The Scripps Research Institute
  News Room Contacts  
  Information for Journalists  
  News  
  Resources  
  Publications  
  Calendar of Events  

 

 

News and Publications



DEPARTMENT OF MOLECULAR BIOLOGY


STAFF

Peter E. Wright, Ph.D. Member and Chairman, Cecil H. and Ida M. Green Investigator in Medical Research
William Balch, Ph.D.* Member
Carlos F. Barbas III, Ph.D. Associate Member
Donald E. Bashford, Ph.D. Associate Member
William H. Beers, Ph.D. Member, Senior Vice President, TSRI
Jeffrey D. Bleil, Ph.D. Assistant Member
Charles L. Brooks III, Ph.D. Member
Dennis R. Burton Ph.D.** Member
David A. Case, Ph.D. Member
Walter J. Chazin, Ph.D. Associate Member
Luis de Lecea, Ph.D. Assistant Member
Bruce S. Duncan, Ph.D. Assistant Member
H. Jane Dyson, Ph.D. Associate Member
John H. Elder, Ph.D. Member
Martha J. Fedor, Ph.D.*** Assistant Member
Terry M. Fieser, Ph.D. Adjunct Assistant Member
Christopher K. Garcia, Ph.D. Assistant Member
Larry Gerace, Ph.D.* Member
Dan D. Gerendasy, Ph.D. Assistant Member
Elizabeth D. Getzoff, Ph.D.** Associate Member
Adam Godzik, Ph.D. Assistant Member
David B. Goodin, Ph.D. Associate Member
David S. Goodsell, Jr., Ph.D. Assistant Member
Joel M. Gottesfeld, Ph.D. Member
Robert Hallewell, D.Phil. Adjunct Associate Member
Karl W. Hasel, Ph.D.**** Adjunct Assistant Member, Digital Gene Technologies, La Jolla, CA
Donald Hilvert, Ph.D.***** Member
Jonathan D. Hirst, Ph.D. Assistant Member
Richard A. Houghten, Ph.D. Adjunct Member
Kim D. Janda, Ph.D.***** Member
John E. Johnson, Ph.D. Member
Gerald F. Joyce, M.D., Ph.D Member
Angray S. Kang, Ph.D. Assistant Member
Peter A. Kast, Ph.D.***** Assistant Member
Ehud Keinan, Ph.D. Adjunct Member
Andrzej Kolinski, Ph.D. Associate Member
James LaClair, Ph.D. Assistant Member
Richard A. Lerner, M.D. Member, President, TSRI, Lita Annenberg Hazen Professor of Immunochemistry, Cecil H. and Ida M. Green Chair in Chemistry
Clare H. McGowan, Ph.D. Assistant Member
Duncan E. McRee, Ph.D. Assistant Member
David R. Milich, Ph.D. Associate Member
David P. Millar, Ph.D. Associate Member
Louis Noodleman, Ph.D. Associate Member
Arthur J. Olson, Ph.D. Member
James C. Paulson, Ph.D. Adjunct Member
Steven I. Reed, Ph.D. Member
Victoria A. Roberts, Ph.D. Assistant Member
Paul Russell, Ph.D. Associate Member
Arnold C. Satterthwait, Jr., Ph.D. Assistant Member
Paul R. Schimmel, Ph.D. Member
Sandra Schmid, Ph.D.* Associate Member
Anette Schneemann, Ph.D. Assistant Member
Charles G. Shevlin, Ph.D. Assistant Member
Subhash C. Sinha, Ph.D. Assistant Member
Gary E. Siuzdak, Ph.D. Assistant Member
Jeffrey Skolnick, Ph.D. Member
Robyn L. Stanfield, Ph.D. Assistant Member
Charles D. Stout, Ph.D. Associate Member
Enrico Stura, D.Phil. Assistant Member
J. Gregor Sutcliffe, Ph.D. Member
John A. Tainer, Ph.D. Member
Ian A. Wilson, D.Phil. Member
Peter M. Wirsching, Ph.D. Associate Member
Curt Wittenberg, Ph.D. Associate Member
Mark Yeager, M.D., Ph.D.* Associate Member
Todd O. Yeates, Ph.D. Adjunct Associate Member


SERVICE FACILITIES

John Chung, Ph.D. Manager, Nuclear Magnetic Resonance Facilities
Michael E. Pique Director, Graphics Development


SENIOR RESEARCH ASSOCIATES

Edelmira Cabezas, Ph.D.
Monica Carson, Ph.D.
Cindy L. Fisher, Ph.D.**** Structural Bioinformatics, Inc., San Diego, CA
Liliane A. Dickinson, Ph.D.
Flavio Grynszpan, Ph.D.
Jian Li, Ph.D.
Tianwei Lin, Ph.D.
Maria A. Martinez-Yamout, Ph.D.
Clifford Dean Mol, Ph.D.
Govinda Sridhar Prasad, Ph.D.
Vijay Sai Reddy, Ph.D.
Vicente M. Reyes, Ph.D.
Isabelle Rooney, Ph.D.****, La Jolla Institute for Allergy and Immunology, La Jolla, CA
Christoph B. Weber, Ph.D.
Douglas B. Williams, Ph.D.
Mason M. Yamashita, M.D., Ph.D.


RESEARCH ASSOCIATES

Helena Almer, Ph.D.
Jennifer Andris-Widhopf, Ph.D.
Rebecca Alexander, Ph.D.
Carlos E. Alvarez, Ph.D.
Nicolaus Albert Bahr, Ph.D.****, University of Bern, Bern, Switzerland
Yawen Bai, Ph.D.
David Barondeau, Ph.D.
Roger R. Beerli, Ph.D.
Bonnie Bertolaet, Ph.D.
Robert Bjornestedt, Ph.D.**** Astra Biotech Lab, Stockholm, Sweden
Alessandra Blasina, Ph.D.
Michael N. Boddy, Ph.D.
Yves Bourne, Ph.D.**** Institut de Recherche Concertee 1, Marseille, France
Jean-Marc Brondello, Ph.D.
Ronald Brudler, Ph.D.
Christopher M. Bruns, Ph.D.
Mary A. Canady, Ph.D.
Yi Cao, Ph.D.
Danilo R. Casimiro, Ph.D.**** Merck Laboratories, West Point, PA
Silvia Cavagnero, Ph.D.
Daphne Chapman-Shimshoni, Ph.D.**** Weizman Institute of Science, Tel Aviv, Israel
Jonathan A. Chappel, Ph.D. MRC Collaborative Centre, London, England
Udayan Chatterji, Ph.D.
Shu-Wen W. Chen, Ph.D.
Joe Chihade, Ph.D.
John Christodoulou, Ph.D.
Duncan J. Clarke, Ph.D.
Thomas Cleveland, Ph.D.
Brian R. Crane, Ph.D.
Orlando Crescenzi, Ph.D.
Aymeric de Parseval, Ph.D.
Massimo Degano, Ph.D.
Genevieve Degols, Ph.D.**** Institut de Génétique Moléculaire, Montpellier, France
Eugene Demchuk, Ph.D.
Valerie Dillet, Ph.D.
Xiao Fan Dong, Ph.D.
David G. Donne, Ph.D.
Elena Dovalsantos, Ph.D.
Brendan Duggan, Ph.D.
David J. Eliezer, Ph.D.
Patricia A. Fagan, Ph.D.
Richard S. Fee, Ph.D.
Karin Maria Flick, Ph.D.
Katrina T. Forest, Ph.D.
Mark P. Foster, Ph.D.
Frederique Gaits, Ph.D.
Carlos Garcia, Ph.D.
Paul Geymayer, Ph.D.**** University of Bern, Bern, Switzerland
Jayant B. Ghiara, Ph.D.
Phalguni Ghosh, Ph.D.**** Hughes Institute, Roseville, MN
Svetlana I. Gramatikova, Ph.D.
Samantha Greasley, Ph.D.
Karl Gruber, Ph.D.
Yue Guan, Ph.D.
Zhuyan Guo, Ph.D.
Steven Blaine Haase, Ph.D.
Stephen G. Hall, Ph.D.**** GenMed Corporation, San Diego, CA
Seungil Han, Ph.D.
Joan Hanley-Hyde, Ph.D.
Matthew R. Haynes, Ph.D.
Peter B. Hedlund, Ph.D.
Andreas Heine, Ph.D.
Tamara Hendrickson, Ph.D.
Ludger Hengst, Ph.D.
Thomas Herzinger, M.D.
Torsten Hoffmann, Ph.D Hoffmann La Roche, Basel, Switzerland
Signe M.A. Holmbeck, Ph.D.
Yuchu G. Hsiung, Ph.D.
Wei-Ping Hu, Ph.D. National Chung-Cheng University, Chia-Wi, Taiwan
Mingdong Huang, Ph.D.
Michael J. Hunter, Ph.D.
Nathalie Jourdan, Ph.D.
Peter Kaiser, Ph.D.
Junko Kanoh, Ph.D.
Daniel Kaufmann, Ph.D.**** University of Bern, Bern, Switzerland
Arto Tapio Kesti, Ph.D.
Randal R. Ketchem, Ph.D.
Robert Konecny, Ph.D.
Xianjun Kong, Ph.D.**** Molecular Simulations, Inc., San Diego, CA
Richard W. Kriwacki, Ph.D.
Gerard Johannes Kroon, M.D.
Gary S. Laco, Ph.D.**** University of Washington Medical School, Seattle, WA
John H. Laity, Ph.D.
Wai Chung Lam, Ph.D.
Stefan Lanker, Ph.D.
Teresa Larsen, Ph.D.
Janet Leatherwood, Ph.D.**** State University of New York, Stonybrook, NY
Brian M. Lee, Ph.D.
Glen Legge, Ph.D.
Ronald D. Lewis II, Ph.D.
Xiang Li, Ph.D.**** Boehringer Ingelheim Pharmaceuticals, Inc., Ridgefield, CT
Ying-Chuan Lin, Ph.D.
Benjamin List, Ph.D.
Gang Liu, Ph.D.
Qiang Liu, Ph.D.**** Abgenix, Inc., Fremont, CA
Oded Livnah, Ph.D.
Terence Pui Kwan Lo, Ph.D.
John J. Long, Ph.D.
Antonia Lopez, Ph.D.
Guang Xiang Luo, Ph.D.+
Thomas J. Macke, Ph.D.
Jens C. Madsen, Ph.D.**** Bruker Spectrospin, Taby, Sweden
Lena E. Maler, Ph.D.
Georges Mer, Ph.D.
Siobhan M. Miick, Ph.D.**** Nanogen, Inc., San Diego, CA
Debasisa Mohanty, Ph.D.
Vincent Moncollin, Ph.D.
Guillaume Mondesert, Ph.D.
Garrett M. Morris, D.Phil.
Rabi A. Musah, Ph.D.
Sangari Mylvaganam, Ph.D.
Padmaja Natarajan, Ph.D.
Tyzoon Nomanbhoy, Ph.D.
Jacek Nowakowski, Ph.D.
Eric H. Olender, Ph.D.**** San Jose State University, San Jose, CA
Phillip Ordoukhanian, Ph.D.
Angel Ramirez Ortiz, Ph.D.
Michael J. Osborne, Ph.D.
Michael R. Otto, Ph.D.
Jean-Luc Pellequer, Ph.D.
Gabriela Perez-Alvarado, Ph.D.
Jeroen A. Pikkemaat, Ph.D.
Stefan Prytulla, Ph.D.**** Institut für Pharmazeutische Chemie, Graz, Austria
Christoph Rader, Ph.D.
Ishwar Radhakrishnan, Ph.D.
Jennifer Lynn Radkiewicz, Ph.D.
Sun Ai Raillard-Yoon, Ph.D.**** Maxygen, Santa Clara, CA
Mohan Srinivasa Rao, Ph.D.
Boris Alexander Reva, Ph.D.
Nicholas Rhind, Ph.D.
Lluis Ribas De Pouplana, Ph.D.
Olaf Ritzeler, Ph.D.**** University of Bern, Bern, Switzerland
Jeffrey Keith Rogers, Ph.D.
Christopher Rosin, Ph.D.
Franziska Ruchti, Ph.D.
Leszek Rychlewski, Ph.D.
Kandasamy Sakthivel, Ph.D.
Michel F. Sanner, Ph.D.
Niranjan Sardesai, Ph.D.
Mallika Sastry, Ph.D.
Kevin Y. Sato, Ph.D.
Sergio D.B. Scrofani, Ph.D.
David Segal, Ph.D.
Marisa Segal, Ph.D.
Brent W. Segelke, Ph.D.**** Lawrence Livermore National Laboratory, Livermore, CA
Doron Yacov Shabatt, Ph.D.
Joan-Emma Shea, Ph.D.
Terry Lee Sheppard, Ph.D.
Kazuhiro Shiozaki, Ph.D.
William Arthur Shirley, Ph.D.
Anjana Sinha, Ph.D.**** University of California, San Diego, CA
Doree F. Sitkoff, Ph.D.
Monique F.M. Smeets, Ph.D.
Velin Slatkov Spassov, Ph.D.
Jeffrey A. Speir, Ph.D.
Charles H. Spruck III, Ph.D.
Jayashree Srinivasan, Ph.D.
Brian Steer, Ph.D.
Peter Steinberger, Ph.D.
Heimo Strohmaier, Ph.D.
David T. Stuart, Ph.D.
Ying Su, Ph.D.
Kwang-Ai Won Szymanski, Ph.D.
Fujie Tanaka, Ph.D.
John Tate, Ph.D.
Maria Julie Thayer, Ph.D.
Elizabeth A. Thomas, Ph.D.
Robert Turner, Ph.D.
Maria Henar Valdivieso, Ph.D.**** University of Salamanca, Salamanca, Spain
Sara Venturini, Ph.D.
Michal Vieth, Ph.D.
John Howard Viles, Ph.D.
Keisuke Wakasugi, Ph.D.
Chien-Chia Wang, Ph.D.
Mark Howard Watson, Ph.D.
Douglas B. Williams, Ph.D.
Pamela A. Williams, Ph.D.
Christer Wingren, Ph.D.
Martin Charles Wright, Ph.D.**** Phylos, Boston, MA
Bin Xia, Ph.D.
Jian Xu, Ph.D.
Jing Xu, Ph.D.
Jian Yao, Ph.D.
William S. Young, Ph.D.**** Molecular Simulations, Inc., San Diego, CA
Ke Zeng, Ph.D.
Baohong Zhang, Ph.D.
Hongyu Zhang, Ph.D.**** Center for Advanced Research in Biotechnology, Rockville, MD
Li Zhang, Ph.D.
Mingzhu Zhang, Ph.D.**** Vanderbilt University, Nashville, TN
Ouwen Zhang, Ph.D.
Ruoheng Zhang, Ph.D.
Xu-Yang Zhao, Ph.D.
Guo-Fu Zhong, Ph.D.
Zhongxiang Zhou, Ph.D.**** Alanex Corporation, San Diego, CA
Leiming Zhu, Ph.D.


SCIENTIFIC ASSOCIATES

Andrew S. Arvai
Diane Marie Kubitz


VISITING INVESTIGATORS

Stephen J. Benkovic, Ph.D. The Pennsylvania State University, University Park, PA
Sture Forsen, Ph.D. University of Lund, Lund, Sweden
Arne Holmgren, M.D., Ph.D. Karolinska Institute, Stockholm, Sweden
Tai-huang Huang, Ph.D. Institute of Biomedical Sciences, Academica Sinica, Taipei, Taiwan
Kenneth Kustin, Ph.D. Brandeis University, Waltham, MA
Johannes Langedijk, Ph.D. Institute for Animal Sciences and Health, Lelystad, the Netherlands
Robert D. Rosenstein, Ph.D. Lawrence Berkeley Laboratory, Berkeley, CA
Michael Taussig, Ph.D. Babraham Research Institute, Cambridge, England


* Joint appointment in Department of Cell Biology
** Joint appointment in Department of Immunology
*** Joint appointment in The Skaggs Institute for Chemical Biology
**** Appointment completed; new location shown
***** Joint appointment in Department of Chemistry
+ Appointment completed

Top of Page


Chairman's Overview

Peter E. Wright, Ph.D.

Research in the Department of Molecular Biology encompasses a wide range of disciplines, extending from structural biology at one extreme to molecular genetics at the other. During the past year, our scientists have continued to make exciting progress toward understanding the fundamental molecular events that underlie the processes of life. These accomplishments have been reported on the pages of the most influential journals and have been rewarded by a higher than ever level of grant support. The faculty members of this department have every reason to be proud of these achievements, as do the exceptionally talented postdoctoral fellows, graduate students, technologists, and support personnel, who are absolutely vital to the success of each and every research program.

Within the confines of this brief overview, I cannot do justice to the many exciting research programs within the department. These are described in detail on the following pages, and a few of the highlights are mentioned briefly here.

A collaboration between Joel Gottesfeld of this department and Peter Dervan at the California Institute of Technology led to the development of small synthetic molecules that bind in the minor groove of DNA and can regulate gene expression. Gottesfeld, Dervan, and their colleagues showed that these compounds, pyrrole-imidazole polyamides, can be designed to target specific DNA sequences and interfere with the transcription of specific genes. Polyamides directed against DNA target sequences within the HIV enhancer and promoter are highly effective inhibitors of viral replication in isolated human blood peripheral lymphocytes. These compounds offer the potential for design of small molecules that can regulate the transcriptional activity of selected target genes in living cells. These results are extremely exciting and could eventually lead to a new class of therapeutic agents directed against a broad spectrum of viral and other diseases.

A different approach to targeted gene regulation is being pursued by Carlos Barbas and his colleagues. Proteins containing multiple zinc-finger domains are being engineered to recognize specific binding sites containing as many as 18 nucleotides. Such proteins can target a unique locus in the human genome and have considerable potential as genetic regulators in a variety of human diseases.

In the area of structural biology, a number of spectacular new advances were made during the past year. Ian Wilson and his coworkers determined the three-dimensional structure of murine CD1, a protein that is distantly related to MHC molecules and that presents antigens to T cells. The overall structure is similar to that of an MHC class I molecule, but CD1 contains a deep, almost entirely hydrophobic, binding cavity. The structure reinforces the view that the role of CD1 is to display lipid and glycolipid antigens.

In another major advance, Elizabeth Getzoff and coworkers determined the atomic resolution structure of a signaling intermediate in the light cycle of a photoreceptor protein by using millisecond time-resolved x-ray crystallography. This achievement is impressive, because the lifetime of the photocycle intermediate is less than 1 second, and the findings provide the first detailed insights into the structure of a light-activated intermediate. The structural changes observed, relative to the ground state, suggest a mechanism for signal transduction and provide a general framework for understanding the structural mechanisms of protein photocycles.

Other important structures determined during the past year include catalytic antibodies; animal and plant viruses; calcium signal transduction proteins; the first structure of a complex between a transcriptional activation sequence and a domain of the coactivator CREB-binding protein; and the structure of a human DNA-repair enzyme, uracil-DNA glycolase. This last structure is of interest for the insights it provides into the mechanism of recognition of damaged bases, which are flipped out from the double helix and into the active site of the enzyme, where they are excised.

Gerald Joyce and Martin Wright developed a new method to evolve catalytic RNA molecules continuously in vitro. This method is a major advance in the evolution of novel RNA catalysts. In addition, continuous in vitro evolution provides a more realistic model of biological evolution and should provide new insights into evolutionary mechanisms. Other work in the Joyce laboratory led to the development of a small and highly efficient DNA enzyme that can be tailored to recognize and cleave different RNA target sequences. Such DNA enzymes have potential applications both as laboratory reagents and in medicine.

In the area of cell-cycle regulation, Curt Wittenberg and his colleagues published several key articles describing the phosphorylation-dependent degradation of cyclins. Cyclins regulate cell-cycle transitions in eukaryotes through association with cyclin-dependent protein kinases. Wittenberg and coworkers showed that phosphorylation of a G1 cyclin, by the kinase that the cyclin activates, provides a signal that promotes rapid degradation of the cyclin by a ubiquitin-dependent pathway and thereby makes activation self-limiting. This work is providing important new insights into the complex molecular mechanisms that govern the major cell-cycle transitions.

Finally, I am delighted to report that we have recruited several outstanding scientists to full or joint appointments in the department. Gary Siuzdak, director of the Mass Spectrometry Facility, joined our faculty during the past year, and Paul Schimmel, Jamie Williamson, and Martha Fedor will soon bring new strengths to our programs in the molecular and structural biology of RNA. In addition, Jonathan Hirst, Chris Garcia, Robyn Stanfield, and Luis de Lecea have recently joined us and strengthen our research in the areas of theory, x-ray crystallography, and molecular neurobiology.

Top of Page


Investigators' Reports


Crystallographic Analyses of Viruses and Macromolecular Assemblies

J. Johnson, A. Schneemann, V. Reddy, T. Lin, W. Wikoff, A. Kumar, P. Natarajan, S. Hall, J. Tate, F. Dong, M. Canady, H. Giesing, B. Sheehan, H. Langedijk*

* Institute for Animal Science and Health, Lelystad, the Netherlands

Our group investigates the structure and function of viruses to elucidate molecular mechanisms of infection. We use this information to determine potential targets for antiviral agents and to use viruses as reagents for understanding and exploiting the biology of the cell. We have used a variety of virus families for our investigations; each family has novel properties appropriate for specific lines of investigation.

For research on assembly, we use virus groups readily studied in vitro and in vivo. An area of specific interest is the formation of quasi-equivalent viral capsids in which the same gene product participates in polymorphic interactions to generate both hexamers and pentamers in the formation of icosahedral shells. These structures are the biological equivalent of Buckminster Fuller's geodesic domes. They provide particles large enough to package and protect the genome and have targeting properties to deliver the genome to host cells for viral replication.

Knowledge of the structure of cowpea chlorotic mottle virus has enabled us to determine the parts of the subunit that control different aspects of assembly. In vitro assembly with structure-based mutations has been studied in collaboration with M. Young at Montana State University, who has developed an Escherichia coli expression system for the cowpea chlorotic mottle virus subunit and can purify, refold, and assemble the protein into viruslike particles. These studies confirm the postulated functional roles, based on crystallographic studies, of different regions of the subunit.

To further understand assembly polymorphism, we examined the different patterns of assembly of the coat protein of alfalfa mosaic virus. In nature, this multipartite RNA virus exists as four bacilliform particles in which capsid protein forms a particle proportional to the size of the RNA genome segment packaged. Purified viral protein can assemble in vitro into an icosahedral particle in the absence of RNA. We determined the structure of this assembly product at 4.0-Å resolution. Using the coordinates of the subunit structure, we created a model of the bacilliform particles that agrees in detail with the hexagonal lattice observed in electron microscopy studies.

In collaboration with A. Schneemann, we are exploring structure-function relationships in the nodavirus and tetravirus families of animal viruses. Our studies have shown that capsid polymorphism is controlled by a protein switch and/or RNA, that the maturation cleavage required for infection is autocatalytic and depends on assembly and the binding of calcium ions, that the release of RNA probably depends on a specific protein-RNA interaction that occurs only once in the context of the symmetric capsid, and that membrane translocation of RNA is probably mediated by a pentameric helical bundle that is rendered covalently independent from the subunit by the maturation cleavage.

Although the quaternary structures and subunits of nodaviruses and tetraviruses differ dramatically in size, our studies have shown an evolutionary relationship between the two families. Using heterologous protein expression systems in which the capsid protein spontaneously assembles to form particles, we crystallized and solved the structures of particles specifically mutated to reveal the atomic details of the phenomena described earlier. Recently, we used computational chemistry methods to explore assembly trajectories and their dependence on intersubunit stabilities.

We are investigating the use of plant viral particles as carriers for genetically inserted, heterologous polypeptides up to 40 amino acids long. These polypeptides have been used to generate neutralizing antibodies to HIV and to present peptide antagonists that trigger cell-surface phenomena. Structures of these "chimeric viruses" have enabled us to design presentations that increase the efficacy of inserted peptides and to better understand factors that affect peptide folding and viral assembly.

We are continuing the first crystallographic study of the capsid of HK97, a -like double-stranded DNA bacteriophage. The T = 7, 620-Å diameter particles were assembled by expressing the gene for the capsid protein in E. coli. These particles were crystallized, and x-ray diffraction patterns beyond 3.5-Å resolution were obtained. A data set has been collected at the Cornell High Energy Synchrotron Source, processed to 7.0-Å resolution, and merged with ultralow resolution data (200- to 16-Å resolution) collected at the Stanford Synchrotron Radiation Laboratory. Phases between 200- and 50-Å resolution were computed with the particle model determined by electron cryo-microscopy and image reconstruction at 35-Å resolution. Phases for the higher resolution data were determined by using extension procedures and the 60-fold noncrystallographic symmetry. The structure shows a classical T = 7 lattice with the subunits formed predominantly of -helices. On the basis of the x-ray structure and a electron cryo-microscopy reconstruction of the 450-Å diameter procapsid, we have proposed a detailed mechanism for particle maturation.

PUBLICATIONS

Baker, T.S., Johnson, J.E. Low resolution meets high: Towards a resolution continuum from cells to atoms. Curr. Opin. Struct. Biol. 6:585, 1996.

Baker, T.S., Johnson, J.E. Principles of virus structure determination. In: Structural Biology of Viruses. Chiu, W., Burnett, R.M., Garcea, R. (Eds.). Oxford University Press, New York, 1997, p. 38.

Bothner, B., Dong, X., Bibbs, L., Johnson, J., Siuzdak, G. Evidence of viral capsid dynamics using limited proteolysis and mass spectrometry. J. Biol. Chem., in press.

Chandrasekar, V., Johnson, J. The structure of tobacco ringspot virus: A link in the evolution of icosahderal capsids in the picornavirus superfamily. Structure, in press.

Chandrasekar, V., Munshi, S., Johnson, J.E. Crystallization and preliminary x-ray analysis of tobacco ringspot virus. Acta Crystallogr. D53:125, 1997.

Flasinski, S., Dzianott, A., Speir, J.A., Johnson, J.E., Bujarski, J.J. Structure-based rationale for the rescue of systemic movement of brome mosaic virus by spontaneous second-site mutations in the coat protein gene. J. Virol. 71:2500, 1997.

Johnson, J.E., Lin, T., Lomonossoff, G.P. Presentation of heterologous peptides on plant viruses: Genetics, structure and function. Annu. Rev. Phytopathol. 35:67, 1997.

Johnson, J.E., Rueckert, R.R. Packaging and release of the viral genome. In: Structural Biology of Viruses. Chiu, W., Burnett, R.M., Garcea, R. (Eds.). Oxford University Press, New York, 1997, p. 269.

Johnson, J.E., Schneemann, A. Nodavirus endopeptidase. In: Handbook of Proteolytic Enzymes. Barret, A.J., Rawlings, N.D.,Woessner, J.F. (Eds.). Academic Press, San Diego, in press.

Johnson, J.E., Speir, J.A. Quasi-equivalent viruses: A paradigm for protein assemblies. J. Mol. Biol. 269:665, 1997.

Kumar, A., Chipman, P., Yusibov, V., Fita, I., Hatta, Y., Loesch-Fries, S., Baker, T.S., Rossmann, M.G., Johnson, J.E. The structure of alfalfa mosaic virus capsid protein assembled as a T = 1 icosahedral particle at 4.0-Å resolution. J. Virol. 71:7911, 1997.

Porta, C., Lin, T., Johnson, J., Lomonossoff, G. The development of cowpea mosaic virus as a potential source of novel vaccines. Intervirology 39:79, 1996.

Reddy, V.S., Giesing, H., Kumar, A., Morton, R., Post, C.B., Brooks, C., Johnson, J.E. Energetics of quasi-equivalence: Computational analysis of protein-protein interactions in icosahedral viruses. Biophys. J., in press.

Schneemann, A., Reddy, V., Johnson, J.E. The structure and function of nodavirus particles: A paradigm for understanding chemical biology. Adv. Virus Res., in press.

Spall, V.E., Porta, C., Taylor, K.M., Lin, T., Johnson, J.E., Lomonossoff, G.P. Antigen expression on the surface of a plant virus for vaccine production. In: Engineering Crops for Industrial End Uses. Shewry, P.R., Napier, J.A., Davis, P. (Eds.). Portland Press, London, in press.

Wikoff, W., Tsai, C.J., Wang, G., Baker, T.S., Johnson, J.E. Crystallographic analysis and cryoelectron microscopy reconstruction of cucumber mosaic virus. Virology 232:91, 1997.

Zlotnick, A., Natarajan, P., Munshi, S., Johnson, J.E. Resolution of space group ambiguity and the structure determination of nodamura virus to 3.3 Å resolution from pseudo R32 (monoclinic) crystals. Acta Crystallogr., in press.

Top of Page


Assembly and Uncoating of Icosahedral Viruses

A. Schneemann, D. Marshall, V. Reddy, F. Dong, J. Johnson

Coat proteins of nonenveloped, icosahedral animal viruses have multiple functions during the course of viral infection and spread, including assembly of the viral capsid, specific encapsidation of the viral genome, binding to a cellular receptor, and uncoating. We are interested in the chemical and structural determinants that endow a single polypeptide chain with such versatility. Our investigations focus on a structurally and genetically well-characterized model system, the T = 3 nodaviruses.

Nodaviruses are assembled from 180 copies of a single coat protein and two strands of messenger sense RNA. To acquire infectivity, assembled particles must undergo a maturation step in which subunits of the coat protein autocatalytically cleave into two polypeptides. We are investigating the precise pathway of viral assembly and the mechanism of the associated cleavage reaction. Our studies are guided by the high-resolution structures of several nodaviruses that have been determined in the laboratory of J. Johnson, Department of Molecular Biology. Analysis of the atomic structures of these viruses has enabled us to detect regions in the coat protein that appear to be critical in regulating viral assembly, stability, and maturation. For example, genetic, biochemical, and biophysical analyses have shown that the N-terminal part of the coat protein regulates the shape of viral particles, whereas the C-terminal part controls specific encapsidation of the viral genome.

A second area of research focuses on viral uncoating. Uncoating and delivery of viral genomes into the cytosol of susceptible cells are poorly understood processes. The viral nucleic acid must cross a cellular membrane at some point during entry, but the mechanistic details of this step remain unknown. The nodaviruses contain a preformed helical bundle just below the surface of the shell that may form a channel within the endosomal membrane after receptor-mediated endocytosis of the virion. This channel might be used for translocation of genomic RNA across the membrane into the cytosol. We have obtained preliminary evidence that supports this model.

The third project in our laboratory is determining and isolating the viral receptor protein. We have generated monoclonal antibodies that inhibit nodavirus infection of cultured cells; these antibodies will be used to isolate the receptor protein. Availability of the viral receptor will enable us to study the initial interactions of the viral particle with the cell surface and will help us to elucidate the mechanism of viral uncoating in detail.

PUBLICATIONS

Johnson, J.E., Schneemann, A. Nodavirus endopeptidase. In: Handbook of Proteolytic Enzymes. Barret, A.J., Rawlings, N.D., Woessner, J.F. (Eds.). Academic Press, San Diego, in press.

Schneemann, A., Reddy, V., Johnson, J.E. The structure and function of nodavirus particles: A paradigm for understanding chemical biology. Annu. Rev. Virus Res., in press.

Top of Page


X-Ray Crystallographic Studies of Immunologically Important Macromolecules, Growth Factors, Receptors, and Enzymes Involved in Human Disease

I.A. Wilson, E.A. Stura, R.L. Stanfield, K.C. Garcia, T.A. Cross, W.L. Densley, M. Degano, S.E. Greasley, K. Gruber, J.B. Ghiara, M.R. Haynes, A. Heine, T. Horton, K. Hotta, M. Huang, D.A. Jewell, H.-M. Li, O. Livnah, G. Luo, E.E. Ollmann, K.A. Renner, V.M. Reyes, C.E.A. Scott, B.W. Segelke, J.A. Speir, R.S. Stefanko, Y. Su, M.J. Taussig,* D.B. Williams, J. Xu, M.M. Yamashita, K. Zeng

* Babraham Institute, Cambridge, England

The main goal of our research program is to understand macromolecular recognition, especially in the immune system. Current projects include x-ray crystallographic studies of antibody-antigen complexes; complexes that consist of T-cell receptors (TCRs), MHC molecules, and peptide antigens; classical and nonclassical MHC molecules and their antigens; human growth factors and their receptors; and enzyme targets for chemotherapy.

T-CELL RECEPTOR

The TCR is a heterodimeric glycoprotein expressed on the surface of T lymphocytes that plays a central role in the recognition of peptide antigens in the context of the MHC. The ability to discriminate between self and foreign antigens depends on the fine specificity of the interaction between TCRs and MHC molecules. We have determined the three-dimensional structure of the murine class I TCR 2C to a resolution of 2.5 Å. Overall, the structure generally resembles an Fab fragment of an antibody. However, significant differences are found in the pairing of the variable and the constant domains and in the unconventional fold of the constant domain of the -chain. The hypervariable loops form a flat surface with a deep central pocket that is involved in the recognition of the peptide antigen bound to the MHC molecule.

To define the molecular basis of the TCR-MHC interaction, we have crystallized and are determining the structures of 2C-MHC complexes with both self and foreign peptide antigens bound to the MHC. The peptide residues contribute about 25% of the surface buried in the formation of the complex. The third complementarity-determining regions of the - and ß-chains of the TCR interact with the central residue of the peptide bound to the MHC. Specific recognition of the MHC molecule is achieved through contacts of complementarity-determining regions 1 and 2 of the - and the ß-chain of the TCR to the helices of the peptide-binding domain.

In the MHC class II system, our focus has been on the production and crystallization of TCRs specific for the murine class II MHC molecule IAd in complex with ovalbumin, hemagglutinin, and insulin peptides. Structural studies of class II TCR-MHC complexes would address the important question of whether TCR recognition of class I MHC molecules differs substantially from TCR recognition of class II molecules. The TCR-MHC studies are being done in collaboration with L. Teyton, Department of Immunology.

CLASS I AND CLASS II MHC MOLECULES

Several cell-surface molecules involved in T-cell activation have been expressed in soluble forms in collaboration with Dr. Teyton and A. Brunmark, R.W. Johnson Pharmaceutical Research Institute. These include murine MHC class I molecules H-2Ld, H-2Kb mutants and murine class II molecule IAd, each complexed with peptides. Evaluation of a number of protein engineering strategies was necessary to produce milligram quantities of IAd for structure-function studies.

H-2Ld is alloreactive with the H-2Kb--specific TCR 2C, and therefore a structural comparison between H-2Ld and H-2Kb is of great value. The structure of H-2Ld has been determined with a mixture of peptides (Fig. 1). The results show that the allogeneic reaction appears to be due mainly to the C-terminal residues of specific peptides that provide the optimal shape and chemistry for 2C reactivity normally present in the syngeneic H-2Kb interaction. The class II molecule IAd is biologically well characterized and is a factor in murine models of autoimmunity. Two crystal structures of IAd with different peptides have been determined at 2.4- and 2.6-Å resolution and reveal novel elements of this particular class II peptide binding.

Other unusual interactions between MHC class I molecules and small hapten antigens were examined by determining the structure of H-2Kb complexed with a glycopeptide. The disaccharide of the ligand is visible on the outer boundaries of the binding groove, where interactions with TCRs would predominate. These studies have furnished detailed new insights on antigen selection and presentation to TCRs and have advanced our understanding of these complex interactions that are critical to cellular immunity.

CATALYTIC ANTIBODIES

The structure of 13G5, an antibody to metallocene that catalyzes the disfavored exo Diels-Alder reaction, has been determined at 1.95-Å resolution for the unbound form and for the antibody in complex with a ferrocene derivative (Fig. 2).
Interactions between antibody side chains and the inhibitor provide the basis for understanding the chemical transformation. Modeling studies with transition-state analogs and substrates suggest hypotheses for the preference for the exo vs the endo pathway.

The structure of 33F12, an aldolase antibody that was generated by reactive immunization, has been determined at 2.15-Å resolution. The particular sequence of residues in the binding pocket and the hydrophobic environment of the pocket can account for the unusual pKa (~6.0) and reactivity of a lysine in the active site. Determination of the structure of a similar natural aldolase enzyme is in progress.

Antibody 5C8 catalyzes the disfavored 6-endo-tet ring closure reaction of trans-epoxy alcohols (violating "Baldwin's rules"). The crystal structures of the Fab fragment of 5C8, unbound (to 2.5 Å) and in complex with the transition-state analog (to 2.0 Å), have been determined (Fig. 3). Further crystallographic and modeling studies are being done to establish the catalytic mechanism and explain the observed regiospecificity and stereospecificity.

The Fab fragments of two antibodies that catalyze another Diels-Alder reaction and a decarboxylation reaction have been crystallized both in their native forms and in complex with their corresponding transition-state analogs. The x-ray data for these two Fabs in both native and complex forms were collected at 2.1 and 2.2 Å, respectively, for each antibody. Refinements of these four structures are in progress. The catalytic antibody projects are being done in collaboration with R. Lerner of the Departments of Chemistry and Molecular Biology; with C. Barbas and C. Shevlin of the Department of Molecular Biology; and with K. Janda, D. Hilvert, and C.-H. Wong of the Department of Chemistry.

NEUTRALIZING ANTIBODIES TO HIV TYPE I

We are studying several antibodies that neutralize HIV type 1. Four of these antibodies were generated by Repligen Corporation (Needham, MA) against a peptide of 40 amino acids that corresponds to the third hypervariable (V3) loop of the HIV type 1 surface glycoprotein gp120. These four antibodies differ in their viral-strain specificity and have overlapping but nonidentical epitopes on the V3 loop. We have determined the structures for Fab fragments of antibody 50.1 in the unliganded and peptide-bound forms, of antibody 59.1 in two peptide-bound forms, and of antibody 58.2 in several peptide-bound forms.

The structures of the peptides bound to the Fabs of 50.1 and 59.1 show that the V3 loop adopts an extended structure, followed by a type II ß-turn around residues Gly-Pro-Gly-Arg. The type II turn is followed by a turn of 310 helix, resulting in a double turn at the tip of this loop. We replaced an alanine residue in the double-turn region of the loop with an -aminoisobutyric acid residue to constrain the secondary structure in this part of the peptide. The crystal structure of Fab 59.1 in complex with the designed peptide containing the -aminoisobutyric acid residue showed no difference, as expected, from the original complex. However, nuclear magnetic resonance studies by H. Dyson, Department of Molecular Biology, of the original and the designed peptides showed that the -aminoisobutyric acid residue does indeed confer additional structure on the free peptide in solution.

The structures of the Fab of 58.2 in complex with linear and cyclic peptides and with peptides containing -aminoisobutyric acid show that the V3 peptide can adopt a conformation that differs from that of the V3 peptide bound to Fabs of 50.1 and 59.1. In the 58.2 complexes, the peptide differs around residues Gly-Pro-Gly-Arg, adopting a type I turn, followed again by another unclassified turn to form a different type of double turn. The linear and cyclic peptides and the peptides containing -aminoisobutyric acid all adopt this alternative conformation. These results indicate that the V3 loop can adopt at least two conformations on the viral surface.

Antibodies b12 and 3B3 are neutralizing antibodies that recognize the CD4 binding site of gp120. These antibodies were developed in the laboratories of C. Barbas and D. Burton, Department of Molecular Biology. Phage display technology was used to obtain the antibodies from a library of antibodies derived from sera from patients with long-term asymptomatic HIV infection. These antibodies potently neutralize more than 75% of American HIV isolates and 50% of foreign isolates tested and strongly neutralize primary viral isolates. Studies to obtain crystals of the Fab fragments of these antibodies in complex with gp120 are under way. In a collaboration with Progenics (Tarrytown, NY), we are also studying complexes of various CD4 constructs with gp120.

ANTISTEROID ANTIBODIES

Refinement is in progress for a complex of testosterone with an Fab. A molecular replacement solution has also been obtained for the free Fab. The refinement of the Fab to estrone-3-glucuronide is nearing completion.

TISSUE FACTOR

The blood coagulation protease cascade initiated by tissue factor (TF) can be greatly inhibited in vivo by 5G9, a potent monoclonal antibody to human TF. This antibody binds the carboxyl module of the extracellular domain of TF at a nanomolar binding constant and inhibits formation of the ternary complex consisting of TF, factor VIIa, and factor X that is essential for the initiation of the protease cascade in blood coagulation.

To study recognition of TF by the antibody, we determined the crystal structures of the extracellular modules of human TF, an Fab fragment of 5G9, and a complex consisting of TF and the Fab at 2.4, 2.5, and 3.0 Å, respectively, and measured the binding constants of a panel of TF mutants with 5G9 in collaboration with T. Edgington and W. Ruf, Department of Immunology. The structure of the Fab complexed with TF (Fig. 4) explains and confirms the mutagenesis results, illustrates the molecular mechanism of antibody-antigen recognition, and provides insights into the mechanism by which 5G9 can inhibit formation of the ternary complex.

IL-2 RECEPTOR

IL-2 is one of the most important regulators of the immune response. It produces its effects by binding to its receptor, which is composed of three distinct chains, , ß, and . We are involved in crystallographic studies of this heterotrimeric IL-2 receptor system in a collaboration with T. Ciardelli at Dartmouth College.

ERYTHROPOIETIN RECEPTOR

The structure of the extracellular domain of the receptor for erythropoietin and an antagonist peptide has been determined to the resolution of 2.8 Å. The antagonist peptide was derived from a set of agonist peptides, which were discovered via phage-display techniques in collaboration with R.W. Johnson Pharmaceutical Research Institute and Affymax, Palo Alto, California. A single amino acid modification (tyrosine to 3,5-dibromotyrosine) reversed the activation properties of the peptide. The structure reveals a 2:2 assembly in which two peptides bind two receptor molecules and generate an asymmetric receptor dimerization, a condition that differs from the perfect twofold symmetry observed for the complex containing the peptide agonist.

Symmetric dimerization is a common feature of other erythropoietin receptor--peptide complexes that have agonistic properties. It is well established that in many systems receptor dimerization is the initial key event in signaling, but in this instance, we showed that a nonproductive mode of dimerization occurs. Such a difference in dimer formation can have a great impact on the understanding of signaling in both cytokine and other receptor systems. We are currently determining structures of the agonist peptide complexes with the erythropoietin receptor in collaboration with L. Jolliffe, R.W. Johnson Pharmaceutical Research Institute.

CANCER TARGETS

Glycinamide ribonucleotide transformylase and aminoimidazole carboxamide ribonucleotide transformylase are folate-dependent enzymes involved in the de novo biosynthesis of purine. These enzymes are potential targets for anticancer and antiinflammatory drugs. Crystal structures are in progress for a number of folate-derived inhibitors synthesized in D. Boger's laboratory, Department of Chemistry, and cocrystallized with wild-type Escherichia coli glycinamide ribonucleotide transformylase. These x-ray structures will guide future development of novel nonfolate analogs specific for glycinamide ribonucleotide transformylase, thereby reducing nonspecific interaction with other folate-dependent enzymes in the cell.

Recently, a novel mutant form of E. coli glycinamide ribonucleotide transformylase that lacks the ability to form dimers was developed by S. Benkovic, Pennsylvania State University, in conjunction with P. Jenning's laboratory, University of California, San Diego. We are determining the x-ray structures of this mutant and of additional nonactive mutant structures of glycinamide ribonucleotide transformylase in a number of complexed forms to provide evidence for the mechanism of transformylation.

GUANINE NUCLEOTIDE DISSOCIATION INHIBITOR

The guanine nucleotide dissociation inhibitor (GDI) regulates the retrieval of Rabs in the vesicular traffic while inhibiting the dissociation of GDP from Rabs, small GTP-binding proteins related to the Ras superfamily. The crystal structure of the bovine -isoform of GDI expressed from E. coli has been determined to 1.8-Å resolution. Mutagenesis studies have shown that a substructure in domain I is possibly involved in Rab association. The structure of the GDI-Rab complex has also been investigated. Meanwhile, a more active form of GDI expressed from insect cells has been crystallized, and x-ray data have been collected at 1.7-Å resolution. The high-resolution structure of this GDI will reveal any posttranslational modifications that this molecule undergoes in higher eukaryotic systems.

CD1

CD1 is an MHC-like antigen-presenting molecule. Recent studies have shown that CD1 may present mycobacterial cell wall lipids or phage display--derived peptides as antigens to T cells. Attempts to elute a natural ligand directly from CD1 molecules have been unsuccessful. The crystal structure of mouse CD1d1 has been determined to 2.8-Å resolution in two different space groups. Determination of the structure revealed a molecule markedly similar to MHC class I molecules but with a larger, mostly hydrophobic, ligand-binding groove (Fig. 5).

A second structure determination was undertaken to elucidate the nature of the interaction between CD1 and a large hydrophobic peptide presented to T cells by mouse CD1d1. Several features of the molecule are now apparent that were poorly ordered in the original crystal structure, including several carbohydrate moieties. Surprisingly, a long unbranched ligand occupying the binding groove is also apparent. The density for this ligand is consistent with density seen in the binding groove from the original structure determination and is not consistent with peptide. Attempts to crystallize the CD1d1-peptide complex are continuing. The studies on mouse CD1d1 have been done in collaboration with P. Peterson's group and others at R.W. Johnson Pharmaceutical Research Institute; those on human CD1, with M. Kronenberg, La Jolla Institute of Allergy and Immunology.

We are also attempting to crystallize the human CD1b isoform. Earlier reports stated that this molecule presents mycolic acid derived from mycobacterial extracts to T lymphocytes. Recent studies detected a series of natural lipoglycans that are ligands for human CD1b and that elicit T-cell responses. The structure of the complex of CD1b with a lipid would provide insights into the presentation of nonpeptide antigens by MHC-like molecules to T cells. The studies on CD1b have been done in collaboration with R. Modlin, University of California, Los Angeles, with ligands provided by M. Brenner and S. Porcelli, Harvard Medical School.

Top of Page


Methods for the Crystallization of Macromolecules

E.A. Stura, S. Ghosh, A. Muhlberg, K. Hotta, P. Shim

The development of methods and techniques for the rapid determination of crystallization conditions with minimal amounts of proteins and nucleic acids has been the focus of the research this current year.

Reverse screening, a method based on the principle that a precipitant or buffer should play no role other than that of inducing supersaturation or maintaining the pH, has been applied to proteins that have a binding site for phosphate moieties. In collaboration with J. Elder and C. Stout, Department of Molecular Biology, we used sodium citrate instead of ammonium sulfate, which is a phosphate mimic, to obtain complex crystals of dUTPase with deoxynucleotide diphosphates and triphosphates. These complexes are not compatible with the 1.8-Å diffracting crystals obtained by using cacodylate, another phosphate mimic, which can accommodate only deoxynucleotide monophosphates.

Reverse screening has been also applied to the crystallization of the humanized catalytic antibody Hu21D8. Crystals of the native antibody were obtained by slightly changing the conditions used for the crystallization of Fab fragments of catalytic antibodies obtained in the P.G. Schultz laboratory at the University of California, Berkeley, by using a similar expression and purification system. In collaboration with C. Stout and W. Chazin, Department of Molecular Biology, we are applying the principles of reverse screening to crystallization of RNA and RNA-DNA enzymes with good success.

Dynamic light scattering can be used to determine the aggregation state and polydispersity of macromolecules. Polydispersity can be considered a measure of heterogeneity and, to a certain extent, of the likelihood of crystallization. This technique has been used to differentiate the dimeric and monomeric states of the erythropoietin receptor in the presence of different ligands and to confirm the trimeric state of dUTPase of feline immunodeficiency virus. In collaboration with S. Schmid, Department of Molecular Biology, we have determined conditions in which dynamin remains relatively monodisperse in its tetrameric form.

Initial crystals of gap junctions have been obtained in collaboration with B. Gilula, Department of Cell Biology.

PUBLICATIONS

Cabezas, E., Stanfield, R.L., Wilson, I.A., Satterthwait, A.C. Defining conformational requirements for the principle neutralizing determinant on HIV-1. In: Peptides: Chemistry, Structure and Biology. Proceedings of the 14th American Peptide Symposium. Kaumaya, P. (Ed.). ESCOM, Leiden, the Netherlands, 1996, p. 800.

Desmet, J., Wilson, I.A., Joniau, M., De Maeyer, M., Lasters, I. Computation of the binding of fully flexible peptides to proteins with flexible side chains. FASEB J. 11:164, 1997.

Garcia, K.C., Scott, C.A., Brunmark, A., Carbone, F.R., Peterson, P.A., Wilson, I.A., Teyton, L. CD8 enhances formation of stable T-cell receptor/MHC class I molecule complexes. Nature 384:577, 1996.

Ghiara, J.B., Ferguson, D.C., Satterthwait, A.C., Dyson, H.J., Wilson, I.A. Structure-based design of a constrained peptide mimic of the HIV-1 V3 loop neutralization site. J. Mol. Biol. 266:31, 1997.

Haynes, M.R., Heine, A., Wilson, I.A. Catalytic antibody structures: An early assessment. Isr. J. Chem. 36:133, 1996.

Haynes, M.R., Lenz, M., Taussig, M.J., Wilson, I.A., Hilvert, D. Sequence similarity and cross-reactivity of a Diels-Alder catalyst and an anti-progesterone antibody. Isr. J. Chem. 36:151, 1996.

Stura, E.A., Ruf, W., Wilson, I.A. Crystallization and preliminary crystallographic data for a ternary complex between tissue factor, factor VIIa, and a BPTI-derived inhibitor. J. Cryst. Growth 168:260, 1996.

White J.M., Hoffman, L.R., Arevalo, J.H., Wilson, I.A. Attachment and entry of influenza virus into host cells: Pivotal roles of the hemagglutinin. In: Structural Biology of Viruses. Chiu, W., Burnett, R.M., Garcea, R. (Eds.). Oxford Press, New York, 1997, p. 80.

Wu, S.K., Zeng, K., Wilson, I.A., Balch, W.E. Structural insights into the function of the Rab GDI superfamily. Trends Biochem. Sci. 21:472, 1996.

Zeng, Z.H., Castano, A.R., Segelke, B.W., Stura, E.A., Peterson, P.A., Wilson,I.A. Crystal structure of mouse CD1: An MHC-like fold with a large hydrophobic binding groove. Science, in press.

Top of Page


Principles of Protein Structure for Recognition, Interaction, and Function

E.D. Getzoff, A.S. Arvai, S.L. Bernstein, I.L. Canestrelli, B.R. Crane, T. Cross, C.L. Fisher, K.T. Forest, U.K. Genick, S. Hartsock, F. Henderson, C.K. Koike, T.P.K. Lo, S.E. Mylvaganam, C.D. Mol, E.H. Olender, J.L. Pellequer, M.E. Pique, M.M. Thayer

We determine the structural basis for protein recognition, function, and interaction by using x-ray crystallography and molecular biology, coupled with new computational and computer graphics approaches and tested by protein design. We focus on crystallographic studies for five proteins that undergo functionally important conformational and spectroscopic changes mediated by protein-cofactor interactions: photoactive yellow protein, to characterize a protein photocycle; sulfite reductase hemoprotein, to define the structural basis for the six-electron reduction of sulfite to sulfide; Root-effect hemoglobin, to study extreme pH effects on allostery and oxygen binding; Cu,Zn superoxide dismutase (SOD) metalloproteins, to understand their unusual stability and efficiency in regulating reactive oxygen; and, most recently, nitric oxide synthase, to understand the mechanism by which it oxidizes arginine to produce nitric oxide for biological signaling and defense.

To understand how a chromophore and protein interact to undergo a light cycle, we are studying photoactive yellow protein, a bacterial blue-light photosensor. Sequence homologies suggest that the fold in this protein is the structural prototype for the superfamily of Per-Arnt-Sim domains found in diverse biological sensors and clock proteins. We are extending our structure of dark-state photoactive yellow protein to 0.82-Å resolution, where individual atoms appear as spheres (Fig. 1). We used millisecond time-resolved Laue crystallography and simultaneous optical spectroscopy to determine the first atomic structure for a protein photocycle intermediate. The structures of the two different states of photoactive yellow protein reveal the synergistic interactions between the chromophore and the protein that tune the spectral and kinetic properties of the light cycle for efficient protein-mediated signal transduction. We made photoactive yellow proteins with site-directed mutations and modified chromophores by recombinant expression and chemical attachment of the chromophore to experimentally test hypotheses for the light-cycle mechanism.

Sulfite and nitrite reductases catalyze fundamental chemical transformations for biogeochemical cycling of sulfur and nitrogen. We determined the 1.6-Å crystallographic structure of sulfite reductase hemoprotein, which catalyzes the concerted six-electron reductions of sulfite to sulfide and nitrite to ammonia, by using multiwavelength anomalous diffraction of the native siroheme and Fe4S4 cluster cofactors, multiple isomorphous replacement, and selenomethionine sequence markers. A distinctive three-domain /ß fold controls assembly and reactivity of the cofactor and contains a sulfite or nitrite reductase repeat common to a redox-enzyme superfamily. Coupled spectroscopy and crystallography of the enzyme in three oxidation states showed that heme activation occurs via reduction-mediated ligand exchange. Examination of refined crystallographic complexes with substrates, inhibitors, intermediates, and products showed how the active site facilitates the reaction and accommodates the varied reaction intermediates without release (Fig. 2).

The fish Root-effect hemoglobins act as acid-controlled molecular oxygen pumps, delivering oxygen against high oxygen pressures to the swim bladder for neutral buoyancy and to the retina for visual acuity. Our 2-Å crystal structures of the ligand-bound hemoglobin from the fish Leiostomus xanthurus show that key residues of the hemoglobin recruit conserved residues to strategically assemble positive-charge clusters that promote the extremely pH-dependent allosteric RT switch with concomitant release of oxygen.

For SOD metalloenzymes, we solved structures of bacterial, bovine, and human mutant enzymes to characterize the structural basis of the enzymes' activity and stability. Structures of the oxidized and reduced states of bovine SOD and active-site mutants of human SOD provide information on the enzymes' mechanism. Our bacterial SOD structure is representative of the class of SODs from bacterial pathogens and provides the potential for drug design. Although the bacterial SOD subunit fold and active-site geometry match those of human SOD, the elements recruited to form the dimer interface, the active-site channel, and the disulfide bond are strikingly different. Our SOD structural results suggest a hypothesis for the mechanism by which single-site mutations in human SOD cause the fatal degenerative disease amyotrophic lateral sclerosis.

Antibody interactions are a final focus for characterizing protein recognition, function, and interaction. Structures of the E8 antibody in the free state and bound to cytochrome c provided information on conformational changes that occur on binding, electrostatic interactions, and the key role of water molecules at the antibody-antigen interface. With V. Roberts, Department of Molecular Biology, and S. Benkovic, Pennsylvania State University, we designed, constructed, and characterized metalloantibodies and catalytic antibodies that exploit the versatility of the sequence-variable but structurally conserved antibody scaffold.

PUBLICATIONS

Boissinot, M., Karnas, S., Lepock, J.R., Cabelli, D.E., Tainer, J.A., Getzoff, E.D., Hallewell, R.A. Function of the Greek key connection analysed using circular permutants of superoxide dismutase. EMBO J. 16:2171, 1997.

Bourne, Y., Redford, S.M., Steinman, H.M., Lepock, J.R., Tainer, J.A., Getzoff, E.D. Novel dimeric interface and electrostatic recognition. Proc. Natl. Acad. Sci. U.S.A. 93:12774, 1996.

Crane, B., Siegel, L., Getzoff, E.D. Probing the catalytic mechanism of sulfite reductase by x-ray crystallography: Structure of the E. coli hemoprotein in complex with substrates, inhibitors, intermediates and products. Biochemistry, in press.

Crane, B., Siegel, L., Getzoff, E.D. Structures of the siroheme and Fe4S4-containing active center of sulfite reductase in different states of oxidation: Heme activation via reduction-gated exogenous ligand exchange. Biochemistry, in press.

Crane, B.R., Bellamy, H., Getzoff, E.D. Multiwavelength anomalous diffraction of sulfite reductase hemoprotein: Making the most of MAD data. Acta Crystallogr. D53:8, 1997.

Crane, B.R., Getzoff, E.D. Determining phases and anomalous scattering models from the multiwavelength anomalous diffraction of native protein metal clusters: Improved MAD phases error estimates and anomalous scatterer positions. Acta Crystallogr. D53:23, 1997.

Crane, B.R., Getzoff, E.D. The relationship between structure and function for the sulfite reductases. Curr. Opin. Struct. Biol. 6:744, 1996.

Devanathan, S., Genick, U.K., Getzoff, E.D., Meyer, T.E., Cusanovich, M.A., Tolin, G. Preparation and properties of a 3,4-dihydroxycinnamic acid chromophore variant of the photoactive yellow protein. Arch. Biochem. Biophys. 340:83, 1997.

Fisher, C.L., Cabelli, D.E., Hallewell, R.A., Beroza, P., Lo, T.P., Getzoff, E.D., Tainer, J.A. Computational, pulse-radiolytic, and structural investigations of lysine-136 and its role in the electrostatic triad of human Cu,Zn superoxide dismutase. Proteins, in press.

Genick, U.K., Borgstahl, G.E.O., Ng, K., Ren, Z., Pradervand, C., Burke, P.M., Srajer, V., Teng, T.-T., Schildkamp, W., McRee, D.E., Moffat, K., Getzoff, E.D. Structure of a protein photocycle intermediate by millisecond time-resolved crystallography. Science 275:1471, 1997.

Genick, U.K., Devanathan, S., Meyer, T.E., Canestrelli, I.L., Williams, E., Cusanovich, M.A., Tollin, G., Getzoff, E.D. Active site mutants implicate key residues for control of color and light cycle kinetics of photoactive yellow protein. Biochemistry 36:8, 1997.

Top of Page


Structural Molecular Biology and Protein Design

J.A. Tainer, A.S. Arvai, D. Barondeau, G.E.O. Borgstahl, S.L. Bernstein, Y. Bourne, C. Bruns, B. Crane, T. Cross, D. Daniels, C.L. Fisher, K. Forest, Y. Guan, R.A. Hallewell, S. Han, F. Henderson, M.J. Hickey, D. Hosfield, C. Koike, T. Lo, T. Macke, C. Mol, S. Parikh, J.L. Pellequer, M.E. Pique, C. Putnam, S.M. Redford, M.M. Thayer

We focus on four classes of proteins at the interface of structural biology and cellular chemistry: (1) enzymes that regulate reactive oxygen and xenotoxins (superoxide dismutases [SODs], catalase, and glutathione transferase), (2) enzymes that control DNA repair and evolution enzymes, (3) pilus fiber motility and adherence proteins of bacterial pathogens, and (4) proteins that control the cell cycle. Current structural and design results are providing an understanding of these systems that should contribute to new treatments for infectious disease, degenerative diseases, and cancer.

BACTERIAL PILUS

We seek to improve our understanding of the many bacterial pathogens that use long fibers called type IV pili for attachment and mobility and to develop new treatments for infections caused by these microorganisms. We solved the crystallographic structure of pilin, the protein that forms the fiber. The fiber is essential to the virulence of these pathogens, which include Neisseria meningitidis, Neisseria gonorrhoeae, Pseudomonas aeruginosa, Dichlobacter nodosus, Moraxella bovis, Vibrio cholerae, and enterotoxogenic Escherichia coli.

The pilin subunit is ladle-shaped with an extended -helix spine wrapped at one end by a ß-sheet. Interactions between a few key residues allow the remaining part of a hypervariable region to undergo extreme antigenic variation to escape the host's immune response. The assembled fiber shows extreme sequence variation plus glycosylation and phosphorylation at the surface. However, we have defined two conserved regions recognized by human, mouse, and rabbit antibodies. Current efforts include the redesign of pilin to develop potential vaccines by replacing the hypervariable region with epitopes from other essential target proteins.

REACTIVE OXYGEN AND XENOBIOTIC CONTROL ENZYMES

Atomic structures of human cytoplasmic copper-zinc SODs, the mitochondrial manganese SODs, and schistosomal glutathione transferases are improving our understanding of the control of reactive oxygen and xenotoxins within cells. SODs are master regulators for reactive oxygen species involved in injury, pathogenesis, aging, and degenerative diseases.

For copper-zinc SODs, we defined the structural chemistry of the active site responsible for the rapid reaction. We are now examining how single-site mutations cause degenerative disease such as Lou Gehrig disease or familial amyotrophic lateral sclerosis. For manganese SOD, we found that single-site mutations can destabilize the tetramer and also reduce stability and activity in ways that may cause degenerative diseases.

Structures of glutathione-S-transferase, which is an essential detoxification enzyme in all organisms, showed how the leading antischistosomal drug praziquantel binds to this enzyme. This information may enable us to design new drugs to overcome the growing resistance to current antischistosomal drugs. Protein design of glutathione-S-transferases includes using random libraries for the loop regions surrounding the active site, thereby developing glubodies as a new class of binding proteins in biotechnology.

DNA REPAIR AND GENETIC EVOLUTION

Cells must balance DNA repair to preserve fidelity with DNA variation that allows evolutionary changes. Because more than 10,000 DNA bases per day are repaired in each human cell, DNA excision-repair enzymes are essential to cell survival and to protection against cancer-causing mutations. Surprisingly, DNA-repair inhibitors may improve current radiation therapy and chemotherapy for cancer by specifically killing cancer cells. Unlike normal cells, cancer cells often undergo DNA synthesis and cell division with unrepaired DNA, resulting in the death of the cells.

Our structures of DNA-repair enzymes show in atomic detail how damaged DNA bases are recognized and removed. These enzymes repair DNA by flipping the DNA nucleotides out from the double helix and into specific binding pockets (Fig. 1), which are ideal for the design of inhibitors for anticancer therapies. We confirmed our understanding of these binding pockets by deliberately altering the specificity of the DNA-repair enzyme uracil-DNA glycosylase. We made mutants that remove cytosine or thymine from normal DNA, resulting in mutator phenotypes in vivo.

We found that the endonuclease III structure is representative of a superfamily of DNA-repair enzymes and a key HhH motif that recognizes DNA backbone. Our new structure of the major DNA-repair abasic-site endonuclease, which cuts DNA at sites where bases are missing, defines the active site of the enzyme and indicates its mechanism for recognizing missing bases.

Human dUTP pyrophosphatase (dUTPase) catalyzes the breakdown of uracil nucleotide triphosphates to keep the RNA base uracil out of DNA and to provide material for the biosynthesis of the DNA building block dTTP. These dUTPase functions prevent cycles of uracil misincorporation and removal that would generate multiple breaks in DNA strands and eventual cell death, a process called thymine-less cell death. Atomic structures of dUTPase with bound nucleotides show that uracil binds within a groove that is then capped when the flexible tail region closes over the bound dUTP substrate. These structures establish how dUTPase recognizes its substrate with exquisite specificity and provide a basis for the design of inhibitors as future anticancer drugs.

CONTROL OF THE CELL CYCLE

Together with S. Reed's group, Department of Molecular Biology, we are working to define the structural basis for control of the cell cycle. Structures of the Cks or suc1 proteins, which are essential to the progression of the cell cycle, provide clues for new mechanisms for regulation of the cycle via a conformational switch that controls two distinct Cks folds and assemblies. A straight ß-hinge conformation of Cks, which forms a dimer of swapped ß-strands, blocks binding to the cell-cycle kinase Cdk2. Formation of a closed, bent ß-hinge conformation creates a single domain fold that promotes Cdk2 binding (Fig. 2). Preliminary experiments by Dr. Reed's group show that blocking expression of Cks results in cell death for several types of cancer cells. This finding suggests that Cks is a useful target for the development of anticancer drugs.

PUBLICATIONS

Boissinot, M., Karnas, S., Lepock, J.R., Cabelli, D.E., Tainer, J.A., Getzoff, E.D., Hallewell, R.A. Function of the Greek key connection analysed using circular permutants of superoxide dismutase. EMBO J., in press.

Bourne, Y., Redford, S.M., Steinman H.M., Lepock, J.R., Tainer, J.A., Getzoff, E.D. Novel dimeric interface and electrostatic recognition in bacterial Cu,Zn superoxide dismutase. Proc. Natl. Acad. Sci. U.S.A. 93:12774, 1996.

Crane, B.R., Arvai, A.S., Gachhui, R., Wu, C., Ghosh, D.K. The structure of NO synthase oxygenase domain and inhibitor complexes. Science, in press.

Fisher, C.L., Cabelli, D.E., Hallewell, R.A., Beroza, P., Lo, T.P., Getzoff, E.D., Tainer, J.A. Computational, pulse-radiolytic and structural investigations of lysine 136 and its role in the electrostatic triad of human of Cu,Zn superoxide dismutase. Proteins, in press.

Forest, K.T., Tainer, J.A. Type IV pilin structure, assembly, and immunodominance: Applications to vaccine design. In: Vaccines. Brown, F., et al. (Eds.). Cold Spring Harbor Press, Cold Spring Harbor, NY, 1997, p. 167.

Forest, K.T., Tainer, J.A. Type IV pilus structure: Outside to inside and top to bottom. Gene 192:165, 1997.

Gorman, M.A., Morera, S., Rothwell, D.G., La Fortelle, E.D., Mol, C.D., Tainer, J.A., Hickson, I.D., Freemont, P.S. The crystal structure of the human DNA-repair enzyme endonuclease HAP1 suggests the recognition of extra-helical deoxyribose at DNA abasic sites. EMBO J., in press.

Mol, C.D., Harris, J.M., McIntosh, E.M., Tainer, J.A. Crystal structures of free and nucleotide-bound complexes of human dUTP pyrophosphatase: Uracil recognition by a ß-hairpin and active sites formed by three separate subunits. Structure 4:1077, 1996.

Roberts, V.A., Nachman, R.J., Coast, G.M., Hariharan, M., Chung, J.S., Holman, G.M., Williams, H., Tainer, J.A. Consensus chemistry and ß-turn conformation of the active core of the myotropic/diuretic insect neuropeptide family. Chem. Biol. 4:105, 1997.

Slupphaug, G., Mol, C.D., Kavli, B., Arvai, A.S., Krokan H.E., Tainer, J.A. A nucleotide flipping mechanism from the structure of human uracil-DNA glycosylase bound to DNA. Nature 384:87, 1996.

Watson, M.H., Bourne, Y., Arvai, A.S., Hickey, M.J., Santiago, A., Bernstein, S.L., Tainer, J.A., Reed, S.I. A mutation in the human cyclin-dependent kinase interacting protein, CksHs2, interferes with cyclin-dependent kinase binding and biological function, but preserves protein structure and assembly. J. Mol. Biol. 261:646, 1996.

Zu, J.S., Deng, H.-X., Lo, T.P., Mitsumoto, H., Ahmed, M.S., Hung, W.-Y., Cai, Z.-J., Tainer, J.A., Siddique, T. Exon 5 is not required for the toxic function of mutant SOD1 but essential for dismutase activity: Identification and characterization of two new SOD1 mutations associated with familial amyotrophic lateral sclerosis. Neurogenetics, in press.

Top of Page


Crystallography of Iron Metalloproteins

D.E. McRee, C. Bruns, Y. Cao, M. Israel, N. Jourdan, V. Shridhar, P. Williams

Our laboratory studies the structure, function, and catalysis of metalloproteins, with particular emphasis on iron metalloproteins. One question being investigated is how proteins give iron diverse biological roles ranging from electron transfer to oxygen activation to iron transport. We are using high-resolution protein crystallography, protein engineering, and biochemistry to answer this question.

IRON-BINDING PROTEIN

Iron is an essential growth requirement of all organisms, and using lactoferrin and transferrin to sequester iron from invading pathogens is one means of antibacterial defense in humans. However, some bacterial pathogens, notably Neisseria and Haemophilus, have evolved to turn adversity into advantage; they have receptors for transferrins that enable the bacteria to steal the host iron. This iron is transported into the pathogens by a bacterial protein, iron-binding protein. We have solved the structure of the iron-binding protein to high resolution (Fig. 1) and have started protein engineering studies. Our eventual goal is to find a way to knock out this protein in bacteria and thus produce a bacteriostatic agent.

ELECTRON TRANSFER

Cytochrome c552 is the physiologic electron transfer partner of cytochrome c oxidase in Thermus thermophilus. We are solving the structure of cytochrome c552 in collaboration with E. Stura, Department of Molecular Biology, and J. Fee, University of California, San Diego. We also have crystals of the CuA fragment of T. thermophilus cytochrome c oxidase, and the combination of this fragment and cytochrome c552 provides a unique opportunity to study an electron transfer system. We will do protein engineering studies in collaboration with J. Fee.

VERY HIGH-RESOLUTION METAL-SITE STRUCTURES

With the advent of freezing devices and synchrotron radiation sources, very high-resolution structures (<1.4 Å) are now routinely accessible. At these resolutions, it becomes possible to refine structures with the rigorous methods used in small-molecule crystallography. This refinement includes adding hydrogens and anisotropic thermal factors and using full-matrix least-squares analyses to determine the standard uncertainties of the atom positions. Because metal centers scatter more than the lighter carbon atoms do, the metal centers are even better determined.

In a high-resolution 1.35-Å refinement of Azotobacter seven-iron ferredoxin, we have determined the positions of the iron and sulfur atoms with a standard uncertainty of 0.01 Å (Fig. 2). This uncertainty is within the limit of molecular orbital calculations and will lead to a better coupling of theoretical calculations on metal centers with structure. We have also refined cytchrome c peroxidase to 1.38-Å resolution. These more accurate coordinates will allow improved calculations of the enzyme's mechanism. In the coming years, we plan to build a database of high-resolution metalloprotein structures in conjunction with the Scripps Metalloprotein Structure and Design Group (http://www.scripps.edu/pub/dem-web/metallo/), of which our group is a member.

COMPUTATIONAL CRYSTALLOGRAPHY

As part of a project funded by the National Science Foundation to improve visualization and computational tools for protein crystallography, we have developed a software package called XtalView (http://www.scripps.edu/pub/dem-web/toc.html). XtalView uses the power of modern workstations to solve protein crystallographic problems by using a visual, graphical user interface. XtalView is available from the Computational Center for Macromolecular Structure (http://www.sdsc.edu/CCMS/) and has been downloaded by more than 1000 groups.

PUBLICATIONS

Cao, Y., Musah, R.A., Wilcox, S.A., Goodin, D.B., McRee, D.E. Protein conformer selection by ligand binding observed with protein crystallography. Protein Sci., in press.

Genick, U.K., Borgstahl, G.E.O., Ng, K., Ren, Z., Pradervand, C., Burke, P.M., Srajer, V., Teng, T., Shildkamp, W., McRee, D.E., Moffat, K., Getzoff, E. D. Millisecond time-resolved Laue crystallography: Structure of a protein photocycle intermediate. Science 275:1471, 1997.

Israel, M., McRee, D.E. XtalView. In: Crystallographic Computing 7. Bourne, P., Watenpaugh, K. (Eds.). Oxford Press, New York, in press. Available at http://www.sdsc.edu/projects/Xtal/IUCr/CC/School96/IUCr.html

Prasad, G.S., McRee, D.E., Stura, E.A., Levitt, D.G., Lee, H.C., Stout, C.D. Crystal structure of Aplysia ADP ribosyl cyclase, a homologue of the bifunctional ectozyme CD38. Nature Struct. Biol. 3:957, 1996.

Prasad, G.S., Stura, E.A., McRee, D.E., Laco, G.S., Hasselkus-Light, C., Elder, J.H., Stout, C.D. Crystal structure of dUTP pyrophosphatase from feline immunodeficiency virus. Protein Sci. 5:2429, 1996.

Top of Page


Crystallography of Biological Macromolecules

C.D. Stout, G.S. Prasad, S.J. Lloyd, P.J. Shim, N. Kresge, A. Muhlberg, V. Sridhar

This laboratory focuses on experimental x-ray crystallography of macromolecules. Several fundamental questions are being addressed through structure determination of key proteins involved in biological processes. This research often involves collaboration with scientists at TSRI and other institutions. The experimental work has several stages, including biochemical preparation, crystallization, and collection and analysis of x-ray diffraction data. Once a structure is solved, experiments are designed to study relationships between structure and function. These experiments entail preparation of site-directed mutants and protein-ligand complexes, structure analysis, and assays of biological function. Several projects focused primarily on iron-sulfur proteins and enzymes, fertilization proteins, and enzymes that use nucleotides are in progress.

In collaboration with J. Elder and E. Stura, Department of Molecular Biology, we have determined at high resolution the structure of dUTPase of feline immunodeficiency virus and of dUTPase in complex with dUMP. The enzyme dUTPase is an important target for drug design, and a series of mutant and inhibitor complexes are being analyzed to understand the affinity and specificity of small-molecule ligands in the active site and the catalytic mechanism. Experiments also are in progress to study the binding of substrates to ADP ribosyl cyclase, an enzyme that synthesizes the secondary messenger cyclic ADP ribose. ADP ribosyl cyclase appears to use a novel dual-site mechanism. Understanding this reaction is relevant to the function of the cyclase homolog, CD38, a widely distributed cell-surface ectozyme in the immune system. A third nucleotide-processing enzyme being studied is dynamin, a critical component in cellular endocytosis. Through collaboration with S. Schmid, Department of Molecular Biology, we are determining the structure of the domains of this complex multifunctional GTPase.

Continuing our focus on proteins involved in fertilization, in collaboration with V. Vacquier, Scripps Institution of Oceanography, we are determining the structure of the 16,000 and 18,000 molecular weight lysins from green abalone. These structures will provide important insights into the molecular details of the interaction between sperm and egg. The 16,000 molecular weight protein is a species-specific homolog of red abalone lysin, for which we have two crystal structures; the 18,000 molecular weight lysin is a novel protein involved in fusion of the gamete plasma membranes. In collaboration with J. Bleil, Department of Molecular Biology, experiments are in progress to determine the structure of the octameric mouse sperm cell-surface protein, sp56, which functions in recognition and binding of egg oligosaccharides.

Detailed study of the mechanism of the iron-sulfur enzyme aconitase, a dehydratase that uses an iron-sulfur cluster in catalysis, is continuing with the determination of the structures of five site-directed mutants of active-site residues in eight complexes with substrates and inhibitors. A large number of crystallization trials are in progress with cytosolic aconitase and the stem-loop RNA molecule, the iron regulatory element, to which aconitase binds.

In collaboration with B. Burgess, University of California, Irvine, an ongoing study of the structure and function of the seven-iron ferredoxin from Azotobacter vinelandii has led to additional structures of mutants. One of the mutants has a cysteine persulfide, a possible intermediate in the biosynthesis of iron-sulfur clusters. A series of six structures of the seven-iron ferredoxin oxidized with ferricyanide have been refined to study the chemical decomposition of iron-sulfur clusters. In collaboration with D. McRee, Department of Molecular Biology, the seven-iron ferredoxin has been refined at 1.35-Å resolution, providing an unprecedented level of accuracy in the structural details of the [3Fe-4S] and [4Fe-4S] clusters (see Fig. 2 in D. McRee's report). The redox partner of ferredoxin is NADPH:ferredoxin oxidoreductase; determination of the crystal structure of this enzyme, nearly complete, will allow docking of the proteins to model the electron transfer reaction.

PUBLICATIONS

Beinert, H., Kennedy, M.C., Stout, C.D. Aconitase as iron-sulfur protein, enzyme, and iron-regulatory protein. Chem. Rev. 96:2335, 1996.

Kemper, M.A., Stout, C.D., Lloyd, S.J., Prasad, G.S., Fawcett, S., Armstrong, F.A., Shen, B., Burgess, B.K. Y13C Azotobacter vinelandii ferredoxin I: A designed [Fe-S] ligand motif contains a cysteine persulfide. J. Biol. Chem. 272:15620, 1997.

Lauble H., Kennedy M.C., Emptage M.H., Beinert H., Stout, C.D. The reaction of fluorocitrate with aconitase and the crystal structure of the enzyme-inhibitor complex. Proc. Natl. Acad. Sci. U.S.A. 93:13699, 1996.

Prasad, G.S., McRee, D.E., Stura, E.A., Levitt, D.G., Lee, H.C., Stout, C.D. Crystal structure of Aplysia ADP ribosyl cyclase, a homologue of the bifunctional ectozyme CD38. Nature Struct. Biol. 3:957, 1996.

Prasad, G.S., Stura, E.A., McRee, D.E., Laco, G.S., Hasselkus-Light, C., Elder, J.H., Stout, C.D. Crystal structure of dUTP pyrophosphatase from feline immunodeficiency virus. Protein Sci. 5:2429, 1996.

Top of Page


Nuclear Magnetic Resonance Investigations of the Three-Dimensional Structure and Dynamics of Proteins in Solution

P.E. Wright, H.J. Dyson, B. Duggan, M. Foster, S. Holmbeck, R. Kriwacki, J. Love, J. Laity, B. Lee, G. Legge, G. Perez-Alvarado, J. Pikkemaat, I. Radhakrishnan, J. Xu, L. Zhu, L.L. Tennant, M. Martinez-Yamout, J. Chung, M. Gearhart, D.A. Case, H.J. Dyson, J. Gottesfeld, U. Hommel,* T. Huang**

* Novartis Pharmaceuticals, Basel, Switzerland
** Institute of Biomedical Sciences, Taipei, Taiwan, Republic of China

We use multidimensional nuclear magnetic resonance (NMR) spectroscopy to investigate the structures and dynamics of proteins in solution. Such studies are essential for understanding the mechanisms of action of these proteins and for elucidating structure-function relationships.

PROTEIN STRUCTURE DETERMINATION IN SOLUTION

Solution three-dimensional structures have been determined for a number of proteins and protein complexes of molecular weight up to more than 20 kD. These include plastocyanin, thioredoxin, the human anaphylatoxin C3a, enzyme IIAglc, myoglobin, rusticyanin, and various DNA-binding domains and protein-DNA complexes. We are beginning heteronuclear three- and four-dimensional NMR experiments with proteins labeled with 2H, 13C, or 15N to extend use of the NMR structure determination method to even larger proteins and protein complexes. In addition, new computational methods are being developed to facilitate structure determination and refinement.

TRANSCRIPTION FACTOR--DNA COMPLEXES

NMR methods are being used to determine the three-dimensional structures and intramolecular dynamics of various DNA-binding motifs from eukaryotic transcriptional regulatory proteins, both free and complexed with the target DNA. Structures determined include that of the HMG domain of the lymphoid enhancer-binding factor 1 (LEF-1) bound to its cognate DNA sequence and the DNA complex formed by the three amino-terminal zinc fingers of transcription factor IIIA (TFIIIA). Ongoing work on LEF-1 has focused on the structural characteristics of the free protein and on the role of the C-terminal basic tail in DNA binding and bending. In addition, studies of the complex of LEF-1 with a Holliday junction are in progress.

Three-dimensional structures have been determined for the complex of zf1-3, a protein containing the three N-terminal zinc fingers of TFIIIA, with the cognate 15-bp oligonucleotide duplex. The three zinc fingers bind in the DNA major groove, thus validating parts of our earlier model of the TFIIIA-DNA complex deduced from biochemical experiments. The structures contain several novel features and show that prevailing models of DNA recognition, which assume that zinc fingers are independent modules that contact bases through a limited set of amino acids, are outmoded.

Each zinc finger contacts 4--5 bp, and the repertoire of base contact residues is expanded in the zf1-3--DNA complex. Lysine and histidine side chains involved in base recognition are dynamically disordered in the solution structure, thereby minimizing the entropic cost of DNA binding. On DNA binding, the protein forms an ordered globular structure with substantial protein-protein interactions between adjacent fingers. Dynamics measurements show that the linker regions that connect the zinc-finger domains lose their intrinsic flexibility on binding to DNA. The linkers appear to play an active structural role in stabilization of the protein-DNA complex. Contributions to high-affinity binding come from both direct protein-DNA contacts and from indirect protein-protein interactions associated with structural organization of the linkers and formation of well-packed interfaces between adjacent zinc fingers in the DNA complex.

In addition to its role in binding to and regulating the 5S RNA gene, TFIIIA also forms a complex with the 5S RNA transcript. The minimal set of zinc fingers required for 5S RNA recognition has been determined and consists of fingers 4--6. The minimal region of the 5S RNA needed to bind these fingers has also been mapped, and NMR studies of the protein-RNA complex have been started to gain insights into the structural basis for 5S RNA recognition. Work is also in progress to determine the structure of finger 6 and examine its interactions with RNA. In a further new direction, NMR studies of the Wilms' tumor zinc-finger protein and the zinc finger--DNA complex have commenced.

In collaboration with R. Evans, Salk Institute, the solution structure of the DNA-binding domain of the nuclear hormone receptor RXR, which is activated by 9-cis retinoic acid, has been refined by using 13C,15N-labeled protein. The refined structures confirm the existence of the novel third helix, which appears to play a role in homodimer formation and DNA binding. NMR structural studies of the DNA-binding domain of ERR, a member of a class of nuclear hormone receptors that bind DNA as monomers, are also in progress in collaboration with Dr. Evans. The protein has been uniformly labeled with 13C and 15N, and resonance assignments have been made for free ERR and for the complex with its cognate DNA.

PROTEIN-PROTEIN INTERACTIONS IN TRANSCRIPTIONAL REGULATION

Activated transcription in eukaryotes relies on protein-protein interactions between DNA-bound factors and coactivators that, in turn, interact with the basal transcription machinery. The nuclear factor cyclic AMP response element binding protein (CREB) activates transcription of target genes, in part through direct interactions with the KIX domain of the coactivator CREB-binding protein (CBP) in a phosphorylation-dependent manner. In collaboration with M. Montminy, Harvard Medical School, we recently determined the structure of the complex formed by the phosphorylated kinase-inducible domain (pKID) of CREB and the KIX domain of CBP (Fig. 1).

The structure reveals that pKID undergoes a coil-to-helix folding transition on binding to KIX, forming two -helices in the process. The amphipathic helix B of pKID interacts with a large hydrophobic patch defined by helices 1 and 3 of KIX. The other pKID helix, A, interacts with a different face of the 3 helix of KIX. The critical phosphoserine residue of pKID is in a position to interact favorably with Y658 of KIX. The structure provides a model for interactions between other transactivation domains and their targets.

PROTEIN-PROTEIN INTERACTIONS

A novel method has been developed, in collaboration with G. Siuzdak, Department of Molecular Biology, for probing protein-protein interactions by using mass spectrometry and isotopic labeling. Application of this method led to detection of the kinase inhibitory domain of the cyclin-dependent kinase inhibitor p21Waf1/Cip1/Sdi1. NMR studies of this important cell-cycle regulatory protein established that it is disordered in the free state but adopts a stable folded structure when bound to cyclin-dependent kinase 2. These observations challenge the generally accepted view that stable tertiary structure is a prerequisite for biological activity and suggest that a broader view of protein "structure" should be considered in the context of structure-function relationships.

INTERACTIONS BETWEEN DOMAINS OF CELL ADHESION MOLECULES

We recently completed resonance assignments for the I (inserted) domain of lymphocyte function--associated antigen-1, and structure calculations are in progress. Our interest in this system is in the understanding of the interactions between lymphocyte function--associated antigen-1 and its physiologic partner intracellular adhesion molecule-1. This research is being done in parallel with a collaborative effort with B. Cunningham, Department of Neuropharmacology, to probe the interactions between the immunoglobulin-like domains of the neural cell adhesion molecule. These projects illustrate the excellent sensitivity of the NMR method as a tool to probe protein-protein interactions. The NMR spectrum can be used directly to map the sites of interaction in a straightforward way, once assignments are available.

PUBLICATIONS

Chen, Y., Case, D.A., Reizer, J., Saier, M.H., Jr., Wright, P.E. High-resolution solution structure of Bacillus subtilis IIAglc. Proteins, in press.

Foster, M.P., Wuttke, D.S., Radhakrishnan, I., Case, D.A., Gottesfeld, J.M., Wright, P.E. Domain packing and dynamics in the DNA complex of the amino-terminal zinc fingers of transcription factor IIIA. Nature Struct. Biol., in press.

Gippert, G.P., Wright, P.E., Case, D.A. Distributed torsion angle grid search in high dimensions: A systematic approach to NMR structure determination. J. Biomol. NMR, in press.

Kriwacki, R.W., Hengst, L., Tennant, L., Reed S.I., Wright, P.E. Structural studies of p21 Waf1/Cip1/Sdi1 in the free and Cdk2-bound state: The p21 NH2-terminus folds on binding to Cdk2. Proc. Natl. Acad. Sci. U.S.A. 93:11504, 1996.

Kriwacki, R.W., Wu, J., Tennant, L., Wright P.E., Siuzdak, G. Probing protein structure using biochemical and biophysical methods: Proteolysis, MALDI mass analysis, HPLC, and gel-filtration chromatography of p21 Waf1/Cip1/Sdi11. J. Chromatogr., in press.

Markley, J.L., Bax, A., Arata, Y., Hilbers, C.W., Kaptein, R., Sykes, B.D., Wright, P.E., Wüthrich, K. Recommendations for the presentation of NMR structures of proteins and nucleic acids. Pure Appl. Chem., in press.

Sem, D.S., Casimiro, D.R., Kliewer, S.A., Provencal, J., Evans, R.M., Wright, P.E. NMR spectroscopic studies of the DNA-binding domain of the monomer-binding nuclear orphan receptor, human ERR2: The carboxy-terminal extension to the zinc-finger region is unstructured in the free form of the protein. J. Biol. Chem. 272:18038, 1997.

Wright, P.E. Smith, J.L. Biophysical methods: Faster and bigger. Curr. Opin. Struct. Biol. 6:583, 1996.

Wuttke, D.S., Foster, M.P., Case, D.A., Gottesfeld, J.M., Wright, P.E. Solution structure of the first three zinc fingers of TFIIIA bound to the cognate DNA sequence: Determinants of affinity and sequence specificity. J. Mol. Biol., in press.

Zhu, L., Dyson, H.J., Wright, P.E. A NOESY-HSQC simulation program, SPIRIT. J. Biomol. NMR, in press.

Top of Page


Folding of Proteins and Protein Fragments

P.E. Wright, H.J. Dyson, Y. Bai, S. Cavagnero, D. Donne, D. Eliezer, C. Garcia-Gonzalez, S. Prytulla, J. Viles, J. Yao, O. Zhang, J. Chung, L.L. Tennant, S. Lahrichi, V. Tsui

The molecular mechanism by which proteins fold into their three-dimensional structures remains one of the most important unsolved problems in structural biology. Nuclear magnetic resonance (NMR) spectroscopy is uniquely suited to provide information on the structure of transient intermediates formed during protein folding. We have used NMR methods to show that many peptide fragments of proteins tend to adopt folded conformations in water solution. The observation of transiently populated folded structures, including reverse turns, helices, nascent helices, and hydrophobic clusters, in water solutions of short peptides has important implications for initiation of protein folding. Formation of elements of secondary structure probably plays an important role in the initiation of protein folding by reducing the number of conformations that must be explored by the polypeptide chain and by directing subsequent folding pathways.

APOMYOGLOBIN FOLDING PATHWAY

A major program in our laboratory is to establish a structural and mechanistic description of the folding pathway of apomyoglobin. We used quenched-flow pulse-labeling methods in conjunction with two-dimensional NMR spectroscopy to map the kinetic folding pathway of the wild-type protein. With these methods, we showed that an intermediate in which the A-, G-, and H-helices adopt hydrogen-bonded secondary structure is formed within 6 msec of the initiation of refolding. Folding proceeds by stabilization of structure in the B-helix and then the C- and E-helices. Using time-resolved small-angle x-ray scattering, in collaboration with S. Doniach and K.O. Hodgson, Stanford University and Stanford Synchrotron Radiation Laboratory, we showed that apomyoglobin folds into a highly compact state within 20 msec after initiation of refolding. We are now using carefully selected myoglobin mutants and both optical stopped-flow spectroscopy and hydrogen-exchange pulse-labeling methods to further probe the kinetic folding pathway. We have detected mutations that increase the rate of folding and apparently influence the folding pathway.

Apomyoglobin provides a unique opportunity for detailed characterization of the incremental development of structure and of the changes in dynamics that accompany compaction of a protein during folding. By careful manipulation of the pH, a number of partially folded states that are directly implicated in folding can be stabilized under conditions suitable for direct study by multidimensional NMR. Backbone resonance assignments have now been completed for an apomyoglobin molten globule intermediate, formed at pH 4.1. Analysis of 13C and other chemical shifts has provided the first "high resolution" insights into the structure of this state at the level of individual amino acid residues. In addition, 15N relaxation measurements have shown that the terminal regions of the polypeptide chain form a tightly packed hydrophobic core, whereas the central parts retain considerable intrinsic flexibility.

In earlier work, we used peptides to model the initiation events in the folding of apomyoglobin. Peptides covering the entire sequence of myoglobin were synthesized, and their propensities for spontaneous formation of secondary structure were examined by using NMR and circular dichroism spectroscopy. We have now fully assigned the polypeptide backbone resonances for the acid-denatured state of apomyoglobin. The NMR data show formation of helical secondary structure in regions that form the A- and H-helices in the folded protein. The acid-denatured state is an excellent model for the fluctuating local interactions that lead to the transient formation of unstable elements of secondary structure and local hydrophobic clusters during the earliest stages of folding.

The view of protein folding that results from our work on apomyoglobin is one in which collapse of the polypeptide chain to form increasingly compact states leads to progressive accumulation of secondary structure and increasing restriction of fluctuations in the polypeptide backbone. Chain flexibility is greatest at the earliest stages of folding, in which transient elements of secondary structure and local hydrophobic clusters are formed. As the folding protein becomes increasingly compact, backbone motions become more restricted, the hydrophobic core is formed and extended, and nascent elements of secondary structure are progressively stabilized. The ordered tertiary structure characteristic of the native protein, with well-packed side chains and relatively low-amplitude local dynamics, appears to form rather late in folding.

PLASTOCYANIN FOLDING

Structural characterization of an unfolded state of the ß-sheet protein apoplastocyanin is also in progress. Apoplastocyanin offers a unique opportunity to study an unfolded protein under nondenaturing conditions, because it forms an unfolded state at neutral pH under low-salt conditions. Extensive resonance assignments have been completed, revealing that the unfolded polypeptide has a marked propensity to populate the ß-region of dihedral angle space. NMR relaxation measurements indicate that the polypeptide backbone is highly fluctional on a timescale shorter than 1 nsec.

FRAGMENTS OF PRION PROTEINS

A number of proteins are unfolded or only partly folded before they bind to their biological receptors. Although scientists have known for many years that this fact applies to peptide hormones, it is now evident that on binding to their receptors, the activation domains of some transcription factors, the p21 cyclin-dependent kinase inhibitors, and certain other proteins undergo transitions to folded states. Furthermore, folding has been implicated in a number of abnormalities, such as Alzheimer's disease and the prion diseases such as bovine spongiform encephalopathy (mad cow disease).

We recently used NMR methods to characterize the structure and dynamics of the full-length prion protein. The N-terminal region, residues 29--124, including the octapeptide repeats, is highly disordered and undergoes rapid backbone fluctuations on a subnanosecond timescale. This highly flexible region probably provides the plasticity required for the conformational transition of the cellular form into the infectious scrapie state.

PUBLICATIONS

Bai, Y., Karimi, A., Dyson, H.J., Wright, P.E. Absence of a stable intermediate on the folding pathway of protein A. Protein Sci. 6:1449, 1997.

Donne, D.G., Viles, J.H., Groth, D., Mehlhorn, I., James, T.L., Cohen, F.E., Prusiner, S.B., Wright, P.E., Dyson, H.J. Structure of the recombinant full-length hamster prion protein PrP(29--231): The N-terminus is highly flexible. Proc. Natl. Acad. Sci. U.S.A., in press.

Eliezer D., Wright, P.E. Is apomyoglobin a molten globule? Structural characterization by NMR. J. Mol. Biol. 263:531, 1996.

Prytulla, S., Dyson, H.J., Wright, P.E. Gene synthesis, high level expression and assignment of backbone 15N and 13C resonances of soybean leghemoglobin. FEBS Lett. 399:283, 1996.

Reymond, M.T., Huo, S., Duggan, B., Wright, P.E., Dyson, H.J. Contribution of increased length and intact capping sequences to the conformational preference for helix in a 31-residue peptide from the C-terminus of myohemerythrin. Biochemistry 36:5234, 1997.

Reymond, M.T., Merutka, G., Dyson, H.J., Wright, P.E. Folding propensities of peptide fragments of myoglobin. Protein Sci. 6:706, 1997.

Yao, J., Dyson, H.J., Wright, P.E. Chemical shift dispersion and secondary structure prediction in unfolded and partly-folded proteins. FEBS Lett., in press.

Top of Page


Nuclear Magnetic Resonance Studies of the Structure and Mechanism of Enzymes

H.J. Dyson, P.E. Wright, J. Chung, S. Huo, G. Kroon, M. Martinez-Yamout, M.J. Osborne, M. Reymond, S. Scrofani, B. Xia, L.L. Tennant, S.J. Benkovic,* A. Holmgren**

* Pennsylvania State University, University Park, PA
** Karolinska Institute, Stockholm, Sweden

We use site-specific information from nuclear magnetic resonance (NMR) spectroscopy to further the understanding of enzyme function through study of structure and dynamics.

THIOREDOXIN

Thioredoxin, a small (108-residue) thiol-disulfide oxidoreductase, has a multitude of functions in the cell, including the vital reduction of ribonucleotides to form deoxyribonucleotides for DNA synthesis. Thioredoxins occur in all living organisms, including viruses. Mammalian thioredoxins have a vital role in cellular control mechanisms and have been implicated in human diseases; serum levels of the enzyme are elevated in AIDS patients. One of the primary functions of thioredoxin in the cell is as a protein disulfide reductase, a function vital for the prevention of misfolded proteins in vivo.

The Escherichia coli thioredoxin system has been fully characterized by NMR, including the calculation of high-resolution structures for both the oxidized (disulfide) and the reduced (dithiol) forms of the protein. Backbone dynamics and amide proton hydrogen exchange enabled us to determine that functional differences in phage systems between oxidized and reduced thioredoxin were due to differences in the flexibility of the molecules, rather than to structural differences.

We have also delineated the mechanism of action of E. coli thioredoxin. The reduction reaction of thioredoxin depends critically on the movement of protons during the two-electron--two-proton transfer reaction as a substrate disulfide is reduced. An extensive series of experiments on the pH-dependence of the NMR spectrum has given us important insights into the complex relationship between the mechanism of action of this enzyme and local structure at the active site, including the presence of conserved, buried charged residues. This project has recently been extended to include a structural study by NMR of a related protein, glutaredoxin 2, whose structure and function are unknown.

DYNAMICS IN ENZYME ACTION

The relationship between dynamics of the polypeptide chain and enzyme catalysis is being explored. We hypothesize that efficient enzyme catalysis requires flexibility at the active site and that enzymes have therefore evolved to incorporate this flexibility. To test these hypotheses, we are using mutagenesis coupled with detailed characterization of changes in enzymatic function, structure, and dynamics in two very different enzyme systems: dihydrofolate reductase, for which we already have clear evidence that dynamics play a role in catalysis, and a metallo-ß-lactamase, for which evidence from the crystal structure indicates flexibility in the region of the active site.

NMR relaxation measurements on substrate and cofactor complexes of dihydrofolate reductase, in collaboration with S. Benkovic, Pennsylvania State University, are providing novel insights into the relationship between dynamics and enzyme activity. Motions on a wide range of time scales (picoseconds to milliseconds) have been detected and can be correlated with enzyme function. Both of these enzymes are clinically important drug targets, for anticancer drugs in the case of dihydrofolate reductase and for prevention of antibiotic resistance in the case of the lactamase. Our approach is leading to new insights into the role of dynamics in catalysis.

DESIGN AND SYNTHESIS OF PROTEINS

We have developed an efficient method of overproducing proteins for NMR that involves designing and synthesizing genes specifically for expression in E. coli. This method is exceptionally useful in the production of several proteins, including Thiobacillus ferrooxidans rusticyanin and soybean leghemoglobin. The method is also easily adaptable for the production of site-specific mutants, an important aspect of the investigation of the relationship between structure and function.

DESIGN OF A CATALYTIC ANTIBODY

Related to the issue of design for catalysis is a long-standing project on the characterization and redesign of the Fv fragment of a catalytic antibody. This work is extremely promising for understanding the influence of structure on function for enzymes. Because they have evolved over millions of years, most enzymes are exquisitely tuned to the reactions they catalyze and may also be tolerant of mutations. By contrast, catalytic antibodies have much lower efficiency and specificity. A knowledge of the local structure and dynamics of the catalytic site will allow novel insights into the mechanisms of antibody catalysis and will guide future work aimed at enhancing the catalytic efficiency. These studies will also provide valuable insights into the structural and functional evolution of enzymes.

PUBLICATIONS

Bender, C.J., Casimiro, D., Dyson, H.J. Electron spin envelope modulation spectra of Thiobacillus ferrooxidans rusticyanin and a mutant lacking one of the copper ligands. J. Chem. Soc. [Faraday], in press.

Casimiro, D., Wright, P.E., Dyson, H.J. PCR-based gene synthesis for protein over-production. Structure, in press.

Dyson, H.J., Jeng, M.-F., Slaby, I., Lindell, M., Cui, D.-S., Holmgren, A. Effects of buried charged groups on cysteine thiol ionization and reactivity in Escherichia coli thioredoxin: Structural and functional characterization of mutants of Asp 26 and Lys 57. Biochemistry 36:2622,1997.

Ghiara, J.B., Ferguson, D., Satterthwait, A.C., Dyson, H.J., Wilson, I.A. Structure-based design of a constrained-peptide mimic of the HIV-1 V3 loop neutralization site. J. Mol. Biol. 266:31,1996.

Zhu, L., Dyson, H.J., Wright, P.E. A NOESY-HSQC simulation program SPIRIT. J. Biomol. NMR, in press.

Top of Page


Structural Biology of Protein and Oligonucleotide Recognition, Binding, and Signal Transduction

W.J. Chazin, M. Akke,* G. Bifulco, D. Boger,** O. Crescenzi, P.A. Fagan, S. Forsén,* L. Gomez-Paloma,*** H. Hidaka,**** B. Huang,***** M.J. Hunter, R.R. Ketchem, S. Linse,* M. Lubienski, J.C. Madsen, L. Mäler, M. Nelson, K.C. Nicolaou,** S. Parikh, B.C. Potts, J. Rydzewski, M. Sastry, J.A. Smith, E. Thulin,* T. Tsudo,**** C. Weber, B.T. Wimberly

* University of Lund, Lund, Sweden
** Department of Chemistry, TSRI
*** University of Naples, Naples, Italy
**** Nagoya University, Nagoya, Japan
***** Department of Cell Biology, TSRI

Our laboratory uses nuclear magnetic resonance spectroscopy to examine the solution structure and dynamics of proteins and oligonucleotides, free in solution and in complexes with cellular targets or drugs. Highlights in the past year include solving the structure of the calcium-activated state of calcyclin, an important calcium sensor from the S100 subfamily of calcium-binding proteins (CaBPs), and characterizing the related heterodimeric S100 protein complex, MRP8/MRP14, which has been implicated in inflammatory rheumatic disease and cystic fibrosis.

CALCIUM SIGNAL TRANSDUCTION

Calcium and CaBPs play a central role in intracellular signal transduction and are associated with a wide range of effects on health and disease. Our goal is to understand the molecular basis for the activation of CaBPs and the concomitant transduction of calcium signals. Ultimately, we hope to control binding properties and interactions with target proteins, thereby enabling the design of specific biological activities and therapeutic strategies relevant to calcium-mediated disease. We are using nuclear magnetic resonance spectroscopy and computational methods to determine the three-dimensional structures and internal dynamics of specific CaBPs in the presence and absence of calcium of cellular targets. Comparative structural analysis and protein engineering experiments are being used to elucidate the fundamental principles that govern the calcium-induced activation of these proteins.

We are determining the structure of several novel CaBPs. One of these is caltractin, an essential component of the microtubule organizing center in the centrosome, which is required for accurate chromosomal segregation during the M stage (mitosis) of the cell cycle. Among CaBPs, this protein has the unique property of being phosphorylated in vivo, apparently in a cell cycle--dependent manner. Interest in caltractin is high because of its potential role in coupling calcium and phosphorylation signaling pathways. In collaboration with B. Huang, Department of Cell Biology, we have developed high-level expression systems for the intact protein and for the two separate domains and have extensively characterized the biophysical properties of all three. Determination of the structure of caltractin in the presence and absence of calcium is in progress.

We have placed a particular emphasis on the S100 subfamily of CaBPs. A growing body of evidence indicates that these proteins provide cell type--specific transduction of calcium signals. Members of the S100 subfamily have been implicated in a variety of disease states and have apparent cellular activities ranging from growth and proliferation to apoptosis. The S100 protein calcyclin is preferentially expressed in the G1 phase of the cell cycle, shows deregulated expression in association with cell transformation, and is found in high abundance in certain cancer cell lines. We determined the structure of native calcyclin in the apo state and found a novel organization of the calcium-binding domains (Fig. 1). We propose that all other S100 proteins will have a similar structure and that their mode of signal transduction is clearly distinct from classical calmodulin-like calcium sensors. The structure of the calcium-activated state of calcyclin has recently been solved and is being refined. We are also characterizing interactions with one of calcyclin's cellular targets, annexin XI.

A new effort involves the MRP8 and MRP14 proteins, which constitute a unique two-chain S100 system that is expressed by myeloid cells during inflammatory reactions. Inflammatory disorders such as chronic bronchitis, cystic fibrosis, and rheumatoid arthritis are associated with elevated levels of MRP8/MRP14. Substantial progress has been made in the characterization of the pairing of the subunits and in the production of isotope-enriched protein samples for structure determination.

The comparison of different CaBPs is highly informative but is not sufficient to provide a fundamental understanding of the driving forces and the molecular basis for calcium activation. For this purpose, we are analyzing all known CaBP structures, determining the structure of calbindin D9k at ultrahigh resolution, and doing protein engineering experiments. In an effort to increase our understanding of calcium activation, we are attempting to rationally remodel calbindin D9k so that it undergoes a calmodulin-like opening when it binds calcium, thereby creating the novel protein calbindomodulin.

STRUCTURE-BASED DESIGN OF ANTITUMOR DRUGS

We are using three-dimensional structures and molecular design to determine the structural basis of the antitumor activity of ligands that bind in the minor groove of duplex DNA. From this foundation, we are establishing a rational basis for the design of new antitumor drugs. Projects involve determining, at the molecular level, the factors that govern binding affinity, specificity, and chemical reactivity of specific antitumor agents. Current efforts are focused on duocarmycin DNA alkylating agents and calicheamicin DNA scission agents, compounds that are distinguished by chemical reactivity with the DNA substrate and by exceptionally high potency.

We recently completed new structures of DNA complexes from each of these families. A high-resolution structure of (+)-duocarmycin-SA bound to a high-affinity binding site has been obtained (Fig. 2). This and related duocarmycin structures reveal a DNA binding--induced twist in the ligand that catalyzes the alkylation reaction, slows the rate of the back reaction, and stabilizes the resulting adduct. The structures of the head-to-head and head-to-tail dimers of the calicheamicin oligosaccharide domain in complex with double-site DNA duplexes have shown the crucial role of the oligosaccharide in recognition and binding. We are probing the molecular basis for the high binding affinity and substantially increased inhibitory activity of these chemically linked oligosaccharide oligomers.

PUBLICATIONS

Eis, P.S., Smith, J., Rydzewski, J., Case, D.A., Boger, D.L., Chazin, W.J. High resolution solution structure of a DNA duplex alkylated by the antitumor agent duocarmycin SA. J. Mol. Biol. 272:237, 1997.

Kördel, J., Pearlman, D.A., Chazin, W.J. Protein solution structure calculations in solution: Solvated molecular dynamics refinement of calbindin D9k. J. Biomol. NMR, in press.

Lee, L.K., Rance, M., Chazin, W.J., Palmer, A.G. III. Rotational diffusion anisotropy of proteins from simultaneous analysis of 15N and 13C nuclear spin relaxation. J. Biomol. NMR 9:287, 1997.

Nelson, M., Chazin, W.J. Calmodulin as a calcium sensor. In: Calmodulin and Signal Transduction. van Eldik, L.J., Watterson, D.M. (Eds.). Academic Press, San Diego, in press.

Nelson, M., Chazin, W.J. An interaction-based analysis of calcium-induced conformational changes in Ca2+ sensor proteins. Protein Sci., in press.

Nelson, M., Weber, C., Chazin, W.J. Calcium-binding proteins. In: Encyclopedia of Molecular Biology. Creighton, T. (Ed.). Wiley, New York, in press.

Potts, B.C.M., Chazin, W.J. Chemical shift homology in proteins. J. Biomol. NMR, in press.

Skelton, N.J., Chazin, W.J. Solution structure determination of proteins using NMR spectroscopy. In: Peptide and Protein Analysis. Reid, R. (Ed.). Marcel Dekker, New York, in press.

Smith, J., Paloma, L.G., Case D.A., Chazin, W.J. Molecular dynamics docking driven by NMR derived restraints to determine the structure of the calicheamicin 1I oligosaccharide domain complexed to duplex DNA. Magn. Reson. Chem. 34:S147, 1996.

Top of Page


Structure and Dynamics of Oligonucleotide Intermediates in Recombination and Repair

W.J. Chazin, D.P. Millar, R. Fee, D. Gonzales, S. Mangahas, G. Mer, S. Miick, M. Otto, A.G. Palmer III*

* Columbia University, New York, NY

The purpose of our multidisciplinary research program is to understand the structural biology of unusual oligonucleotide structures formed in the course of genetic recombination and repair. Our efforts to date have focused on a specific four-arm DNA crossover structure, the Holliday junction, which is formed as a transient intermediate during genetic recombination and repair. The structure and flexibility of these DNA crossovers appear to be critical to their recognition and processing into parental or recombinant products (Fig. 1). The principal goals of our research are to determine the three-dimensional structure of Holliday junctions, to establish the dependence on sequence, and, ultimately, to understand how the cellular recombination machinery distinguishes Holliday junctions of different sequence.

We use nuclear magnetic resonance (NMR) and time-resolved fluorescence resonance energy transfer (tr-FRET) spectroscopies to study 32-bp model Holliday junctions. Computation and molecular modeling are used in data interpretation and analysis. Our studies have shown that the key sequence-dependent structural property of a Holliday junction is the equilibrium distribution between the two crossover structural isomers (Fig. 2).
In the past year, this research was highlighted by the discovery of a previously unrecognized dependence of the structure and dynamics of Holliday junctions on the sequence adjacent to the branching point. In addition, we developed new methods for the production of DNA oligomers and adapted a protocol that uses a DNA polymerase strategy to produce duplexes and Holliday junctions enriched in NMR active isotopes (13C, 15N). These labeled DNA molecules are being used to facilitate NMR analysis of complexes formed with recombination and repair enzymes.

NMR studies have provided complete 1H assignments for three synthetic 32-bp Holliday junctions and have established base pairing, local conformation in the region of the junction, and the presence and relative ratio of crossover isomers. Using tr-FRET, we have determined the global folding arrangement of the four duplex arms and characterized the unique conformational flexibility of these structures. An NMR selective isotope-based strategy and a new implementation of tr-FRET data analysis have been developed to quantitate the ratio of crossover isomers. Molecular dynamics calculations with restraints for NMR- and tr-FRET--derived distance ranges are being used to generate representations of the three-dimensional structures in solution. As a first step toward a biophysical analysis of Holliday junctions that can undergo branch migration, tr-FRET is being used to examine monomobile Holliday junctions designed to permit a single step of branch migration.

Together, these results are providing evidence to distinguish between various structural models proposed for Holliday junctions and to determine how the sequence at the branching point modulates the global folding, local molecular conformation, and internal dynamics of these four-arm DNA crossover structures.

PUBLICATIONS

Miick, S.M., Fee, R.S., Millar, D.P., Chazin, W.J. Crossover isomers bias is the primary sequence-dependent property of immobilized Holliday junctions. Proc. Natl. Acad. Sci. U.S.A. 94:9080, 1997.

Top of Page


Protein--Nucleic Acid Interactions in Transcriptional Regulation

J.M. Gottesfeld, L.A. Dickinson, L. Neely, J.J. Long, S.J. McBryant, R.W. Kriwacki, R. Winston, A. Leresche,* J. Xu, P.E. Wright, R.J. Gulizia,** D.E. Mosier,** E. Baird,*** J. Trauger,*** P.B. Dervan***

* Baxter Pharmaceuticals, Neuchatel, Switzerland
** Department of Immunology, TSRI
*** California Institute of Technology, Pasadena, CA


PROTEIN--NUCLEIC ACID INTERACTIONS

Transcription factor IIIA (TFIIIA) from Xenopus oocytes is unique among the Cys2His2 class of zinc-finger proteins: it functions as both a 5S RNA gene-specific transcriptional activator and as a 5S rRNA storage particle protein. Our studies have focused on the domains of this protein that confer specific binding to 5S DNA and to 5S RNA and on the DNA and RNA sequences that form the binding sites for individual zinc fingers.

DNA binding is mediated by major groove interactions with two sets of three fingers (fingers 1--3 and 7--9) and with finger 5. A combination of mutagenesis and footprinting experiments showed that fingers 4 and 6 each bind in or across the minor groove to bridge the major groove--binding fingers. The central zinc fingers 4--7 are largely responsible for RNA binding. We have designed and synthesized small RNA molecules that retain TFIIIA binding activity to use in nuclear magnetic resonance structural studies. Selection and amplification binding assays are in progress to determine the key nucleotides within the 5S rRNA molecule that are critical for zinc-finger interactions. Additional protein-DNA binding studies concern determination of the amino acid residues of the Drosophila basic helix-loop-helix regulatory protein Deadpan that mediate protein dimerization.

SEQUENCE-SPECIFIC INHIBITION OF GENE TRANSCRIPTION WITH DESIGNED LIGANDS

Our collaborators, P.B. Dervan and colleagues at the California Institute of Technology, have established rules for sequence-specific recognition of DNA in the minor groove with polyamides containing N-methylimidazole and N-methylpyrrole amino acids. These molecules have affinities and specificities for their target DNA sequences comparable to the affinities and specificities of natural DNA-binding transcriptional regulatory proteins. A series of pyrrole-imidazole polyamides were designed and synthesized to target the minor groove of the DNA-binding site for zinc finger 4 of TFIIIA. These polyamides inhibit binding of TFIIIA to 5S DNA and also inhibit the transcriptional activity of the 5S rRNA genes in cell-free extracts. Similarly, polyamides selectively disrupt transcription complexes on the 5S RNA genes in Xenopus fibroblasts in culture.

Polyamides have also been designed and synthesized to target DNA sequences within the HIV enhancer and promoter. Because HIV uses the cellular transcription machinery for viral RNA synthesis by RNA polymerase II, sequences adjacent to the binding sites for the cellular transcription factors LEF-1/Ets-1 and the TATA-box binding protein (TBP) were chosen as polyamide binding sites. In vitro, these polyamides inhibit (1) interactions between the transcription factors and DNA, (2) transcription by RNA polymerase II, and (3) replication of HIV in isolated human peripheral blood lymphocytes. These data show that pyrrole-imidazole polyamides are cell-permeable DNA ligands and that molecules of this class can be designed to inhibit the transcriptional activity of selected genes in living cells.

TRANSCRIPTIONAL REGULATION OF CELL GROWTH AND DEVELOPMENT

Many aspects of transcriptional regulation in eukaryotic cells involve reversible phosphorylation events. One example is the global repression of nuclear transcription that occurs when cells enter mitosis. Active transcription by RNA polymerase III occurs in extracts prepared from unfertilized Xenopus eggs in interphase but is strongly repressed by the addition of B1 cyclin, which induces the mitotic kinase cdc2/cyclin B.

Repression of transcription in this in vitro system does not involve chromosome condensation, nucleosome assembly of the template DNA, or nonspecific DNA-binding proteins. Rather, the target of the kinase is a component of transcription factor IIIB, the general pol III initiation factor, which consists of TBP and pol III--specific TBP-associated factors. Highly purified transcription factor IIIB, but not TBP, is sufficient to restore full transcriptional activity to a kinase-inhibited transcription system. Protein-labeling experiments showed that TBP and a polypeptide of 92 kD are targets of the mitotic kinase in protein fractions containing transcription factor IIIB. Future studies will determine whether phosphorylation of this 92-kD protein is responsible for transcriptional repression.

As for pol III genes, purified cdc2/cyclin B kinase is sufficient to inhibit transcription of several test promoters by RNA polymerase II in both nuclear extracts and a reconstituted transcription system. Repression of transcription in these systems is due to protein phosphorylation by the cdc2 kinase, because both the general protein kinase inhibitor 6-dimethylaminopurine and the cdc/ckd inhibitor p21 prevent inhibition. Transcription rescue and inhibition experiments with each of the basal factors and polymerase suggest that for activator-dependent transcription, the general transcription factor IID is the target of the kinase and that for a basal promoter, transcription factor IIH (TFIIH) is inactivated by the cdc2/cyclin B kinase. Protein-labeling experiments indicate that the p62 and p37 (cyclin H) subunits of TFIIH are in vitro substrates for mitotic phosphorylation. The cdk7/cyclin H-associated kinase of TFIIH phosphorylates the carboxy-terminal domain of the large subunit of RNA polymerase II, and this kinase activity of TFIIH is inhibited by the cdc2/cyclin B kinase.

PUBLICATIONS

Foster, M.P., Wuttke, D.S., Radhakrishnan, I., Case, D.A., Gottesfeld, J.M., Wright, P.E. Domain packing and dynamics in the DNA complex of the amino-terminal zinc fingers of transcription factor IIIA. Nature Struct. Biol. 4:605, 1997.

Gottesfeld, J.M., Forbes, D.J. Mitotic repression of transcription. Trends Biochem. Sci. 22:197, 1997.

Gottesfeld, J.M., Neely, L., Trauger, J.W., Baird, E.E., Dervan, P.B. Regulation of gene expression by small molecules. Nature 387:202, 1997.

Leresche, A., Wolf, V.J., Gottesfeld, J.M. Nobel Symposium: Repression of RNA polymerase II and III transcription during M phase of the cell cycle. Exp. Cell Res. 229:282, 1996.

McBryant, S.J., Gottesfeld, J.M. Differential kinetics of transcription complex assembly distinguish oocyte- and somatic-type 5S RNA genes of Xenopus. Gene Expr., in press.

Neely, L., Trauger, J.W., Baird, E.E., Dervan, P.B., Gottesfeld, J.M. Importance of minor groove binding zinc fingers within the transcription factor IIIA-DNA complex. J. Mol. Biol., in press.

Wuttke, D.S., Foster, M.P., Case, D.A., Gottesfeld, J.M., Wright, P.E. Solution structure of the first three zinc fingers of TFIIIA complexed with its cognate DNA sequence: Determinants of affinity and sequence specificity. J. Mol. Biol., in press.

Top of Page


Time-Resolved Spectroscopic Studies of Nucleic Acid Structure and Nucleic Acid--Protein Interactions

D.P. Millar, R.S. Fee, W.C. Lam, M.R. Otto, M. Auer,* C.M. Joyce,** L.C. Sowers***

* Novartis Research Institute, Vienna, Austria
** Yale University, New Haven, CT
*** City of Hope National Medical Center, Duarte, CA

Fluorescence spectroscopy is a powerful tool for investigating the structure and dynamics of biological macromolecules under solution conditions that mimic the intracellular environment. In our laboratory, we use picosecond time-resolved fluorescence techniques to analyze a variety of nucleic acid--protein interactions and to probe the solution structure of DNA and RNA.

DNA-PROTEIN INTERACTIONS

The advent of automated methods for the chemical synthesis and site-specific labeling of oligonucleotides has led to an increasing interest in the use of fluorescence spectroscopic techniques for the quantitative analysis of DNA-protein interactions. Using a combination of different oligonucleotide-labeling schemes and fluorescence methods, we have developed new procedures that can be used to determine thermodynamic and kinetic parameters that describe the binding of proteins to DNA and to probe localized changes in the structure of DNA.

These methods have been particularly useful for investigating the mechanism of template-directed synthesis of DNA by DNA polymerases and the exonucleolytic proofreading that helps maintain a low prevalence of errors during DNA replication. During proofreading, the 3´ end of the nascent DNA strand must shuttle from the polymerase site to the 3´-5´ exonuclease site, located in a separate structural domain of the enzyme, and the last few base pairs of DNA must melt in readiness for the exonuclease reaction. By using DNA substrates labeled at different positions with fluorescent probes, we have been able to directly monitor these translocation and localized melting steps. These experiments have helped elucidate the mechanism by which the polymerase recognizes and selectively removes misincorporated nucleotides from the nascent DNA strand. During the past year, we investigated the role of specific protein side chains in mediating this mismatch recognition.

In collaboration with C. Joyce, Yale University, we characterized the effect of protein mutations on the partitioning of DNA substrates between the polymerase and the 3´-5´ exonuclease sites. Site-directed mutagenesis was used to change each of the protein side chains of the Klenow fragment of DNA polymerase I that crystallographically are close to the 3´ terminus of a DNA substrate bound at the 3´-5´ exonuclease site. From the spectroscopic analysis of these mutant enzymes, we could directly quantify the energetic contribution of each side chain to the binding of DNA at the exonuclease site, thereby providing a direct correlation between structural properties and thermodynamics.

In another study, we examined the effect of sequence-directed DNA structure on polymerase-DNA interactions. We found that the presence of an A-tract sequence element in the DNA substrate strongly promoted the binding of the primer terminus at the 3´-5´ exonuclease site, despite the absence of mismatched bases. These results show that the polymerase can also recognize irregularities in the helix geometry of DNA.

RNA-PROTEIN INTERACTIONS

Replication of HIV type 1 requires the production of two virally encoded regulatory proteins, Tat and Rev. Rev induces the cytoplasmic accumulation of full-length RNA transcripts that code for viral structural proteins, resulting in the subsequent production of active viral particles. Rev achieves its regulatory function by binding to a specific mRNA sequence from HIV type 1 known as the Rev responsive element (RRE). We are using fluorescence resonance energy transfer (FRET) methods to probe the three-dimensional structure of the Rev-RRE complex, as part of a collaboration between TSRI and the Novartis Research Institute in Vienna. In addition, the single tryptophan residue of Rev, located within the RNA-binding domain of the protein, has been used as a spectroscopic probe of Rev-RNA interactions and of Rev-Rev multimerization.

Donor and acceptor dyes have been incorporated into a synthetic fragment of the RRE containing the minimal Rev-binding site. Intramolecular distances within the RRE obtained from FRET measurements are used as distance constraints to model the overall structure of the RNA, both in its uncomplexed form and bound to Rev. Time-resolved measurements of FRET are used to determine the distribution of distances between each pair of labeled sites, which provides information on the conformational flexibility of the RNA. Similar methods are being used to investigate the interaction of the RRE with carbohydrates that inhibit binding of Rev.

PUBLICATIONS

Carver, T.E., Millar, D.P. Recognition of sequence-directed DNA structure by the Klenow fragment of DNA polymerase I. Biochemistry, in press.

Lam, W.C., Seifert, J.M., Amberger, F., Graf, C., Auer, M., Millar, D.P. Structural dynamics of HIV-1 Rev and its complexes with RRE and 5S RNA. Biochemistry, in press.

Lam, W.C., Van der Schans, J.C., Joyce, C.M., Millar, D.P. Effect of mutations on the partitioning of DNA substrates between the polymerase and 3´-5´ exonuclease sites of DNA polymerase I (Klenow fragment). Biochemistry, in press.

Millar, D.P. Time-resolved fluorescence spectroscopy. Curr. Opin. Struct. Biol. 6:637, 1996.

Yang, M., Millar, D.P. Fluorescence resonance energy transfer as a probe of DNA structure and function. Methods Enzymol. 278:417, 1997.

Top of Page


Applications of Solution Single-Molecule Spectroscopy

J.J. La Clair, R.A. Lerner, M. Eigen*

* Max Planck Institute for Biophysical Chemistry, Göttingen, Germany

Current understanding of chemical and biological transformations is based primarily on assays of populations of molecules. In recent years, a number of techniques have been developed that allow the detection and monitoring of individual molecules. In our research, we use the recently described confocally adjusted fluorescence correlation spectroscopy to examine chemically and biologically relevant properties that cannot be investigated by using classical methods.

Traditionally, the outcome of a chemical reaction is determined by isolating and characterizing the products of the reaction. The relative energetics of each pathway are then expressed in terms of the relative portions of each material produced. However, with highly disfavored processes, the amount of the minor material produced is so small that the material cannot be isolated. Furthermore, detection of this material is complicated by the presence of a large excess of similar substances. Therefore, the energetics of these pathways must be estimated indirectly.

Recently, we devised a method wherein only the product of a highly disfavored path contains the appropriate functionality to allow an intramolecular charge transfer. When the transfer occurs, the absorption and fluorescence properties of the minor product are dramatically different from those of the other materials in the reaction. Using this scheme, we have been able to detect products from reaction pathways that were disfavored by more than 8 kcal/mol (Fig. 1).

Inane by unity, the properties of a single molecule are not required to mimic those seen in an average population of molecules. We have attempted to determine what can be learned by examining individual molecules. In particular, we have focused on monitoring the recognition between carbohydrates and their binding proteins. To accomplish this task, we developed a new fluorescent probe called SENSI that has extreme restrictochromaticity and solvochromaticity, in part because of the generation of numerous twisted intramolecular charge transfer states. These states are generated by internal rotation upon photoexcitation from a planar conformation to states in which parts of electronic functionality are aligned orthogonally. Typically, twisted intramolecular charge transfer states are sensitive to solvent properties and are characterized by weak fluorescence.

In commonly used buffers, the fluorescence from SENSI-labeled maltoside and glucoside is extremely weak (Fig. 2). As the fluorescent label moves from solution into or near the vicinity of a protein-binding pocket, the ability of the label to rotate diminishes, and its solvent shell is altered. This change leads to a decrease in the number of twisted intramolecular charge transfer states and ultimately enhances the fluorescence. We used this scheme to selectively detect single bound complexes of concanavalin A and maltoside. Since then, we have prepared derivatives of this fluorophore that can be used to efficiently attach the fluorescent label to amines, aldehydes, and thiols, providing a new means to examine molecular interactions. We are using these derivatives to monitor protein-carbohydrate aggregation, protein-protein aggregation, low-affinity binding, and membrane interactions.

PUBLICATIONS

La Clair, J.J. Analysis of highly disfavored processes through pathway-specific correlated fluorescence. Proc. Natl. Acad. Sci. U.S.A. 94:1623, 1997.

La Clair, J.J. Selective detection of the carbohydrate-bound state of concanavalin A at the single molecule level. J. Am. Chem. Soc. 119:7676, 1997.

Top of Page


Characterizing Biomolecular Interactions

G. Siuzdak, J. Boydston, B. Bothner, T. Hollenbeck, J. Wu, K.S. Chatman, K. Harris, A. Aparicio, R. North, D. Owens

In the late 1980s, mass spectrometry emerged as an important tool in biomolecular research, and since then the usefulness of this method has grown significantly. The value of mass spectrometry lies in the unique ability to accurately characterize the structure and dynamics of biomolecules by using only femtomoles to picomoles of material. Although this evolution could be termed groundbreaking, its full potential is far from being realized.

Our interests focus on understanding biomolecular complexes and on developing novel methods to examine the interactions and reactions of the complexes. For instance, the goal in our studies of protein structure is to determine specific sites of protein interactions and activity, with special emphasis on viral capsid proteins, enzymes, and regulatory proteins. Through our unique application of mass spectrometry to viral analysis, we have examined both local and global viral structure, thus gaining interesting insight into the structure of proteins on the viral surface.

Despite the detailed structural models created for viruses, little is known about the mechanisms used by the capsid proteins that facilitate binding to host cells, penetration of the cell membrane, and release of nucleic acid. Matrix-assisted laser desorption/ionization mass spectrometry combined with limited proteolysis has been used to examine the viral capsid of flock house virus. Using only microgram quantities of virus, we have obtained novel information about the dynamic nature of the viral surface (Fig. 1).

We are also developing mass spectrometry methods for assaying inhibitors and ligands (Fig. 2) in an effort to better understand activity, reaction dynamics, and the structure-inducing features of these molecules and their receptors. For instance, with the development of a novel quantitative mass spectrometry assay, we can detect new enzyme inhibitors without using a chromophore or radiolabeling. In addition, we are applying the sensitivity offered by mass spectrometry to detect novel initiator molecules in biofluids. Clearly, the usefulness of mass spectrometry extends far beyond molecular weight characterization and is now offering greater prospects for understanding dynamic biomolecular interactions.

Publications

Bothner, B., Dong, X.F., Bibbs, L., Johnson, J.E., Siuzdak, G. Evidence of viral capsid dynamics using limited proteolysis and mass spectrometry. J. Biol. Chem., in press.

Fitzgerald, M.C., Siuzdak, G. Biochemical mass spectrometry: Worth the weight? Chem. Biol. 3:707, 1996.

Kriwacki, R., Wu, J., Siuzdak, G., Wright, P. Probing protein-protein interactions by mass spectrometry: Analysis of the p21/Cdk2 complex. J. Am. Chem. Soc. 118:5320, 1996.

Kriwacki, R., Wu, J., Tennant, L., Wright, P., Siuzdak, G. Probing protein structure using biochemical and biophysical methods: Proteolysis, MALDI mass analysis, HPLC, and gel-filtration chromatography of p21. J. Chromatogr. 777:23, 1997.

Siuzdak, G. Mass spectrometry. In: The Encyclopedia of Molecular Biology. Creighton, T.E. (Ed.). Wiley, New York, in press.

Siuzdak, G., Henriksen, S. New directions in the analysis of brain substances related to sleep and wakefulness. In: Molecular Regulation of Arousal States. Lydic, R. (Ed.). CRC Press, Boca Raton, FL, in press.

Siuzdak, G., Lewis, J.K. Innovations in combinatorial chemistry and mass spectrometry. Combinatorial Chem., in press.

Takayama, S., Martin, R., Wu, J., Laslo, K., Siuzdak, G., Wong, C.-H. Chemoenzymatic preparation of novel cyclic imine sugars and rapid biological activity evaluation using electrospray mass spectrometry and kinetic analysis. J. Am. Chem. Soc. 119:8146, 1997.

Wu, J., Chatman, K., Harris, K., Siuzdak, G. An automated MALDI mass spectrometry approach for optimizing cyclosporin analysis. Anal. Chem. 69:3767, 1997.

Wu, J., Takayama, S., Wong, C.-H., Siuzdak, G. Quantitative electrospray mass spectrometry for the rapid assay of enzyme inhibitors. Chem. Biol. 4:653, 1997.

Top of Page


Computer Modeling of Proteins and Nucleic Acids

D.A. Case, A. Dejaegere,* E. Demchuk, G.P. Gippert, M. Johnson,** T. Macke, D. Sitkoff, J. Smith, J. Srinivasan, V. Tsui, N. White

* Universite Louis Pasteur, Strasbourg, France
** University of Illinois, Chicago, IL

Computer simulations offer an exciting approach to the study of many aspects of biochemical interactions. In our group, we focus primarily on molecular dynamics simulations (in which Newton's equations of motions are solved numerically) to model the solution behavior of biomacromolecules. Recent applications include a detailed analysis of electrostatics interactions in folded and unfolded short peptides, proteins, and oligonucleotides in solution. In addition, molecular dynamics methods are useful in refining solution structures of proteins by using constraints derived from nuclear magnetic resonance spectroscopy, and we are continuing to develop new methods in this area.

PROTEIN AND NUCLEIC ACID STRUCTURE AND DYNAMICS

An understanding of the internal dynamics of proteins is essential for the interpretation of relaxation behavior and for generating the most information from experimental spectra. We have been using normal mode analysis and molecular dynamics simulations to probe the dynamics of C-H, N-H, and H-H interactions in proteins, RNA hairpins, and DNA and RNA duplexes. We are also using these models to develop better methods for determining biomolecular structures in solution from nuclear magnetic resonance data, going beyond distance constraints to generate closer connections between calculated and observed spectra. Recent structural studies include the blue-copper protein rusticyanin, DNA oligonucleotide duplexes (both free and bound to potential anticancer drugs), a phospho-transfer protein, and complexes of zinc-finger domains with DNA. We are also using quantum chemistry approaches to study ways in which chemical shifts and spin-spin coupling constants are related to protein and nucleic acid structure.

A recent project centers on the development of novel computer methods to construct models of "unusual" nucleic acids that go beyond traditional helical motifs. We are using these methods to study circular DNA; small RNA fragments; and three- and four-strand DNA complexes, including models for recombination sites. Continuum solvent methods are used to provide a powerful evaluation of the relative energies of different conformers. Examples include comparisons of A and B forms of DNA and RNA, and a hairpin-duplex transition in RNA.

SIMULATION OF PEPTIDE CONFORMATION IN SOLUTION

Detailed molecular dynamics simulations on peptides that form turns in solution (such as APGD and AYPYD) illustrate ways in which early events in protein folding can be studied. A potential of mean force for the folding-unfolding transition in APGD shows that the unfolded and folded forms have nearly equal free energies; continuum electrostatic calculations (in collaboration with D. Bashford and C. Brooks, Department of Molecular Biology) provide a better understanding of various components of such energy differences.

ANALYSIS OF MUTANT HEMOGLOBIN STRUCTURES

Using both continuum electrostatic models and molecular dynamics simulations to study the effects of charge mutations on oxygen binding and subunit dissociation, we have constructed an extensive model for the dependence of hemoglobin oxygen affinity on pH; we considered more than 150 titrating sites. These results offer new insights into the ways in which protons and anions act as allosteric ligands in modulating the functional properties of oxygen-binding proteins. New directions include models of the effects of protein flexibility, specific ion-binding events, and studies of inhibitor binding to factor Xa.

PUBLICATIONS

Bashford, D., Case, D.A., Choi C., Gippert, G.P. A computational study of the role of solvation effects in reverse turn formation in the tetrapeptides AGPD and APGN. J. Am. Chem. Soc. 119:4964, 1997.

Beroza, P., Case, D.A. Including sidechain flexibility in continuum electrostatic calculations of protein titrations. J. Phys. Chem. 100:20156, 1996.

Botuyan, M.V., Toy-Palmer, A., Chung, J., Blake, R.C., Beroza, P., Case, D.A., Dyson, H.J. NMR solution structure of Cu(I) rusticyanin from Thiobacillus ferrooxidans: Structural basis for the extreme acid stability and redox potential. J. Mol. Biol. 263:752, 1996.

Case, D.A. NMR spectral simulation and structure refinement. In: Encyclopedia of Computational Chemistry. Grant, D.M., Harris, R.K. (Eds.). Wiley, New York, in press.

Case, D.A. Normal mode analysis of biomolecular dynamics. In: Computer Simulations of Biomolecular Systems, Vol. 3. van Gunsteren, W.F., Weiner, P.K., Wilkinson, A.J. (Eds.). ESCOM, Leiden, in press.

Chen, Y., Case, D.A., Reizer, J., Saier, M.H., Jr., Wright, P.E. High resolution structure of Bacillus subtilis IIAglc. Proteins, in press.

Demchuk, E., Bashford D., Case, D.A. Dynamics of a type VI reverse turn in a linear peptide in aqueous solution. Folding Design 2:35, 1997.

Demchuk E., Bashford, D., Gippert, G.P., Case, D.A. Thermodynamics of a reverse turn motif: Solvent effects and sidechain packing. J. Mol. Biol. 270:305, 1997.

Eis, P.S., Smith, J.A., Rydzewski J.M., Case, D.A., Boger, D.L., Chazin, W.J. High resolution solution structure of a DNA duplex alkylated by the antitumor agent duocarmycin SA. J. Mol. Biol. 272:237, 1997.

Foster, M., Wuttke, D., Radhakrishnan I., Case, D.A., Gottesfeld, J.M., Wright, P.E. Domain packing and dynamics in the DNA complex of the amino-terminal zinc fingers of transcription factor IIIA. Nature Struct. Biol. 4:605, 1997.

Li, J., Beroza, P., Noodleman, L., Case, D.A. Quantum mechanical modeling of active sites in metalloproteins. In: Molecular Modeling and Dynamics of Biological Molecules Containing Metal Ions. Banci, L., Comba, P. (Eds.) Kluwer, Boston, in press.

Richardson, W.H., Peng, C., Bashford, D., Noodleman, L., Case, D.A. Incorporating solvation effects into density functional theory: Calculation of absolute acidities. Int. J. Quant. Chem. 61:207, 1997.

Smith, J.A., Paloma, L.G., Case, D.A., Nicolaou, K.C., Chazin, W.J. Molecular dynamics docking driven by NMR derived restraints to determine the structure of the calicheamicin 1I oligosaccharide domain complexed to duplex DNA. Magn. Reson. Chem. 34:S147, 1996.

Top of Page


Electronic Structure Calculations for Active-Site Models of Iron-Sulfur Proteins and for Biologically Relevant Transition-Metal Complexes

L. Noodleman, D.A. Case, J. Li, D.N. Hendrickson,* M.J. Knapp,* J.M. Mouesca,** B. Lamotte,** K. Kustin***

* University of California, San Diego, CA
** Centre d'Etudes Nucléaires, Grenoble, France
*** Brandeis University, Waltham, MA

We use modern methods of quantum chemistry (density functional methods) to investigate the electronic structures, energetics, and spectroscopic characteristics of transition-metal complexes at the active sites of metalloproteins and related synthetic analogues. These systems include iron-sulfur and manganese-oxo complexes and active-site models for manganese, iron, and copper-zinc superoxide dismutases. Many of these complexes lie at the active sites of important metalloproteins, where the complexes are involved in electron transfer, catalysis, and regulation of enzyme activity.

In collaboration with E. Keinan, Department of Molecular Biology, we have also investigated the binding and geometric distortion of transition metal complexes by antibodies. With D. Hendrickson, University of California, San Diego, we have studied the physical properties of synthetic transition metal--quinone systems that show internal transfer of electrons and changes in geometry and spin state that are sensitive to pressure, temperature, and ligands (valence tautomerism).

Because iron-sulfur proteins are often electron transfer agents, we seek to understand which features of cluster electronic structure and what interactions with the protein environment determine the wide range of redox potentials in these systems. We have examined the role of cluster electronic structures in redox potentials in synthetic iron-sulfur clusters. The solvent was modeled as a continuous dielectric medium. The predicted changes in redox potentials correlated well with experimental trends. This work has been extended to a full self-consistent reaction field method for clusters in solvent environments, and further extensions to redox potentials of 2Fe2S clusters in protein environments are in progress. This work is done in collaboration with D. Bashford, Department of Molecular Biology. With M. Knapp, University of California, San Diego, and D. Hendrickson, we are studying the effect of arylthiolate conformation on the electronic structure and spectra of [Fe4S4(SR4)4]2-,3- complexes.

Manganese, iron, and copper-zinc superoxide dismutases are important detoxifying agents for the superoxide radical anion. We calculated redox potentials and pKa's for the active site of manganese superoxide dismutase and assessed related mechanistic issues. We are extending our model to include an electrostatic representation of the protein environment.

We have used the self-consistent reaction field method to calculate redox potentials and pKa's for the transition-metal cations Mn2+, Mn3+, Fe2+, and Fe3+ in aqueous solution. These simpler systems provide a starting point for analyzing redox and protonation-deprotonation events in more complex systems, including manganese and iron superoxide dismutases. Acidity calculations have also been made for an extensive series of organic acids that resemble all the common amino acid side chains; correlation with experimental pKa's was good.

Multielectron transfer catalysis is shown by dinuclear and polynuclear manganese sites in manganese catalases and in the water oxidation complex of photosystem II. The water oxidation complex is critical to photosynthesis in plants and cyanobacteria and is the major source of molecular oxygen in the atmosphere. We are exploring the electronic structures of related synthetic dinuclear manganese-oxo complexes to gain insight into the ease of multielectron transfer and coupled proton transfer in these systems.

With E. Keinan, we have compared our calculations of the red shift of an optical metal-to-ligand charge transfer band with that observed when a copper(I)-bipyridine complex binds to an antibody. Comparison of experimental and calculated shifts indicates that the Cu(I) complex is being compressed toward the geometry displayed by the more "ideal" hapten (a silicon complex). This finding is a rare demonstration that an antibody can transduce a substantial geometric change in a transition-metal complex. In the longer term, this finding should have implications for enhancing reactivity of and catalysis by transition-metal complexes bound in antibody cavities. We plan to extend these studies to other transition-metal complexes that can bind to antibodies. For example, valence tautomeric complexes such as the cobalt-di-semiquinone complex we have studied can undergo internal metal-to-ligand electron transfer (or the reverse) as a function of temperature and pressure, and related conversions may occur as a response to antibody binding.

PUBLICATIONS

Adams, D.A., Noodleman, L., Hendrickson, D.N. Density functional study of the valence tautomeric interconversion low-spin Co(III)(SQ)(Cat)(phen) <=> high-spin Co(II)(SQ)2(phen). Inorg. Chem. 36:3966, 1997.

Li, J., Beroza, P., Noodleman, L., Case, D.A. Quantum mechanical modeling of active sites in metalloproteins: Electrostatic coupling to the protein/solvent environment. In: NATO ASI-Molecular Modeling and Dynamics of Biological Molecules Containing Metal Ions. Banci, L., Comba, P. (Eds.). Kluwer, Boston, in press.

Li, J., Noodleman, L., Case, D.A. Electronic structure calculations: Density functional methods with applications to transition metal complexes. In: Inorganic Electronic Structure and Spectroscopy. Vol. 1, Methods. Solomon, E.I., Lever, A.B.P. (Eds.). Wiley, New York, in press.

Noodleman, L., Li, J., Zhao, X.G., Richardson, W.H. Density functional studies of spin coupled transition metal dimer and tetramer complexes In: Density Functional Methods: Applications in Chemistry and Materials Science. Springborg, M. (Ed.). Wiley, New York, 1997.

Richardson, W.H., Peng, C.Y., Bashford, D., Noodleman, L., Case, D.A. Incorporating solvation effects into density functional theory: Calculation of absolute acidities. Int. J. Quantum Chem. 61:207, 1997.

Zhao, X.G., Richardson, W.H., Chen, J.-L., Li, J., Noodleman, L., Tsai, H.L., Hendrickson, D.N. Density functional calculations of electronic structure, charge distribution, and spin coupling in manganese-oxo dimer complexes. Inorg. Chem. 36:1198, 1997.

Top of Page


Theoretical Studies of Electrostatic Fields and Artificial Pore Transport

D. Bashford, E. Demchuk, W. Chen, V. Spassov, V. Dillet

We are developing and applying macroscopic dielectric models of the macromolecule-solvent system. The protein is treated as a low-dielectric medium immersed in a high-dielectric solvent, and the electric potential is determined by using the Poisson-Boltzmann equation, which is solved by finite-difference methods. The details of the atomic structure are incorporated into the placement of charges and dielectric boundaries. We call the model macroscopic electrostatics with atomic detail (MEAD).

We recently completed studies of APGD and AYPYD, members of classes of peptides that have nascent secondary structure in solution: a type I or type II turn in the case of APGD, and a type VI turn in the case of AYPYD. Such sequences in proteins could act as nucleation sites in protein folding. The MEAD model has been used to calculate the solvation contribution to the potential of mean force for a large number of conformers of these peptides. When combined with a standard empirical potential for the intramolecular contribution, the resulting potential of mean force gives a conformational energetics that is consistent with both experimental and more detailed and costly calculations of solvent effects. This work was carried out in collaboration with the laboratories of D. Case, C. Brooks, and P. Wright in the Department of Molecular Biology. We plan to extend this method to studies of the folding of small proteins.

Determination of pKa values is a key test of electrostatic calculations, as well as an important application. We recently completed a study of the highly perturbed pKa values of histidine residues in a low molecular weight protein, tyrosine phosphatase, and in several mutants. The calculations are in general agreement with experimental values and provide insights into both the causes of the pKa shifts and the stabilization of the active-site loop. We are currently studying the cysteine residue, which serves as the active-site nucleophile, and its environment. Other current subjects of pKa calculations are the active site of thioredoxin, in which cysteine and aspartate residues have unusual titration behavior, and calbindin, in which titrating groups appear to play a role in the cooperativity of calcium binding. In much of this work, the combined quantum mechanics/MEAD method developed in collaboration with L. Noodleman, Department of Molecular Biology, will be used to obtain a more detailed and accurate picture of the active sites.

We are continuing our studies of the diffusion of water and ions through peptide nanotubes of the type developed in the laboratory of M. Ghadiri, Department of Chemistry. Previous calculations showed that water in the nanotubes has a layered structure with deviations that allow water molecules to pass one another, in contrast to the single-file structure of water in the gramicidin pore, which has slower transport properties. Current work includes the development and application of "hopping models" of diffusion and molecular dynamics calculations of ion transport.

PUBLICATIONS

Bashford, D. Macroscopic electrostatics with atomic detail (MEAD): Applications to biomacromolecules. In: Biomacromolecules: From 3-D Structure to Applications. Proceedings of the 34th Hanford Symposium on Health and the Environment. Ornstein, R.-L. (Ed.). Battelle Press, Columbus, OH, 1997, p. 53.

Bashford, D., Case, D., Choi, C., Gippert, G. A computational study of the role of solvation effects in reverse turn formation in the tetrapeptides (APGD) and (APGN). J. Am. Chem. Soc. 119:4964, 1997.

Demchuk, E., Bashford, D., Case, D. Dynamics of a type (VI) reverse turn in a linear peptide in aqueous solution. Folding Design 2:23, 1997.

Demchuk, E., Bashford, D., Gippert, G., Case, D. Thermodynamics of a reverse turn motif: Solvent effects and side-chain packing. J. Mol. Biol., in press.

Fisher, C.L., Chen, J.-L., Bashford, D., Noodleman, L. Density-functional and electrostatic calculations of a model of a manganese superoxide dismutase active site in aqueous solution. J. Phys. Chem. 100:13498, 1996.

Richardson, W., Peng, C.-Y., Bashford, D., Noodleman, L., Case, D. Incorporating solvation effect into density functional theory: Calculation of absolute acidities. Int. J. Quant. Chem. 61:207, 1997.

Top of Page


Simulations of Protein Folding, Structure, and Dynamics

J. Skolnick, A. Kolinski, J. Fetrow,* W.-P. Hu, A.R. Ortiz, R.P. Rotkiewicz,** H. Zhang, L. Zhang

* State University of New York, Albany, NY
** University of Warsaw, Warsaw, Poland

Prediction of the native conformation of a protein is one of the most important unsolved problems in molecular biology. A full solution to the protein-folding problem, likened to cracking the second half of the genetic code, would enable us to predict not only the native conformation of the protein but also the mechanism of assembly from the random, unfolded state. Once we have established the models that qualitatively mimic reality, we can use them in computer experiments to glean information not readily accessible from real experiments. Using a broad spectrum of computer simulation approaches, we are developing techniques to predict the native conformation and folding pathways of water-soluble globular proteins and macromolecular assemblies of chains.

The feasibility of predicting the global fold of small proteins by incorporating predicted secondary and tertiary restraints into ab initio folding simulations was shown with a test set of 20 nonhomologous proteins. These proteins contained 37--100 residues and represented all secondary structural classes and a wide variety of global topologies. Restraints on secondary structure were provided by the PHD secondary-structure prediction algorithm that incorporates multiple sequence information. Predicted tertiary restraints were derived from multiple sequence alignments via a two-step process: First, seed side-chain contacts were determined from correlated mutational analysis, and then a threading-based algorithm was used to expand the number of these seed contacts.

The result was a lattice-based reduced protein model and a folding algorithm especially designed to incorporate these predicted restraints at the appropriate level of spatial resolution. Depending on fold complexity, it is possible to assemble nativelike topologies with a coordinate root-mean-square deviation (RMSD) from the native protein of 2.5--6.5 Å.

Our results suggest that the requisite level of accuracy in side-chain contact map prediction can be roughly 25% on average, provided that about 60% of the contact predictions are correct within plus or minus one residue and that 95% of the predictions are correct within plus or minus four residues. Precision in predicting tertiary contacts is more critical than absolute accuracy. Furthermore, only a subset of the tertiary contacts, on the order of 25% of the total, is sufficient for successful topology assembly. Overall, the findings suggest that the use of restraints derived from multiple sequence alignments combined with a fold assembly algorithm holds considerable promise for the solution of the protein-folding problem.

A new, efficient method for the assembly of protein tertiary structure from known secondary structure (in the form of a conventional three-letter code) and sparse information about side-chain contacts has been proposed and evaluated. In comparison with existing algorithms, the new method requires a smaller number of tertiary restraints (on average, one per seven residues). At the same time, the obtained results are at least as accurate. For small structures such as the B domain of protein G, the accuracy range is within 3 Å RMSD from experimental structure; for myoglobin, within 4.5 Å; and for the 247-residue triose phosphate isomerase barrel, within 6 Å. With a larger number of tertiary restraints, the accuracy of assembled structures improves.

The method is based on a new, simple reduced model of protein structure and dynamics. The conformational space of model proteins is restricted to a discrete collection of lattice chains with a single interaction center per residue. Despite such a simple representation, the model has built-in implicit multibody correlations simulating short- and long-range packing preferences, the cooperativity of hydrogen bonding, and the mean force potential of hydrophobic interactions. Because of the simplicity of representation and the simple definition of the model, the Monte Carlo algorithm is at least an order of magnitude faster than previously published Monte Carlo algorithms for structure assembly. With its simplicity, reliability, and powerfulness, the new method can be used routinely in model building based on various (also sparse) experimentally derived structural restraints.

Many existing derivations of knowledge-based statistical pair potentials use the quasi-chemical approximation to estimate the expected frequency of side-chain contacts if there are no pair-specific interactions. At first glance, the quasi-chemical approximation that treats the residues in a protein as being disconnected and expresses the probability of side-chain contacts as being proportional to the product of the mole fractions of the pair of residues would appear to be rather severe.

To investigate the validity of this approximation, we have introduced two new reference states that retain the connectivity of the protein chain. Both reference states effectively permit the factorization of the probability of side-chain contacts into sequence- and structure-dependent terms. Then, because the sequence distribution of amino acids in proteins is random, the quasi-chemical approximation to each of these reference states is excellent. Thus, the range of validity of the quasi-chemical approximation is determined by the magnitude of the side-chain repacking term, which is currently unknown.

We examined the effect of tertiary interactions on the observed secondary structure found in the native conformation of globular proteins in the context of a reduced protein model. Short-range interactions are controlled by knowledge-based statistical potentials that reflect local conformational regularities seen in a database of three-dimensional protein structures. Long-range interactions are approximated by mean field, single residue--based, centrosymmetric, hydrophobic burial potentials. Even when pairwise specific long-range interactions are ignored, the inclusion of such burial preferences noticeably modifies the equilibrium chain conformations, and the observed secondary structure is closer to that seen in the folded state.

For a test set of 10 proteins (belonging to various structural classes), the accuracy of prediction of secondary structure was about 66%, and it increased by 9% with respect to a related model based on short-range interactions alone. The increased accuracy is due to the interplay between the short-range conformational propensities and the burial and compactness requirement built into the current model.

Other activities include the development of self-consistent field approaches to threading, an improved method for the prediction of U-turns in which the chain reverses global direction, detailed studies of the ability to derive statistical potentials from structural databases, and a series of simulations of the mechanism of assembly of viral coat proteins.

Top of Page


Analysis and Predictions of Protein Structures

A. Godzik, B. Zhang, L. Rychlewski

Continual efforts to improve the methods used to recognize protein folds and predict protein structure led us to reexamine the question of structural similarity between pairs and groups of proteins that have related folds but no recognizable sequence similarity. We found that topologically similar proteins are stabilized by mostly overlapping sets of interactions. The overall percentage of interactions conserved in such proteins is within the range of 50--60%, almost independent of the presence or absence of homology between both proteins. This ratio is higher by about 10% or 15%, respectively, for nonlocal interactions and for the residues in the core.

Accordingly, using an interaction set from the template protein with the target sequence provides a good approximation of the target's own energy, as calculated with its own sequence and structure. However, the updated interaction partners must be used. Using interaction partners from the template protein in an approximation that is formally equivalent to the frozen approximation used in the threading calculations increases the error to the point where the two-body interaction energy has almost no information.

Software tools developed in our group are made available to the TSRI community via the group's World Wide Web server. The first program available on the server is a new method for predicting secondary structure. The method is based on finding similarities between sequence segments from the target sequence and segments contained in the database of proteins with known structures. The accuracy of the predicted three-state secondary structure reaches 72.4% on a nonhomologous (maximum sequence identity <25%) data set derived from the Protein Data Bank and has been reproduced on two independent test sets.

The prediction method was developed with simplicity and open architecture in mind, allowing for an easy extension to other types of predictions and to the analysis of the contributions to the local formation of structure. For instance, the design of the method enables us to trace back segments of the database that contributed to the prediction. It can be shown that those segments came from various structural classes and that even the complete exclusion of related folds from the database does not result in a significant drop in prediction accuracy. Other tools, including the fold prediction algorithm, will be made available soon.

PUBLICATIONS

Godzik, A. Counting and classifying possible protein folds. Trends Biotechnol. 15:147, 1997.

Hu, W.-P., Godzik, A., Skolnick, J. Sequence-structure specificity: How does an inverse folding approach work? Protein Eng. 10:317, 1997.

Keasar, C., Elber, R., Skolnick, J. Simultaneous and coupled energy optimization of homologous proteins: A new tool for structure prediction. Folding Design 2:247, 1997.

Kolinski, A. Collapse transitions in protein-like polymers: The effect of sequence patterns. Biopolymers, in press.

Kolinski, A., Galaska, W., Skolnick, J. On the origin of the cooperativity of protein folding: Implications from model simulations. Proteins 26:271, 1996.

Kolinski, A., Skolnick, J. Determinants of secondary structure of polypeptide chains: Interplay between short range and burial interactions. J. Chem. Phys. 107:953, 1997.

Kolinski, A., Skolnick, J. High coordination lattice models of protein structure, dynamics and thermodynamics. Acta Biochim. Pol., in press.

Kolinski, A., Skolnick, J. Lattice Models of Protein Folding, Dynamics and Thermodynamics. R.G. Landes, Austin, TX, 1996.

Kolinski, A., Skolnick, J., Godzik, A., Hu, W.-P. A method for the prediction of U-turns and transglobular connections in small proteins. Proteins 27:290, 1997.

Milik, M., Kolinski, A., Skolnick, J. An algorithm for rapid reconstruction of a protein backbone from carbon coordinates. J. Comp. Chem. 18:80, 1997.

Ortiz, A., Hu, W.-P., Kolinski, A., Skolnick, J. Method for low resolution prediction of small protein tertiary structure. In: Proceedings of the Pacific Symposium on Biocomputing 97. Altman, R.B., et al. (Eds.). World Scientific, River Edge, NJ, 1997, p. 316.

Ortiz, A., Hu, W.-P., Kolinski, A., Skolnick, J. A method for prediction of the low resolution tertiary structure of small proteins. J. Mol. Graph., in press.

Pawlowski, K., Jaroszewski, L., Bierzynski, A., Godzik A.

Multiple model approach: Dealing with alignment ambiguities in comparative protein modeling. In: Proceedings of the Pacific Symposium on Biocomputing 97. Altman, R.B., et al. (Eds.). World Scientific, River Edge, NJ, 1997, p. 328.

Rychlewski, L., Godzik, A. Local sequence similarity secondary structure prediction. Protein Eng., in press.

Skolnick, J. A Monte Carlo model of fd and Pf1 coat proteins in membranes. Chemtracts 10:242, 1997.

Skolnick, J., Jaroszewski, L., Kolinski, A., Godzik, A. Derivation and testing of pair potentials for protein folding: When is the quasi-chemical approximation correct? Protein Sci. 6:676, 1997.

Skolnick, J., Kolinski, A. Monte Carlo approaches to the protein folding problem. Adv. Chem. Phys., in press.

Skolnick, J., Kolinski, A. Monte Carlo lattice dynamics and the prediction of protein folds. In: Computer Simulations of Biomolecular Systems: Theoretical and Experimental. van Gunsteren, W.F., Weiner, P.K., Wilkinson, A.J. (Eds.). ESCOM Science, Leiden, the Netherlands, in press.

Skolnick, J., Kolinski, A. Protein modelling. In: Encyclopedia of Computational Chemistry. Schleyer, P., Kollman, P. (Eds.). Wiley, New York, in press.

Skolnick, J., Kolinski, A., Ortiz, A. MONSSTER: A method for folding globular proteins with a small number of distance restraints. J. Mol. Biol. 265:217, 1997.

Skolnick, J., Milik, M. Modeling of membrane proteins and peptides. In: Membrane Proteins Assembly. Part IV, Modeling and Simulation. von Heijne, G. (Ed.). R.G. Landes, Austin, TX, 1997, p. 201.

Zhang, B., Jaroszewski, L., Rychlewski, L., Godzik, A. Similarities and differences between nonhomologous proteins with similar folds: Evaluation of threading strategies. Folding Design, in press.

Top of Page


Molecular Dynamics of Proteins and Peptides

C.L. Brooks III, B. Dominy, Z. Guo, J.D. Hirst, X. Kong, R.T. Morton, Y. Nochomovitz, J. Radkiewicz, F.B. Sheinerman, W.A. Shirley, M. Vieth, W.S. Young,* Z. Zhou

* Molecular Simulations, Inc., San Diego, CA

Understanding the atomic-level forces that determine the structure of proteins, peptides, and protein-peptide complexes and the processes by which these structures are adopted is essential for complete knowledge of protein and peptide structure and function. To address such questions, we use statistical mechanics, molecular simulation, statistical modeling, and quantum chemistry.

Building atomic-level models to simulate biophysical processes (e.g., protein folding or ligand binding to a biological receptor) requires (1) the development of potential functions that accurately represent the atomic interactions and (2) the use of quantum chemistry to aid in characterizing these models. Calculation of thermodynamic properties requires the development and implementation of new theoretical and computational approaches that connect averages over atomistic descriptions to experimentally measurable thermodynamic and kinetic properties.

Interpreting experimental results at a more atomic level leads to the development of theoretical models for these processes. Additionally, the massive computational resources needed to produce atomic-level descriptions of proteins, peptides, and protein-peptide complexes in solution motivate efforts aimed at the efficient use of new computer architectures, including large supercomputers. Each of the tools or areas mentioned represents ongoing areas of development in our research program in computational biophysics.

RAPID SCREENING OF ENZYME INHIBITORS

Quantitative and semiquantitative assessment of relative binding affinities of ligands for biological receptors is a key objective for computational scientists involved in rational drug design. We are developing new methods for free-energy simulations that may provide a powerful and flexible framework for exploring interactions between proteins (or nucleic acid) and ligands. During the past year, we have extended our novel synthesis of methods that enable us to do such calculations in a more controlled and systematic manner. Our new approach, termed -dynamics, not only enables us to make conventional pairwise comparisons between putative inhibitors for a targeted receptor but also facilitates calculations of "competitive" binding, in which multiple ligands are simultaneously compared for their binding ability.

The use of -dynamics for competitive binding in computational drug design is illustrated in Figure 1. The population of the various para compounds and the parent benzamidine in the binding pocket is modulated by the free-energy perturbation variables, i, which evolve throughout the course of the calculation to separate the best binding inhibitors from the poorer binders, thereby providing a relatively rapid computational screen of putative inhibitors.

These developments are complemented by ongoing work in simplified treatment of solvation that uses continuum electrostatics and surface area--based solvation methods and by efforts to develop peptide and peptide-mimetic conformational databases for the development and assessment of pharmacophores. Our efforts are focused on a number of specific biological targets, including inhibitors for HIV protease.

MECHANISMS OF PROTEIN FOLDING

Understanding the means by which a linear amino acid sequence adopts its functional three-dimensional structure is a key challenge for scientists in many disciplines, from biology to physics. We have been using statistical mechanics and computer simulation to elucidate the principles that govern this process. Our exploration of the links between protein topology and the overall mechanism of protein folding is providing new insights that fuel further development of theoretical models and experiments. Our specific focus during the past year has been the folding mechanism of simple single-domain proteins with different overall topology. We have computed first-principles free-energy landscapes for the folding of two distinct proteins, an all-helical protein and a protein with a mixed /ß motif.


Figure 2 shows the molecular topologies and the corresponding folding free-energy surfaces from our calculations on fragment B of staphylococcal protein A and segment B1 of streptococcal protein G. These free-energy surfaces present a landscape for folding that connects topology with mechanism. For the helical protein, folding, which is dominated by local interactions, is downhill with concomitant collapse and formation of native tertiary structure as indicated by the "diagonal" nature of the free-energy surface projected onto the Rg and CNat reaction coordinates. The folding of protein G involves initial collapse, with only small amounts of native tertiary structure being formed, and then a "search" through compact conformational states for the native structure.

On the basis of these findings, and other ongoing work in our laboratory, it appears that this general picture holds. That is, for proteins with more delocalized topologies--ß-sheet structure--folding includes a collapse phase before the acquisition of native structure; more localized structures can fold with concomitant collapse and formation of native structure. Our work in this area is coupled to the theoretical and experimental developments ongoing in other laboratories at TSRI and in the La Jolla area.

PUBLICATIONS

Boczko, E.M., Brooks, C.L. III. On the use of information functions in cluster analysis: Application to clustering protein structures. Proteins, in press.

Guo, Z., Boczko, E.M., Brooks, C.L. III. Exploring the folding free energy surface of a three-helix bundle protein. Proc. Natl. Acad. Sci. U.S.A. 94:10161, 1997.

Guo, Z., Brooks, C.L. III. Thermodynamics of protein folding: A statistical mechanical study of a small ß protein. Biopolymers, in press.

Hirst, J.D., Hirst, D.M., Brooks, C.L. III. Multireference configuration interaction calculations of electronic states of N-methylformamide, acetamide, and N-methylacetamide. J. Phys. Chem. 101:4821, 1997.

Karpen, M.E., Brooks, C.L. III. Modelling protein conformation by molecular mechanics and dynamics. In: Prediction of Protein Structure: A Practical Approach. Sternberg, M.J.E. (Ed.). Oxford University Press, New York, 1996, p. 229.

Reddy, V.S., Giesing, H.A., Morton, R.T., Kumar, A., Post, C.B., Brooks, C.L. III, Johnson, J.E. Energetics of quasi-equivalence: Computational analysis of protein-protein interactions in icosahedral viruses. Biophys. J., in press.

Sheinerman, F.B., Brooks, C.L. III. A molecular dynamics simulation study of segment B1 of protein G. Proteins, in press.

Sheinerman, F.B., Brooks, C.L. III. Molecular picture of folding of a small /ß protein. Proc. Natl. Acad. Sci. U.S.A., in press.

Shirley, W.A., Brooks, C.L. III. Curious structure in 'canonical' alanine-based peptides. Proteins 28:59, 1997.

Young, W.S., Brooks, C.L. III. A reexamination of the hydrophobic effect: Exploring the role of the solvent model in computing the methane-methane potential of mean force. J. Chem. Phys. 106:9265, 1997.

Top of Page


Biological Applications of Computational Chemistry

J.D. Hirst, C.L. Brooks III

The amide bond is the fundamental constituent of proteins. Consequently, knowledge of the optical properties of the amide bond is key to understanding the absorption and circular dichroism spectra of proteins. Development of this knowledge requires a theoretical description of the electronic states of amides. We recently completed multireference configuration interaction calculations on a series of small amides. These calculations have provided us with the most reliable description to date of the electronic excited states of amides.

Figure 1 shows the permanent charge distributions of the n* and * states of N-methylacetamide. Both obvious and subtle differences between our calculations and older semiempirical calculations are evident. We are currently assessing the applicability of these charge distributions in calculations of the circular dichroism of proteins. This work represents an exciting bridge between state-of-the-art quantum chemical methods and the understanding of an important biophysical phenomenon.

Circular dichroism spectroscopy is one of the most widely used analytical tools in the study of protein structure and folding. It provides a coarse measure of the secondary structure of a protein in solution. We are developing the theoretical basis of circular dichroism to further our understanding of how protein conformation determines the circular dichroism spectra. Circular dichroism arises from the electronic transitions of chiral compounds.

We are using quantum chemistry calculations to investigate the electronic absorption spectra of small amides. In parallel with the ab initio approach, we are also investigating an empirical derivation of parameters to describe the electronic states of the amide chromophore. These parameters will be used to develop a quantitative method to calculate the circular dichroism of given protein structures and thus provide the connection between the atomic-detailed structure of proteins and their circular dichroism spectra.

We are also pursuing research in the area of computer-aided drug design. The advent of combinatorial chemistry technologies demands the development of fast and accurate quantitative structure-activity relationships, with the goal of directing the robotic synthesis of compounds in real time. We have developed a fast and reliable method for deriving nonlinear quantitative structure-activity relationships. In a test case on the inhibition of dihydrofolate reductase by pyrimidines, the method performed as well as other state-of-the-art methods such as neural networks, but with much greater speed and algorithmic simplicity. Efficiency will be a key issue for the analysis of data sets of thousands of molecules.

PUBLICATIONS

Hirst, J.D., Hirst, D.M., Brooks, C.L. III. Multireference configuration interaction calculations of electronic states of N-methylformamide, acetamide, and N-methylacetamide. J. Phys. Chem. 101:4821, 1997.

Top of Page


Molecular Graphics and Computation

A. Olson, B. Duncan, D. Goodsell, M. Pique, C. Carrillo, R. Huey, T. Larsen, T. Macke, G. Morris, M. Rao, B. Reva, M. Sanner, R. Rosenstein, C. Rosin,* A. Tchekirov, W. Walker**

* University of California, San Diego, CA
** University of California, Los Angeles, CA

The combined exponential growth of computational power and biostructural information presents an expanding frontier of research opportunities and challenges. This laboratory is addressing these challenges by developing novel software technologies and applying these approaches to problems in biology and medicine. During the past year, we significantly extended our computational and visualization tools and our analysis and prediction of important biological systems. This report describes our advances in the application of docking technologies to the understanding and design of biomolecular interactions, the development of new approaches to the extension of molecular structural to higher order complexes, and novel technical approaches to the problem of extending the usability of molecular software.

DOCKING TECHNOLOGY

Proteins rarely perform their biological function in isolation; they typically operate in conjunction with other molecules. Consequently, understanding protein interactions is an important task. We have continued our development of computational methods for characterizing, predicting, and visualizing these interactions.

Two complementary techniques for predicting molecular interaction are under study: AutoDock, a program for docking flexible ligands to macromolecules, and SurfDock, a program for predicting protein-protein interactions by using a hierarchical surface definition based on spherical harmonics. AutoDock is a mature technique that has been distributed to more than 200 research sites worldwide. SurfDock is rapidly becoming a generally useful and useable tool that will have a significant impact on the characterization of drug targets and other macromolecular assemblies. Our HARMONY package computes multiresolution representations of molecular surfaces, and the SurfDock program uses these representations to analyze and predict interactions between proteins of known structure. We have extended our computational methods by developing multiresolution pattern-matching algorithms for predicting interaction sites and have constructed a multiresolution database of protein complexes so that molecular interactions can be analyzed at different spatial resolutions.

To aid in the docking of flexible molecules by pairwise domain docking and in building up large molecular complexes interactively, we have developed tools that use an efficient sphere hierarchy to allow fast bump checking between molecules. Pairwise overlap is determined by using this sphere hierarchy. The bump-checking algorithm traverses the hierarchy and stops at the first level where no overlap is detected. If the final level is reached and the sets of spheres are intersecting, the molecules are considered intersecting, and the volume of intersection between the sphere sets is quantified. Barns-Hut Trees are used to increase the speed of the calculation, making an overlap check O(nlogn) rather than O(n2), where n is the number of spheres at the given level.

In other efforts, we are focusing on applying SurfDock and AutoDock to problems of biomedical and structural biological interest and on developing methods to combine the strengths of both approaches to address each problem.

BLOOD COAGULATION FACTORS IN TREATMENT OF THROMBOTIC DISEASES

We have predicted the ternary complex that consists of human tissue factor (TF) plus coagulation factors VIIa and X, the complex that leads to the conversion of factor X to active factor Xa. Earlier we proposed a model for the TF-VIIa complex (Fig. 1), which closely resembled the subsequently reported crystal structure. Recently, we used SurfDock to predict structures for the ternary complex. Factor X was docked by parts to the previous TF-VIIa model; the crystal structure of the fragment consisting of protease plus epidermal growth factor 2 was docked, as well as homology models of the EGF1 and Gla domains of epidermal growth factor. In our predicted ternary complex, both EGF1 and Gla domains of factor Xa extensively interact with tissue factor.

In collaboration with M. Johnson, University of Illinois at Chicago, we have designed twelve 2,7-bis-(4-amidinobenzilidene)-cycloheptan-1-one derivatives to bind with factor Xa with high affinity by optimizing their interaction with the factor. In these complexes, both the s1 and s4 pockets of factor Xa are filled by benzamidine groups, and the inhibitor heptane ring interacts with the catalytic triad. These inhibitors are being synthesized and tested for antithrombotic activity.

Protease inhibitors of HIV type 1 are showing remarkable success in the treatment of AIDS, but resistance mutation remains a major obstacle. We are using the protease of feline immunodeficiency virus as a model for mutation, searching for inhibitors that bind strongly the proteases of both the feline virus and the human virus. Using AutoDock, we are designing a series of keto-amide peptidomimetics (Fig. 2) in collaboration with C.-H. Wong in the Department of Chemistry.

DESIGN OF OPTIMAL INHIBITORS

We are studying the more general problem of how to create an optimal inhibitor for a given system. HIV protease is a challenging example: it binds and cleaves a diverse collection of peptide substrates, not a single well-defined substrate as most enzymes do. The optimal type of inhibitor for this type of "fuzzy" recognition is not clear. Do we want to find inhibitors that closely mimic one of the substrates, as one might do with a typical enzyme, or an inhibitor that somehow matches the "average" size of the active site? The question becomes even more difficult to answer in the case of HIV protease, because the enzyme mutates rapidly, with consequent shifting of binding affinities of different substrates and inhibitors. In collaboration with R. Belew, University of California, San Diego, we are approaching this problem by using competitive coevolution simulations. The simulation pits a population of inhibitors against a population of mutating proteases. The inhibitors strive to inhibit all the proteases, and the proteases attempt to evade the inhibitors while remaining competent to cleave the enzymes' natural substrates.

We are also studying the theoretical limits of sequence-specificity in DNA-binding molecules. J. Gottesfeld, Department of Molecular Biology, and P. Dervan, Caltech, have recently shown that polyamides that bind in the minor groove of DNA effectively block transcription in living cells and thus are potentially useful in cancer chemotherapy. Currently, the polyamides under study are partially sequence-specific: they can distinguish cytosine, guanine, and adenine plus thymine, but they cannot distinguish adenine from thymine. In collaboration with E. Landaw, University of California, Los Angeles, we are studying the limits of specificity possible with the current types of inhibitors and are defining the modifications that are needed to provide full sequence specificity.

DOCKING POTENTIALS

We are currently developing and testing new energy functions for evaluating protein-protein complexes. The potentials are derived by using a theory of Boltzmann-like statistics from a database of protein structures based on -carbon or ß-carbon positions. A parallel program that performs a six-degrees-of-freedom search with local minimization has been written to test the potentials. For a test set of 12 protein complexes, approximately 1 million dockings are generated for each complex, and the rank of the native is observed. We found that these potentials narrow the search space by roughly four orders of magnitude.

We have developed a new force field for AutoDock that predicts binding free energies for ligand-protein complexes with a standard error of about 2 kcal/mol. This level of error is sufficient to discriminate between millimolar, micromolar, and nanomolar inhibition constants. The force field incorporates a measure of the change in free energy of solvation on binding that is easily incorporated into our precalculation of atomic affinity grids. This new force field successfully predicted the free energy of binding of 20 HIV type 1 protease inhibitors and was able to separate binders from nonbinders in a series of 35 nitrogen-heterocyclic ligands and the W191G-mutant of cytochrome c peroxidase.

FROM ATOMS TO CELLS

A new focus of the Molecular Graphics Laboratory is the simulation of large macromolecular ensembles. An experimentally invisible size range exists between the level of molecular structure and the level of cellular ultrastructure. Techniques of microscopy are useful for probing the ultrastructure of cells and large molecular assemblies, but the rigors of sample preparation and radiation damage limit the resolution to the molecular level, obscuring atomic details. X-ray crystallography and nuclear magnetic resonance spectroscopy, on the other hand, yield atomic information, but only for single purified subjects or defined molecular assemblies. By computationally simulating the atomic structure of cellular ensembles, we seek to bridge this gap between atoms and cells.

As a first step, we have developed a modular "symmetry server" to facilitate the simulation of complex multimolecule assemblies (Figs. 3 and 4). The server is composed of a collection of symmetry-generating modules and a set of basic transformation modules that can be combined to form custom applications. Currently, the server is available in a stand-alone version and as a set of AVS modules. The diverse tools available within the AVS data-flow environment allow the use of hierarchical surface representations and property-based texture mapping, all within an interactive environment. In collaboration with J. Johnson and J. Tainer, Department of Molecular Biology, and R. Milligan and M. Yeager, Department of Cell Biology, we are applying the symmetry server to the assembly of macromolecular complexes.

ALGORITHMIC AND PROGRAMMING ADVANCES

One of the major problems in a computational laboratory is the difficulty of having a multitude of approaches and algorithms work together to solve a scientific problem. We have taken the approach of developing much of our graphics applications within the programmable data-flow environment, AVS. We have now expanded this effort by exploring interpreted environments to provide the "glue" that links our computational modeling tools together.

PYTHON is a high-level object-oriented interpreted language that is freely available. It is easy to make existing programs available in this interpreter by writing an interface to the different functions that one wants to be able to call up. It is also easy to extend the language by creating new data types to describe complex data structures. These qualities make PYTHON an excellent tool to create an homogeneous environment in which one can design scripts calling up a large variety of computational tools to address problems that were extremely difficult in the past because of the lack of interoperability of these different tools.

We have started to interface several widely used packages with PYTHON: (1) MSMSLIB, our molecular surface computation library, including a technique for partial reconstruction of the surface after moving a set of atoms; (2) AMBER, a package for performing molecular mechanics minimizations and molecular dynamics; (3) AutoDock, our program for predicting docking of flexible substrates to protein targets; (4) PDB-READER, a PYTHON module that provides flexible entry and manipulation of protein-structure information; (5) AVSPython, a link to our visualization environment that provides an embedded PYTHON interpreter in a general-purpose AVS module that can be connected in a network with the extensive library of other AVS modules; and (6) MATRICES, a module to provide general structural coordinate transformation capability. These initial efforts have already begun to catalyze new modeling prototypes within the laboratory.

PUBLICATIONS

Goodsell, D.S. Our Molecular Nature: The Body's Motors, Machines and Messages. Springer-Verlag, New York, 1996.

Laco, G.S., Fitzgerald, M., Morris, G.M., Olson, A.J., Kent, S.B.H., Elder, J.H. Molecular analysis of the FIV protease: Generation of a novel form of the protease by autoproteolysis and construction of cleavage resistant proteases. J. Virol. 71:5505, 1997.

Laco, G.S., Schalk-Hihi, C., Lubkowski, J., Zdanov, A., Morris, G.M., Olson, A.J., Elder, J.H., Wlodawer, A., Gustchina, A. Crystal structures of the inactive D30N mutant of FIV protease complexed with a substrate and an inhibitor. Biochemistry 36:10696, 1997.

Olson, A.J., Goodsell, D.S. Automated docking and the search for HIV protease inhibitors. SAR QSAR Environ. Res., in press.

Olson, A.J., Pique, M.E. Visualizing the future of molecular graphics. SAR QSAR Environ. Res., in press.

Reva, B.A., Finkelstein, A.V., Sanner, M.F., Olson, A.J. Accurate mean-force pairwise residue potentials for discrimination of protein folds. In: Pacific Symposium on Biocomputing '97. Altman, R.B., et al. (Eds.). World Scientific Press, River Edge, NJ, 1997, p. 373.

Reva, B.A., Finkelstein, A.V., Sanner, M.F., Olson, A.J. Recognition of protein structures on coarse lattices with residue-residue energy functions. Protein Eng., in press.

Reva, B.A., Finkelstein, A.V., Sanner, M.F., Olson, A.J. Residue-residue mean-force potentials for protein structure recognition. Protein Eng., in press.

Sanner, M.F., Olson, A.J. Real-time surface reconstruction for moving molecular fragments. In: Pacific Symposium on Biocomputing '97. Altman, R.B., et al. (Eds.). World Scientific Press, River Edge, NJ, 1997, p. 385.

Walker, W.L., Kopka, M.L., Dickerson, R.E., Goodsell, D.S. Design of stapled DNA-minor-groove-binding molecules with a mutable atom simulated annealing method. J. Comput. Aided Mol. Des., in press.

Walker, W.L., Landaw, E.M., Dickerson, R.E., Goodsell, D.S. Estimation of the DNA sequence discriminatory ability of hairpin-linked lexitropsins. Proc. Natl. Acad. Sci. U.S.A. 94:5634, 1997.

Top of Page


Computer Modeling of Protein and Peptide Structure and Intermolecular Interactions

V.A. Roberts, E.H. Olender, J.L. Pellequer, M.E. Pique, M.M. Thayer, S.J. Benkovic,* A.E. Karu,** R.J. Nachman,*** L.F. Ten Eyck****

* The Pennsylvania State University, University Park, PA
** University of California, Berkeley, CA
*** U.S. Department of Agriculture, College Station, TX
**** San Diego Supercomputer Center, San Diego, CA

The rapid increase in the number of known protein sequences and structures is fueling the need for methods to predict protein structure and intermolecular interactions. We use computational and computer graphics techniques in conjunction with site-directed mutagenesis, peptide synthesis, and protein crystallography to develop testable hypotheses and direct protein engineering.

HIGH-RESOLUTION ANTIBODY MODELS

Our database of superimposed crystallographic antibody structures reveals the structural conservation of both the antibody backbone fold and the side chains that shape the antigen-binding pocket. Using the database, we constructed a three-dimensional model of the antibody 43C9, which efficiently catalyzes the hydrolysis of specific amides and esters. On the basis of the model, we hypothesized that two amino acid side chains were key for catalysis, and we designed two metal-binding mutants. The roles of the two amino acid side chains were verified by site-directed mutagenesis, but neither metal-binding mutant had catalytic activity.

In collaboration with E. Getzoff, Department of Molecular Biology, crystallographic structures of 43C9 and the metal-binding mutants are being determined. In the crystallographic structure of 43C9, one loop that makes up part of the antigen-binding site has a different conformation than predicted by the model. The rest of the structure agrees well with the model, including placement of the two catalytic side chains.

In another project, models were built for two antibodies that bind polyaromatic hydrocarbons, which are significant environmental contaminants. The antigen-binding sites have two positively charged side chains (Fig. 1) , which may enhance the binding specificity of the antibodies.

PREDICTING MACROMOLECULAR COMPLEXES

The computer program DOT has been developed to predict intermolecular interactions. DOT performs a complete, six-degree-of-freedom search of all configurations between two molecules. The search algorithm is fast, and the time required for computation does not depend directly on the size of the molecules being investigated, allowing the algorithm to be applied to proteins.

DOT is currently being tested on two types of systems: electron-transport proteins, which represent transient protein/protein interactions, and the DNA-repair enzyme uracil-DNA glycosylase (UDG), which forms an irreversible complex with an inhibitor, the protein UGI (Fig. 2). Results from DOT are being verified by comparison with the crystallographic structures of the complexes.

ACTIVE CONFORMATIONS OF INSECT NEUROPEPTIDES

Neuropeptides control diverse functions in living organisms. Surprisingly, neuropeptides found in insects often have sequence similarity to mammalian neuropeptides, suggesting the existence of neuropeptide superfamilies with shared conformational determinants. Conformational preferences determined by molecular dynamics simulations combined with structure/activity studies have led to the design of constrained cyclic analogs, highly active simplified linear analogs, and the first nonpeptidal ligand for an insect neuropeptide/receptor system.

PUBLICATIONS

Nachman, R.J., Roberts, V.A., Lange, A.B., Orchard, I., Holman, G.M., Teal, P.E.A. Active conformation and mimetic analog development for the pyrokinin-PBAN-diapause-pupariation and myosuppressin insect neuropeptide families. In: Phytochemicals for Pest Control. Hedin, P.A., et al. (Eds.). American Chemical Society, Washington, DC, 1997, p. 277.

Roberts, V.A., Nachman, R.J., Coast, G.M., Hariharan, M., Chung, J.S., Holman, G.M., Williams, H., Tainer, J.A. Consensus chemistry and ß-turn conformation of the active core of the myotropic/diuretic insect neuropeptide family. Chem. Biol. 4:105, 1997.

Top of Page


Metalloenzyme Engineering

D.B. Goodin, D.E. McRee, R.A. Musah, M.M. Fitzgerald, G.M. Jensen, Y. Cao, S.K. Wilcox, R.J. Rosenfeld, S.W. Bunte,* T.M. Loehr,** R.D. Guiles,*** T.G. Spiro****

* U.S. Army Research Laboratory, Aberdeen, MD
** Oregon Graduate Institute, Portland, OR
*** University of Maryland, Baltimore, MD
****Princeton University, Princeton, NJ

Our research focuses on engineering novel function into oxidative metalloenzymes. We are interested in defining the physical, spectroscopic, and functional properties of heme peroxidases and in understanding how the protein environment confers diverse chemical functions on the active site. These studies encompass the techniques of molecular modeling, protein engineering, x-ray crystallography, electron paramagnetic resonance, electrochemistry, calorimetry, and kinetics. Such information has helped produce unifying themes that suggest that oxidative reactions catalyzed by some heme enzymes might be recruited into the scaffold of other enzymes.

We use this information to introduce new substrate binding sites designed to promote specific catalysis by electron-transfer or oxotransfer reactions. Artificial enzymes designed for specific activities could be widely useful because of their diagnostic, therapeutic, and biosynthetic capabilities. Heme-containing enzymes with artificially modified activities might be applied as novel catalysts for biochemical synthesis, breakdown of environmental carcinogens, and active-site elements of biosensors.

In the past year, we expanded our development of cavity complementation as a tool to introduce binding sites for small molecules into a heme peroxidase. The results show the feasibility of this new approach for placing potential substrates near the heme active site. They also provide a framework for answering some more general questions about weak interactions in ligand-protein complexes and the movement of surface loops that were not originally anticipated.

We have solved x-ray crystal structures of seven examples of such artificial cavities. At least four of these cavity-containing mutants bind small molecules that reflect the size, shape, hydrogen bonding, and electrostatic complementarity of the protein cavity. Two of the mutants catalyze the oxidation of small molecules bound at the artificial binding site by different mechanisms. In one case, 2-aminothiazole binds to a proximal cavity near the heme and acts as an electron donor to the oxidized heme center. In another, aromatic olefins bind to a cavity on the distal heme face and are epoxidized by an oxotransfer mechanism that may be analogous to the reactions catalyzed by cytochrome P-450. In both cases, we have obtained crystal structures of substrates bound to the artificial binding sites.

The crystal structures (Fig. 1), thermodynamic binding parameters, and kinetics for more than 25 compounds bound to one such cavity have provided a unique database for ongoing studies and collaborations aimed at understanding the importance of various interactions that contribute to the binding specificity of these compounds. These computational studies will address more general questions about the role of weak interactions in ligand-protein complexes. For example, we have characterized the binding of two similar thiazole compounds (Fig. 2) that differ only in the polarity of a CH bond that is involved in an unusual CH-to-O hydrogen bond with the protein; this difference results in a 1 kcal/mol difference in binding energy.

We have shown that partially filling an existing surface cavity with designed mutations affects the specificity of natural and nonnatural substrates that serve as electron donors, but in different ways. These studies provide support for proposals in which different electron-transfer pathways are used for different substrates. Finally, in a resonance Raman study, we showed that a ligand-deficient mutant of the enzyme exists as a pH-dependent equilibrium of five-coordinate aquo and hydroxo hemes, coordination states that have not been seen before in a protein. We plan to extend our studies of these cavity mutants to provide models for the coordination states of the medically important enzymes guanylyl cyclase and nitric oxide synthase.

PUBLICATIONS

Fitzgerald, M.M., Musah, R.A., McRee, D.E., Goodin, D.B. A ligand-gated, hinged loop rearrangement opens a channel to a buried artificial protein cavity. Nature Struct. Biol. 3:626, 1996.

Goodin, D.B. When an amide is more like histidine than imidazole: The role of axial ligands in heme catalysis. J. Biol. Inorg. Chem. 1:360, 1996.

Musah, R.A., McRee, D.E., Goodin, D.B. Introduction of novel substrate oxidation into a heme peroxidase by cavity complementation: Oxidation of 2-aminothiazole and covalent modification of the enzyme. Biochemistry, in press.

Sarma, S., DiGate, R.J., Goodin, D.B., Miller, C.J., Guiles, R.D. Effect of axial ligand plane reorientation on electronic and electrochemical properties observed in site-directed mutants of rat cytochrome b5. Biochemistry, in press.

Smulevich, G., Hu, S., Rodgers, K.R., Goodin, D.B., Smith, K.M., Spiro, T.G. Heme-protein interactions in cytochrome c peroxidase revealed by site-directed mutagenesis and resonance Raman spectra of isotopically labeled hemes. Biospectroscopy 2:365, 1996.

Sun, J., Fitzgerald, M.M., Goodin, D.B., Loehr, T.M. The solution and crystal structures of the H175G mutant of cytochrome c peroxidase: A resonance Raman study. J. Am. Chem. Soc., in press.

Wilcox, S.K., Jensen, G.M., Fitzgerald, M.M., McRee, D.E., Goodin, D.B. Altering substrate specificity at the heme edge of cytochrome c peroxidase. Biochemistry 35:4858, 1996.

Top of Page


Peptide Engineering and Vaccines

A.C. Satterthwait, E. Cabezas, R. Zhang, E. Dovalsantos, Y. Xu, G. Liu, L.-C. Chiang,* P. Berman**

* NuMega Resonance Laboratories, San Diego, CA
** Genentech, San Francisco, CA

We continue to develop synthetic methods and new strategies for shaping peptides to the conformations they occupy at the active sites of proteins. Conformation is the key to biological activity. Learning to control it at the synthetic level provides (1) methods for converting inactive peptides to bioactive peptides and (2) routes to drugs, vaccines, and biocatalysts.

The approach we take is based on the observation that regular and irregular structures in native proteins share a common feature: hydrogen bonds that link amino acids and define structure. However, the hydrogen bond is weak and insufficient for folding isolated peptides. To regain structure and activity, we replace structure-defining hydrogen bonds (NH...O=CRNH) with covalent mimics (NN=CHCH2CH2, NC=N(+)RCH2CH2), which are inserted into peptides by using automated multiple solid-phase peptide syntheses. These procedures have provided us with hundreds of peptides constrained to reverse turns, loops, and -helices that constitute the majority of structures found at protein active sites.

With ready access to folded peptides, we have been able to explore the role that conformation plays in enhancing immune responses. Immune responses provide the first and best defense against disease. In principle, the binding pockets of antibodies formed against shaped peptides should complement the ordered surfaces of proteins much better than antibodies formed against disordered peptides do. This situation has been difficult to produce and evaluate, because we must not only mimic a relevant conformation but also establish enhanced immune responses and rigorous structural correlates. In the past several years, however, we and our collaborators have begun to show the remarkable improvements in antigenicity and immunogenicity that can be achieved by folding peptides. This accomplishment opens the way for the conversion of immunologically inactive peptides to active peptides, which will no doubt be key to the usefulness of synthetic vaccines.

The determination of vaccine targets and relevant conformations is a considerable problem. Studies have revealed, contrary to hope and expectation, that few antibodies are capable of blocking pathogens. Thus, it appears that neutralization sites and vaccine targets are far more limited than previously anticipated. HIV type 1 is a case in point. After a decade of research and the examination of thousands of monoclonal antibodies formed against the virus, only a few antibodies show the broad and potent activities expected in an effective immune response. These antibodies, however, are critically important and provide keys to the development of synthetic vaccines. Our approach has been to simultaneously map and mimic the three-dimensional sites recognized by these antibodies with constrained peptides. The mimetics that emerge from these studies then become vaccine candidates with a potential for focusing immune responses to the most vulnerable sites on pathogens.

We have been concentrating our attention on several monoclonal antibodies that broadly and potently neutralize HIV type 1 by binding targets on the glycoproteins gp120 and gp41 that coat the virus. In our first study, we discovered a series of cyclic peptides that bind up to 200-fold better than corresponding linear peptides to a V3-directed antibody. Better binding is an indicator that folding had succeeded. Because R. Stanfield and I. Wilson, Department of Molecular Biology, have solved the crystal structures for peptides bound to the antibody, we have been comparing solution structures from nuclear magnetic resonance studies with the antibody-bound structures. Interestingly, these studies reveal that different cyclic peptides adapt different types of reverse turns in water, with the best binding loop forming, to a better degree, the same set of reverse turns bound by the antibody. Thus, the structure of the V3 sequence can be manipulated for immunologic testing with the goal of relating immune enhancement to specific structures.

Extensive work has been carried out with a monoclonal antibody against the C4 region of gp120 that blocks binding of HIVMN to CD4. Systematic screening of linear and cyclic peptides have indicated a bicyclic peptide that begins to define and mimic the conformation of a complex loop adjacent to the CD4 binding site. Exhaustive studies with systematic scans with amino acids, modified amino acids, and cyclic peptides have indicated conformational preferences and structural hypotheses for a gp41 neutralization site. Each of these studies has led to refinements in our strategy and new perspectives on structure and how to manipulate structure.

PUBLICATIONS

Cabezas, E., Satterthwait, A.C. The NMR structure of a V3 loop peptide that binds tightly to a monoclonal antibody that potently neutralizes HIV-1. In: Peptides: Chemistry, Structure and Biology. Proceedings of the 15th American Peptide Symposium. Tam, J.P., Kaumaya, P.T.P. (Eds.). ESCOM, Leiden, the Netherlands, in press.

Dovalsantos, E.Z., Cabezas E., Satterthwait, A.C. -helix nucleation between peptides with a covalent hydrogen bond mimic. In: Peptides: Chemistry, Structure and Biology. Proceedings of the 15th American Peptide Symposium. Tam, J.P., Kaumaya, P.T.P. (Eds.). ESCOM, Leiden, the Netherlands, in press.

Ghiara, J.B., Ferguson, D.C., Satterthwait, A.C., Dyson, H.J., Wilson, I.A. Structure-based design of a constrained-peptide mimic of the HIV-1 V3 loop neutralization site. J. Mol. Biol. 266:31, 1997.

Satterthwait, A.C., Cabezas, E., Zhang, R., Xu, Y., Dovalsantos, E.Z.

Conformational mapping of neutralizing epitopes with peptide mimetics. In: Peptides: Chemistry, Structure and Biology. Proceedings of the 15th American Peptide Symposium. Tam, J.P., Kaumaya, P.T.P. (Eds.). ESCOM, Leiden, the Netherlands, in press.

Zhang, R., Xu, Y., Nakamura, G.R., Berman, P.W., Satterthwait, A.C. Conformational mapping of a neutralizing epitope on the C4 region of HIV-1 gp120 with cyclic and bicyclic peptides. In: Peptides: Chemistry, Structure and Biology. Proceedings of the 15th American Peptide Symposium. Tam, J.P., Kaumaya, P.T.P. (Eds.). ESCOM, Leiden, the Netherlands, in press.

Top of Page


Directed Evolution of RNA and DNA Enzymes

G.F. Joyce, S.A. Raillard, M.C. Wright, T.L. Sheppard, J.K. Rogers, J. Nowakowski, X. Dai, J. Tsang, R.K. Bruick, S.W. Santoro, M. Anderson

The principles of darwinian evolution can be applied to a large, heterogeneous population of RNA or DNA molecules to obtain particular molecules that have desired biochemical properties, including the ability to catalyze a particular chemical reaction. A population of variant molecules is subjected to repeated rounds of selective amplification in the test tube. Only those molecules that perform a chosen catalytic task are amplified, so that through successive rounds, the population adapts to the task at hand. We can select from among trillions of variant molecules in less than 1 hour. This ability enables us to evolve nucleic acids rapidly, compared with the rate at which whole organisms evolve in nature.

In one study, we began with a naturally occurring RNA enzyme that cleaves single-stranded RNA. We then used in vitro evolution to derive novel RNA enzymes that cleave single-stranded DNA. The enzymes were selected for their ability to cleave a target DNA in a sequence-specific manner under conditions that resemble those of the cellular environment. After 63 "generations" of test-tube evolution, we obtained RNA enzymes that cleave DNA at a rate of about one per minute. This compares with a rate of about one per 200 million years for the uncatalyzed reaction. The evolved RNA enzymes were modified so that they recognize and cleave various 12-nucleotide target sequences within the DNA replication intermediate of HIV type 1. Cleavage occurs efficiently and with high sequence specificity, taking advantage of tertiary contacts within the catalytic center of the enzyme.

In another study, we sought to develop catalytic DNAs that cleave a target RNA, again in a sequence-specific manner under simulated cellular conditions. In this case, however, we began with a population of 1014 random-sequence DNAs. After 10 rounds of selective amplification, we obtained a variety of magnesium-dependent DNA enzymes, including ones that could be generalized to cleave almost any target RNA.

One such DNA enzyme has a catalytic domain of 15 deoxynucleotides flanked by two substrate-recognition domains of 8 deoxynucleotides each. Its activity depends on a divalent metal cation, which may be calcium, magnesium, or manganese. The RNA substrate is bound through Watson-Crick base pairing and is cleaved at a particular phosphodiester located between an unpaired purine and a paired pyrimidine residue. Despite its small size, the DNA enzyme has a catalytic efficiency of more than 109 . M-1 . min-1 under multiple-turnover conditions, exceeding that of other known nucleic acid enzymes and even exceeding that of the protein enzyme RNase A.

In addition to producing interesting new nucleic acid catalysts, our in vitro evolution studies have enabled us to follow the course of darwinian evolution at the molecular level, correlating specific changes in genetic sequence with the consequences of the changes at the level of catalytic function. This situation allows us to investigate evolutionary mechanisms that previously could only be inferred by examining extant organisms. Recently, we developed a method to evolve catalytic function continuously in vitro. During continuous evolution, RNA molecules that catalyze a target chemical reaction are immediately eligible for amplification, and newly produced RNAs are immediately eligible to catalyze another reaction.

Thus, we can maintain laboratory "cultures" of evolving RNA enzymes, analogous to the way one would maintain a culture of bacteria or some other organism. The RNAs are perpetuated by a simple serial transfer procedure; amplification occurs indefinitely so long as an ongoing supply of substrate and other reaction materials is made available (Fig. 1) . During one run of continuous in vitro evolution, the RNA enzymes were amplified by a factor of 10298 in 52 hours. By the end of this process, new "generations" of progeny RNA molecules were being produced approximately every 5 minutes.

The study of RNA-based evolving systems is relevant to understanding the early history of life on Earth. It is thought that an RNA-based genetic system, often referred to as the "RNA world," preceded the DNA- and protein-based genetic system that has existed on this planet for the past 3.5 billion years. One aim of our research program is to recapitulate the biochemistry of the RNA world in the laboratory. We are using in vitro evolution to explore the catalytic potential of RNA and, in particular, to search for RNA enzymes that can catalyze their own replication. In vitro evolution, most notably continuous in vitro evolution, provides a laboratory tool for recreating functional aspects of the RNA world.

PUBLICATIONS

Bruick, R.K., Koppitz, M., Joyce, G.F., Orgel, L.E. A simple procedure for constructing 5´-amino-terminated oligodeoxynucleotides in aqueous solution. Nucleic Acids Res. 25:1309, 1997.

Hendrix, M., Priestley, E.S., Joyce, G.F., Wong, C.-H. Direct observation of aminoglycoside-RNA interactions by surface plasmon resonance. J. Am. Chem. Soc. 119:3641, 1997.

Joyce, G.F. Evolutionary chemistry: Getting there from here. Science 276:1658, 1997.

Joyce, G.F., Orgel, L.E. The origins of life: A status report. Am. Biol. Teacher, in press.

Joyce, G.F., Still, W.C., Chapman, K.T. Combinatorial chemistry: Searching for a winning combination. Curr. Opin. Chem. Biol. 1:3, 1997.

Santoro, S.W., Joyce, G.F. A general-purpose RNA-cleaving DNA enzyme. Proc. Natl. Acad. Sci. U.S.A. 94:4262, 1997.

Wright, M.C., Joyce, G.F. Continuous in vitro evolution of catalytic function. Science 276:614, 1997.

Top of Page


From Catalytic Asymmetric Synthesis to the Transcriptional Regulation of Human Genes and the Neutralization of HIV-1

C. Barbas III, R. Lerner, H. Almer, J. Anderson, R. Beerli, K. Bower, D. Burton,* M.G. Finn,** R. Fuller, J. Ghiara, S. Gramatikova, T. Hoffmann, R. Lewis, B. List, Q. Liu, C. Rader, K. Sakthivel, D. Segal, D. Shabat, P. Steinberger, S. Sweeney, F. Tanaka, S. Venturini, J. Widhopf, G. Zhong

* Department of Immunology, TSRI
** University of Virginia, Charlottesville, VA

Our interest lies in the directed evolution of proteins as a route to producing therapeutic proteins and understanding the principles of molecular recognition and catalysis. To explore this area, we have developed filamentous phage display as a format for molecular evolution. The surface of the filamentous phage physically links the genotype and phenotype of the protein under study. This linkage is necessary for studies that go beyond RNA libraries in which genotype and phenotype may be encoded and expressed by the same molecule. Three areas are currently being examined: catalytic antibodies, therapeutic human antibodies, and zinc-finger recognition of DNA and RNA.

Using the concept of reactive immunization, we produced monoclonal antibodies to a 1,3-dicarbonyl hapten designed to act as a chemical and entropic trap. The hapten selects for antibodies with a lysine residue in the active site by forming a covalently bound enaminone. This reaction mechanism for the formation of the hapten-antibody complex was used to catalyze aldol reactions after a similar mechanism known from natural class I aldolase enzymes: formation of an enamine between the donor carbonyl compound and the -amino group of the essential lysine residue in the binding pocket of the antibody is followed by the nucleophilic attack of this enamine at the carbonyl acceptor substrate and final hydrolytic release of the aldol product.

We found antibodies that catalyze aldol as well as retro aldol reactions of a wide variety of aliphatic open-chain and aliphatic cyclic ketones to various aromatic and aliphatic aldehydes (Fig. 1). More than 100 different substrate combinations--cross aldol and intramolecular aldol reactions--were detected. Most remarkable is the enantioselective formation of the Wieland Miescher ketone. Typical values for kcat range from 10-3/min to 10/min and show a ratio of kcat/kuncat of 105--108. Moreover, the same antibodies catalyze the decarboxylation of ß-keto acids by using the lysine residue and thus mimic the natural enzyme acetoacetate decarboxylase. Structural and chemical characterization of these catalysts has revealed the mechanism whereby both scope and catalytic efficiency can be selected. Further study of these catalysts is providing insight into the evolution and diversification of catalytic proteins.

The ability to manipulate large libraries of human antibodies has immediate implications for the development of therapeutic antibodies. For some time, we have been characterizing the antibody response induced by infection with HIV-1. One of our goals is to create a potent therapeutic cocktail of antibodies specific for HIV-1 that can inhibit mother-to-child transmission of the virus and, in combination with existing HIV treatments, have a therapeutic effect on infected patients. We intend to translate our knowledge of HIV neutralization into new vaccine strategies. Our most successful antibody has shown efficacy in protecting animals from primary challenge with HIV-1. In vitro evolution strategies have been used to increase the binding and neutralization activity of this antibody.

The discovery of a second class of targets for HIV-1 neutralization, a subset of the chemokine receptor family, has led to a new program in the laboratory. We have developed antibodies that bind to the HIV-1--permissive chemokine receptors and block viral entry into the cell. The strategy we used includes selection of antibodies against cells expressing one or more of the coreceptors. We hope that the design and development of antibodies to these chemokine receptors will lead to a better understanding of the process involved in viral entry.

Our third area of investigation involves the selection of novel zinc-finger DNA-binding proteins. Zinc-finger proteins are particularly well suited for this purpose because of their modularity and well-defined structural features. Each finger forms an independently folded domain that typically recognizes three nucleotides of DNA. We and others have shown that proteins can be selected or designed that contain zinc fingers that recognize novel DNA sequences. These studies are aiding the elucidation of rules for sequence-specific recognition within this family of proteins.

Recently, we showed that selected zinc fingers can be appropriately linked to form polydactyl proteins capable of recognizing an 18-nucleotide site and thus can target a unique locus in the human genome (Fig. 2). These proteins bind with subnanomolar affinity and are highly specific in cell culture assays. Attachment of a nuclear localization signal and domains enabled targeted gene repression and, in contrast with other antisense or antigene targeting methods, targeted gene activation. These results suggest that zinc-finger proteins can potentially be used as genetic regulators for a variety of human aliments. We are constructing multifinger proteins for the specific recognition of HIV-1 DNA and of genes important for HIV infection, angiogenesis, and tumor malignancy.

PUBLICATIONS

Barbas, C.F. III, Burton, D.R. Selection and evolution of high-affinity human anti-viral antibodies. Trends Biotechnol. 14:230, 1996.

Bjornestedt, R., Zhong, G., Lerner, R.A., Barbas, C.F. III. Copying nature's mechanism for the decarboxylation of ß-ketoacids into catalytic antibodies by reactive immunization. J. Am. Chem. Soc. 118:11720, 1996.

Crowe, J.E., Jr., Firestone, C.-Y., Crim, R., Beeler, J.A., Coelingh, K.L., Barbas, C.F. III, Burton, D.R., Chanock, R.M., Murphy, B.R. Monoclonal antibody resistant mutants selected with an respiratory syncytial virus (RSV) neutralizing human antibody Fab fragment (Fab19) define a unique epitope on the fusion (F) glycoprotein. J. Virol., in press.

Ditzel, H.J., Parren, P.W.H.I., Binley, J.M., Sodroski, J., Moore, J.P., Barbas, C.F. III, Burton, D.R. Mapping the protein surface of HIV-1 gp120 using human monoclonal antibodies from phage display libraries. J. Mol. Biol. 267:684, 1997.

Hu, D.D., Barbas, C.F. III, Smith, J.W. Evidence for an allosteric Ca2+ binding site on the ß3-integrins that regulate the dissociation rate for RGD ligands. J. Biol. Chem. 36:21745, 1996.

Kessler, J.A. II, McKenna, P.M., Emini, E.A., Chan, C.P., Patel, M.D., Gupta, S.K., Burton, D.R., Barbas, C.F. III, Mark, G.E. III, Conley, A.J. The recombinant human monoclonal antibody IgG1b12 neutralizes diverse human immunodeficiency virus type 1 primary isolates. AIDS Res. Hum. Retroviruses 13:575, 1997.

Lang, I.M., Barbas, C.F. III, Schleef, R.R. Recombinant rabbit Fab to type 1 plasminogen activator inhibitor derived from a phage-display library against human -granules. Gene 172:295, 1996.

Lang, I.M., Chuang, T.L., Barbas, C.F. III, Schleef, R.R. Purification of storage granule protein-23. J. Biol. Chem. 271:30126, 1996.

Lerner, R. A., Barbas, C.F. III. Using the process of reactive immunization to induce catalytic antibodies with complex mechanisms: Aldolases. Acta Chem. Scand. 50:672, 1996.

Liu, Q., Segal, D.J., Ghiara, J.B., Barbas, C.F. III. Design of polydactyl zinc finger proteins for unique addressing within complex genomes. Proc. Natl. Acad. Sci. U.S.A. 94:5525, 1997.

Mo, H., Stamatatos, L., Ip, J.E., Barbas, C.F. III, Parren, P.W.H.I., Burton, D.R., Moore, J.P., Ho, D.D. Human immunodeficiency virus type 1 mutants that escape neutralization by human monoclonal antibody IgG1b12. J. Virol., in press.

Parren, P.W.H.I., Fisicaro, P., Labrijn, A.F., Binley, J.M., Yang, W.-P., Ditzel, H.J., Barbas, C.F. III, Burton, D.R. In vitro antigen 'challenge' of human antibody libraries for vaccine evaluation: The human immunodeficiency virus type 1 envelope. J. Virol. 70:9046, 1997.

Rader, C., Barbas, C.F. III. Phage display of combinatorial antibody libraries. Curr. Opin. Biotechnol., in press.

Wagner, J., Lerner, R.A., Barbas, C.F. III. Synthesis of five enantiomerically pure haptens designed for in vitro evolution of antibodies with peptidase activity. Bioorg. Med. Chem. 4:901, 1996.

Zhong, G., Hoffmann, T., Lerner, R.A., Danishefsky, S., Barbas, C.F. III. Antibody catalyzed enantioselective Robinson annulation. J. Am. Chem. Soc., in press.

Top of Page


Reaction Coordinate Manipulation and the Potential Treatment of Metabolic Disease Through Catalytic Antibodies

C. Shevlin, R. Lerner

One of the goals of catalytic antibody research is to provide tailor-made catalysts for the synthetic chemist. By exploiting nature's own combinatorial system, it has been possible to elicit immunoglobulins that catalyze reactions that are difficult or impossible to carry out by other means. With this powerful tool at hand, it is not unreasonable to suggest that methods that use catalytic antibodies may also be pivotal in treating a number of metabolic diseases.

REACTION COORDINATE MANIPULATION

One major focus of research in our laboratory is the development of antibody-assisted disfavored reactions. We have elicited immunoglobulins capable of catalyzing chemical reactions along an energetically unfavorable reaction pathway, providing an overwhelming enhancement of disfavored products relative to the uncatalyzed reaction (Fig. 1).

Recently, we generated two new antibodies, 5C8 and 14B9, that can catalyze the disfavored enantioselective 6-endo-tet closure of a -hydroxy epoxide. The crystal structures of both native and hapten-bound 5C8 have been solved by I. Wilson's group, Department of Molecular Biology. These results, along with the sequences of other homologs, have provided pertinent information toward a mechanistic description of the action of this outstanding group of catalytic antibodies.

The ability to modulate the path taken by a substrate along its reaction coordinate through de novo catalyst design has inspired us to investigate other types of reactions. One area that we are exploring is the use of catalytic antibodies to probe and manipulate the reaction pathway of carbocation rearrangements associated with the "memory effect." We anticipate that through judicious hapten design, an antibody binding pocket can be programmed that will alter the course of a substrate along the reaction coordinate so that an intermediate carbocation will rearrange via an alternate pathway, resulting in an inverse memory effect. Information gained from these investigations will provide insight into the still-unresolved basis of this phenomenon.

GAUCHER'S DISEASE

Gaucher's disease is one of several inborn errors of metabolism characterized by multisystem abnormalities, including degeneration of the skeletal system and enlargement of the liver and spleen. These effects are due to a deficiency in glucocerebrosidase that causes the accumulation of glucosylceramide in macrophages. We have begun synthesizing haptens designed to mimic the transition state for the cleavage of the glycosidic bond of glucosylceramide. Immunization with these hapten conjugates should elicit an immune response, resulting in the generation of antibodies with glucocerebrosidase activity. Because IgG can traverse the cell membrane of macrophages, antibodies with glucocerebrosidase activity may be clinically valuable in the treatment of Gaucher's disease. Design of systems for related disorders such as Fabry's disease and Krabbe's disease are also being pursued.

Top of Page


Organic Synthesis and Transformations Via Antibody Catalysis

E. Keinan, S.C. Sinha, F. Grynszpan, D. Shabat, H. Itzhaky, A. Haskel, S. Nimri, O. Levy, S. Saphier, H. Shulman, A. Shulman

Catalytic antibodies catalyze a wide variety of organic reactions with high levels of chemoselectivity, regioselectivity, and stereoselectivity. These reactions include difficult and disfavored transformations. We focus our attention on the following questions: Can antibodies catalyze nonnatural reactions? Can antibodies catalyze reactions otherwise disfavored in an aqueous environment? Can antibodies be used as cost-effective catalysts in the total synthesis of natural products?

ORGANOMETALLIC CHEMISTRY

The larger and variable coordination numbers and geometries of transition metals, compared with those of the carbon atom, allow the creative design of haptens closely related to the postulated structure of the transition state of a given organometallic transformation. Furthermore, many organometallic transformations are of special interest because they have no enzymatic counterparts.

Metalloenzymes include metal atoms with chemical properties finely tuned by the surrounding protein environment. According to the "entatic state" hypothesis, the enzyme imposes steric and electronic constraints on the metal ion within the active site. These constraints enhance the enzyme's catalytic activity. Modeling this behavior to prove or disprove the hypothesis has been difficult.

We used antibodies to alter the geometry and spectroscopic properties of an organometallic complex. Specifically, antibodies designed to bind an organosilicon compound (1 in Fig. 1) also bind a geometrically similar but larger Cu(I) complex (2 in Fig. 1). Upon binding, the absorption spectrum of the Cu(I) complex is red shifted. This shift is consistent with the idea that the coordination sphere of the copper complex is slightly compressed to fit into the antibody's binding site. In addition, cyclic voltammetry measurements of the complex in the presence of the antibodies suggest that the initial oxidation product is a high-energy tetrahedral Cu(II) complex. These findings strongly support the entatic state hypothesis and are an important step toward antibody-catalyzed organometallic transformations.

Our next step includes the design of platinum complex haptens (3 and 4 in Fig. 1). Antibodies elicited against these haptens are expected to catalyze some of the most fundamental processes of organometallic chemistry, such as ligand exchange, oxidative addition, and migratory insertion reactions.

REACTIONS OTHERWISE DISFAVORED IN WATER

Many methods in organic synthesis require the manipulation of water-sensitive intermediates. We have already shown that such chemistry might be accessible by using catalytic antibodies in an aqueous media. Antibody 14D9, which has been elicited against a quaternary ammonium hapten, catalyzes the cyclization of a hydroxyethyl enol ether to the corresponding ketal with absolute enantioselectivity. This reaction cannot be carried out in an aqueous medium because the generated oxocarbonium ion rapidly reacts with a water molecule to give the corresponding ketone. Other transformations that involve water-sensitive intermediates should now be considered for antibody catalysis.

TOTAL SYNTHESIS OF A NATURAL PRODUCT

Application of catalytic antibodies in organic synthesis is one of the most important and vital issues in total synthesis of a natural product. We recently completed the total synthesis of (-)--multistriatin via catalytic antibodies. The key step in our synthetic strategy was the antibody-catalyzed enantioselective protonolysis of an enol ether (Fig. 2) . Overall, (-)--multistriatin was prepared in an 11-step synthesis with all four asymmetric centers originating from the chirality achieved in the antibody-catalyzed step.

CATALYTIC ANTIBODIES IN A CONTINUOUS FLOW REACTOR

One obvious need in antibody catalysis, particularly for practical applications in organic synthesis, is to increase the antibodies' cost-effectiveness. We reported the first successful noncovalent entrapment of catalytic antibodies in a sol-gel matrix. Antibodies entrapped directly within a tetramethoxysilane-derived glass retained their activity for long periods. This finding suggests that this approach is the method of choice for preparative-scale organic synthesis. We envisage that the catalytic reactor will allow convenient changes of reaction conditions, the substrate, and even the reaction type.

EMULATING ENZYMES

In the past, we showed that catalytic antibodies elicited against a metalloporphyrin hapten mimic some of the enzymatic features of cytochrome P-450 (i.e., epoxidation of styrene). We are now designing and producing catalytic antibodies that have most of the key functions of cytochrome P-450, such as high levels of catalytic activity, substrate specificity, enantioselectivity, and regioselectivity. Reaching this goal requires a rather sophisticated hapten (5 in Fig. 1) that presents the metalloporphyrin cofactor element together with an organic substrate moiety in the correct orientation to mimic the transition state of the oxygenation reaction.

The intensive research on vitamin A and its related compounds (e.g., retinal and retinoic acid) within the past few years indicates the special importance of retinoid substrates in a plethora of biological processes. For example, 11-cis-retinal covalently binds to an opsin receptor. Its isomerization to all-trans within the protein triggers a cascade of biochemical events resulting in vision. Another example is the important role retinoic acids play in transcription.

Two haptens (6 and 8 in Fig. 1) were designed on the basis of the reported agonist activity of a series of retinoid analogs in biological tests. Another hapten (7 in Fig. 1) was designed to elicit antibodies that mimic the binding site of opsin proteins. We envisage that the antibodies elicited via homologous and heterologous immunization against these haptens will mimic both the binding and chemical features of the different retinoid receptors.

PUBLICATIONS

Keinan, E., Lerner, R.A. The first decade of antibody catalysis: Perspective and prospects. Isr. J. Chem. 36:113, 1996.

Shabat, D., Grynszpan, F., Saphier, S., Turniansky, A., Avnir D., Keinan, E. An efficient sol-gel reactor for antibody catalyzed transformations. Chem. Mat., in press.

Shabat, D., Sinha, S.C., Reymond, J.-L., Keinan, E. Catalytic antibodies as evolutionary probes: Modeling of a primordial glycosidase. Angew. Chem. Int. Ed. Eng. 35:2628, 1996.

Sinha, S.C., Keinan, E. -Multistriatin: The first total synthesis of a natural product via antibody catalysis. Isr. J. Chem. 36:185, 1996.

Top of Page


Chemical Libraries of Annonaceous Acetogenins

S.C. Sinha, A. Sinha, A. Yazbak, T. Doundoulakis, S.C. Sinha, E. Keinan

Annonaceous acetogenins, particularly those with adjacent bis-tetrahydrofuran (bis-THF) rings, have remarkable cytotoxic, antitumor, antimalarial, immunosuppressive, pesticidal, and antifeedant activities. Many of these fatty acid derivatives have similar carbon skeletons; their striking diversity originates mainly from the relative and absolute configuration of their various stereogenic oxygen functions. More than 300 different acetogenins have already been isolated, and the naturally occurring repertoire of these compounds is probably in the order of thousands.

We have used two different approaches to generate chemical libraries of adjacent and nonadjacent bis-THF acetogenins. One approach is based on the enantioselective oxidation of a "naked carbon skeleton." It uses enantioselective olefin oxidation methods in combination with ligand-assisted chirality transfer techniques based on the chemistry of rhenium(VII) and vanadium(V) oxides. For example, we have developed an efficient method for the synthesis of poly-THF compounds based on oxidative cyclization of polyalkene bis homo allylic alcohols. Highly stereoselective tandem oxidative cyclizations have been achieved by using Re2O7 with periodic acid and by using CF3COReO3. This method has been applied for the total synthesis of several naturally occurring acetogenins (Fig. 1), including mono-THF structures, such as solamin and reticulatacin; bis-THF compounds, such as asimicin, bullatacin, trilobacin, trilobin, and rolliniastatin; and even rarer, tris-THF molecules, such as goniocin and its isomers.

With the second approach, a complete 64-member library of the adjacent bis-THF acetogenins is produced by rapidly synthesizing mixture sublibraries of four diastereomers and then separating the products chromatographically. Each set of diastereomeric building blocks is prepared by epoxidation of an enantiomerically pure diene-diol skeleton followed by acid-catalyzed cyclization. Exploratory work into more complex, newly discovered acetogenins is rapidly leading to the completion of the asymmetric synthesis of mucocin and many stereoisomers of these cytotoxic molecules.

PUBLICATIONS

Sinha, S.C., Keinan, E. Total synthesis of (+)-aspicilin: The naked carbon skeleton strategy vs. the bioorganic approach. J. Org. Chem. 62:377, 1997.

Sinha, S.C., Sinha, A., Sinha, S.C., Keinan, E. Tandem oxidative cyclization with rhenium oxide: Total synthesis of a goniocin isomer. J. Am. Chem. Soc., in press.

Keinan, E., Sinha, A., Yazbak, A., Sinha, S.C., Sinha, S.C. Towards chemical libraries of annonaceous acetogenins. Pure Appl. Chem. 1997:69, 423.

Top of Page


Assembly of the Mouse Sperm's Egg Recognition Protein

A. Cheng, I. Rooney, F. Suzuki,* J.D. Bleil

* Chiba University, Chiba, Japan

This laboratory studies mammalian sperm-egg recognition at the molecular level. The first step in fertilization is recognition between the head of the sperm and the zona pellucida, an extracellular glycoprotein coat surrounding the egg. Fertilizing mouse sperm bind to an O-linked oligosaccharide (functional domain oligosaccharide) of ZP3, one of three glycoproteins in the zona pellucida. We have discovered a mouse sperm lectin, a 43-kD polypeptide called sp56, that has high and specific affinity for ZP3 and for ZP3's functional domain oligosaccharide. This lectin is part of a large protein found on the surface of the sperm, the mouse egg recognition protein (mERP).

This recognition protein is a homooctamer of pro-sp56 (Fig. 1). The 61-kD polypeptide pro-sp56 is composed of six, contiguous "sushi" domains, or short homology repeats (S1--S6); followed by a stretch of unique polypeptide, called the sp56-specific sequence; followed by a seventh sushi domain (S7); and terminating with a basic, C-terminal tail. In mERP, pro-sp56 monomers are cross-linked by disulfide bonds within the basic, C-terminal tail. Under most conditions, detergent extraction of mERP from sperm results in autoproteolytic cleavage at specific sites in the sp56-specific sequence, releasing sp56 (sushi domains S1--S6 and part of the sp56-specific sequence) from a disulfide--cross-linked homooctamer composed of eight C-terminal fragments.

Recent studies suggest that this same autoproteolytic cleavage of mERP occurs in vivo: binding of ZP3 to mERP on the surface of the sperm causes release of sp56 dimers (still bound to ZP3) from sperm-associated homooctameric C-terminal fragments. We hypothesize that ZP3-triggered autoproteolysis of mERP induces the sperm's acrosome reaction, a membrane fusion event that enables the sperm to penetrate the zona pellucida and fuse with the egg plasma membrane.

In mice, mERP is expressed exclusively in spermatogenic cells. The homooctamer is assembled from pro-sp56 monomers in the endoplasmic reticulum of early, round spermatids. Native mERP is transported from the endoplasmic reticulum to the Golgi complex, accumulating in the "Golgi zone" of late-stage round spermatids. During spermatid elongation, mERP is transported to the cell surface, where it is confined to that region of the sperm head overlying the acrosome. The mERP is tethered to the sperm head by an anchoring protein, which may play a role in signal transduction and the acrosome reaction.

Structural studies on mERP, in collaboration with the laboratory of D. Stout, Department of Molecular Biology, have been enabled by heterologous expression of the protein. Cell lines transfected with cDNA encoding mERP express and secrete the protein as a homooctamer.

PUBLICATIONS

Rooney, I.A., Cheng, A., Bleil, J.D. Biochemical characterization of the mouse sperm's egg recognition protein (mERP), a homo-octamer of pro-sp56. Development, in press.

Top of Page


Control of Cell Division

S. Reed, B. Bertolaet, E. Bailly, D. Clarke, S. Haase, J. Hanley-Hyde, L. Hengst, M. Henze, T. Herzinger, P. Kaiser, G. Mondesert, B. Niculescu, K. Sato, M. Segal, M. Smeets, C. Spruck, C. Tang, M. Watson, M. Wolfe, K. Won

Biological processes of great complexity can be approached by beginning with a systematic genetic analysis in which the relevant components are first identified and the consequences of their selective elimination by mutation are investigated. We have used yeast, which is uniquely tractable to this type of analysis, to investigate control of cell division. In recent years, it has become apparent that the most central cellular processes throughout the eukaryotic phylogeny are highly conserved in terms of both the regulatory mechanisms used and the proteins involved. Thus, it has been possible in many instances to generalize from yeast cells to human cells.

CONTROL IN YEAST

Most of our work in recent years has focused on the role and regulation of the Cdc28 protein kinase. Initially identified by means of a mutational analysis of the yeast cell cycle, this protein kinase is ubiquitous in eukaryotic cells and central to a number of aspects of control of cell-cycle progression. Considerable progress has been made in understanding the regulation of the kinase in terms of its function at two points in the cell cycle: the transition from G1 to S and the transition from G2 to M. It appears that the same protein kinase catalytic subunit may perform these different functions by associating with other specific regulatory proteins known as cyclins to form active protein kinase complexes. Much of our recent effort has been devoted to identifying critical targets of these Cdc28 kinase activities. We showed that many of the cell-cycle effects of Cdc28 kinase activity are mediated by regulating proteolysis of specific proteins.

We also found that a small Cdc28-associated protein, known as Cks1, appears to regulate the proteasome. Proteasomes are complex proteases that target ubiquitinated proteins, including important cell-cycle regulatory proteins. Our genetic analysis of the proteasome and ubiquitination pathways has yielded a complex web of interactions.

CONTROL IN MAMMALIAN CELLS

We have shown that the human homologs of the Cdc28 protein kinase are so highly conserved, structurally and functionally, relative to the yeast protein kinase, that they can function and be regulated properly in a yeast cell. Analyzing control of the cell cycle in mammalian cells, we produced evidence for the existence of regulatory schemes, similar to those elucidated in yeast, that use networks of both positive and negative regulators. We found that in human cells, agents that have positive or negative effects on proliferation exert these effects at the level of regulating cyclins and cyclin-dependent kinases. Of particular interest is the induction of cyclin-dependent kinase inhibitory proteins by antiproliferative signals. These proteins include not only members of the recently discovered Cip/Kip family of inhibitors but also p130, a relative of the retinoblastoma protein, implicated in human cancer.

As is the case in yeast, we have shown that proteolysis plays a key regulatory role in control of the cell cycle in mammalian cells. We are determining if defects in proteolyis are implicated in human cancer. Finally, we are producing mice with defects in cell cycle--regulated proteolysis to observe the effect of such defects on development.

PUBLICATIONS

Covini, G., Chan, E.K., Nishioka, M., Morshed, S.A., Reed, S.I., Tan, E.M. Immune response to cyclin b1 in hepatocellular carcinoma. Hepatology 25:75, 1997.

Dulic, V., Stein, G., Far, D.F., Reed, S.I. Nuclear accumulation of p21Cip1/Waf1 at the onset of mitosis: A role in G2/M transition? Mol. Cell. Biol., in press.

Kaiser, P., Moncollin, V., Watson, M.H., Bertolaet, B.L., Clarke, D.J., Reed S.I., Bailly, E. Cyclin-dependent kinase and Cks/Suc1 interact with the proteasome in yeast. Science, in press.

Li, H., Lahti, J.H., Valentine, M., Saito, M., Reed, S.I., Look, A.T., Kidd, VI. Molecular cloning and chromosomal localization of the human cyclin C (CCNC) and cyclin E (CCNE) genes: Deletion of the CCNC gene in human tumors. Genomics 32:253, 1996.

Lukas, J., Herzinger, T., Hansen, K., Moroni, M.C., Resnitzky, D., Helin, K., Reed, S.I., Bartek J. Cyclin E-induced S phase without activation of the pRb/E2F pathway. Genes Dev. 11:1479, 1997.

Mondesert, G., Clarke, D.J., Reed, S.I. Indentification of genes controlling growth polarity in the budding yeast Saccharomyces cerevisiae: A possible role of N-glycosylation and involvement of the exocyst complex. Genetics 147:421, 1997.

Niculescu, A.B., Chen, X., Smeets, M., Hengst, L., Prives, C., Reed, S.I. p21Cip1/Waf1 regulates both the G1/S and the G2/M cell cycle phase transitions: pRb as a critical determinant of p21 function. Mol. Cell. Biol., in press.

Reed, S.I. Control of the G1/S transition. Cancer Surv. 29:7, 1997.

Reed, S.I. Cyclin E: In mid-cycle. BBA, Reviews on Cancer [on-line serial]. 1287:151, 1996. Available by e-mail at sreed@scripps.edu or online at http://www.elsevier.nl/inca/publications/store

Reed, S.I. G1/S regulatory mechanisms from yeast to man. Prog. Cell Cycle Res. 2:15, 1996.

Stratton, H., Zhou, J., Reed, S.I., Stone, D.E. The mating-specific Ga protein of Saccharomyces cerevisiae down regulates the mating signal by a mechanism that is dependent on pheromone and independent of Gbg sequestration. Mol. Cell. Biol. 16:6325, 1996.

Tiefenbrun, N., Melamed, D., Levy, N., Resnitzky, D., Hoffmann, I., Reed, S.I., Kimchi A. Interferon- suppresses cyclin D3 and cdc25A genes leading to a reversible G0-like arrest. Mol. Cell. Biol. 16:3934, 1996.

Wittenberg, C., Reed, S.I. Plugging it in: Signaling circuits and the yeast cell cycle. Curr. Opin. Cell Biol. 8:223, 1996.

Top of Page


Control of Initiation of the Cell Cycle in Yeast

C. Wittenberg, D. Chapman-Shimshoni, K. Flick, M. Guaderrama, Y. Hsiung, T. Kesti, S. Lanker, D. Stuart

In budding yeasts, as in most eukaryotic cells, regulation of proliferation is exerted primarily during the G1 phase of the cell cycle. Progression through the cell cycle is governed by the cyclin-dependent protein kinases (CDKs). CDKs are heterodimeric enzymes that consist of a catalytic subunit and a positive regulatory subunit called a cyclin. They drive the major transitions in the cell cycle and are the targets of the negative regulatory signals that mediate checkpoints in the cell cycle.

Budding yeasts have a single CDK catalytic subunit, encoded by the gene CDC28, that collaborates with no fewer than nine different cyclins to regulate events in the cell cycle. The cyclins can be categorized into two major classes, G1 and B, and these classes can be divided into subclasses on the basis of both primary sequence and pattern of expression. Each class appears to be optimized for a specific function. This specificity is derived from the differences between the cyclins themselves and from the differences between the CDK complexes in which the cyclins participate. First, the patterns of cyclin accumulation are differentially regulated as a consequence of the combined effects of cell cycle--dependent transcription and regulated proteolytic degradation. Next, specific cyclin-CDK complexes differ in their substrate specificity. Finally, complexes composed of CDKs and G1 or B cyclins can be inhibited by distinct CDK inhibitors.

We have focused our attention on regulation of initiation of the cell cycle during G1 in the budding yeast Saccharomyces cerevisiae. Initiation of the cycle involves a complex choreography of events driven by the sequential activation CDKs that is just beginning to be fully elucidated. The pattern of CDK activity is a consequence of the interplay of cell cycle--dependent transcription and regulated proteolysis.

Cln3, which accumulates during the early part of G1, activates transcription of the genes CLN1, CLN2, CLB5, and CLB6. The Cln1 and Cln2 proteins then accumulate, activating the Cdc28 CDK for several critical functions. First, these cyclin-CDK complexes promote formation of a bud, the presumptive daughter cell, and duplication of spindle poles, which occurs during the late part of G1 in fungal cells. Next, they inactivate the machinery for B type cyclin proteolysis. Finally, they drive degradation of several key proteins via phosphorylation-dependent ubiquitination, including Sic1, a specific inhibitor of B cyclin--CDK complexes and the G1 cyclins themselves. These events lead to activation of the Clb5 and Clb6 forms of the CDK, which promote entry into S phase, and set the stage for the accumulation of subsequent waves of CDK activity that drive cells into mitosis. Efforts to understand the mechanisms that govern G1-specific transcriptional activation and phosphorylation-dependent ubiquitination form the core of our current research.

In addition to their importance in promoting progression of the cell cycle, cyclin-CDK complexes are targets for checkpoint regulation. Both internal and environmental signals are integrated with the cell-cycle machinery during G1. Cells monitor the availability and quality of nutrients and modulate the rate of progression of the cell cycle and cell size accordingly. If diploid cells are limited for nutrients, they can undergo meiosis and sporulation, an interesting and important modification of the mitotic cell cycle. Finally, cells respond to peptide mating pheromones secreted by cells of the opposite mating type by arresting during G1 in preparation for conjugation. All these processes provide unique opportunities to biochemically and genetically dissect mechanisms of regulation of the cell cycle that we are exploiting in our ongoing research.

PUBLICATIONS

Lanker, S., Valdivieso, M.H., Wittenberg, C. Rapid degradation of the G1 cyclin Cln2 is induced by CDK-dependent phosphorylation. Science 271:1587, 1996.

Willems, A.R., Lanker, S., Patton, E.E., Craig, K.L., Nason, T.F., Mathias, N., Kobayashi, R., Wittenberg, C., Tyers, M. Cdc53 targets phosphorylated G1 cyclins for degradation by the ubiquitin proteolytic pathway. Cell 86:453, 1996.

Wittenberg, C., Reed, S. Links between signal transduction cascades and the cell cycle. Curr. Opin. Cell Biol. 8:223, 1996.

Top of Page


Regulation of the Cell Cycle

P. Russell, R. Aligue, N. Boddy, G. Degols, B.A. Furnari, F. Gaits, J. Kanoh, J. Leatherwood, O. Mondesert, N. Rhind, M. Shiozaki, K. Shiozaki, L. Wu

The minimal cell cycle consists of two phases: S phase, the period of DNA replication, and then M phase, the period of nuclear division. DNA replication and nuclear division occur in a fraction of the time that it takes for a mass doubling of somatic cells. Thus, the S and M phases are separated by gap intervals in the order G1 S G2 M. Our efforts concentrate on the control of mitosis.

We use the fission yeast Schizosaccharomyces pombe, a rod-shaped eukaryote that grows by length extension and divides by medial fission (Fig. 1, upper panel). Two classes of mutants have been most useful. Temperature-sensitive cdc mutants are unable to execute essential events in the cell cycle. They are not defective in cellular growth; therefore they become highly elongated when incubated at the restrictive temperature (Fig. 1, middle panel). Some cdc mutants, such as cdc25-22, are defective in initiating mitosis. A second class of mutants initiates mitosis prematurely. These are called wee mutants because they are smaller than wild-type cells (Fig. 1, lower panel). Recessive wee mutations, such as wee1-50, identify genes that inhibit the onset of mitosis.

Figure 2 shows key elements of the mitotic control network. A Cdc2--cyclin B kinase induces mitosis. The proteins Wee1 and Mik1 inhibit Cdc2--cyclin B by phosphorylating tyrosine-15 in the ATP-binding region of Cdc2. Cdc25 activates Cdc2--cyclin B by dephosphorylating tyrosine-15. The balance of Cdc25 and Wee1 activities determines the timing of mitosis. Another important mitotic regulator is Nim1 protein kinase. Nim1 promotes mitosis by inactivating Wee1.

A common cellular response to DNA damage is arrest of the cell cycle. This checkpoint control, which actively prevents mitosis while DNA damage is being repaired, has been the subject of intensive genetic investigation, but the biochemical mechanism that prevents mitosis after DNA damage was unknown. In fission yeast, the inhibitory tyrosine-15 phosphorylation of Cdc2 is used to couple mitosis with completion of DNA replication, but the role of Cdc2 tyrosine phosphorylation in the DNA-damage checkpoint was uncertain.

Our recent studies established that the DNA-damage checkpoint that governs arrest of the cell cycle in G2 in S. pombe depends on the inhibitory tyrosine phosphorylation of Cdc2 carried out by Wee1 and Mik1. Furthermore, we found that the rate of Cdc2 tyrosine dephosphorylation is reduced by irradiation. This result implicates regulation of Cdc2 tyrosine dephosphorylation, mainly carried out by the Cdc25 tyrosine phosphatase, as an important part of the mechanism by which the DNA-damage checkpoint induces Cdc2 inhibition and arrest of the cell cycle in G2.

Spc1 is a member of the stress-activated protein kinase family, an evolutionary conserved subfamily of MAP kinases (MAPKs). Spc1 is activated by a MAPK kinase homolog, Wis1, and negatively regulated by Pyp1 and Pyp2 tyrosine phosphatases. Mutations in the spc1+ and wis1+ genes cause a delay in G2 in the cell cycle that is exacerbated during stress. We recently found that Wik1 and Mcs4 are two upstream regulators of the Wis1-Spc1 cascade. Wik1 encodes a MAPK kinase. Cells that lack wik1 are impaired in stress-induced activation of Spc1 and show a delay in G2 and osmosensitive growth. Moreover, overproduction of a constitutively active form of Wik1 induces hyperactivation of Spc1 in a wis1+-dependent manner, suggesting that Wik1 regulates Spc1 through activation of Wis1.

Mutations of mcs4+ (mitotic catastrophe suppressor) were originally isolated as suppressors of the mitotic catastrophe phenotype of a cdc2-3w wee1-50 double mutant. We found that mcs4- cells are defective in activation of Spc1 in response to various forms of stress. Epistasis analysis placed Mcs4 upstream of Wik1 in the Spc1 activation cascade. Mcs4 is a response regulator protein. These results indicate that Mcs4 is part of a sensor system for multiple environmental signals that modulates the timing of entry into mitosis by regulating the Wik1-Wis1-Spc1 kinase cascade.

PUBLICATIONS

Aligue, R., Wu, L., Russell, P. Regulation of Schizosaccharomyces pombe Wee1 tyrosine kinase. J. Biol. Chem. 272:13320, 1997.

Degols, G., Russell, P. Discrete roles of Spc1 kinase and Atf1 transcription factor in the UV response of Schizosaccharomyces pombe. Mol. Cell. Biol. 17:3356, 1997.

Furnari, B., Rhind, N., Russell, P. Cdc25 mitotic inducer targeted by Chk1 DNA damage checkpoint kinase. Science, in press.

Furnari, B., Russell, P., Leatherwood, J. pch1+, a second essential C-type cyclin in Schizosaccharomyces pombe. J. Biol. Chem. 272:12100, 1997.

Gaits, F., Shiozaki, K., Russell, P. Protein phosphatase 2C acts independently of stress-activated kinase cascade to regulate the stress response in fission yeast. J. Biol. Chem. 272:17873, 1997.

Rhind, N., Furnari, B., Russell, P. Cdc2 tyrosine phosphorylation is required for the DNA damage checkpoint in fission yeast. Genes Dev. 11:504, 1997.

Shiozaki, K., Russell, P. Stress-activated protein kinase pathway in cell cycle control of fission yeast. Methods Enzymol. 283:506, 1997.

Shiozaki, K., Shiozaki, M., Russell, P. Mcs4 mitotic catastrophe suppressor regulates the fission yeast cell cycle through the Wik1-Wis1-Spc1 kinase cascade. Mol. Biol. Cell 8:409, 1997.

Wu, L., Russell, P. Nif1, a novel mitotic inhibitor in Schizosaccharomyces pombe. EMBO J. 16:1342, 1997.

Wu, L., Russell, P. Roles of Wee1 and Nim1 protein kinases in regulating the switch from mitotic division to sexual development in Schizosaccharomyces pombe. Mol. Cell. Biol. 17:10, 1997.

Wu, L., Shiozaki, K., Aligue, R., Russell, P. Spatial organization of the Nim1-Wee1-Cdc2 mitotic control network in Schizosaccharomyces pombe. Mol. Biol. Cell 7:1749, 1996.

Top of Page


Control of the Cell Cycle in Human Cells

C.H. McGowan, A. Blasina, E.S. Paegle, H.M. Strohmaier

Discovering the molecular mechanisms that regulate progression of the cell cycle is a major goal in cell biology and is essential for developing an understanding of how cell proliferation is governed in human disease. In mammals, the cell cycle is driven by the sequential activation of a number of cyclin-dependent kinases. Each of these kinases phosphorylates and alters proteins required for specific events in the cell cycle. The activity of cyclin-dependent kinases is regulated by changes in the abundance of the cyclin subunit, by inhibitory phosphorylation, and by association with inhibitory proteins. These regulatory mechanisms combine to delay progression of the cell cycle when cells are exposed to adverse conditions. For example, at least two delays in the cycle are involved in coordinating progression of the cycle after treatment with therapeutic agents that damage DNA.

The delays that allow coordination of different functions of the cell cycle are called checkpoints. The checkpoint that controls the transition from G1 to S depends on the induction of a cyclin-dependent kinase inhibitor and prevents cells from replicating damaged DNA. A second checkpoint is involved in the transition from G2 to M and prevents segregation of damaged DNA. In human cells, the mitosis-inducing kinase Cdc2/cyclin B is inhibited by phosphorylation of threonine-14 and tyrosine-15. Although many of the gene products that control the phosphorylation state of Cdc2 have been determined, the mechanism by which they respond to checkpoint controls is not well characterized.

Work in our laboratory focuses on the regulation of the G2-to-M transition in human cells. We have shown that phosphorylation of Cdc2 is essential for the operation of the checkpoint that prevents activation of Cdc2 in the presence of unreplicated or damaged DNA. We are using biochemical analysis of known proteins to define the pathway through which the checkpoint affects Cdc2 phosphorylation. Genetic screens in yeasts and human sequence databases are being used to detect novel molecules involved in the checkpoint control of the transition from G1 to M. The information gained from these studies may provide a rational basis for the improvement of radiation and chemotherapy.

Top of Page


Feline Immunodeficiency Virus

J.H. Elder, A.P. de Parseval, G.S. Laco, U. Chatterji, D.L. Lerner, T.R. Phillips,* A.J. Olson, C.-H. Wong,** A. Wlodawer,*** C.D. Stout, G.S. Prasad, E.D. Stura

* Department of Neuropharmacology, TSRI
** Department of Chemistry, TSRI
*** Frederick Cancer Research and Development Center, Frederick, MD

Feline immunodeficiency virus (FIV) is a lentivirus that causes an immunodeficiency-like syndrome in domestic cats. It is related to HIV, the causative agent of AIDS in humans. Therefore, FIV infection in cats is a natural animal model for study of the lentivirus life cycle and for the development of intervention strategies that will block viral infection and spread. Our laboratory studies the molecular biology of FIV, with particular emphasis on defining regulatory mechanisms that dictate the level and range of infection. These studies include defining the functions of both virus-encoded proteins necessary for replication and host cell proteins that influence viral expression and spread. Our ultimate goal is to develop rational approaches to the development of antiviral agents.

An area that continues to be a major emphasis is elucidation of the structure and function of FIV protease. This enzyme is responsible for accurately processing itself, and other enzymes and structural proteins, from long polyproteins that are translated from virus-encoded mRNAs. The action of the protease is necessary for the production of infectious virus particles. Thus, blocking its function effectively blocks spread of virus. Several protease inhibitors are now available that reduce viral load in HIV-positive patients. However, a major problem is the development of resistance, which inevitably limits the usefulness of the drugs. Toxic effects are also a problem, particularly for certain patients. Our goal is to better understand the nature of substrate and inhibitor specificities so that we can develop drugs against the protease that will be less susceptible to the development of resistance and also be less toxic.

We have molecularly cloned, purified, and defined the three-dimensional structure of both native and mutant FIV proteases. Using the information obtained from these studies, we have determined discrete amino acids most likely to be involved in direct interaction with the polyprotein substrates and thus most likely to be residues critical for targeting by inhibitors. We have also begun to test specific inhibitors developed by our colleagues, to establish the relative effectiveness of these compounds as a function of protease structure. By preparing "HIVinized" versions of FIV protease wherein specific amino acid residues of HIV protease have been substituted in the equivalent positions for FIV protease, we can directly analyze the influence of specific sites on inhibitor and substrate interactions.

During the past year, C.-H. Wong's laboratory, Department of Chemistry, synthesized an inhibitor that is effective against both FIV and HIV proteases, although it is more efficacious against the human enzyme. Certain FIV-HIV chimeric proteases are intermediate in sensitivity, supporting the contention that we have detected target amino acid residues that interact with the inhibitor. The inhibitor prevents acute infections in tissue culture, and treatment of chronically infected cells with the inhibitor results in the production of noninfectious FIV. We are currently monitoring cultures for development of resistance. We will molecularly analyze any virus that can replicate in the presence of the inhibitor. The results thus far are promising; after 1 month of treatment with the inhibitor, no FIV escape variants have been detected. We think that these studies ultimately will yield compounds useful in the treatment of viral infections in both humans and cats.

We are also defining the mechanisms by which FIV enters and exits the cell. Understanding receptor interactions may allow the development of ways to block viral infection. Likewise, understanding how virus traffics to the cell membrane and is released from the host cell may offer potential targets for blocking viral spread. We recently carried out a study with a monoclonal antibody developed by B. Willett and colleagues at the University of Glasgow. This antibody recognizes the cellular protein CD9, and interaction of the antibody with infected cells results in marked reduction in viremia in tissue culture.

We had originally supposed that CD9 might be the receptor for FIV and that the antibody might block viremia by interfering with entry of the virus into the host cell. However, careful examination revealed that the virus could bind to and infect the target cell in the presence of antibody. Furthermore, the viral RNA was transcribed into DNA, integrated into the host DNA, and transcribed and translated into viral proteins in a manner similar to that of early events that occur in the absence of antibody. However, the antibody blocked egress of virus from the cell, a condition that caused buildup of unreleased viral proteins in the target cell. This phenomenon appears to be specific for FIV, implying that all budding viruses do not leave the host cell by the same mechanism. Thus, in addition to offering a possible antiviral strategy, these studies open an interesting avenue to study the poorly defined mechanisms by which viruses exit cells.

Other ongoing studies include the molecular characterization of deoxyuridine triphosphatase and the Orf2 gene product. We are also continuing collaborative studies to define the mechanisms by which FIV induces lesions in the CNS (see the reports of Drs. Phillips, Henriksen, and Fox). We hope that all these studies will lead to a better understanding of the life cycle of FIV, and that this information will lead to the development of effective interventions.

PUBLICATIONS

de Parseval, A., Lerner, D.L., Borrow, P., Willett, B.J., Elder, J.H. Blocking of FIV infection by a monoclonal antibody to CD9 is via inhibition of virus release, rather than interference with receptor binding. J. Virol., in press.

Jacobsen, S., Henriksen, S.J., Prospero-Garcia, O., Phillips, T.R., Elder, J.H., Bloom, F.E., Fox, H.S. Cortical neuronal cytoskeletal changes associated with FIV infection, J. Neurovirol., in press.

Laco, G.S., Fitzgerald, M.C., Morris, G.M., Olson, A.J., Kent, S.B.H., Elder J.H. Molecular analysis of the feline immunodeficiency virus protease: Generation of a novel form of the protease by autoproteolysis and construction of cleavage resistant proteases, J. Virol., in press.

Phillips, T.R., Prospero-Garcia, O., Wheeler, D.W., Wagaman, P.C., Lerner, D.L., Fox, H.S., Whalen, L.R., Bloom, F.E., Elder, J.H., Henriksen, S.J. Neurologic dysfunctions caused by a molecular clone of feline immunodeficiency virus, FIV-PPR. J. Neurovirol. 2:388, 1996.

Prasad, G.S., Stura, E.A., McRee, D.E., Laco, G.S., Hasselkus-Light, C., Elder, J.H., Stout, C.D. Crystal structure of dUTP pyrophosphatase from feline immunodeficiency virus. Protein Sci. 5:2429, 1996.

Top of Page


Immunobiology of Hepatitis B and C Virus Infections

D.R. Milich, F. Schödel,* D.L. Peterson,** J.L. Hughes, J.E. Jones, M. Sällberg,*** M. Chen,*** A. Birkett****

* INSERM U80, Lyon, France
** Virginia Commonwealth University, Richmond, VA
*** Huddinge University Hospital, Huddinge, Sweden
**** Immune Complex Corp., San Diego, CA

HEPATITIS B

Our recent focus has been the question of why some patients infected with hepatitis B virus (HBV) clear the infection relatively efficiently, whereas others cannot clear it and become chronically infected. In a series of studies with human sera and in mouse experimental systems (i.e., transgenic mice), we have developed evidence suggesting that the characteristics of the patient's T helper (Th)--cell response may determine the outcome of the infection. Two major types of Th cells have been described: Th1 cells mediate cellular immunity (i.e., graft rejection, tissue injury, inflammation), and Th2 cells mediate humoral immunity (i.e., antibody production). Theoretically, a correct balance between HBV-specific Th1 and Th2 cells would lead to viral clearance and a minimum of liver disease.

Recent serologic studies indicate an imbalance in Th1 and Th2 cells in patients with chronic HBV infection. For example, acute, self-limited HBV infection is characterized by efficient cellular immunity, T cell--mediated liver injury, and low-level production of antibodies, a Th1-like response pattern. In contrast, chronic, symptomatic HBV infection is characterized by reduced cellular immunity and significantly greater antibody responses, a Th2-like response pattern.

We examined the possibility that the two structural forms of the viral nucleoprotein, the particulate HBV core (HBcAg) and the nonparticulate HBeAg, may preferentially elicit different subsets of Th cells. The immunogenicity of HBcAg, in contrast to that of HBeAg, did not require the use of adjuvants. Furthermore, HBcAg elicited primarily IgG2a and IgG2b antibodies and no IgG1 antibodies. In contrast, the antibody response to HBeAg was dominated by the IgG1 isotype. HBcAg-primed Th cells efficiently produced IL-2 and IFN- and low levels of IL-4. Conversely, efficient production of IL-4 and lesser amounts of IFN- were elicited by immunization with HBeAg. The results indicate that HBcAg preferentially, but not exclusively, elicits Th1-like cells and that HBeAg preferentially, but not exclusively, elicits Th0 or Th2-like cells. These findings show that Th cells with the same specificity can develop into different Th subsets depending on the structural form of the immunogen.

In addition to inducing Th2 cells, HBeAg may modulate the immune response during HBV infection. To study the effects of continuous exposure of the immune system to circulating HBeAg, we bred transgenic mice that expressed HBeAg with mutant mice defective in Fas or in the ligand for Fas. Interactions between Fas and its ligand mediate activation-induced apoptosis and are involved in eliminating autoreactive Th cells and normal Th cells responding to foreign antigens. The production of autoantibodies to HBeAg in transgenic mice expressing HBeAg and in the cross-bred mice was compared.

Transgenic mice expressing HBeAg had high titers of autoantibody to HBeAg that was exclusively IgG1 (i.e., Th2-like profile). The cross-bred mice produced significantly lower titers of autoantibodies, but the antibodies included IgG2a, IgG2b, and IgG3 isotypes as well as IgG1 (i.e., mixed Th1/Th2-like profile). These results indicate an important role for peripheral depletion of Th cells in maintaining tolerance in transgenic mice that express HBeAg and suggest that HBeAg-specific Th1 cells are preferentially depleted by apoptosis mediated by interactions between Fas and its ligand.

Because serum HBeAg preferentially depleted Th1 cells, we examined the effect of secreted HBeAg on HBcAg-specific Th1 cells by transferring HBeAg- and HBcAg-specific Th cells into transgenic mice that expressed both HBeAg and HBcAg. The presence of HBeAg in the serum eliminated the expected Th1 cell--mediated antibody response to HBcAg and shifted the response toward a Th2 phenotype. These results suggest that in the context of an HBV infection, circulating HBeAg can preferentially deplete inflammatory HBeAg- and HBcAg-specific Th1 cells necessary for viral clearance and thereby promote persistence of HBV.

HEPATITIS C

Most patients infected with hepatitis C virus (HCV) cannot clear the virus and become chronically infected. To better understand the immune mechanisms that may influence HCV clearance, we are doing serologic studies of chronically infected patients and immunogenicity studies in a mouse model.

We examined the humoral immune response to HCV in patients with chronic HCV infection by using an enzyme-linked immunoassay and recombinant HCV antigens. The IgG antibody responses to the HCV structural and nonstructural proteins in these patients were relatively low and were defective in terms of IgG subclass switching. These data suggest a suboptimal activation of Th cells in chronic HCV infection rather than a skewing of the Th cell subsets as observed in chronic HBV infection.

To determine if HCV proteins are intrinsically poor immunogens, we immunized mice with recombinant and peptide versions of HCV proteins and analyzed the T- and B-cell responses. Among other findings, it was clear that HCV proteins were not intrinsically poor immunogens when injected into mice in sufficient dose and with adjuvant. Therefore, it will be necessary to examine other characteristics of the HCV-specific immune response to explain the defective humoral responses observed during chronic HCV infection.

PUBLICATIONS

Hultgren, C., Milich, D.R., Sällberg, M. ntibodies to hepatitis B e (HBeAg) can be induced in HBeAg transgenic mice by adoptive transfer of a specific T-helper 2 cell clone. Clin. Diagn. Lab. Immunol., in press.

Milich, D.R. The immune response to the hepatitis B virus: Infection, vaccination, animal models. Viral Hepatitis Rev., in press.

Milich, D.R. Influence of T helper cell subsets and crossregulation in HBV infection. J. Viral Hepatitis, in press.

Milich, D.R., Chen, M., Hughes, J.L., Jones, J.E. The secreted hepatitis B e antigen can modulate the immune response to the nucleocapsid: A mechanism for persistence. J. Immunol., in press.

Milich, D.R., Schödel, F., Hughes, J.L., Jones, J.E., Peteson, D.L. The hepatitis B virus core and e antigens elicit different Th cell subsets: Antigen structure can affect Th cell phenotype. J. Virol. 71:2192, 1997.

Sällberg, M., Townsend, K., Chen, M., O'Dea, J., Banks, T., Jolly, D.J., Chang, S.M., Lee, W.T., Milich, D.R. Genetic immunization using amphotropic retroviral vectors with different forms of the hepatitis B virus core and e antigens: Characterization of humoral and CD4+ cellular responses. J. Virol. 71:5295, 1997.

Sällberg, M., Zhang, Z.-X., Chen, M., Jin, L., Birkett, A., Peterson, D.L., Milich, D.R. Immunogenicity and antigenicity of the ATPase/helicase domain of the hepatitis C virus nonstructural 3 protein. J. Gen. Virol. 77:2721, 1996.

Townsend, K., Sällberg, M., O'Dea, J., Banks, T., Hughes, J., Milich, D.R., Jolly, D.J., Chang, S.M., Lee, W.T. Characterization of CD8+ cytotoxic T lymphocyte responses after genetic immunization with retroviral vectors expressing different forms of the hepatitis B virus core and e antigens. J. Virol. 71:3365, 1997.

Zhang, Z.-X., Milich, D.R., Peterson, D.L., Birkett, A., Schwarcz, R., Weiland, O., Sällberg, M. Interferon- treatment induces delayed CD4+ proliferative responses to the hepatitis C virus nonstructural 3 protein regardless of the outcome of therapy. J. Infect. Dis., in press.

Top of Page


Molecular Parasitology

A.S. Kang, Y. Su, J.A. Chapple, T. Stratmann,* R. Xu,** M. He***

* University of Marburgh, Marburgh, Germany
** Morgan Tan Institute, Fudan, People's Republic of China
*** The Babraham Institute, Cambridge, England

This laboratory studies means of intervention in the life cycle of parasites such as Plasmodium falciparum in humans. The life cycle of Plasmodium in the human host is complex and offers a number of potential targets for attack by the immune system. Exposure to irradiated sporozoites confers immunity to subsequent challenge with live parasites. In the past, efforts to duplicate this immunity have focused on recombinant vaccines and peptides.

When active immunity cannot be generated, one approach is to use passive immunization. Studies have shown that transfer of immune -globulin is effective against malaria. We used peripheral blood leukocytes derived from a person immune to malaria to construct combinatorial libraries of antibodies to P. falciparum. We selected an antibody to the repeat epitope of the circumsporozoite surface protein for further study. The variable regions of this antibody have been transferred into vectors to produce whole human IgG1 and IgG4. Current efforts are directed toward isolating antibodies to regions flanking the repeat epitope on the circumsporozoite surface protein and antibodies against the parasite directly.

In a related project, we are designing molecules to block the transmission of malaria at the mosquito stage. With the support of the McKnight Foundation, we are exploring interventions in arthropod-transmitted viral diseases in plants. Symbiotic bacteria that reside in an insect are engineered to express antibody fragments to impair the ability of the insect to transmit a pathogen.

Engineered antibodies have usefulness beyond diagnosis and treatment of disease. We have used the technology to develop novel cell-surface receptors that permit specific pathogens to enter usually nonhost cells. To enter cells, viruses use a number of different cell-surface molecules. Foot-and-mouth disease virus uses a cell-surface integrin via an arginine--glycine--aspartic acid motif in the capsid protein. A mutant of the virus in which this motif has been deleted is noninfectious.

To propagate the mutant virus, we created a new receptor by fusing a virus-binding single-chain antibody to intracellular adhesion molecule-1. Cells not normally susceptible to infection with foot-and-mouth disease virus became infected after being transfected with the plasmid encoding the fusion protein. The mutant virions that are noninfectious in animals and other cell types grew to high titre and were able to form plaques on transfected cells containing the fusion protein.

These studies were the first production of a totally synthetic cell-surface receptor for a virus. This novel approach will be useful for the development of safer vaccines against viral pathogens of animals and humans. This concept should be useful in research on a number of viral and parasitic pathogens that are difficult to study because no appropriate in vitro cultivation systems exist. Using engineered cells for uptake and propagation of pathogens may enable scientists to study the early events in host-pathogen interaction in detail to determine targets for intervention.

In collaboration with M. He, we have continued our studies on a recombinant antibody to progesterone. The bacterially produced antibody fragment has binding affinity and specificity identical to those of the parent molecule. Antibodies with a mutation in complementarity-determining region 3 of the heavy chain have been analyzed for improved steroid recognition and discrimination.

PUBLICATIONS

Mason, P.W., Berinstein, A., Baxt, B., Parsells, R., Kang, A., Rieder, E. Cloning and expression of a single-chain antibody fragment specific for foot-and-mouth disease virus. Virology 224:548, 1996.

Rieder, E., Berinstein, A., Baxt, B., Kang, A., Mason, P. W. Propagation of an infectious virus by design: Engineering a novel receptor for a noninfectious foot-and-mouth disease virus. Proc. Natl. Acad. Sci. U.S.A. 93:10428, 1996.

Top of Page


Molecular Neurobiology

J.G. Sutcliffe, L. de Lecea, D.D. Gerendasy, M.J. Carson, E.A. Thomas, C. Alvarez, P.E. Danielson, P.E. Foye, M. Calbet-Murtro, M. Ingraham, M. Neal

SLEEP-INDUCING PEPTIDE

Cortistatin is a presumptive neuropeptide; 11 of its 14 amino acids are identical to amino acids in somatostatin. In contrast to administration of somatostatin, administration of cortistatin into rat brain ventricles specifically enhances slow wave sleep, apparently by antagonizing the effects of acetylcholine on cortical excitability. The concentration of cortistatin mRNA increases in response to sleep deprivation, suggesting that the neuropeptide creates a pressure for sleep.

The preprocortistatin mRNA is expressed in the cortex and hippocampus in a subset of cells containing -aminobutyric acid that partially overlaps with a subset of cells containing somatostatin. A significant percentage of cortistatin-positive neurons is also positive for parvalbumin. In contrast, no colocalization occurs between cortistatin and calretinin, cholecystokinin, or vasoactive intestinal peptide. During development, a transient increase in cortistatin-expressing cells occurs in the second week after birth in all cortical areas and in the dentate gyrus. A transient expression of preprocortistatin mRNA in the hilar region at P16 is paralleled by electrophysiologic changes in dentate granule cells, a finding that indicates the presence of a cortistatin receptor.

Analysis of the predicted amino acid sequences of rat and mouse preprocortistatin showed that the 14 C-terminal residues, the sequence that is similar to somatostatin, are conserved between species. Lack of conservation of other dibasic amino acid residues whose cleavage by prohormone convertases would give rise to additional peptides suggests that cortistatin 14 is the only active peptide derived from the precursor. Like rat preprocortistatin mRNA, mouse preprocortistatin mRNA is present in interneurons in the cerebral cortex and hippocampus that contain -aminobutyric acid.

The preprocortistatin gene maps to mouse chromosome 4, in a region that shows conserved synteny with human 1p36. The human putative cortistatin peptide has an arginine where the rat and mouse products have a lysine, and the N-terminal end has three more amino acids than the rat and mouse products do.

HOMEOSTATIC PEPTIDES

The hypothalamus acts as a major regulatory center for autonomic and endocrine homeostasis. Structurally, it is a confederation of nuclei that regulate a broad array of physiologic and behavioral activities. For some of these activities, particular peptides are major products of individual nuclei. We used directional tag polymerase chain reaction subtraction to detect rat mRNAs selectively expressed within the hypothalamus.

In situ hybridization studies showed that one of these mRNAs was expressed exclusively by a bilaterally symmetric structure within the posterior part of the hypothalamus. This mRNA encodes the putative precursor of a pair of small peptides that share substantial amino acid identities with each other and with the gut hormone secretin. The mRNA, which accumulates primarily more than 3 weeks after birth in rats, is restricted to neuronal cell bodies of the dorsal and lateral hypothalamic areas.

Antisera to its protein product detected these cell bodies and their fibers, which are located throughout the posterior part of the hypothalamus and in diverse projection targets in other areas, including brainstem and thalamus. Immunoelectron microscopy indicated that the protein is associated with secretory vesicles. Infusion of a synthetic version of one of the peptides into brain ventricles lowered body temperature. The same peptide had excitatory activity when applied to hypothalamic neurons in culture but not when applied to hippocampal neurons, suggesting a specific receptor. Cumulatively, these observations suggest that the mRNA encodes peptides that act endogenously within the CNS in energy homeostasis and reveal part of the circuitry through which that action is accomplished.

MECHANISM OF DRUG ACTION

Neuroleptic drugs such as haloperidol are used for managing various psychiatric disorders, such as schizophrenia, Huntington's disease, and dementia related to Alzheimer's disease. These drugs antagonize dopamine D2 receptors in the striatum immediately after administration, but a few weeks of treatment are required before beneficial effects occur. In collaboration with Digital Gene Technologies, La Jolla, California, we used their automated gene expression system, TOGA (TOtal Gene expression Analysis), to investigate the effects of haloperidol on the expression of all mRNAs in mouse striatum.

With the TOGA method, each mRNA, known or unknown, was assigned a unique digital identity, defined by an eight-nucleotide sequence and a precise distance from the 3´ end of the mRNAs. In a series of 256 polymerase chain reactions, performed by robots using striatal extracts, the presence and relative concentrations of the majority of the striatal transcripts were determined and viewed in an electronic database.

During 2 weeks of treatment with haloperidol, alterations in the expression patterns of 70--90 mRNAs were revealed. cDNA clones of the affected mRNAs were isolated and identified. These findings provide leads for further studies on the mechanism of action of neuroleptic drugs and clues to the molecular pathologic changes that underlie these psychiatric diseases.

MICROGLIA

Because of the difficulties of isolating adequate numbers of microglial cells from tissue from adults, much of the understanding of the function of the microglia is based on characterizations of microglial cells that develop in mixed glial cultures. We compared the phenotypes of murine microglial cells isolated from adults, neonates, and mixed glial cultures with the phenotypes of spleen cells from fetuses, neonates, and adults.

In the CNS in adults, the only resident population of cells that express CD45, a protein tyrosine phosphatase, are the F4/80+ and FcR+ cells: the microglia. Unlike the situation in all other differentiated cells of hemopoietic origin, CD45 levels in microglial cells do not increase from the neonatal period through adulthood. Rather, their levels are indistinguishable from the low levels found on a small population of embryonic day 16 liver cells. Conversely, the F4/80 levels of microglial cells are higher than those of splenic macrophages. Microglial cells that develop in mixed glial cultures have a more activated phenotype, with low F4/80 values, weak expression of MHC class II molecules, and the appearance of a subset of cells positive for the dendritic cell marker NLDC145. Additionally, the level of CD45 is similar to that found on microglial cells activated in vivo, intermediate between the level of CD45 found on microglial cells from adults and the level found on spleen cells from adults.

Consistent with this activated phenotype, treatment with indomethacin showed that microglial cells from mixed glial cultures can present a peptide antigen to naive T cells expressing a defined T-cell receptor. Although microglial cells from adults did express costimulatory molecules B7.2, intracellular adhesion molecule-1, and CD40 and could be induced to express MHC class II molecules, they did not present antigen in the same assay. Interestingly, these same cells could stimulate T-cell proliferation in a mixed lymphocyte reaction but not in an allogeneic specific manner. Taken together these data suggest that compared with other tissue macrophages, microglial cells in adults remain in a relatively immature and unactivated state of differentiation.

OLEAMIDE AND ANANDAMIDE DEGRADATION

Fatty acid amide hydrolase is the membrane-bound enzyme that degrades neuromodulatory fatty acid amides, including oleamide and anandamide. The mRNA of the hydrolase is distributed throughout the CNS in rats and accumulates progressively between embryonic day 14 and postnatal day 10 and then decreases into adulthood. In situ hybridization showed widespread distribution of the mRNA in neuronal cells and some glial populations throughout the CNS. The most prominent concentrations were detected in the neocortex, hippocampal formation, amygdala, and cerebellum. The distribution in the CNS suggests that degradation of neuromodulatory fatty acid amides at their sites of action influences their effects on sleep, euphoria, and analgesia.

LEARNING AND MEMORY

A large body of work suggests that protein kinase C (PKC) is crucial for the induction of long-term potentiation and long-term depression and for hippocampus-dependent forms of learning and memory. Surprisingly, mice containing a null mutation in the gene encoding the neuron-specific isoform (PKC) have only mild deficits in hippocampus-dependent learning and memory.

We examined the levels of different PKC isoforms and substrates in PKC knockout mice and found that GAP-43 and PKC are upregulated, whereas RC3 and PKCß levels are unaffected. Upregulation of GAP-43 suggests that changes in presynaptic gene expression can compensate for postsynaptic deficits (PKC is thought to be expressed postsynaptically), whereas an increase in PKC suggests that this isoform can compensate for decreased production of PKC. These observations suggest that the mechanisms underlying many physiologic processes in mammals are redundant and that homeostatic mechanisms exist to allow organisms to compensate for mutations that cause loss of function.

PUBLICATIONS

Burton, F.H., Forss-Petter, S., Battenberg, E., Bloom, F.E., Sutcliffe, J.G. Complex neurological disorder in mice caused by a neural cholera toxin transgene. Transgenics, in press.

Carson, M.J., Reilly, C.R., Sutcliffe, J.G., Lo, D. Mature microglia resemble immature antigen-presenting cells. Glia, in press.

de Lecea, L., del Rio, J.A., Criado, J.R., Alcantara, S., Morales, M., Danielson, P.E., Henriksen, S.J., Soriano, E., Sutcliffe, J.G. Cortistatin is expressed in a distinct subset of cortical interneurons. J. Neurosci. 17:5868, 1997.

de Lecea, L., Ruiz-Lozano, P., Danielson, P.E., Peelle-Kirley, J., Foye, P.E., Frankel, W.N., Sutcliffe, J.G. Cloning, mRNA expression and chromosomal mapping of mouse and human preprocortistatin. Genomics 42:499, 1997.

de Lecea, L., Sutcliffe, J.G. Peptides, sleep and cortistatin. Mol. Psychiatry 1:349, 1996.

Kilduff, T.S., de Lecea, L., Usui, H., Sutcliffe, J.G. Isolation and identification of specific transcripts by subtractive hybridization. In: Molecular Regulation of Arousal States. Lydic, R. (Ed.). CRC Press, Boca Raton, FL, in press.

Watson, J.B., Margulies, J.E., Coulter, P.M., Gerendasy, D.D., Sutcliffe J.G., Cohen, R.W. Functional studies of single-site variants in the calmodulin-binding domain of RC3/neurogranin in Xenopus oocytes. Neurosci. Lett. 219:183, 1996.

Top of Page


 

 







Copyright © 2004 TSRI.