Donald M. Engelman - US grants

We propose to use the tools of structural biology to develop a perspective that will lead to fundamental understanding of the structures and interactions of the proteins of the human immunodeficiency virus (HIV). The information will be used as the basis for the chemical design of agents that may be pharmacologically active. Our approach is conceptually straightforward: a) We will produce proteins by expression in E. coli and cultured cells, and peptides by solid phase synthesis. b) We will carry out biochemical analysis of catalytic activities, binding characteristics and other properties. c) We will study the structures and interactions of the molecules using physical methods including x-ray crystallography, neutron diffraction, NMR, and optical spectroscopy. Some strategies of chemical modification will be used as well. d) We will develop a conceptual framework for the study of electrostatic and steric interactions of molecules in the aqueous phase and in the hydrophobic region of lipid bilayer envelopes. Thus, we will have the materials methods, and concepts needed to explore the fundamental structural biology of the HIV proteins and to move toward the use of that understanding in the design of pharmacologically active agents.

The hypothesis that transmembrane helices are independently stable folding units will be examined by studies of fragments of bacteriorhodopsin (bR) and intact bR reconstituted in lipid vesicles. The consequences of sequence modifications will be studied in a collaborative project with Dr. Khorana at MIT. The interaction of fragments with each other will be followed using vesicle fusion techniques, neutron and X-ray diffraction structural methods, spectroscopic measurements and chemical modification techniques. The protein folding problem represents one of the most important scientific challenges facing biochemists, biophysicists and physical chemists as well as the major gap in the understanding of the flow of information in living systems. Understanding the three-dimensional implications of the linear, abstract information in DNA can be regarded as a true understanding of the genetic code. Important principles remain to be established and any attempt to engineer proteins de novo or to understand medical problems in terms of altered genetic information must benefit enormously from insights concerning macromolecular folding. However, in spite of the practical importance and intellectual fascination of the problem, structural prediction and the determination of pathways of folding remain elusive goals.

The remarkable recent advances in genetic engineering now allow industrial-scale production of therapeutic proteins for human health care. With the introduction of new production methods have come immense purification challenges. In order to avoid untoward side effects of contaminants (many of which are as yet unknown), therapeutic proteins must be ultra-purified to levels beyond the scope of present technology. Thus, there is a great need for novel separation and purification techniques which may be applied on an industrial scale. This multi-investigator study is an effort to integrate several innovative technologies ?such as High Performance Liquid Chromatography, (HPLC), electrophoresis, and supercritical-fluid based separations! to provide a basis for industrial ultrapurification of biomacromolecules. Applications of such new purification and separation technologies will be tested on recombinant tissue plasminogen activator, a life-saving therapeutic used in the treatment of heart attack victims.

9406983 Engelman We conceptualize membrane protein stability in terms of a two-stage model: many transbilayer helices can be regarded as independently stable folding units which interact specifically in a side-to-side fashion to generate tertiary and quaternary structures. This contention is supported by experiments and suggests that a detailed understanding of membrane helix interactions will lead to greater understanding of membrane protein structure and to more powerful prediction methods. During the past grant period, we established the independent stability of many (but not all) of the helices of bacteriorhodopsin (BR), evolved a model system permitting detailed computational and experimental studies of the dimerization of glycophorin A (GpA) transmembrane helices, studied the possible roles of polypeptide links and retinal in bacteriorhodopsin stability, and discovered that one bacteriorhodopsin helix is capable of reversible spontaneous insertion across lipid bilayers. These studies have now permitted a more refined investigation to be designed: 1. Spontaneous insertion will be studied using the reversible transbilayer insertion of BR helix C as a starting point. Which sequence features permit or abolish such insertion? Can a pH difference translocate the peptide across a lipid bilayer? Can an experimental delta delta G be assigned to each amino acid? 2. Folding intermediates and alternative domain structures will be studied using BR as a model. Is the helix F - helix G hairpin a stable domain? Do helices A or B interact separately with the rest of the molecule? 3. Structural studies exploiting NMR will be used to define the detailed structure of the glycophorin dimer. Is it as predicted from mutagenesis and computation? 4. Measurement of the energy of helix-helix interactions will be made using sequence variation in the glycophorin helix. These will be used to test computational approaches and to seek chemical principles in the interactio ns. 5. Effects of the lipid environment on the stability of helix- helix interactions will be studied. Possible insights into the relative contributions of lipid-lipid, lipid-helix and helix-helix interactions will be elucidated. %%% Many important biological events happen within membranes. The responses of cells to their environment, the conversion of energy from one form to another, the regulation of the living state of a cell, and many other vital processes involve the interactions of molecules inside the membrane structure. Our work to this date and our planned experiments are aimed at understanding how specific interactions occur and describing these in chemical terms. Since so many important events happen in this environment, and since it has been shown that the specificity of different binding events can be very high, it may well be that new pharmacological agents can be found that will act within membranes to modulate cell activity. Indeed, it may prove to be the case that some drugs already discovered act in this way, but we do not yet understand their mode of action. A higher level of understanding will permit the design of search methods to find new drugs and, perhaps, the de novo design or modification of drugs for specific effects. ***

9317832 Brunger Computational approaches are important for structure determination by X-ray crystallography and solution NMR spectroscopy, studies of macromolecular structure/function relationships, protein folding, structure prediction, and drug design. The proposed workstation cluster will enable us to extend the limitations of these methods and to apply them to more difficult and important biological problems. The workstation cluster will also serve as a platform for testing advanced algorithms for the parallelization of our programs. The main computational tool for the proposed projects is the program X-PLOR (Brunger, 1992b). We routinely use X-PLOR for structure determination and refinement of X-ray and solution NMR structures, as well as for free energy perturbation calculations, structure prediction and certain aspects of structure-based drug design. It is proposed to acquire a workstation cluster which provides the most cost-effective and flexible solution to satisfy our needs. X-PLOR performs very efficiently on the proposed workstations. We intend to primarily use the cluster in a mode where multiple and independent jobs are run independently on the processors. Thus, we can readily make use of the proposed cluster without extensive software development.

We seek to establish the structures of several simple oligomeric complexes of transmembrane helices using several strategies, including a combination of mutagenesis and modeling, the use of multi-dimensional heteronuclear NMR, and, perhaps, crystallography. At the same time, we will use sequence variation to probe the chemical interactions responsible for stable association of helical structures. Measurements using small angle x-ray scattering, calorimetry, and electrophoresis will define differences in stability arising from specific amino acid substitutions. These will be compared with predicted differences based on computational modeling via energy minimization, molecular dynamics, and simulated annealing. Initially, work will focus on pentamers formed by phospholamban transmembrane helices, the dimers of glycophorin helices, and transmembrane segments from the T cell receptor. Combining knowledge of the structures with an understanding of the chemical principles of their formation will lead to improved predictive strategies for the structures of helical membrane proteins.

structural biology; membrane proteins; protein structure; thermodynamics; intermolecular interaction; chemical stability; glycophorin; protein engineering; synthetic peptide; mathematical model; lipid bilayer membrane; chemical association; conformation; computer program /software; dimer; molecular weight; chimeric proteins; SDS polyacrylamide gel electrophoresis; peptide chemical synthesis; computer simulation; analytical ultracentrifugation; site directed mutagenesis; protein purification;

The computational facility will be used in the crystallographic work to determine and refine structures based on x-ray crystallographic and NMR studies. Computational modelling will also be used to explore issues in drug design and in modelling the interactions of transmembrane domains.

The central idea of our project is that understanding intra-membrane protein-protein interactions, and finding specific agents for their disruption may define a novel strategy for combatting enveloped virus infections. Our objective is to characterize, understand, and predict the structural interactions that occur in the transbilayer region when membrane proteins oligomerize. Our findings in two membrane protein systems (bacteriorhodopsin (BR) and glycophorin A (GpA)) show that such interactions can be important factors in driving oligomerization and can have a high degree of specificity. Preliminary results suggest that the transbilayer regions of HIV gp41 and influenza hemagglutinin (HA) are also capable of oligomerization without their respective endo- and ecto-domains. If either the oligomerization or the function of viral envelope proteins is mediated to some extent by association of their transmembrane domains, then disruption of these interactions by pharmacological agents may prove useful. We will use mutagenesis, disruptive agents, thermal denaturation and variations in the lipid environment to define interactions driving oligomerization in gp41 and HA transbilayer domains. We will then express the altered whole protein in cultured cells to explore the effect of disruptive mutations on oligomerization and fusion activity. Structures of the transmembrane portions will be explored using NMR and other spectroscopic methods, and thermal disruption will be used to probe interaction energies. Computational modeling will be exploited to lead toward generalization of ideas, with the aims of predicting structure and serving as a basis for drug design enhancement. Finally, we hope to exploit our assays in high capacity screens to find agents specific for disruption of transmembrane interactions.

The general perspective of this project is that an understanding of the motifs and chemical principles that govern helix-helix interactions in membrane proteins will serve as a basis for understanding and predicting structure, oligomerization, and function. Given the genomic abundance of sequences that appear to encode helical membrane proteins and the paucity of detailed high resolution structures, our alternative approach should prove valuable in providing working models for structures as well as in guiding thoughts concerning functional states. The key is to develop connections between structural and thermodynamic descriptions of helix-helix interactions. Additionally, motifs may be identified that allow simple approaches using straightforward sequence analysis. Progress during the first two years of the program has resulted in a number of promising directions, which we plan to exploit during the next grant period. An important development that serves as a cornerstone for our plan is the emergence of two genetic screens for helix interaction in E. coli membranes. These will be exploited to search the E. coli genome for naturally occurring interactions and to screen random libraries to obtain global views for interacting sequences. A second platform for our future studies is provided by the determination of the glycophorin A transmembrane helix dimer structure. Using the structure, we are testing ideas concerning the energy terms important in the interaction by redesigning the interface and studying structural and energetic properties of different designs. Additionally, we are exploring the use of natural motifs, such as the leucine zipper, to design interacting transmembrane helices. Computational chemistry and genomic database studies will be used to refine chemical ideas and document the occurrence of specific interactions in naturally occurring membrane proteins. The properties of helix interaction interfaces appear encouraging with regard to new avenues for drug discovery.

Engelman
MCB 9905671

This research is aimed at providing a connection between the subjects of membrane protein structure, folding and oligomerization on the one hand, and lipid bilayer properties on the other. The approach is to measure helix-helix association-dissociation equilibria in lipid bilayers and in membranes, observing changes in the dissociation constant resulting from changes in the lipid environment. General effects, such as those arising from bilayer thickness and cholesterol content, and specific lipid association effects, in which dilute quantities of a lipid in a bilayer have disproportionate effects on the helix dissociation will be investigated. Two key assays are Forster Resonance Energy Transfer and a new genetic assay for helix association. These will allow studies in synthetic lipid bilayer vesicles and in the inner membrane of living E. coli, respectively.

These studies are relevant to such important areas as transmembrane signaling events (which often involve changes in helical membrane association), selective transport in membrane trafficking (where evidence suggests bilayer thickness effects to segregate retained and transported proteins in cytoplasmic compartments), and modulation of many membrane functions. A fundamental understanding of the principles of membrane protein folding should also prove enlightening in attempts to interpret
information in genomic databases.

The CORE facility is designed to provide support in several technical areas used by all or most of the participants in the program. It offers capabilities in NMR spectroscopy with facilitated access to 500, 600, and 800 MHz instruments, analytical ultracentrifugation using the Beckman XL-I ultracentrifuge, Fourier transform infrared spectroscopy, peptide synthesis, and computational capabilities that include three dimensional structural displays, modeling using molecular dynamics simulation, and database searching. In the period since the initiation of the grant a little more than two years ago, each of these facilities has been created and brought to operational status. While some desirable upgrades are requested, it is felt that the CORE has basically evolved to an excellent state, fully capable of supporting the research proposed for the next grant period if it is appropriately staffed.

[unreadable] DESCRIPTION (provided by applicant): It is often assumed that bilayers in membranes have the thicknesses that they would have in vesicles made from the lipids alone. Many ideas rest on this view, including hydrophobicity scans to find helices, notions of protein sorting between membranes, and the rationale for the huge literature on pure lipids. Yet almost no measurements exist as points of reference for the thicknesses of actual membranes. Similarly, there are few data to inform us of the proportion of the area of a natural membrane that can be regarded as an unperturbed bilayer. Our preliminary studies suggest that the classical view, first established in the Fluid Mosaic Model, may be in need of revision: proteins rather than lipids set membrane dimensions. A consequence is that strain exists between protein and lipid, and the implications will be fascinating to explore. The principal themes of this proposal are to use diffraction to measure the mutual influences of membrane proteins, phospholipid and cholesterol on membrane thickness, to establish values for the areas occupied by these components at the membrane center, and to use model systems that combine biochemistry and molecular dynamics simulations to discern the underlying chemical principles. [unreadable] [unreadable]

DESCRIPTION (provided by applicant): Membrane proteins are the targets of a majority of all drugs. Many of these drugs bind deeply within the non-polar region of the membrane, where transmembrane helix (TM) interactions are a dominant factor in the stability and activity of target proteins. We have developed assays that might be useful in finding molecules that specifically influence TM-TM interactions, either strengthening or weakening the association of these key elements of structure. If such molecules can be found in screens, they will provide leads for the discovery of new pharmacological agents. To investigate this possibility, we will set up screens using important proteins that we have worked with and already know to interact through their TM domains: the PDGF receptor TM, the TM of the E5 protein from Papilloma virus, the interaction of PDGFR TM with E5, and the li TM from the MHC. We will test specificity by cross comparison among different targets, by positive and negative controls with Glycophorin A TM, and by TM mutational analysis of the most specific and effective hits, where it is expected that modification of a specific binding site will alter affinity and efficacy.

DESCRIPTION (provided by applicant): Interactions between transmembrane helices are a major feature of the architecture of membrane proteins. Understanding the principles that guide such interactions will illuminate membrane protein folding, stability, and oligomer formation. Since membrane proteins are coded by more than 20% of all genes and are the targets of more than half of all drugs, such an understanding will be both enlightening and useful. We will work towards understanding the motifs and interactions that give rise to transmembrane helix (TM) interactions that stabilize protein structures in biological membranes, and to begin to connect our understanding to biological functions. We have developed a diverse toolkit for the study of TM interactions, including: (a) genetic and biochemical assays in micelles, bilayers and membranes, (b) computational approaches, and (c) structural studies using optical spectroscopy and NMR. Our approach will have several major divisions: a study of TM trimers in HIV gp41 TMs and MHC li;an effort to find the link between helix dimerization and signaling;a study of interaction motifs;an improvement of TM-TM computational methods;and tests of the principles and motifs that can be identified in helix associations.

[unreadable] DESCRIPTION (provided by applicant): Our project is based on the use of a water-soluble membrane peptide, pHLIP, which we have shown, by whole-body fluorescence and PET imaging, to selectively target acidic solid tumors in vivo and to translocate polar cargo molecules into the cytoplasms of cultured cancer cells. pHLIP inserts unidirectionally across the lipid bilayer of a cell membrane as a monomer under mildly acidic conditions, as are found in tumors and forms a transmembrane alpha helix, whereas there is practically no insertion across the membranes of cells with the normal extracellular pH of healthy tissue. To date, no toxic effects of pHLIP exposure have been observed either for cells in culture or for mice. Here we propose to develop a nanotechnology platform for selective delivery of imaging and therapeutic agents to tumors based on the use of the pHLIP-bionanosyringe. By attaching cargo molecules to the end of pHLIP that stays outside of the membrane, we can anchor imaging or therapeutic probes to the surfaces of cancer cells, facilitating diagnosis, treatment and therapeutic monitoring. By attaching cargo to its inserted end via cleavable links, pHLIP can be used for the selective translocation of polar, cell-impermeable molecules into cancer cells. By combining the efforts of three laboratories, a broad development of this promising technology will be possible. We will use pHLIP targeting to test cancer models and establish how tumor growth and development correlate with tumor acidity. To improve pHLIP technology, we will design, synthesize and test various dendrimeric-pHLIP constructs to enable delivery of multiple therapeutic and/or imaging probes to tumors. We will introduce a synthetic scheme of simultaneous conjugation of cargo molecules and fluorescent dyes to the C-terminus of pHLIP via a cleavable S-S bond and establish the properties (polarity, shape, charge and size) of cargo molecules that pHLIP can translocate through the lipid bilayer of a membrane, defining a new, polar class of therapeutic molecules that can be delivered for tumor treatment. We will test pHLIP for the intracellular delivery of two functional cell-impermeable molecules in vivo: a toxin (phalloidin) and a gene regulation agent (Peptide Nucleic Acid). Importantly, we will attempt the simultaneous detection and treatment of tumors by labeled pHLIP-phalloidin, which is our first lead for a potential antimetastatic drug. Further, we will develop a two-step delivery scheme for the specific tethering and assembly of nanoparticles at the surfaces of cancer cells in vivo: 1) targeting tumors using pHLIP with a binding domain, which will be tethered to the surface of cancer cells and 2) targeting the pHLIP with liposomes containing therapeutic and/or imaging payloads and having a surface-exposed complementary binding domain. Inspired by the properties of pHLIP in its current version, we will further evaluate the effect of pHLIP sequence variation on peptide insertion into a membrane, enabling the design of a second generation of the nanosyringe with a range of useful properties. pHLIP nanotechnology offers a new approach for the disease-specific imaging and treatment of cancers. Our ultimate goal is to improve the diagnosis and treatment of cancer, which is responsible for about 25% of all deaths in the USA and other developed countries. There are several aspects of the problem where our technology development could be useful, but the major concept is the selective delivery of therapeutic and imaging agents to cells in tumors. Another aspect of the technology is that it permits the use of a new class of therapeutic agents: cell-impermeable molecules that would be translocated into cells only in diseased tissue while not affecting healthy cells. A therapy based on these concepts would exhibit much higher efficacy and/or significantly reduced side effects. Such improvements are especially important for cancer treatment, since the majority of anti-cancer drugs are poisons that damage normal cells. [unreadable] [unreadable] [unreadable]

DESCRIPTION (provided by applicant): It is a central idea of the NIH that basic research will lead to new approaches in medicine, and we believe that we have found one. As a result of earlier funding of this grant, we have discovered a peptide that (1) targets acidic tissues in vivo, including tumors, (2) can deliver polar molecules into cells, releasing them in the cytoplasm, and (3) gives an opportunity to better understand how peptides can insert across membranes. We now plan to explore both the basic and applied aspects of this discovery. The peptide, which we call pHLIP (for pH (Low) Insertion Peptide) is soluble as an unstructured monomer in aqueous solution, binds as an unstructured monomer to the surface of a bilayer or membrane, and inserts across the bilayer as a trans-membrane helix (TM) when the pH is lowered. We have established the basic energetics and kinetics of peptide insertion. We have shown that a labeled version of pHLIP targets and images tumors as small as 1 mm in mice, and that the imaging accurately identifies tumor borders. We have also established that large, polar cargo molecules (M ~ 1000 Da, log P ~ -2) attached to the inserting end of pHLIP by a disulfide are delivered across membranes and released in the cytoplasms of cultured tumor cells at low pH. By continuing our basic research we hope to frame the technology for use in the clinic. We will study the process of and sequence requirements for insertion of water soluble peptides into membranes, find improved ways to target tumors and other acidic tissues, and develop expanded ways to deliver polar molecules into cells, releasing them into the cytoplasm by disulfide or ester cleavage. Targeting imaging agents to tumors with pHLIP could aid in diagnosis or act as a guide for surgery, and delivering therapeutics could assist in treatment. Using biophysical, biochemical, and biological approaches, we will seek answers to the following questions: 1. What are the kinetic intermediates, energetics and structures of the bilayer and peptide during insertion? 2. Which sequence features allow a water-soluble peptide to insert spontaneously to form a TM? 3. What role(s) do lipids play in TM insertion? 4. Can pHLIP be used to image cargo delivery in vivo? 5. What is the range of polar molecules that can be delivered to cells?