About half of potential pharmaceutical targets are membrane proteins (ion channels and GPCR’s), which have been difficult to approach with modern, structure-based approaches to drug design. Traditionally, there have been several problems associated with working with membrane proteins:

1. Because of their location in membranes, they are frequently difficult to purify and characterize.
2. They are very difficult to obtain in large quantities, and recombinant DNA methods – which work well for obtaining large quantities of water-soluble proteins – generally fail to provide large amounts of properly folded membrane proteins.
3. Even when available in significant amounts, it is much more difficult to crystallize and determine the structures of membrane proteins as compared to water-soluble proteins.

PolyMedix’s method addresses all these problems by converting the receptor to water- soluble proteins while preserving their correct three-dimensional structures and biological activities. We call our computational work in the area of crystallization of transmembrane protein receptors SUCCEED. This stands for Statistical United Combinatorial Computational Environmental Energy Design algorithm. As was successfully demonstrated with phospholamban, the potassium channel, and the influenza M2 proton channel, PolyMedix’s computational technologies allow the water- solubilization and crystallization of transmembrane receptors. Dr. DeGrado continues to advance work in this area.

This technology utilizes both proprietary computational as well as experimental methods, and involves new computational approaches for both:

• Residue based potential, and
• Atomic scoring

Proof of principle has successfully been demonstrated for three membrane-bound receptors to date: phospholamban, the potassium channel (K_CSA), and more recently, the influenza M2 proton channel.

Phospholamban

Calcium ions have long been known to play a critical role in the contraction of all three types of muscle cells-- smooth, skeletal, and cardiac. Muscle cells relax when calcium is moving into the sarcoplasmic reticulum (SR) and contract when the SR releases calcium. For the cells to relax following a contraction, calcium ions must be pumped back into the SR. In cardiac muscle cells, this pumping is accomplished via an enzyme known as sarco(endo)plasmic reticulum Ca2+-ATPase (SERCA). SERCA, in turn, is inhibited by the protein phospholamban, also found in the sarcoplasmic reticulum.

Phospholamban (PLB) is a 6-KD phosphorylatable transmembrane protein of 52 amino acid residues that has been sequenced and characterized from mammals. Phospholamban is the regulator of the SERCA2 activity in cardiac sarcoplasmic reticulum (SR). Phospholamban is a prototype for a very large class of membrane proteins that change their degree of association in response to a biological response, for example phosphorylation or binding of a ligand to an extracellular region of the receptor. Dephosphorylation inhibits the affinity of SERCA2 for Ca2+, whereas phosphorylation removes its inhibition, facilitating Ca2+ transport into the SR lumen and enhancing cardiac function.

The physiological importance of PLB was recently elucidated using genetically-altered mouse models. Activation of contraction in cardiac muscle is dependent on the entry of Ca2+ from the extracellular fluid and the release of Ca2+ from the intracellular Ca2+ store viz the sarcoplasmic reticulum (SR). Ablation of PLB is associated with increased SERCA2 Ca2+ affinity, enhanced contractility and attenuated beta-agonist responses. Phosphorylation occurs in response to beta-adrenergic agonists. These alterations are not due to compensatory mechanisms in the Ca2+-handling proteins or the beta-receptor signaling pathway. However, there is a down-regulation of the ryanodine receptor in an attempt to maintain Ca2+ homeostasis in the hyperdynamic PLB knockout hearts. Re-introduction of wild-type (pentameric) PLB in the knockout heart is capable of reversing its hyperdynamic function. However, monomeric (C41F) PLB is less effective than pentameric PLB in inhibiting contractility, assessed in isolated cardiomyocytes, perfused hearts or intact animals. These findings suggest that PLB is a key regulator of cardiac function and pentameric assembly of PLB is necessary for its optimal regulatory effects in vivo.

Thus, phospholamban is a highly relevant and biologically important target, which could be the basis of therapeutic drug discovery programs for novel treatments for congestive heart failure and cardiomyopathies.

De novo design of a water-soluble analogue of phospholamban: A working three-dimensional model for the pentameric form of the protein was first constructed. This model was used as the framework in the first step of de novo design. Next, nine hydrophobic sidechains on the outside of the protein (green) that are responsible for the solubility of the protein in membranes were computationally changed to polar amino acids (red, acidic; blue, basic; grey, neutral), which would provide aqueous solubility. A Monte Carlo method was used to optimize the nature and arrangement of these sidechains to provide both water-solubility as well as stability.

The water-soluble phospholamban (WSPL) protein was cloned and expressed in bacteria. Strikingly, the water-soluble analogue expressed at very high yield and water-soluble form in bacteria – very different from the wild type membrane-soluble form that expresses in low yield and as an insoluble aggregate. Nevertheless, WSPL retained the essential features of the wild type protein. For example, WSPL adopts the same helical secondary structure as phospholamban, and it exists in a monomer-pentamer equilibrium, which is also the case for the native protein. Furthermore, phosphorylation of the water-solubilized receptor resulted in an increase in the stability of the pentameric form, just as is the case in the membrane-bound form of this protein. Thus, the water-soluble protein retained the essential hallmark of the natural protein.

We have obtained diffraction quality crystals of the transmembrane portion of the water-solubilized phospholamban analogue. These crystals diffract to high resolution, and a complete set of data has been obtained (2.1 Å resolution). We have experimentally confirmed that the crystals are indeed formed from the designed protein, and are now in the process of determining its structure by MAD or isomorphous replacement techniques.

Crystals of the transmembrane region of WSPL (A), and diffraction pattern (B).

These results are unprecedented, and were published in the January 22, 2003 edition of Protein Science. For this work Dr. DeGrado earned the prestigious Merryfield Award in June 2003.

Potassium Channel (K_CSA)

Potassium channels are a diverse and widespread family of ion channels. The voltage-gated K channels of excitable cells have diverse biological roles, in particular in the nervous system. A bacterial K channel, KcsA, for which there is an X-ray structure, provides a structural paradigm for the more complex K channels of the human nervous systems. Simulation and modeling studies based on the KcsA structure are being used to address a number of aspects of the atomic resolution physiology of this and related K channels.

Potassium channels regulated by intracellular ATP (K_ATP channels) have been associated with important cellular functions like vasodilation, hormone secretion, cardiac action potential duration and neurotransmitter release. Thus, potassium channel openers have gained increasing actuality in various therapeutic applications like hypertension, bronchial disorders or acute myocardial ischemia. The potassium channel has been implicated in numerous disorders and diseases including anxiety, depression, ischemia, epilepsy, demyelinating diseases such as multiple sclerosis, immunosuppression, cardiac arrhythmias, neurodegenerative and psychiatric diseases. Potassium channel dysfunction in fibroblasts has also been shown to identify patients with Alzheimer’s disease.

Thus, the potassium channel is a highly relevant and biologically important target, which could be the basis of therapeutic drug discovery programs for novel treatments for a wide variety of major disorders.

PolyMedix used our de novo design technology to prepare a water-soluble version of a potassium channel, K_CSA. Because the fold of this protein is more complex than that of phospholamban, we found that a more fine-grained potential function was required for this project, than that used in the phospholamban project.

As was the case for phospholamban, the water-soluble analogue of this channel reproduced the helical secondary structure, as well as the aggregation (tetrameric in this case) state of the native receptor. Further, although K_CSA is a bacterial channel, we have “humanized” this protein by introducing the corresponding residues from the human sequence into our protein. The resulting protein bound a toxin, which is highly selective for the native tertiary and quaternary structure of this channel. Thus, in summary:

• The entire ion channel was successfully water solubilized.
• The channel retains the correct alpha-helical secondary structure and tetrameric quaternary structure as the native receptor.
• The solubilized receptor binds a toxin that is exquisitely sensitive to the biologically active three-dimensional structure of the protein.

A manuscript has been published in PNAS. ‘Computational design of water-soluble analogues of the potassium channel KcsA.’ DeGrado, et.al., PNAS, February 17, 2004, vol. 101, no. 7, 1828-1833.

PolyMedix hopes to continue to refine this computational technology, and apply it to other pharmaceutically relevant membrane proteins, particularly other ion channel targets. Furthermore, we hope to initiate a program specifically aimed at G-Protein-Coupled Receptors (GPCR’s). PolyMedix plans to use some of the structures generated for internal drug discovery programs, and form collaborations with pharmaceutical and biotechnology companies to design drugs for membrane-bound receptors.

GPCR’s

GPCR’s and other cell-membrane bound receptors are some of the most important – and difficult – targets in medicine. GPCR’s mediate cell signaling and thus control many important human biological functions. GPCR abnormalities are believed to be the underlying causes of many diseases, such as cancers, neurological disorders, and others. Approximately half of currently-marketed pharmaceutical drugs target a GPCR. There are estimated to be approximately 2,200 known GPCR’s, with potentially an even greater number of unknown GPCR’s. To date, it is believed that ligands have been identified for less than 500 GPCR’s.

GPCR’s belong to a family of receptors known as 7-transmembrane bound receptors. Unlike other receptors which are found on the surface of cell membranes, GPCR’s are intimately integrated into cellular membranes with a complex 7-member looping region. Given that important domains of the receptor structure are integral with the cell membrane, it has generally not been possible to reliably water solubilize and thus crystallize a GPCR for X-ray diffraction or high-field NMR studies. It is for this reason that to date it has been very difficult to apply rational drug design approaches for membrane-bound receptors. Thus, the true structure of most membrane-bound receptors is unknown, and most drug discovery has been by serendipitous random screening.

return to top ^