The twisted hydrophobic channels in CD1d and many other lipid-binding proteins inflict energetically unfavourable conformations on bound HCs. This mismatch between the shape of the lipid-binding protein channel and the preferred minimum energy conformation of the HC can be accentuated or neutralised by chemical modifications of the HC. We will employ fluorine atom modification of HCs which can exert a strong impact on HC conformation, while leaving the flexibility of the HC fully intact.

The overarching objective of the proposed research is to provide a framework for the rationalisation of chemical HC modifications to manipulate their interaction with lipid-binding proteins. This may eventually lead to the development of new HC-containing drugs. The project embraces a cross-disciplinary approach employing sophisticated computational, synthetic and molecular biology techniques to achieve the following objectives:
1) To analyse the stabilising and destabilising effects of selective fluorine atom modifications on CD1d protein binding of a model HC. The HC-containing CD1d-binding antigen KRN7000, which is currently examined as a possible new drug in human clinical trials, will be used as a template for these fluorine modified HCs. Fluorine atom-modified KRN7000 analogues will be generated by chemical synthesis. The novel compounds will then be compared for their ability to bind and stabilise CD1d protein, and to switch on the immune function of CD1d.
2) In theory, a large number of fluorine-atom modified KRN7000 analogues could be compared. However, chemical synthesis of so many different lipids is not feasible. Therefore, we will employ high-end computational methods integrating empirical force fields and quantum mechanical calculations to guide the synthesis of a limited number of chemical KRN7000 analogues.
3) We expect that integration of the "real-life" CD1d-HC binding data from this project and the computational data will enable a deeper understanding of the mechanistic basis for the observed effects of fluorine-HC modifications on CD1d protein binding. This would be an important step towards rationalisation of protein-HC interactions and the development of a future computational modelling platform for HC-based drug design.

The data generated in the proposed project will benefit different groups of academics, in particular: researchers working on protein-binding lipids; immunologists working on lipid-specific immune responses; synthetic chemists with interests in fluorine chemistry or glycolipid chemistry; computational chemists interested in complex systems modelling.

If successful, the results of the proposed research will be of great interest for the pharmaceutical industry. KRN7000 and certain analogues of KRN7000 are currently being tested in human clinical trials. The proposed research may lead to rational design of novel KRN7000-derived analogues with therapeutic properties.

Crystal structures of ligands with hydrocarbon chains (HC) bound to proteins reveal that the hydrophobic binding channels of these proteins enforces non-linear, and therefore destabilised, HC conformations upon the bound HC ligands. The central tenet of this project is that rational selective HC fluorination can be used to modulate lipid-protein affinity through conformational preorganisation and fluorophilic protein-lipid interactions. As a model system for these studies we will make use of the human protein CD1d and the synthetic CD1d-ligand KRN7000, because: 1) The CD1d-bound acyl HC in KRN7000 adopts a "destabilised" non-linear conformation; 2) Presentation of KRN7000 via CD1d to a conserved regulatory human T cell population, called iNKT cells, is of therapeutic interest; 3) Variations in the acyl HC of KRN7000 impacts on the CD1d/KRN7000 complex stability and iNKT cell function.
A cross-disciplinary approach will be used to achieve the objectives of the research. The conformational stability of the acyl HC in KRN7000 will be modified via the introduction of (two) fluorine atoms on the HC and recombinant single ligand containing CD1d proteins and human iNKT receptors will be used to compare the CD1d binding parameters of these analogues and their ability to mediate iNKT receptor binding and iNKT function. To rationalise the selection of difluorinated KRN7000 analogues for chemical synthesis, we will employ a combined computational approach of classical force fields and large scale quantum mechanical calculations. The programme will deepen our insight into structure-function relationships of protein-HC binding, offer new rationales for designing optimised lipid ligands, and provide novel immunomodulatory compounds for future therapeutic studies.

The overarching misssion of the proposed research project is to enable future rational design of protein-targeting hydrocarbon chain-based compounds for the development of novel medicinal substances and other, non-medicinal protein-targeting molecules. Non-academic beneficiaries of the proposed research therefore most prominently include stakeholders from the Pharmaceutical and Agrochemical industry.

Fluorines are already incorporated into drug and agrochemical compound design to enhance their biophysical properties and performance. An exciting recent promising development is the exploitation of weak fluorophilic interactions in drug design to enhance ligand-protein binding. In addition, recent insights have indicated the possibility to achieve fluorine-mediated conformational control of hydrocarbon segments for enhance ligand binding. The demonstration of the reliability (or otherwise) of the proposed computational simulation approaches regarding selection of fluorination positions on hydrocarbons to modulate their interactions with proteins, both through conformational pre-organisation and fluorophilic interactions, will therefore be an important step forward towards the rational use of fluorine modifications for bioactive compound design.

The chosen lipid-binding protein model system CD1d, and the chosen CD1d-binding ligand, KRN7000, have potential clinical applications as novel therapies for a range of conditions, including chronic infections, inflammation and cancer. Industry, e.g. biotech companies such as "NKT THERAPEUTICS INC" (http://www.nktrx.com/main_page.html, owners: AstraZeneca/MedImmune group), which are trying to target iNKT cells to "treat autoimmune diseases and inflammatory diseases, cancer, asthma and dermatitis" could therefore benefit considerably from the new insights provided by our studies. The results of the proposed studies may therefore attract specific interest for these KRN7000 analogues from the pharmaceutical industry.
Furthermore, the results generated in this project will be highly relevant for all other human CD1 proteins, including CD1c and CD1b which are involved in the immune defence against tuberculosis, because the acyl HC-binding A'channel of CD1d is highly conserved in the other human CD1 proteins. Indeed, the general HC-protein binding mode of CD1 proteins is not restricted to this protein family but is similarly seen in other human lipid-binding proteins which exert important biological functions. Thus, the proposed studies will of more general interest for the pharmaceutical industry. Furthermore, since lipid-binding proteins are ubiquitous in nature we anticipate that the results of these studies may also be of interest for the agrochemical industry.

In summary, understanding the structure-activity relationships of protein-binding lipids could, in the longer-term future, allow for enhanced design and testing of new potential drugs or compounds for agriculture, thereby potentially strengthening UK industry and benefitting society.