Molecular Targets for Drug Development KBF Symposium* September 4 - 6, 1996, Braunschweig, Germany Organizer: GBF, Gesellschaft fL~r Biotechnologische Forschung mbH, Mascheroder Weg 1,38124 Braunschweig, Germany

*The Biomedical Research Association, the "Verbund Klinisch-Biomedizinische Forschung (KBF)", is composed of four national biomedical research centers which belong to the Hermann von Helmholtz Gemeinschaft Deutscher Forschungszentren (Hermann von Helmholtz Association of German Research Centers) as well as nine other German research institutions with medical research programs. The purpose of the KBF is to improve clinical research by providing a network which fosters the exchange of information and the establishment of joint projects. The annual scientific meeting of the KBF is held in different cities in Germany and serves as a forum for the discussion of biomedical and clinical research. Previous symposia were devoted to "Gene Therapy", "Genetic Instability" and "Complex Genetic Diseases".

Abstracts

Springer

J Mol Med (1996) (Abstracts): B13-B20

B14

Inhibitors of DNA-topoisomerases - when good enzymes turn bad

Connecting architecture and function in the m a m m a l i a n genome

A. H. Andersen, H. Biersack, B. R. Knudsen, K. Christiansen, O. Westergaard

R. Berezney, H. Ma, X. Y. Wei, J. Samarabandu, M. J. Mortillaro, S. Kim, W. S. Liou, S. J. Pan, P.-C. Cheng

Department of Molecular and Structural Biology, Aarhus University, C.E Moilers All6, Building 130, DK-8000 Aarhus C., Denmark

Departments of Biological Sciences and Electrical and Computer Engineering, State University of New York (SUNY), Buffalo, NY 14260, USA

In most biological systems DNA is topologically constrained due to intertwiningof the two DNA strands. This property of DNA allows for higher order structures, which can dictate essential biological functions. Thus, unwinding facilitates processes such as DNA replication, transcription and recombin-ation. Regulation of DNA topology is therefore of critical importance to essential cellular functions. In eukaryotic cells the topological structure of DNA is modulated by the enzymes called topoisomerases. These enzymes can be divided into two classes, type I and type lI, which both interconvert different topological forms of DNA by creating transient breaks in the DNA backbone. Beyond their normal physiological functions, topoisomerases are important cellular targets in the treatment of human cancers. Considerable effort has therefore been invested to clarify the molecular mechanism of interaction between drugs and their enzyme target. Data will be presented, which demonstrate that: (i) DNA topoisomerases are the primary cellular targets for a number of clinically employed anticancer agents; (ii) Both the catalytic and drug interacting mechanisms of eukaryotic DNA topoisomerases can be investigated by employing a new DNA substrate system; (iii) the domain structures and the drug interacting domains of human DNA topoisomerases can be studied by complementation studies in yeast; (iv) DNA domains (gene loops) can be mapped by topoisomerase II/drug mediated DNA cleavages within the scaffold attachment regions, corroborating the radial loop model of the 30 nm chromatin fiber.

Application of state of the art microscopy and computer imaging approaches along with highly sensitive molecular probes for detecting genes, chromosomes and their functions, has revealed a remarkable degree of functional compartment-alization inside the nucleus of the intact mammalian cell. Appropriate computer programs have enabled us to define higher orders of chromatin arrangement in correlation with genomic function. Combining these in situ structural techniques with biochemical fractionation demonstrate that individual chromosome territories along with the sites of replication, transcription and RNA splicing factors are anchored to the nuclear matrix architecture. These results have profound implications for elucidating drug-target interactions and activities inside the nucleus of functioning cells. There are growing indications, for example, that various anti-cancer drugs affect proliferation and survival of cancer cells by targeting to the genome at the sites of anchoring to the nuclear matrix. Studies of the proteins which compose the nuclear matrix architecture are in progress. Microsequencing and molecular cloning has suggested that a significant portion of the major nuclear matrix proteins correspond to known pre-mRNA proteins, transcriptional or RNA splicing factors. Additional proteins are involved in other regulatory properties or are newly identified proteins whose functions remain to be resolved. Identifying and characterizing hormonally-induced and tumor specific nuclear matrix proteins will provide a basis for studying their relationships to genomic functions and to drug-target sites associated with the nuclear matrix.

Insights into D N A polymerization mechanism and drug design from structural studies of the AIDS virus reverse transcriptase

Design and properties of influenza neuraminidase inhibitors

E. Arnold P. M. Colman Center for Advanced Biotechnology and Medicine, and Rutgers University Chemistry Department, 679 Hoes Lane, Piscataway, New Jersey 08854-5638, USA In a collaboration with the laboratory of Dr. Stephen Hughes at the ABL-Frederick Cancer Research and Development Center, and scientists at Janssen Research Foundation (Beerse, Belgium) and Hoechst Pharma AG (Frankfurt, Germany), we have been pursuing structural studies of HIV-1 reverse transcriptase (RT). Using the techniques of X-ray crystallography, we have obtained structures of HIV-1 RT with a bound nucleic acid template-primer (JacoboMolina et al., PNAS 90:6320-6324, 1993), with bound nonnucleoside inhibitors (Ding et al., Structure 3:365-379, 1995; Ding et al., Nature Struct. Biol. 2:407-415, 1995; Das et al., submitted) and without bound ligands (Hsiou et al., submitted). These results, together with structures obtained by several other laboratories, have provided insights into the conformational flexibility of HIV- 1 RT, as well as mechanisms of polymerization, inhibition, and drug resistance. Implications for structure-based drug design and for therapeutic strategies will be discussed.

Biomolecular Research Institute, 343 Royal Parade, Parkville, 3052, Victoria, Australia The three-dimensional structure of influenza virus neuraminidase was used to direct the design of novel inhibitors. Analogues of the product of the enzyme reaction, sialic acid, substituted at the 4position are potent and selective inhibitors of influenza virus neuraminidase, inhibit virus replication in cells and are potent antivirals when applied topically in man. New strains of virus with decreased sensitivityto the inhibitors can be selected in tissue culture. They have changes in either the haemagglutininor the neuraminidase gene sequence.

B15

Identification of MHC associated peptides that comprise tumor and transplantation antigens: immunological and therapeutic implications V. H. Engelhard Beirne Carter Center for Immunology Research, University of Virginia Health Sciences Center, MR4 Box 4012, Charlottesville, VA 22908, USA A major focus of research in this laboratory has been to identify peptides that are presented by MHC molecules and ~ecognized by T-lymphocytes. The approach utilized involves direct extraction of the peptides from cell surface MHC molecules and RP-HPLC fractionation. Subsequently, individual fractions are assessed for their ability to reconstitute the epitope recognized by antigen specific T lymphocytes, and the active peptides are identified by sequencing using tandem mass spectrometry. This methodology is particularly powerful when applied to the identification of peptides that comprise tumor or transplantation antigens, or antigens derived from complex intracellular parasites, where the identity of the source protein is not apparent. To date, 10 different peptides that give rise to such antigens have been identified. Identification of the source proteins of these peptides from searches of protein and DNA sequence databases has led to insights into the nature of antigens recognized by the human immune system on melanoma tumors, and the nature of minor histocompatibility antigens that are associated with graftversus-host disease. In addition, while this analysis suggests that many peptides are generated by the conventionalantigen processing mechanism, we have identified an additional mechanism that may be relevant for the production of peptides from membrane and secreted proteins. Finally, this technology has enabled us to demonstrate that MHC-associated peptides may be post-translationally modified in ways that significantly alter their recognition by T-lymphocytes.

Cellular regulation by protein phosphorylation: a historical overview E. H. Fischer Department of Biochemistry, University of Washington, Seattle, Washington, 98195-7350, USA Protein phosphorylation can be considered to be the most prevalent mechanism by which eukaryotic cellular events are regulated. The first part of this talk will be devoted to a brief historical account of how this field has originated.The second part will cover cell regulation by tyrosine phosphorylation as initiated by mitogenic hormones and growth factors. These act on receptors that trigger signalling events leading to cell growth, differentiation and transformation. Mutations can result in a number of pathological conditions such as insulin-independent diabetes (NIDDM) or oncogenicity. Of course, regulation must also involve protein phosphatases that catalyze the reverse reaction. Protein tyrosine phosphatases (PTPs) represent an expanding family of transmembrane and intracellular enzymes. Most receptor forms contain two cytoplasmic catalytic domains and highly variable external structures often containing immunoglobnlin-likeand/or fibronectin type 3 repeats that display the properties of adhesion molecules involved in cell-cell or cell-matrix interaction.The intracellularPTPs

also display a diversity of regulatory/localizationsegments, attached to conserved catalytic domains. The structural motifs involved in the localization of a human T-cell PTP and the function of this enzyme as compared to a truncated form in which the localization segment has been eliminated, will be discussed. The data indicate that kinases and phosphatases cannot be viewed as simply providing on-off signals: depending on their type and where they localize within the cells, the phosphatases can act either positively or negatively in eliciting a particular response.

Functional and biochemical consequences of T cell receptor recognition of partial agonists and antagonists R. N. Germain, E. Levine, J. Madrenas, C. Reis e Sousa, L Samelson, R. Wange, N. Isakov, R. Schwartz, I. Stefanova Lymphocyte Biology Section, Laboratory of Immunology, National Institute of Allergy and Infectious Diseases, and Cell Biology and Metabolism Branch, National Institute of Child Health and Development, National Institutes of Health, Bethesda, MD 20892, USA Antigen-specific activation of T-lymphocytes depends on signals generated upon binding of clonally distributed receptors (TCR) to peptide-major histocompatibility complex (MHC) molecule ligands on an opposing cell membrane. This binding initiates a signaling cascade that in its earliest phases typically involves activation of two src family kinases (p59 fyn and p56 lck) that contribute to tyrosine phosphorylation of several sites on invariant chains (~ and CD3 y, 5, E) that are physically associated with the TCR. This is followed by recruitment of the syk family kinase ZAP-70 to the phosphorylated ~ chains via paired SH2-domains, followed by its phosphorylation and enzymatic activation, which appears critical for additional downstream events such as release of intracellular Ca ++. This picture of TCR signaling, derived initially based on antibody cross-linking studies, has been largely confirmed using physiological peptide-MHC ligands with agonist properties. Recently two other classes of TCR ligands have been identified among ligands composed of peptide-MHC molecule combinations with limited amino acid replacements compared to the original agonist. These partial agonists and antagonists have been shown to induce altered signaling through theTCR, characterized by a change in the forms of phosphorylated ~ chain produced, and the recruitment of ZAP-70 to these phosphorylated ~ chains without its visible phosphorylation or enzymatic activation. This novel pattern is not seen at any tested concentration of agonist, indicating that it represents a qualitative change in the pattern of signal transduction events accompanying TCR recognition of these variant ligands. These changes have profound consequences on the behavior of the T-cell, ranging from differentiation without death in the thymus to peripheral clonal anergy. Possible molecular mechanisms involved in this change in TCR signal transduction will be discussed, as will be the implications of these findings for the design of drugs intended to modulate T-cell function.

BI6

H o w to immortalize a h u m a n B-cell - lessons from Epstein-Barr virus

Cyclin Dependent Kinases: Structure, Function and Destruction

W. Hammerschmidt

To Hunt, M. Brandeis, A. Klotzbiicher, H. Yamano

Institut ftir Klinische Molekularbiologie und Tumorgenetik, H~imatologikum der GSF, Marchioninistral3e 25, 81377 Mfinchen, G e r m a n y

Imperial Cancer Research Fund, Clare Hall Laboratories, South Mimms, Herts EN6 3LD, U K

Epstein-Barr virus (EBV) is a prototypic human tumor virus. It is causally associated with a variety of B-cell lymphomas and at least one carcinoma. Much research indicates that EBV is a human tumor virus because it affects the control and proliferation of infected cells. In particular, EBV induces and maintains proliferation of infected primary human B-lymphocytes efficiently. This process in cell culture is termed immortalization. Biochemical and genetic analyses of the genes of this human herpesvirus have identified five of them as contributing fundamentally to the process of immortalization. The best c h a r a c t e r i z e d gene p r o d u c t s i n v o l v e d in B-cell immortalization are EBNA1, EBNA2 and LMP1. EBNA1 is necessary for the episomal replication of the viral genome by binding to the viral origin of replication, oriP. EBNA2 and LMP1 are essential for the immortalization potential of EBV EBNA2 acts as a transcriptional activator of cellular and viral genes. It transactivates the viral promoters of a number of viral and cellular promoter elements. LMP1 is the only viral protein that is not only essential for B-cell transformation but has also transforming effects in nonlymphoid cells. LMP1 acts as an oncogene in established rodent cells, blocks differentiation in epithelial cells, and induces the expression of B-cell activation markers. LMPI is an integral membrane protein with six transmembrane domains separated by short reverse turns. Both the 23 amino acid amino-terminal and 200 amino acid long carboxy-terminal parts are cytoplasmatic. Little is known about L M P I ' s function but the recent finding by Mosialos and colleagues (Cell, 80, 389-399, 1995) suggests that LMP1 interacts with components of the signal transduction pathway which is used by members of the tumor necrosis factor receptor family. The contributions of both EBNA2 and LMP1 to the control of Bcell proliferation will be discussed.

Mitotic cyclins activate p34cdc2 as a protein kinase by forming stable 1:1 complexes, and the destruction of the cyclin subunit by proteolysis is required for passage from mitosis to interphase, although other proteins (such as Cut2 and Pds I in fission and budding yeast respectively) must also be degraded, probably by the same machinery, in order to bring about the metaphase to anaphase transition. There is also very good evidence that the proteolysis machinery comprises a specific ubiquitiation complex, known as the "cyclosome" or "anaphase promoting complex" acting together with the multicatalytic proteasome. Different cyclins are degraded with slightly different timing. To test if this timing depended on the N-terminal regions of cyclinsA, B 1 and B2, we made domain swaps between them. While the N-terminus of the B-type cyclins confers mitotic instability on cyclin A and other proteins, such as Gal4 or CAT, the cyclin A N-terminus does not function as a destruction signal in any other context save cyclin A:cdc2 (even free cyclin A is not a substrate for destruction). This failure of destruction does not appear to be caused by a failure of ubiquitination, as the indestructible chimeras are well ubiquitinated, suggesting that ubiquitination per se is not a sufficient signal for destruction, although the evidence that it is necessary is extremely strong. In the rapid cleavage cell cycles of early amphibian and molluscan embryos, cyclin destruction is active only briefly, starting just before the metaphase-anaphase transition and ending about 5 minutes later. But in cultured fibroblasts, we find that destruction of B-type cyclins remains constitutively active for several hours, and is not switched off until the onset of S-phase. It seems that this post-translational control mechanism plays a major role in keeping the levels of B-type cyclins low during G 1 phase. We would like to know how many other targets for this cell-cycle phase specific proteolysis system exist, as well as to understand its mechanism and control.

The tyrosine kinase inhibitor drug discovery cascade for the development of novel therapeutic agents

Signal transduction by protein-tyrosine kinases and phosphatases

P. Hirth

T. Hunter, A. Bilwes, M. Broome, J. den Hertog, J. Noel, D. Schlaepfer

SUGEN, Inc., 515 Galveston Drive, Redwood City, California 94063-4720, U S A Activated tyrosine kinases have been implicated as causative factors in the pathomechanism of a variety of diseases, particularly those that involve hyperproliferation of certain cell types. The result of hyperproliferation can manifest itself in cancer, angiogenesis, restenosis, or psoriasis. Therefore, therapeutic intervention on these targets seems attractive. Strategies to find the best modes of intervention are discussed. In more detail, key elements of a drug discovery cascade are discussed which include aspects of hithroughput screening systems, as well as downstream assays and animal models. Pharmacological requirements and other key parameters for active compounds as drug candidates will be discussed.

The Salk Institute, La Jolla, CA 92037, USA Mitogenic signalling by the PDGF receptor protein-tyrosine kinase (RPTK) requires not only its intrinsic PTK activity, but also that of the c-Src nonreceptor PTK. In PDGF-treated cells c-Src binds to the PDGFR via its SH2 domain, and is itself activated. Overexpression of a catalytically inactive SH3/SH2-containing c-Src fragment, Y138F or Y133F SH3 mutant mouse c-Src proteins in Src- mouse fibroblasts all block PDGF- and E G F - i n d u c e d mitogenesis implying that both c-Src catalytic activity and its SH3 function are critical for RPTK signalling. Thus, Src-specific PTK inhibitors may be useful in abrogating RPTK-induced cell growth, for instance in certain types of cancer. Receptor protein-tyrosine phosphatase a (RPTPa), like other RPTPs, has two cytoplasmic catalytic domains, DI and DII. The crystal structure of RPTPa DI shows it can form an inactive dimer, suggesting that RPTPa and

B17 other RPTPs are negatively regulated by ligand-induced dimerization. The structure of the dimer reveals how specific inhibitors might be designed. Integrins are cell surface receptors for matrix proteins, such as fibronectin, localized in focal adhesions. Integrins lack intrinsic PTK activity, but fibronectin binding increases tyrosine phosphorylation of the associated FAK PTK, leading to the sequential binding of c-Src, and the SH3/SH2 adaptor protein, Grb2, through their SH2 domains, to P.Tyr397 and RTyr925, respectively, in FAK. The binding of Grb2/Sos leads to activation of ERK2 MAP kinase. This is blocked by expression of dominant-negative Ras. This suggests a model in which c-Src, once bound to activated FAK, phos-phorylates additional sites in FAK, which recruit SH2 domain proteins, such as Grb2, thus providing a mechanism for integrinmediated signalling. Activation of ERK2 also occurs in SrcP cells expressing a dominant negative form of Src, under conditions where Grb2 binding to FAK is not observed. This alternative mechanism of ERK2 activation may involve Nck binding to Cas, which is phosphorylated by FAK itself.

Cyclin dependent kinase 2 as drug targetCDK inhibitors, senescence and lymphoid neoplasia A. A. Jagasia, F. Fazal, C. Pan, D. Sher, M. O. Diaz Loyola University Chicago, Oncology Institute, 2160 S. First Avenue, Maywood Illinois 60637, USA. The CDK inhibitor p 16(INK4A) is not expressed in normal mature T lymphocytes, B lymphocytes, thymocytes, or in CD34+ hematopoietic precursors. The INK4A gene is frequently deleted or mutated in lymphoblastic leukemia cells. The deletion or mutation of INK4A suggests that selection pressure against the gene developed during leukemogenesis. We proposed the hypothesis that the gene is expressed as part of the normal mechanism of senescence in transformed lymphocytes, and that its deletion allows the neoplastic clone to continue its expansion to give rise to a leukemia. In order to test this hypothesis we have studied the expression of the INK4A gene in normal lymphocytes induced to proliferate continuously in vitro after stimulation with PHA and interleukin-2. These cells irreversibly arrest after 20-24 population doublings, and they express both INK4A mRNA, and pl6 protein at the time of senescence. Expression of pl6 in senescent cells occurs in the presence of hypophosphorylated RB, which in young cells represses INK4A transcription. INK4A mRNA is expressed in most leukemia and lymphoma dell lines that have not deleted the gene. But, in these cases, the p 16 protein is absent. We postulate that post-transcriptional downregulation of p16 is a frequent mechanism of inactivation of INK4A expression in neoplastic cells.

Cyclin dependent kinase 2 as drug target S-H. Kim, U. Schulze-Gahmen, J. Brandsen*, W. Filgueira de Azevedo, Jr. Department of Chemistry and E. O. Lawrence Berkeley National Laboratory, University of California, Berkeley, California 94720 USA; *Current address: Department of Crystallography, University of Utrecht, The Netherlands Cell cycle progression is tightly controlled by the activity of cyclindependent kinases (CDKs). Deregulation of CDK activities have been associated with many cancers, and there is strong interest in the discovery of chemical inhibitors of CDKs. Since ATP is the authentic cofactor of CDK2 it can be considered as a "lead compound" for discovery of CDK2 inhibitors. However there are two major concerns: adenine containing compounds are common ligands for many enzymes in cells, thus, any adenine derivatives may inhibit many enzymes: second, any highly charged groups such as phosphates in ATP will prevent uptake by the cells. Our studies provide the structural basis for overcoming the two difficulties by appropriate modification of a common base such as adenine to endow specificity and cell uptake of the derivatives. In this presentation crystal structures of the complexes between human CDK2 and four different ligands are summarized: ATP, olomoucine, isopentenyladenine, and flavopiridol.

Structural aspects of binding specificity anct regulation in hematopoietic hormone-receptor complexes A. A. Kossiakoff, A. de Vos, L. M. Randal Department of Protein Engineering, Genentech, Inc., 460 Point San Bruno Blvd., South San Francisco, CA 94080, USA We are studying the relationships between structure and the associated biology of cytokine hormones and receptors. The systems include the endrocrine family of hormones and receptors (growth hormone, placental lactogen and prolactin) and interferon-'/(IFN1,). The sequence homology of the extracellular domains (ECD) of these receptors classifies them as belonging to the hematopoietic receptor superfamily. This family is subdivided into Class I, which includes the endrocrine receptors, and Class II, characteristic of the IFN-y receptors. Biological response is triggered by hormone binding that drives receptor aggregation. The aggregated receptor complex binds one or several tyrosine kinases belonging to the JAK family, which then transphosphorylate elements on themselves, receptors, and associated transcription factors. The stoichiometry of aggregation is different between the endrocrine systems and IFN-y. The endrocrine hormones bind two copies of the same receptor to give a 1:2 active complex, whereas IFN-7, which is homodimeric, binds two copies of two different receptors, the so-called ~ and [~ chains to give a complex with a 1:2:2 stoichiometry. We have determined the structures of five endrocrine hormone-ECD receptor complexes and IFN-y bound to the ECD of the a chain receptor. We also have the complementary mutagenesis data that relates the energetically important resides in the binding epitope with those residues defining the structural interface. An important finding is that only about 25% of the residues in the interface contribute significantlyto the binding energy. In fact, there are groups involved in good hydrogen bonds and salt-bridges that are energetically null. We have also produced a mutant of IFN-7 that enforces a stoichiometry of 1:1:1 indicating that the active complex needs only a single copy of the ct and ~ chains.

B18

Regulation and quality control of peptide loading onto M H C class II molecules H. Kropshofer, A. B. Vogt, G. Moldenhauer, G.J. Hiimmerling Division of Molecular Immunology, Tumor Immunology Program, German Cancer Research Center, Im Neuenheimer Feld 280, 69120 Heidelberg, Germany MHC class II molecules are peptide receptors and specialized in presenting antigenic peptides to T helper cells. After biosynthesis a and b chains of class II molecules assemble together with the Invariant chain (Ii) in the endoplasmic reticulum forming a nonameric complex that is sorted to endosomal/lyso-somal compartments. There, Ii is proteo-lytically digested, but a fragment of Ii designated CLIP remains in the class II peptide binding groove and needs to be removed before loading with antigenic peptides and exit to the cell surface can proceed. An additional molecule encoded by the MHC-class II region, HLA-DM, appears to be required for efficient CLIP release and acquisition of cognate peptides. Using in vitro approaches, we have observed that DM acts as a catalyst and chaperone for peptide loading. The experiments show that DM transiently interacts with DR molecules. This interaction not only increases the dissociation rate of CLIP from DR molecules but also accellerates subsequent loading by antigenic peptides, following Michaelis Menten kinetics, with turnover rates of 5-20 molecules per minute. Importantly, DM releases not only CLIP but also any other peptide that is suboptimally bound to the DR molecules. In contrast, peptides with optimal anchors that are firmly bound to DR are resistant to removal by DM. Thus, one can envisage that in the cellular loading compartments DM molecules continuouslyprobe DR:peptide complexes for release of the peptide, with subsequent replacement by high affinity peptides. This peptide editing and quality control process is an important biological function of HLA-DM and ensures that a large percentage of stable class II:peptide complexes are formed that persist at the cell surface and increase the probability of being recognized by T lymphocytes.

Inhibition of protein kinases and the cell cycle machinery as prospective drugs

rats (4). (III) EGF receptor blocker was shown to inhibit squamous cell carcinoma growth in mice engrafted with this tumor which overexpresses EGFR (5). (IV) Tyrosine kinase blockers inhibit sepsis in mice infected withE. Coli (6,7) and EAE in mice (8). Recently, we demonstrated the efficacy of an EGFR kinase blockerAG 1478 and a blocker of cdk2 activation in inhibiting the growth of BPVI6 infected human kertinocytes. These results identifies these compounds as potential antipapilloma agents (9). References 1. A. Levitzki, Eur. J. Biochem., 226, 1-13, 1994. 2. A. Levitzki and A.Gazit, Science, 267,1782-171, 1995. 3. N. Maydan et al., Nature, 379, 645-648, 1996. 4. G. Golomb, et al., Atherosclerosis, in press, 1996. 5. T. Yoneda et al., Cancer Res.,51, 4430-4435,199 1. 6. A. Novogrodsky et al., Science, 264, 1319-1322,1994. 7. A. Vanichkin et al., J. Infect. Dis., in press, 1996. 8. T. Brenner et al., submitted. 9. H. Ben-Bassat et al., Submitted.

Farnesyltransferase inhibitors as agents to treat ras-dependent tumors C. A. Omer, N. J. Antony, M. W. Conner, S. J. deSolms, J. S. Gibbs, S. L. Graham, G. D. Hartman, K. S. Koblan, N. E. Kohl, A. Oliff, T. M. Williams Departments of Cancer Research, Medicinal Chemistry and Safety Assessment, Merck Research Laboratories, West Point, PA 19486, USA Mutationally activated Ras is found in greater than 20 % of all human cancers. For Ras to transform mammalian cells, it must be post-translationally modified with a farnesyl group in a reaction catalyzed by the enzyme farnesyl-protein transferase (FTase). Inhibitors of FTase have thus been proposed as anti-cancer agents. We have genetically and biochemically characterized FTase and identified selective inhibitors of its activity. One of these Ftase inhibitors, L-744,832, is capable of regressing tumors in transgenic mice that express a mutationally activated H-ras gene. Mice treated with L-744,832 show no systemic toxicity suggesting that such agents may be safe and effective anti-tumor agents for some cancers.

A. Levitzki Dept. of Biological Chemistry, Institute of Life Sciences, The Hebrew University of Jerusalem, Givat Ram, Jerusalem 91904, Israel Numerous diseases, such as atherosclerosis, psoriasis and inflammatory diseases result form aberrations in signal transduction pathways, mainly because of the enhanced activities of protooncoproteins and oncoproteins. In inflammatory diseases as rheumatoid arthritis (RA) and multiple sclerosis (MS), the enhanced signaling of certain cytokines as TNFc~ and IL-1 seem to be the causative element. We, therefore, initiated the development of selective agents to attenuate signaling pathways which are enhanced in a variety of diseases and formulated the general concept of signal transduction therapy (1,2). We demonstrated the efficacy of a number of tyrosine kinase inhibitors (tyrphostins) in vivo: (I) A JAK-2 blocker was shown to eradicate recurrent human pre-B acute lymphoblastic leukemia (ALL) engrafted in SCID mice (3) (II) A PDGF receptor kinase blocker was shown to inhibit restenosis in

Alternative targets for HIV antiviral drugs: target identification and assay development G. N. Pavlakis National Cancer Institute, ABL-Basic Research Program, Frederick Cancer Research and Development Center, Frederick, MD 21702, USA Knowledge of the function of two HIV enzymes, reverse transciptase and proteases led to the development of drugs that inhibit enzyme function and HIV replication. Development of drugs against other HIV targets is highly desirable, and may give the opportunity to develop synergistic combination of drugs for clinical applications. HIV integrase is another viral enzyme that is an attractive target. Non-enzymatic viral proteins may also provide targets. An example is the attempt to develop compounds inhibiting the RNA-protein

B19 interactions mediated by the unique Zn-finger domains in nucleocapsid part of Gag protein. The Rev protein of HIV is an essential regulatory protein, which binds specifically to a unique RNA target within the env coding region, named Rev responsive element (RRE). This interaction leads to nucleocytoplasmatic mRNA transport and increased expression of structural viral proteins. Under all circumstances, inhibition of Rev function leads to inhibition of viral expression. Since Rev is an essential viral protein with well-defined functions, it is an attractive target for drug development. Development of high thruput assays is essential for this purpose, and the simplification of assays is a key for this. We have developed approaches that can be adopted to such assays for detecting lead compounds with anti-Rev activities. Research sponsored by die National Cancer Institute, DHHS, under contract with ABL.

New protein kinases as therapeutic targets: clues from yeast and flies G. Plowman SUGEN, Inc., 515 Galveston Drive, Redwood City, CA 940634720, USA Focused cloning strategies have lead to a dramatic expansion of the known members of the protein kinase family. Understanding the biology of these proteins and validation of their role in human disease presents an enormous challenge. The validation process for a new gene can be dramatically accelerated by early recognition of a homologue in organisms where genetic manipulation is both rapid and informative. The tremendous growth of DNA sequence and biologic relational databases for yeast and invertebrates increases the likelihood of making such connections.The "collision of databases" from yeast and humans will play an ever more important role in the drug discovery process. Using such strategies, we have recently identified several new families of mammalian kinases which are structurally related to cervesiae, drosophila, and elegans genes. One group of kinases are required for chromosome segregation and centrosome replication in yeast. We will present data on the effect of over-expression of dominant negative and activated forms of these human kinases on centrosome biology, microtuble organization, and nuclear division, as compared with the phenotype of the yeast mutants. We have further correlated the expression of these genes, in normal vs. diseased tissue, suggesting that one of these novel kinases may serve as a viable target for drug discovery.

Cellular signaling by tyrosine phosphorylation J. Schlessinger New York University Medical Center, Department of Pharmacology, 550 First Avenue, New York, NY 10016 Protein tyrosine kinases and phosphatases play an important role in the control of cell growth and differentiation. Dimerization of growth factor receptors is driven by ligand binding to the extracellular domain (1). Receptor dimerization is responsible for tyrosine kinase activation and for autophosphorylation. Signal transmission by receptor tyrosine kinases is mediated by a group of signaling molecules containing SH2, SH3 and pleckstrin homology (PH) domains. The PH domain of phospholipase Cdl binds with high

affinity and sterospecificity to PtdlnsP2 containingregions in plasma membrane. The PH/PTB domain of the adaptor protein Shc binds with high affinity to NPXpY containing sequences in receptor tyrosine kinases. The amino acid sequences of PH domains are poorly conserved. Yet, PH domains exhibit a common fold and mediate cell surface localization in cellular signaling (2,3).We have discovered a novel non-receptor protein tyrosine kinase termed Pyk2 that is highly expressed in neuronal cells and is activated by a variety of extracellular signals that elevate intracellular calcium concentrations. Pyk2 appears to function as an intermediate that links a variety of calcium signals with ion channel function and with the Ras/MAP kinase signaling pathway (4). Pyk2 is also activated by LPA, Thrombin, Angiotensin II as well as by TNFa, UV- irradiation and by changes in osmolarity. We have demonstrated that Pyk2 overexpression leads to activation of the C-Jun N-terminal kinase (JNK). By using dominant negative mutants of Pyk2 and dominant negative mutants of CDC42, Racl, and Ras we have demonstrated that Pyk2 can function as a cell type specific, upstream mediator of the JNK or MAPK signaling pathways. Pyk2 has the potential to activate the MAPK signaling pathway, the JNK signaling pathway or both signaling pathways in the context of different cells and in response to different extracellular stimuli (5). Receptor protein tyrosine phosphatase b (RPTPb) is expressed in the developing nervous system and contains a carbonic anhydrase domain (CAH) as well as a fibronectin type II repeat in its extracellular domain. We have demonstrated that RPTPb expressed on the surface of glial cells binds to the glycosylphosphatylinositol (GPI) anchored recognition molecule contactin on neuronal cells by means of the CAH domain leading to neurite outgrowth and differentiation (6). We have cloned a novel contactin associated transmembrane receptor (pl90/Caspr) containing a mosaic of domains implicated in protein- protein interactions. The extracellular domain of Caspr contains a neurophilin/ coagulation factor homology domain, a region related to fibrinogen b/g, EGF-like repeats, neurexin motifs as well as unique PGY repeats found in a molluscan adhesive protein. The cytoplasmic domain of Caspr contains a proline-rich sequence capable of binding to a subclass of SH3 domains of signaling molecules. Caspr and contactin exist as a complex in rat brain and are bound to each other by means of lateral (cis) interactions in the plasma membrane.We propose that Caspr functions as a signaling component of contactin enabling recruitment and activation of intracellular signaling pathways in neurons. The binding of RPTPb to the contactin/Caspr complex may provide a mechanism for cellcell communication between glial cells and neurons during development (7).

Rho-family GTP-binding proteins as drug targets for Ras-mediated malignancies M. Symons*, R-G. Qiu*, B. Anand-Apte*, A. Viswanathan*, J. Onton*, J. Chen*, M. Ruggieri*, Do Kirn*, B. Zetter* *ONYX Pharmaceuticals, 3031 Research Drive, Richmond, CA 94806; USA, *Departments of Cell Biology and Surgery, Children's Hospital, Harvard Medical School, Boston, MA 02115, USA The small GTP-binding proteins Rac and Rho play a central role in the reorganization of the actin cytoskeleton induced by growth factors or oncogenic Ras. We showed that Rac is essential for Ras transformation and that expression of constitutively active Rac 1V 12 in Ratl fibroblasts is sufficient to cause malignant transformation,

B20 indicating that Ras drives both the Rac and MAP kinase pathways to induce transformation. We also showed that Rho is essential for Ras transformation, and although expression of constitutively active RhoAV 14 is not sufficient to cause malignant transformation on its own, it strongly synergizes with constitutively active Raf in focus formation assays. Together these results indicate that the Rac and Rho pathways are a source for novel drug targets for Ras-mediated malignancies. We are currently investigating the role of Cdc42 in Ras transformation. Tumors produced by Rac 1V 12-expressing Ratl fibroblasts in nude mice display strong local invasion. These lines also show increased in vitro invasion through Matrigel, whereas R a t l fibroblasts expressing dominant negative R a c l N 1 7 are inhibited in Matrigel invasion, confirming a role for Rac in invasion. We also studied the role of Rac in migration stimulated by PDGF and LPA, p r o c e s s e s w h i c h we p r e v i o u s l y s h o w e d to be Ras-dependent. PDGF- and LPA-stimulated migration was inhibited in RaclN17-expressing Ratl fibroblasts, but was normal in Ratl fibroblasts expressing a dominant negative mutant of MEK1. Migration stimulated by soluble fibronectin was unaltered in the Rac 1N 17-expressing Rat 1 fibroblasts, showing that not all migratory behavior is controlled by Rac. These results demonstrate that the Rac and MAP kinase pathways mediate different signals which are controlled by Ras. Interestingly, in contrast to PDGF, LPA does not stimulate ruffling, indicating that Rac regulates cell migration independent of its control of lamellipodia formation.

Ras and its regulators and effectors as target molecules A. Wittinghofer, R. Ahmadian, C. Block, C. Herrmann, G. Horn, R. Mittal, N. Nassar, K. Scheffzek Max-Planck-Institut f~ir molekulare Physiologie, Rheinlanddamm 201, 44139 Dortmund, Germany The product of the Ras oncogene is involved in the transduction of signals from the plasma membrane to the nucleus. It is a GTP-binding protein with functions as a switch between the GTP-bound ONstate and the GDP-bound OFF-state. Through the action of many different cell surface receptors Ras becomes loaded with GTP and thus activated. In its GTP-bound state it interacts with effector proteins and through this interaction the effectors pick up the signal originating from the cell surface and transmit it to the nucleus. Structural and mechanistic studies on the interaction of Ras with its effectors will be presented. The GTPase reaction of p21ras is of crucial importance to its biological function because a blockage of this reaction is the primary cause for the oncogenic behaviour of Ras mutants involved in tumour formation in humans. To understand the mechanism of GTP hydrolysis we are involved in kinetic, thermodynamic and structural investigations of the GTPase reaction. We have put forward the suggestion that the g-phosphate of GTP itself acts a general base for the reaction. This has been verified by various experimental approaches which show in effect that there is a linear free energy relationship which relates the catalytic activity with the pKa of the g-phosphate. GTPase activating proteins (GAPs) stimulate the GTPase reaction by several orders of magnitude. The catalytic domains of p 120-GAP and neurofibromin were expressed in E. coli and their interactions were tested. Equilibrium and stoppedflow fluorescence measurements were used to analyze the role of GAP in the GTPase reaction. The three-dimensional structure of the catalytic domain of p 120-GAP was solved and its implications for understanding the mechanism will be presented.

Structures of retroviral enzymes as guides in drug design A. Wlodawer Macromolecular Structure Laboratory, NCI-Frederick Cancer Research and Development Center, ABL-Basic Research Program, Frederick, MD 21702, USA Crystal structures of all three enzymes which are encoded by the human immunodeficiency virus (HIV) are now available. These enzymes include protease (PR), integrase (IN), and reverse transcriptase (RT). The first two enzymes have been extensively studied in this laboratory, together with related enzymes encoded by other retroviruses. This combination of approaches yielded information of major importance in drug design. Crystals structures of the inhibitor complexes of HIV PR were used by pharmaceutical companies in the discovery of new drugs, three of which have now been approved for use against AIDS. Structures of inhibitor complexes of enzymes from feline immunodeficiency virus and equine infectious anemia virus are helping in understanding of the e m e r g i n g problem of drug resistance, and may lead to the development of therapeutical agents targeted at animal diseases. The structure of the Y 185K mutant of the catalytic domain of HIV IN, solved by E Dyda and D. Davies (NIH), elucidated many details of the architecture of the enzyme, but part of the active site remained obscure, since it was disordered in the available crystals. The active site of a related enzyme from the avian sarcoma virus, however, was fully ordered, and yielded important clues important in drug design. It also suggested that a Y185H mutant of HIV IN could be crystallizable. Most recently, we solved the structure of this mutant of HIV IN, and found that in this case the complete active site could be traced. This structure will shed light on the relationships of the enzymes from different retroviral sources, and will provide important information to drug designers. Research sponsored by the National Cancer Institute, DHHS, under contract with ABL.