Fred Cohen - US grants

The amino acid sequence of a protein uniquely determines its tertiary structure. Deciphering this relationship, the protein folding problem, has become increasingly important to molecular biologists. DNA sequencing has become routine, but structural experiments remain difficult. Two semi-empirical approaches to the folding problem have evolved: detailed energy calculations and the hierarchic condensation model. This grant focuses on the hierarchic condensation model which presumes that secondary structure is a useful computational intermediate in protein folding. We propose to study known structures. Initial efforts will be directed at deducing the rules which govern alpha-helix and beta-strand formation. Secondary structure prediction will follow an "expert systems" approach which has been successfully applied to locating turns in a variety of proteins (Cohen et al (1986) Biochemistry 25, 266-275). This algorithm will be combined with procedures for exploring possible tertiary structure through the packing of secondary structure units (e.g. Cohen et al (1979) J. Molec. Biol. 132, 275-288). Methods to sort amongst the alternative structures will be developed. These methods hopefully will evolve from the study of the spatial, electrostatic and catalytic properties of proteins of known structure. Finally, these mathematical models will be applied to a variety of biologically interesting macromolecules with known amino acid sequence but unknown tertiary structure. Collaborative ventures will be established to experimentally test the predicted structures for Interleukin-1 (IL1), Interleukin-2 (IL2), and Parathyroid Hormone (PTH). Fragments of the IL1 sequence which correspond to predicted loop regions will be designed. Dr. Dinarello at Tufts will investigate the immunologic relevance and potential antagonist activity of these fragments. With Drs. Ciardelli and Smith at Dartmouth, the structure of IL2 will be examined through site directed mutagenesis. The theoretical effect of a mutation on the protein's stability and activity will be modelled and tested by circular dichroism spectroscopy and receptor binding assays. With Drs. Kuntz, Stewler and Arnaud at UCSF, the structure of PTH and chemical variants of PTH will be characterized. It is hoped that structural modelling will lead to the design of pharmacologically relevant mutant proteins.

The amino acid sequence of a protein uniquely determines its tertiary structure. Deciphering this relationship, the protein folding problem, has become increasingly important to molecular biologists. DNA sequencing has become routine, but structural experiments remain very difficult. Computational strategies are needed to help address this problem. This proposal describes a strategy to identify the location of alpha-helices and beta-strands throughout the sequence. A rationale is offered for employing neural networks and pattern based algorithms to address the secondary structure prediction problem. Once secondary structure is located, computational methods exist for generating plausible tertiary structures. However, these combinatorial strategies give rise to a large number of alternative structures which are difficult to distinguish from the correct fold. Simplified potential functions are proposed as a method for overcoming this structure evaluation problem. The properties of a non-lattice based simplified representation of a polypeptide chain will be explored to aid in the construction of an appropriate simplified potential function. Collaborative ventures are planned to experimentally test the merits of existing algorithms for predicting protein structure. In collaboration with Dr. Bunn at Harvard, the relationship of the erythropoietin sequence to its structure and function will be explored. In collaboration with Dr. Wang at UCSF, the merits of a proposed structure of hypoxanthine guanine phosphoribosyl transferase will be studied using site directed mutagenesis. An exploration of the possibility of grafting the active site of one enzyme onto the structural scaffold provided by another protein will be studied in collaboration with Dr. Craik at UCSF and Dr. Wells at Genentech.

The award will support the development of a new mathematical model for intrapolymer diffusion and will apply it to kinetics of protein folding. Protein folding is one of the important problems in computational structural biology. Understanding how a protein folds, and what the final state (conformation) a protein will have has tremendous implications for many areas in biology and medicine. In addition, this award will help support the retraining of a mathematician, David Anick, to work at the interface between biology and mathematics. This award is being co-funded with the Biochemistry and Molecular Structure and Function program.

9317845 O'Shea This proposal seeks funding for circular dicroism (CD) and UV VIS spectrophotometers. These instruments will be used in a series of biochemical and biophysical studies concerned with how proteins fold and how they interact with one another and with other macromolecules. The proposed research includes studies in the following broad research areas: (1) protein structure, structure prediction and protein folding, (2) protein protein and protein nucleic acid interactions. (1) Studies on protein structure, prediction and folding will involve both model peptides and a bacterial protease, a lytic protease. Computational predictions of protein structure will be tested using synthetic polypeptides. The structure of designed peptides will be studied to investigate fundamental aspects of protein stability and structure. Structural studies of a stable intermediate in the folding of a lytic protease will be carried out to understand the forces which stabilize this intermediate. Conversion of the intermediate to the native enzyme is catalyzed by an additional polypeptide sequence. Thus, a further goal of these studies is to understand the mechanism by which folding to the native state is catalyzed. (2) Studies on protein protein interactions will involve a group of proteins that play important roles in transcriptional regulation. These include the glucocorticoid receptor from mammalian cells and the SYMBOL 97 \f "Symbol" 2, al, MCM1, PHO2, PHO4, and PHO80 proteins from yeast. A goal of these studies is to identify domains that comprise autonomous folding units, as well as domains that are responsible for inte ractions between proteins of multisubunit transcription complexes. An additional goal is to investigate changes in conformation of these molecules upon interaction with one another. These studies will provide an essential first step in obtaining three dimensional information on multi subunit protein complexes. The CD and UV VIS spectrophotometers are essential instruments for carrying out these studies. CD provides a simple and informative measure of secondary structure. It also provides a way to assess conformational changes in mixtures of macromolecules, such as two different proteins (or domains) or a protein and DNA. The UV VIS spectrophotometer provides a sensitive assay for measuring protein concentrations, a necessary component of a CD experiment. Additionally, UV VIS spectroscopy provides a method for measuring the stability of a molecule by monitoring absorbance. For proteins, the information derived from such a study is often complementary to that obtained from CD because changes in the UV spectrum are typically reflections of tertiary structure while CD is most sensitive to secondary structure. Finally, UV VIS spectroscopy provides a very sensitive way to monitor changes in nucleic acid stability and structure. These studies shall be carried out as part of ongoing, active research programs in which training of graduate students and postdoctoral fellows is a strong component. u ~ 9317845 O'Shea This proposal seeks funding for circular dicroism (CD) and UV VIS spectrophotometers. These instruments will be . / 0 x k m I K ' ) v x ! ! ! ! ! ! F x N x ; CG Times Symbol & Arial 1 Courier 0 MS LineDraw 9 N h = ABSTRACT Deseree King, BIR Deseree King, BIR

We hypothesize that prion diseases are a result of changes in the conformation and oligomerization state of PrPc, such that PrPc PrPsc Oligomer. The physical-chemical properties of PrPc have made it difficult to conduct experimental biophysical studies to determine the structure of this protein. We propose a series of theoretical studies to model the structure of the possible conformers of PrP. New methods for secondary structure prediction will be applied to the family of known PrP sequences. Heuristic models for secondary structure packing will be used to propose plausible tertiary structures. Peptides have been designed based on the regions of likely secondary structure in PrP. Recent experiments by our colleagues in this Program Project Grants suggest that these peptides can adopt an a-helical structure under some conditions and a beta-structure under other conditions. In fact, some of these peptides are capable of forming amyloid. We plan to use molecular dynamics calculations to stimulate the conformational behavior of these peptides. Simulations will be carried out in water and a variety of mixed solvents that could mimic the membrane environment. The results of these studies should suggest mutations that could effect the conformational preference of the peptide and perhaps the conformational preferences of the intact protein. Models of peptide aggregation will be developed from molecular dynamics simulations of two or more peptides. Again, a close correlation with experimental results from the Prusiner group will be critical to the progress of the computational effort. From the results of the peptide simulation, it may be possible to develop a useful model of the intact PrP molecule. We anticipate that alternative models will be developed and mutagenesis experiments will be designed to help sort between alternative models will be developed and mutagenesis experiments will be designed to help sort between alternative models. Insertion and deletion mutations as well as the introduction of new disulfide bridges could provide structural information that would confirm or refute the alternative model structures.

prions; chimeric proteins; protein structure function; physical model; computer simulation; protein sequence; scrapie; conformation; model design /development;

proteins; nervous system; biomedical resource; biomedical equipment development; biological products;

technology /technique development; proteins; biomedical resource; biological products;

The amino acid sequence of a protein uniquely determines its tertiary structure. Deciphering this relationship, the protein folding problem has become increasingly important to molecular biologists. DNA sequencing has become routine, but structural experiments remain very difficult. Computational strategies are needed to help address this problem. This proposal describes a strategy to identify the location of alpha- helices and beta-strands throughout the sequence. A method for using off- lattice simulations of a polypeptide chain to identify secondary structure preferences in the ensemble average is proposed. Once secondary structure is located, computational methods exist for generating plausible tertiary structures. However, these combinatorial strategies give rise to a large number of alternative structures which are difficult to distinguish from the correct fold. Experimental and theoretical methods for clarifying the distinction between correctly folded structures and their misfolded counterparts will be considered. In a new direction, we propose to develop a multiple sequence analysis strategy to relate sequence and structure to function. In particular, we will focus on identifying the binding sites on the G-alpha family of GTPases for the relevant G-protein coupled receptors, G-beta-gamma and downstream effectors. We plan to continue to develop a genetic algorithm for the construction of polypeptide loops subject to a series of constraints. This method will be used to model the loop regions of G- protein coupled receptors involved in the interaction with peptide ligands and the hetero-trimeric G-protein complex. Finally, we propose to develop a new method to compare structures based on the area of the minimal "soap film" that could join them following the appropriate rotation and translation of one structure relative to another. This provides a natural way to circumvent the gap penalty problem that plagues current structure alignment algorithms.

A program for the development of new agents to treat T. Cruzi, the parasite responsible for Chagas' Disease, is presented. Cruzain, a cysteine protease is identified as a target for a structure based drug discovery program. Models of the three-dimensional structure of cruzain built by homology to other cysteine proteases of known structure and the recently determined x-ray structure of cruzain will be used as a template for the in compuo screening of the fine chemicals directory, a data base of about 70,000 purchaseable small molecules. Compounds selected in this computer based screen will be characterized experimentally. We expect that 2-10% of the computer selected inhibitors will be active experimentally against the enzyme at concentrations less than 100 microM. Already, we have identified cruzain inhibitors active at 2-4 microM concentrations. Free energy perturbation calculations will be used to help the synthetic chemists develop more active compounds. Some will be based on the non-peptidic leads derived from the computer data base searches, while others will be analogs of the nanomolar mechanistic inhibitor Z-Phe-Arg-FMK (Z= carbobenzoxy, FMK = fluoromethyl ketone) that have been altered to eliminate peptide character. A close interaction between molecular biology, structural biology, synthetic chemistry and computational chemistry will be used to exploit a structure based drug design cycle.

Perhaps the most unorthodox feature of prion disease is the co-existence of infectious, genetic and sporadic forms of this class of neurodegenerative diseases. With the recent occurrence of new variant Crretzfeldt-Jakob disease in the United Kingdom and the link to Mad Cow disease, it is important to understand the molecular basis of this class of diseases and begin to develop therapeutic agents. The tools of structure based drug design will be used to identify plausible inhibitors of key step(s) in prion replication from an analysis of the structure of PrPc. Preliminary work has identified a collection of leads studied for their ability to terminate multi-merization mediated by intermolecular beta- structure. Finally, peptides that are capable of inducing prion neuropathology in a conformation dependent fashion will be studied to better understand the molecular requirements for prion activity. These efforts should yield improved assays for inhibitors of prion replication. A close collaboration between biologists, synthetic chemists, computational biophysicists and biochemists, computational biophysicists and biochemists is proposed to achieve these goals. Success in the development of therapeutic strategies should have implications for the treatment of other neurodegenerative diseases that involve protein aggregation.

The insolubility of the scrapie prion protein (PrPSc) has frustrated all attempts to solve its structure by X-ray crystallography or NMR spectroscopy. Recently, we reported the discovery of two-dimensional (2D) crystals of the N-terminally truncated PrPSc (PrP 27-30) and a redacted miniprion (PrPSc106). Analyzing the 2D crystals by electron crystallography allowed us to map the differences between PrP 27-30 and PrpSc106. These data were used to constrain structural models of PrPSc. We propose to investigate the parameters that govern growth of the 2D crystals in order to obtain specimens suitable for low dose, cryo-electron crystallography. These crystals will be used to collect higher resolution data to generate a three-dimensional reconstruction of the structure of PrPSc. We will also pursue various labeling techniques to localize different parts of the molecule in order to position computational models of PrPSc on the crystal lattice. The experimental data will be used to refine structural models of PrPSc that will be based on structures of known proteins or domains of proteins. In particular, we will focus on the parallel Beta-helix as a motif that can account for the secondary structure constraints implied by FTIR spectroscopy data and spatial constraints determined by electron microscopy. While fiber diffraction results have historically contributed only low-resolution structural data to our analysis of the structure of PrPSc, we believe that when coupled with the models and electron crystallography data, we will be able to extract more meaningful information out of these studies. Therefore, we want to revisit fiber diffraction studies using more advanced methods of fiber alignment coupled with synchrotron-based X-ray and electron diffraction approaches. Furthermore, we want to explore which of the well-established prion strains are most suitable for structural analysis by either electron crystallography or by fiber diffraction. We are also interested in studying PrP sequence truncations that can support PrPSc formation in transgenic animals. This work will include PrP 89-231, a homolog of PrP 27 - 30, and variants of PrP106 that extend or shrink the internal deletion from 141 - 175 yet still support PrPSc formation.

The in vivo conversion of the cellular prion protein (PrPC) into an alternate conformer, the infectious prion (PrPSc), is the etiologic event in prion replication and pathology. Until recently, in vitro conversion protocols have failed to yield infectious prions from recombinant sources of wild-type PrP or segmental deletion variants of PrP. Recently, we have shown that in vitro converted, recombinant wild-type mouse PrP(89-230) can cause a historically accurate neurodegenerative disease that is serially transmissible. This observation creates several opportunities for advancing biophysical studies of prions. In vitro prions show unusual strain characteristics, and by inference structural properties, that are maintained when passaged in wild-type or transgenic mice. We propose to study the structure of these synthetic prions by fiber diffraction, electron crystallography, and molecular modeling. Currently, the in vitro conversion is still fairly inefficient. For this reason, we will focus our initial investigations on synthetic prions that have been passaged in mice. Once the efficiency of the in vitro conversion process is improved, we plan to carry out additional experiments to characterize synthetic prions directly from recombinant sources. Our initial experiments are designed to analyze synthetic prions that form amyloid fibers and two-dimensional (2D) crystals and to compare these results with those for naturally occurring prions. The experimental data will be used to refine structural models of PrPSc that are based on structures of known proteins or domains of proteins. In particular, we will focus on the left-handed parallel beta-helix as a motif that can account for the secondary structure constraints implied by optical spectroscopy data and spatial constraints determined by electron crystallography. Historically, fiber diffraction results have contributed only low-resolution data to our analysis of the structure of PrPSc. However, we believe that when coupled with modeling and electron crystallography data, we will be able to extract more meaningful information from these studies. Therefore, we plan to revisit fiber diffraction studies using more advanced methods of fiber alignment coupled with synchrotron-based X-ray and electron diffraction approaches.

The prion diseases are characterized by aggregation of a mis-folded isoform of the prion protein, termed PrP (Sc). No treatment is available to halt progression of prion diseases, and as such, clinically viable therapies are being sought. Potent pyridine-based compounds have been designed to mimic dominant negative PrP and hence bind to auxiliary molecules involved in the conformational conversion of PrP(C) to PrP (Sc). These compounds have been shown to inhibit PrP (Sc) replication and have been analogued using SAR techniques. While the EC(50)s of the original lead compounds were in the 15-30 micromolar range, the most potent new analogues are active at 300 nanomolar concentrations. Independently, we have identified the acridine-based compound, quinacrine as a potent anti-prion agent in a cell-based assay of PrP(Sc) replication (EC(50)=300 nanomolar), and developed bis-quinacrine analogs that are even more potent (EC(50)=30 nanomolar). Given the improved activity of the bis- analogs, subsequent research efforts will focus on developing this class of compound. Refinement of bioactive pyridine- and bis-acridine-based analogs using synthetic libraries and computational methods could furnish potent non-toxic heterocyclic compounds that could be viable candidates for the treatment of prion disease. The cellular and molecular mechanism of action of the pyridine- and acridine-based compounds is being examined to better understand how these compounds reduce PrP(Sc) replication in a cell-based model of prion disease. The trafficking of PrP(C) and PrP(Sc) between cellular compartments will be determined using a neuronal imaging system employing confocal microscopy. Subsequently, fluorophore labeled pyridine- and acridine-based analogs will be introduced to the imaging system to determine if these compounds exert their effect on PrP(Sc) replication by interfering with the trafficking of the prion protein. Photoactivatable labeled analogs will be synthesized for receptor cross-linking studies. These probes will be used to identify the molecular targets of the heterocyclic-based compounds. Once the molecular targets have been isolated and characterized, the structural and functional information pertaining to the targets will be used for the structure-based design of potent and selective compounds, active against PrP(Sc) replication.