Our research focuses on the study of the physical basis of immune system function (immunophysics), the design of immune system proteins with tailored properties (immunoengineering), and the design of low-molecular mass inhibitors that target immune system proteins (drug discovery). We use: (i) nuclear magnetic resonance (NMR) spectroscopy to study the structure and dynamics of peptides and proteins; (ii) computational methods, such as molecular dynamics simulations and electrostatic calculations, to study structure-dynamics-interactions-function correlations; and (iii) structural bioinformatics methods to classify the spatial arrangement of physicochemical properties of protein families, based on their structure and dynamics.

Most of our current work focuses on the complement system, which is an evolutionary ancient form of innate immunity. It is also a link between innate and adaptive immunities. The complement system consists of more than 30 serum proteins and cell-bound receptors and regulators. It is activated via complex and interacting pathways, during which several functional protein fragments and multi-component assemblies are formed. The end-results of complement activation are the formation of (i) protein fragments which tag foreign pathogens for subsequent recognition and elimination by phagocytosis, (ii) protein fragments which promote inflammation at the point of infection, and (iii) a protein assembly that is capable of lysing pathogen membranes.

The proper function of the complement system is finely regulated by the so-called complement control proteins (CCPs), the function of which assures that complement discriminates “self” from “non-self” and attacks only foreign pathogens. However, in certain autoimmune diseases and other pathological situations, complement regulation breaks down with the result of attacking own cells. CCPs are chain-like modular proteins formed by the repeated linear arrangement of the so-called CCP modules (or short consensus repeats). We call CCPs invariably all regulators, receptors, and inhibitors made of CCP modules, without distinction for their specific origin or function.

We have performed Poisson-Boltzmann electrostatic calculations using structure coordinates at atomic resolution. We have studied natural CCPs, such as CR1 and CR2, and viral CCPs, such as VCP and SPICE. The latter are secreted proteins from the pox viruses, which imitate natural CCPs to evade the immune system. We have shown that electrostatics drive the recognition between CCPs and their target complement proteins. We have performed molecular dynamics simulations to study the intra-modular flexibility and inter-modular mobility of CCPs. We have also calculated electrostatic potentials during the MD trajectories to determine the effect of inter-modular mobility on the spatial arrangement of the electrostatic potentials and we have performed covariance analyses to delineate the contributions of the distinct electrostatic properties of each CCP module to the correlated modular motions. We have used electrostatic calculations to design series of CCPs with predicted variable association properties. We use structural bioinformatics analyses to classify active from inactive CCPs, based on the calculation of electrostatic similarity indices (ESIs) and clustering methods.

We have performed Poisson-Boltzmann electrostatic calculations to separate the effects of long- and short-range electrostatic interactions in the formation of the complex between complement protein C3d and its receptor CR2. We propose a two-step model for the recognition and binding of C3d with CR2. Long-range interactions between protein macrodipoles dominate recognition and the process of formation of the encounter complex and short-range electrostatic interactions between residue pairs are essential for binding and the formation of the final complex. We have constructed several theoretical mutants, which we have used to predict their association properties and to explain the large amount of available experimental data. We have used electrostatics to predict the pH and ionic strength dependencies of C3d-CR2 association, which are in good agreement with experimental data. We have also predicted the titration properties for a number of C3d and CR2 mutants, for which there are available association data but not complete titration data.

We have used electrostatics and structural analyses to predict the stabilities and pKa values of ionizable residues for the 10th type III domain of human fibronectin and previously known mutants with higher stabilities. Because of its similarity to immunoglobulins, fibronectin has been previously used as a scaffold for the experimental design of active antibody-like fragments. We will critically discuss the validity of our data in view of the available experimental data for pKa values and the pH and ionic strength dependencies of protein stabilities.

We will discuss how the type of analyses described above, and the interplay between computation and experiment, can be used for the design of proteins with desired stabilities and binding properties, using the native proteins as scaffolds.

Finally, we will present a brief overview of our design efforts for peptide-complement inhibitors, using NMR spectroscopy, structure determination by restrained molecular dynamics-based simulated annealing with NMR-derived or deduced restraints, molecular dynamics simulations, and quasi-dynamic pharmacophore methods.

Extended Abstract Status: Not Uploaded