Research
The Stroud Lab is researching a variety of topics to help understand molecular mechanisms of certain key biological processes. To find out more about what specifically we are seeking to understand, please use the pull-down menu below or browse by scrolling down. Topics are listed from most current.
Ammonia Channel | Aquaporins | Signal Recognition Particle | Glycerol Channel | Polyketide Synthase | RNA Processing Enzymes | HIV Integrase | HIV Protease | KSHV Protease | Pro-Chymase Protease | Erythropoietin Receptor | Thymidylate Synthase | Isocitrate Dehydrogenase | Bacteriorhodopsin | Cholesterol Esterase | Colicin | Trypsin | Acetylcholine Receptor | Drug Design
Ammonia Channel
Aquaporins GlpF/ AQP0/ AQPM/ AQPZ
Signal Recognition Particle (SRP)
The ubiquitous cytoplasmic ribonucleoprotein signal recognition particle (SRP) mediates targeting of nascent polypeptides to the translocation apparatus in the endoplasmic reticulum membrane of eukaryotes and the inner membrane of bacteria. In eukaryotes, this particle consists of six polypeptides complexed with a 7S RNA with the 54 kD subunit of the SRP (SRP54) involved in recognition and binding of a signal sequence on a nascent polypeptide emerging from the ribosome. This interaction produces a cytosolic complex comprised of the nascent polypeptide chain, the translating ribosome and the SRP. The complex is then targeted to the membrane through an interaction with the SRP receptor in a GTP-dependent manner. Hydrolysis of GTP causes the release of the SRP and the insertion of the nascent chain into the membrane through the translocon pore.
Homologues of SRP54, SRP RNA, and the a subunit of the SRP receptor have been identified and isolated in E. coli and other bacteria. In E. coli, the homologues Ffh, 4.5S RNA, and FtsY, respectively, have been shown to function in a parallel pathway to eukaryote SRP. The 37 kD 4.5S RNA and the 47 kD Ffh form a complex that interacts with the 33 kD receptor FtsY in a manner that reciprocally stimulates the GTPase activity of both the Ffh and FtsY. The 4.5S RNA although much smaller than the eukaryote SRP RNA, stabilizes the interaction between these enzymes.
The mechanism of protein translocation across the endoplasmic reticulum membrane of eukaryotic cells and the plasma membrane of prokaryotic cells appear to be evolutionarily related. We seek to understand the mechanism by which SRP54, SRP RNA and the SRP receptor coordinate binding and release of signal sequences during targeting of the nascent chain to the translocation apparatus of the cell.
Glycerol Channel
Cell membranes contain membrane channel proteins that act as molecular filters. Humans have ten different variants of this channel protein family. They are located in the kidneys, heart, liver, brain, tear and saliva glands and just about every cell in the body that needs to be permeable for water or glycerol. For example, the cell membranes of fat cells contain glycerol channels that pass through the sweet-tasting nutrient molecule glycerol. This occurs when fat molecules are broken down into their components fatty acids and glycerol, fuels for our bodies. Since the cell membrane is relatively glycerol tight these specialized channels are employed because they can facilitate the passage of glycerol molecules from the cell interior into the blood stream. These channels allow the passive but selective passage of small molecules such as water or glycerol, but they block the passage of large or charged molecules such as amino acids.
Studying this filtration process on the molecular level in humans is very difficult, but bacterial cells offer a similar and available alternative. Bacterial E.coli cells utilize two types of such molecular filters. One channel is specific for the permeation of water the other channel conducts both, glycerol and water. Since this glycerol channel is very similar to the glycerol channels in human cells and since it is easy to produce large and pure quantities of it, my research is directed towards the glycerol channel from E.coli. I hope that by unraveling its structure-function relationship the closely related human channel proteins can also be understood.
Determining how glycerol channels work and what they look like is difficult since the channels are too small (< 10 nm) to see with a microscope. Therefore I am performing what are referred to as "crystallographic experiments" which allow the atomic structure of this channel to be seen. For such experiments, protein crystals that look just like salt crystals are grown. Exposing the crystals to X-ray radiation results in diffraction patterns and by computer-based analysis of them a model of the channel protein can be generated. The model describes the precise positions of most of the atoms of the protein, the bound water and the glycerol. Such atomic models are the basis of figuring out how glycerol and water pass through the channel.
In a team effort, we have recently investigated the function and determined the atomic structure of the bacterial E.coli glycerol channel. As expected, the protein contains a narrow tunnel, demonstrating that the glycerol channel works like a sieve. It allows small molecules to fit through and blocks larger molecules. We also found a chain of three glycerol molecules that bind at three different sites inside the channel. We consider this an overlay of multiple snapshots, resembling a glycerol molecule hopping from one site to another during its passage through the channel. The intermediate binding and unbinding events allow the filter to distinguish different types of small molecules, i.e. to reject potentially toxic molecules and let pass nutrient molecules. Consequently, the details of interaction of the channel lining atoms with those of the glycerol and water molecules are crucial to understanding how the channel works. Specifically, my research goal is to understand the route a glycerol molecule takes through the channel. This requires determining where exactly the glycerol and water molecules are located and how strong they hold on to the binding sites in the channel. Researchers in the field are puzzled that these channels conduct water molecules (H2O) but not their smaller components, i.e. the hydrogen ions (H+). We suspect that several water molecules line up in the channel to keep it from collapsing. But such a chain of water molecules is known to pass along hydrogen ions. Therefore, the precise arrangement of the water molecules inside the channel that IÍm about to determine will be particularly interesting. Ultimately, I want to understand the function of the channel by identifying its minute structural changes that occur as either glycerol or water molecules pass though the channel.
The results of these studies will contribute to the general understanding of how small molecules pass through protein channels. Detailed information will likely not be available for many of the human versions of this channel protein. For example, the recently discovered glycerol channel that is located in the cell membrane of human fat tissue cells plays an important role in the release and uptake of glycerol from the bloodstream. Since glycerol production in fat tissue is increased in cases of obesity, affecting this channel may relieve the harmful effects of the associated hyperglycemia. Likewise, several diseases of the kidney, the eye, and the lungs are related to defective channel proteins of the same family. By close analogy to the bacterial glycerol channel structure and function these channel proteins may be understood better. I hope that this will ultimately form the basis for developing rational treatments.
Polyketide Synthase: DEBS Thioesterase
RNA Processing Enzymes
Written by Sanjay Agarwalla
Stable RNAs contain post-transcriptionally modified nucleotides which are derivatives of the four common nucleotides. Most rRNA modifications are located near the functionally important regions of ribosome and thus speculated to have an essential function. Likewise, several of the t-RNA modifications are highly conserved. The enzymes that carry out modifications on RNA are often position specific, and approximately 1% of the bacterial genome is devoted to encoding RNA modification enzymes. We are pursuing a structural genomics approach, which focuses on determining the structures of RNA pyrimidine methyltransferases and pseudouridine synthases. We are in the process of solving structures for several of these enzymes in apo-form and also in complex with RNA. These structures will provide important information on reaction mechanism as well as RNA-protein interaction. In addition, we are carrying out biochemical studies for identifying the nature and the site of modifications and their physiological role, which has been poorly understood.
HIV Integrase
HIV integrase is a viral enzyme responsible for ligating a DNA copy of the viral genome into the host cell's genome. Integration is essential for viral replication, and therefore is an attractive therapeutic target. The DOCK program is being used to search for small molecule inhibitors of HIV integrase. Inhibitors are tested using a number of HIV integrase activity assays, and are also used for co-crystallization experiments.
HIV Protease
KSHV Protease
Pro-Chymase Protease
Human chymase is a protease involved in physiological processes ranging from inflammation to hypertension. As are all proteases of the trypsin fold, chymase is synthesized as an inactive zymogen with an N-terminal pro-region that prevents the transition of the zymogen to an activated conformation. The 1.8 Å structure of pro-chymase is the first zymogen with a dipeptide pro-region (Glycine-Glutamate) to be characterized at atomic resolution. Three segments of the pro-chymase structure differ from that of the activated enzyme: the N-terminus (Gly14-Gly19), the autolysis loop (Gly142-Thr154), and the 180s loop (Pro185A-Asp194). The four N-terminal residues (Gly14-Glu15-Ile16-Ile17) are disordered. The autolysis loop occupies a position up to 10Å closer to the active site than is seen in the activated enzyme, thereby forming a hydrogen bond with the catalytic residue Ser195 and occluding the S1’ binding pocket. Nevertheless, the catalytic triad (Asp102-His57-Ser195) is arrayed in a geometry close to that seen in activated chymase (all atom rmsd of 0.52Å). The 180s loop of pro-chymase is, on average, 4 Å removed from its conformation in the activated enzyme. This conformation disconnects the oxyanion hole (the amides of Gly193 and Ser195) from the active site and positions only ~35% of the S1-S3 binding pockets in the active conformation. The backbone of residue Asp194 is rotated 1801⁄4 when compared to its conformation in the activated enzyme, allowing a hydrogen bond between the main chain amide of residue Trp141 and the carboxyl of Asp194. The side chains of residues Phe191 and Lys192 of pro-chymase fill the Ile16 binding pocket and the base of the S1 binding pocket, respectively. The zymogen positioning of both the 180s and autolysis loops are synergistic structural elements that appear to prevent premature proteolysis by chymase and, quite possibly, by other dipeptide zymogens.
Erythropoietin Receptor
Structure & Recognition: Structure-Based Drug Design of Thymidylate Synthase
TMP synthetase is the target for fluorouracyl, a major clinically used anti-cancer drug. Since Janet Finer-Moore determined the structure of TMP sythetase from L. casei in our laboratory, the structure of TS from human, E. coli, and T4-phage have been determined; those from Leishmania major, pneumocystis, and p. falciparum are to be determined. Since TS provides the sole de novo synthesis for dTMP from dUMP, it is essential for DNA synthesis and for life. As such, it is a key target for drug design of either parasite-specific drugs, or anti-cancer drugs.
Drug design based on matching structure and electronic environment provides the quintessential test of how well one understands molecular interactions; there is no better assay than one where the results are of interest to therapy. Elusive chemistry of methyl group transfer has been subject to intensive investigation. Based on structures of enzyme-substrate-drug complexes, we propose a new mechanism. Mutants are now readily made and TS expressed in collaboration with Dan Santi's group. Of particular advantage is the selection prqduced by cloning in thy-minus cells tbat lack the TS gene, and so only live if fed on uracyl, or if the mutant TS is at least 1.0% active.
Eric Fauman and Earl Rutenber ask how structure of mutants affect function at the structural level for mutants containing replacements of key residues in the active site. Comparison of substrate plus cofactor-bound and unbound TS reveals a concerted closure of the active site which sequesters the reactants within a large pocket. Understanding this transition in conformation offers a rich avenue for design of new anti-cancer drugs that prevent this essential conformation change.
Comparison of (currently) five species of TS by Janet Finer-Moore, Celia Schiffer, Eric Fauman, and Kathy Perry shows how proteins accommodate changes in sequence while retaining the same three-dimensional structure. The structures and properties of the human enzyme and TS from certain parasites, including malaria, and from viruses, such as herpes virus, provide avenues to clinically important drug design. TS is a symmetric dimer of identical monomers. The interface consists primarily of sidechains contributed by beta sheets. Conditions for the reversible in vitro refolding of TS have been determined, and Kathy Perry ingeniously uses this to investigate the influence of point mutations in the interface, on the overall stability and folding of the dimer; with noncovalent interactions there is the unique advantage that one can measure the change in dissociation of the assembly, something quite impos- sible with monomeric proteins. Thus, it is the ideal system for studying the folding and stability of domains, unfettered by covalent forces. Kinetics of refolding reveal that dimerization occurs late in the folding process. Thus, a goal is to understand how protein subunits assemble to form multimeric pro- teins inside the cell. Paul Foster has purified the tRNA methylating enzyme that specifically methylates U54 in t-RNAs for structure analysis by crystallography, aimed at understanding tRNA recognition.
Regulation of Enzyme Action: Isocitrate Dehyrdrogenase
How Does Protein Phosphorylation Regulate Activity?
To understand how phosphorylation—a ubiquitous means for biological signaling and control in the cell affects protein activity, we analyzed the structural consequences of phosphorylation. We show, using site-directed mutagenesis and electrostatic calculation, that regulation by phosphorylation is due to a direct electrostatic effect. A recently crystallized complex of protein kinase, an ATP analog, and the target protein are beginning to define the specificity of phosphorylation networking inside the cell.
Isocitrate dehydrogenase (IDH) from E. coli is completely inactivated by phosphorylation at serine 113, which inhibits binding of the substrate isoci- trate. To understand the principles underlying this important regulatory mechanism, the structures of phosphorylated and dephosphorylated IDH were determined at atomic (2.5 A) resolution. The structure of the enzyme-substrate complex and structures of site-directed mutants, in which the phosphorylatable serine has been replaced by an aspartate or glutamate, have also been solved and refined to R < 0.18. This implies a new mechanism for control of protein activity by phosphorylation, in which covalently bound phosphate interferes with ligand binding through direct electrostatic interaction.
Structures of S113D and S113E, which place negative charge at the site of phosphorylation, are inactive, whereas S113N, S113Q, and S113Y, which are active, and electrostatic potential calculations using the Poisson-Boltzmann equation, confirm that electrostatic interactions provide direct control. The coordination of magnesium by the a- carboxylate and a-hydroxyl moieties of isocitrate in the enzyme substrate complex clarify the critical role of the Mg++ ion in activating both steps in the reaction catalyzed by IDH. AminQ acid residues 160, 230, and 283 are implicated in cataysis. Site-directed mutagenesis, carried out with Dr. Daniel Koshland and his group at University of California, Berkeley, coupled with structure analysis, will test these roles. We now seek to understand the interaction with the specific protein kinase/phosphatase.
The complex formed by ICDH and its kinase, with a nonhydrolyzable analog of ATP, ATPS, has been crystallized by Diana Cherbavaz. This opens the way to understanding how specificity within the cell is generated, and so, how networking within the ceu, mediated by phophorylation, takes place. This is the first, best opportunity to look directly at a protein- kinase interaction structurally.
Bacteriorhodopsin
Cholesterol Esterase
Colicin
Trypsin
Cellular Signaling: Acetylcholine Receptor
Transduction in the Nervous System
The nervous system is rich in dialogue between cells involved in processes of thought, memory, and movement. We focus on two key examples of neuro-active cell surface proteins transistors of the nervous system as a means of understanding mechanisms of cellular signaling: potassium channels, and the acetylcholine receptor.
Several of the voltage-gated human potassium channels have been cloned and expressed with one goal being to crystallize one of these proteins for three-dimensional molecular structure determination. The channel, detected by recording of ionic currents through single channels, is formed from four copies of the same subunit in symmetric rosette arrangement. This exciting avenue has impact both on fundamentals of neurochemistry, and on drug design aimed at anti-arhythmic.
The nicotinic acetylcholine receptor, the best understood of any cell-surface receptor, is a five subunit complex, composed of four different, but evolutionarily related, chains in the stoichiometry. We showed in the most detailed structural analysis of any eukaryotic cell-surface receptor that the complex of five transmembrane subunits surrounds a water-filled ion-conducting channel. The acetylcholine receptor (AChR) responds to acetylcholine, a small neurotransmitter, by opening a conducting channel, during which time 10,000 sodium ions per millisecond pass through the channel into the target cell. The resulting electrical depolarization translates the chemical signal released by the nerve cell into the electrical language of the target cell.
How is the conducting region constructed and regulated‹since it is a water-filled channel, across a hydrophobic lipid bilayer? How is the closed channel, insulating a poten- tial field of 100,000 volts per centimeter, opened by binding a small neurotransmitter? We showed that the effector-binding site lies at the top crest of the 110 A thick, funnel-shaped complex. Fully 60 A away lies the entrance to the 40 A long transmembrane-conducting channel. Using an ion that displaces calcium competitively from the receptor complex, we showed that there are ion-binding sites within the channel, and that the channel is effectively opened throughout the remainder of its length even in the closed state. Our hypothesis is that the gate is composed of tight ion-binding sites rather than swinging residues of protein. We originated a scheme to map the topography of the chains unambiguously using chemical labels, and so determined the structures of all bound oligosaccharides, intrinsic sidespecific labels, using mass spectrometry. Collaborating with Al Burlingame's laboratory, this powerful new approach that we developed can determine topography of any membrane protein.
We have crystallized the five subunit, 300,000 Dalton AChR. The crystals are cubic with unit cell size a = 257 A, in the space group I23, with 24 com- plexes per unit cell. This offers the best prospect of any for determining the three-dimensional structure of any transmembrane ion channel. Improved crystal growth so far depends on understanding the physical chemistry of protein, lipid, and detergent interactions, as defined by Michael Shuster. The crystals provide a means of obtaining the highest resolution structural detail so far for the AChR. A 9 A three-dimensional diffraction has been recorded. We hope to exploit the low-radiation damage induced in proteins frozen in vitreous ice, at liquid nitrogen temperature. Under these conditions, protein crystals last almost indefi- nitely. Our long-term aim is a highresolution structure for the AChR, a great improvement on what has been attained by electron microscopic image analysis. How do the five subunits correctly come together post-translationally? The process involves only one minute for translation and glycosylation, but three hours for functional assembly. We seek to determine the processes involved in maturation of this model membrane protein.
Structure at 2.5A of a Designed Peptide that Maintains Solubility of Membrane Proteins
A 24-amino acid peptide designed to solubilize integral membrane proteins has been synthesized by Chris Schafmeister. The design was for an amphipathic alpha helix with a flat hydrophobic surface that would interact with a transmembrane protein as a detergent. When mixed with peptide, 85% of bacteriorhodopsin and 60% of rhodopsin remained in solution over a period of two days in their native forms. The crystal structure of peptide alone showed it to form an antiparallel four- helix bundle in which monomers interact, flat surface to flat surface, as predicted.
Transmembrane Channel, Function, and Structure
Colicin Ia, a 70,000 Dalton bactericidal protein, forms voltage-dependent ion channels. It is synthesized and released in soluble form and kills by receptormediated attachment, forming transmembrane ion channels. By determining the structures of the soluble and transmembrane forms of the molecule, we seek to unravel the mechanisms of membrane insertion and ion-channel formation at the molecular level. To this end, Partho Ghosh and Michael Wiener have trekked to Paris, Ithaca, and Palo Alto to employ the unique advantages of synchrotron X-ray sources, and so obtained crystallographic data for colicin Ia to 2.3A, enough to see how even the amino acid sidechains are disposed. Michael has solved several heavy metal containing derivatives necessary to phase the patterns, and has now built over 70% of the molecular structure. This opens the way to understanding the process of insertion into membranes.
To define the internal structure and the structural changes upon insertion of colicin Ia into membranes, Stephanie Mel noticed that the circular dichroism spectra in both soluble form, and in detergent, which mimics the membrane environment, were only very slightly different for such obviously different functional states! Colicin is approximately 60-65% alphahelix, in both buffer and detergent. This suggests that colicin does not undergo any change in the amount of secondary structure upon insertion into membranes, but alternatively rearranges its helices, turning inside.
Natural Insecticide Tests Mechanisms of Membrane Pore Formation
Strains of Bacillus thuringiensis contain transmembrane pore-forming toxins specific to killing certain insects, that are naturally crystalline (of note to a crystallographer!) within the bacilli. The carcass of the target insect is used for growth of the bacilli, which form the lethal crystalline toxin during sporu- lation. Working on the strain Bacillus thuringiensis israeliensis (Bti), the 27,000 Dalton protein, the most toxic part of the complex, has been purified, recrystallized, and used for structure analysis to high resolution. An oligomeric complex assembles to form transmembrane channels, like the complex of C9 proteins formed in complement lysis of cells.
The structure will provide the essential first step in understand- ing target specificity for the specific insect host, and for the membrane pore-forming function of Bti. V. Ramalingam is trying to show how these processes work from x-ray data recorded to 2.8 A resolution from the recrystallized protein and heavy metal derivatives. Protein engineering of the toxin using three-dimensional structure will help define a transmembrane ion channel, and not insignificantly lead to an environmentally safe insecticide. In addition, V. Ramalingam has also crystallized, and solved, the structure of one of the insect- specific strains of Bt. This will open the way to determining how natural insecticides can be engineered, and how membrane insertion takes place.
Drug Design
Proteins of HIV and SIV
We apply our drug design methods to structures of three of the seven proteins encoded by the viral genome of HIV/SIV. From these, we seek to design and iteratively improve drugs that bind to and alter those activities. Specific drug target proteins include HIV and SIV protease. Studies with SIV are because the monkey virus SIV permits an animal model for testing of designed drugs. We design improved affinity and properties of new inhibitors for use as drugs. The compounds will be pursued through toxicology in animal trials. This extremely exciting avenue in the most fundamental aspect of basic science is of potentially enormous value to humanity. Bob Rose has crystallized SIV protease with a covalent inhibitor, EPNP (1.2-epoxy-(p- nitrophenoxy) propane). The integration enzyme for the reverse tran- script, a superb anti-HIV target, is especially exciting as it is an activity not found in uninfected human cells. A specific inhibitor for it will prevent replica- tion of HIV, a goal we pursue in collaboration with Harold Varmus. We have developed a purification scheme for HIV integrase and designed an oligonucleotide aimed at binding in a unique fashion to the enzyme. It mimics the transition state of the reaction. The REV protein is involved in specific recognition of RNA in the nucleus, where HIV RNA is transported out into the cytoplasm. Complexes of HIV-REV with RNA have been produced for structure analysis, aimed at drug design. Our drug design principles are based on precisely matching compound chemistry to target crevices on the protein structure, truly the best test of our understanding of mo}ecular interfaces. The goal and assay of our success in the basic understanding of molecular dialogue is to prevent replication of HIV in cells, and so in HIV-infected individuals.
