Chiral transition metal complexes have been designed which target and probe double helical DNA. These synthetic transition metal complexes have been useful in exploring principles in molecular recognition, as mimics for gene regulatory proteins, and as photophysical and photochemical probes of nucleic acids. In particular, luminescent characteristics of these metal complexes have been exploited in probing DNA-mediated electron transfer chemistry, or how the DNA double helix could function as a molecular "wire."

The nicotinic receptor is a transmembrane hetero-oligomer which carries the binding sites for acetylcholine (ACh), and the ion channel as the elements which account for its fast opening and slow desensitization by the neurotransmitter. The amino acids which compose the principal component of the ACh binding sites have been identified, on the native [[alpha]]-subunit form Torpedo marmorata, with a photolabile competitive antagonist p-(N,N)-dimethyl aminobenzene diazonium (DDF), as [Trp 86, Tyr 93], [Trp 149, Tyr 151], and [Tyr 190, Cys 192, Cys 193, Tyr 198]. They belong to three distinct loops within the larg NII2-terminal hydrophilic domain of the [[alpha]]-subunit. Amino acids belonging to the complementary component have also been identified by site-directed mutagenesis. Furthermore, in T. marmorata, the relative contribution of these various loops change upon stabilization of the high-affinity desensitized state by the allosteric effector meproadifen. The channel blocker chlorpromazine covalently labels, upon UV irradiation, all the subunits when bound to its unique high-affinity site located in the ion channel. The labeled amino acids belong to the hydrophobic segment MII and form three superimposed rings assuming MII [[alpha]] helixal. The functional relevance of the DDF and chlorpromazine-labeled amino acids, on the pharmacological and gating properties of the receptor, are investigated by site-directed mutagenesis. Mutations within MII of [[alpha]]7 neuronal nicotinic receptor selectively alter the closing of the desensitized state, the ionic selectivity for Ca⁺⁺ vs. Na⁺/K⁺ and for cations vs. anions. Furthermore, obtention of functional chimaera between [[alpha]]7 and 5HT3 receptor reveals a functional autonomy of the neurotransmitter binding and channel domains.

Plausible molecular mechanisms of ion channel opening, function, and desensitization are presented and the structure of anionic and cationic ligand-gated ion channels discussed in terms of common backbone structure and side-chain diversity. Implication for mechanisms of short-term memory are discussed and a molecular model of the IIebb synapse presented.

We consider the chemistry of p-RNA relevant to a chemical etiology of the natural nucleic acids, the quest for a chemical rationalization of Nature's evolution of the structures of RNA and DNA. An experimental approach to the problem consists in the chemical synthesis of potential nucleic acid alternatives and the analysis of their functional properties (base pairing, nonenzymic replication, phenotype catalysis) in comparison to those of natural RNA. Such nucleic acid alternatives are chosen on the basis of the criterion of whether their potential for constitutional self-assembly under preenzymic conditions is judged to be comparable to, or greater than, that of RNA. Alternatives that turn out to be nonfunctional could not have functional alternatives; on the other hand, could have, but have not (or may have only temporarily) been chosen by Nature as genetic system(s). p-RNA is the fourth in the series of nucleic acid alternatives studied in our laboratory in this context. Its chemical properties in comparison to those of RNA are such that we are confronted with the question of whether p-RNA could have been a forerunner of RNA.

Proteins control expression of particular genes by recognizing DNA elements near their transcriptional start sites. Structural analyses of the resulting protein/DNA complexes have turned the notion of specific gene regulation, first formulated on the basis of genetic experiments by Jacob and Monod, into concrete, three-dimensional molecular pictures. The lecture will review what we have learned about DNA recognition in transcriptional control, beginning with the prokaryotic repressors and including a variety of eukaryotic regulatory proteins.

At the level of amino acid side chains and DNA base pairs, there is no fixed recognition "code," but there are recurrent kinds of interactions. Generally several amino acid residues contact a critical base pair. These residues are presented to DNA bases, often in the major groove, by a folded structure anchored to the sugar-phosphate backbone by positioning contacts from other residues. Proteins can impose strikingly altered conformations on bound DNA, and DNA conformational propensities therefore influence specificity. Tightly bound water molecules can act as specific adapters in bridging protein and DNA functional groups.

At the level of intact proteins and complete DNA binding sites, a relatively modest number of structural classes accounts for a large fraction of the known DNA recognition domains. There are a number of mechanisms for extending recognition over 10 to 20 base pairs. Oligomerization, including combinatorial heterodimerization, and concatenation of modules within a polypeptide chain are examples. The nature of a protein dimer interface can determine the relative orientation of two half-sites, their spacing, and even some aspects of the intervening nucleotide sequence.

At the level of multiple transcription factors and complex regulatory elements, the proteins bound at nearby sites on DNA can modify each other's specificity and affinity. The interactions often involve the very domains that also bear DNA recognition elements.

Bacteriorhodopsin and the related Halobacterial membrane proteins halorhodopsin and sensory rhodopsins I and II are membrane proteins containing a bundle of seven trans-membrane alpha-helices and the covalently bound chromophore, retinal. Light absorption causes the retinal to isomerise from all-trans to 13-cis and this triggers a sequence of conformational changes in the protein. In bacteriorhodopsin, this is coupled to the proton pumping activity of the protein, whereas in the homologous proteins, it is likely that similar conformational changes are coupled to chloride ion pumping (halorhodopsins) or to signal transduction and bacterial motility (sensory rhodopsins). We have been investigating the nature of the conformational changes and the coupling to proton pumping using electron crystallographic methods. The current status of the work will be presented.

The regulation of protein function through selective phosphorylation of certain amino acids is ubiquitous in both eukaryotic and prokaryotic cells. Control by phosphorylation occurs in metabolic pathways, cell growth and cell division, gene replication and gene transcription, membrane transport, muscle contraction, hormonal response, viral oncogene response, and learning and memory. How does addition of a phosphate group lead to either activation or to inhibition of protein function?

Two enzymes are understood in detail. In mammalian glycogen phosphorylase, phosphorylation on a serine residue acts as an allosteric activator. Phosphorylation promotes conformational changes (shifts of up to 50 Å) local to the site of phosphorylation, and these changes are associated with conformational changes elsewhere in the molecule that alter the subunit-subunit contacts of the dimeric molecule, the constellation of groups at the catalytic site, and the surface properties of the protein. In the bacterial enzyme isocitrate dehydrogenase, phosphorylation occurs in the vicinity of the catalytic site. The enzyme is inhibited by an electrostatic steric blocking mechanism in which the bulk and the negative charge on the phosphate prevent the binding of the negatively charged substrate to the catalytic site.

Structural studies on other systems, especially the protein kinases, are revealing other mechanisms. The results from crystal structures of eight protein kinases, three in the active conformation and five in the inactive conformation, have shown how phosphorylation of a residue (most commonly a threonine residue) in the activation loop can result in new interactions that may promote the correct orientation of the loop to facilitate correct domain-domain orientation, the correct orientation of the g-phosphate of ATP to effect phosphoryl transfer, the correct substrate recognition site, and a favorable electrostatic environment for the catalytic aspartate that acts as a base in the catalytic transfer.

The recent results will be reviewed with special emphasis on the work of my laboratory on glycogen phosphorylase and phosphorylase kinase. A unifying feature of the current mechanisms is based on the property conferred by the dianionic phosphate group. A double negative charge is not available to any of the naturally occurring amino acids. It can confer special effects either by electrostatic repulsion or by promoting neutralizing interactions with positively charged groups, especially arginine residues. The ability of cells to regulate their functions depends on this simple physical-chemical property.

In his lecture to the Sorbonne in April 1864, Louis Pasteur provided strong evidence to refute the notion of spontaneous generation (1). Even the simplest forms of life derive from preexisting life, while nonliving matter is incapable of organizing spontaneously to form living organisms. Yet, life itself must have had a beginning. Since the time of Darwin and Wallace, it has been recognized that biological organization does originate spontaneously, not in terms of the individual, but through successive populations of individuals undergoing Darwinian evolution based on natural selection. The principles of Darwinian evolution also apply to populations of molecules, which can organize themselves to exhibit complex chemical behaviors. Such events are thought to have led to the origin of the "RNA world," a time during the early history of life on earth when RNA served as both the carrier of genetic information and the agent of catalytic function. RNA-based life eventually gave rise to DNA and protein-based life, just as molecular life gave rise to microbial life.

Our laboratory has developed methods for in vitro Darwinian evolution of a large, heterogeneous population of an RNA or DNA molecules. We are able to obtain particular molecules that have desired biochemical properties, including the ability to catalyze a target chemical reaction. In one study, we began with 10¹³ variants of the Tetrahymena group I ribozyme, which catalyzes the sequence-specific cleavage of RNA. Through 63 successive "generations" of test-tube evolution, we produced ribozymes that have the ability to cleave DNA (kcat=0.2 min^-1, Km=20nM) (2). During the first 27 generations, we adjusted the reaction time and substrate concentration to direct improvement of catalytic rate and substrate binding affinity, respectively. Beginning with generation 28, we introduced RNA as a competitive inhibitor, resulting in enhanced specificity for DNA relative to RNA substrates. Ribozymes obtained at the 27th generation (but not the wild-type or those obtained at the 63rd generation) catalyze the hydrolysis of an unactivated, alkyl amide, presented in the context of either an oligonucleotide or an oligonucleotide-peptide chimera (3).

In another study, we sought to develop catalytic DNAs that promote the metal-dependent cleavage of an RNA phosphoester. In this case, without benefit of existing DNA enzymes, we were forced to begin with a population of 10¹⁴ random-sequence DNAs. After several rounds of selective amplification based on activity in the presence of either 1 mM Pb²⁺ or 1 mM Mg²⁺, the population as a whole attained the ability to cleave the target phosphoester (4). Guided by common features of representative individuals isolated from either the Pb²⁺- or Mg²⁺-dependent lineages, we constructed simplified versions of the catalytic domains that enable cleavage in an intermolecular format with turnover. The Pb²⁺- dependent motif is a 30mer DNA with a catalytic rate of 1 min^-1, while the Mg²⁺-dependent motif is a 36mer DNA with a catalytic rate of 0.04 min^-1. This corresponds, in both cases, to a rate acceleration of about 10³-fold compared to the uncatalyzed reaction.

The coiled-coil motif consists of two, three or four helices wrapped about one another with a left-handed superhelical twist (1; and references therein). A "spring-loaded" model, in which a loop region of the influenza hemagglutinin protein forms a three-stranded coiled coil, was proposed for the conformational change that allows the protein to become membrane-fusion active (2). The salient feature of this model was confirmed by X-ray crystallography studies of the low-pH form of hemagglutinin (3). Recent studies of the membrane-fusion protein from the Moloney murine leukemia virus, that suggest a similarity between the retrovirus and orthomyxovirus proteins, will be presented.

All coiled coils that have been characterized in structural detail have a left-handed superhelical twist. Recent progress in the design of coiled coils with a right-handed superhelical twist will be described. An essential component of these protein-design efforts is to consider sidechain packing explicitly, using a procedure for repacking the cores of coiled coils with backbone freedom (4) that takes advantage of an algebraic parameterization for coiled coils proposed by Crick (5).

Genetically encoded libraries of peptides and oligonucleotides are unsurpassed in their ability to lead to a rapid identification of ligands for a variety of macromolecules. The major drawback of these techniques is that the resultant ligands are subject to degradation by naturally occurring enzymes, limiting their use as pharmaceutical leads and as tools in biological research. A method will be illustrated that uses the advantages of a biologically encoded library for the isolation of (D)-amino acid peptide ligands, which should be insensitive to degradation by natural proteases.

A quiet revolution that may rival the astonishing developments of genetic chemistry and genetic engineering in its impact on the biological and medical sciences is the extraordinary coalescence of these sciences into a single discipline. This coalescence is coming about because diverse disciplines -- from anatomy to genetics, from botany to biochemistry -- are becoming expressible in a single, universal language, the language of chemistry.

Much of life can be understood in rational terms if expressed in the language of chemistry. It is an international language without dialects, a language for all of time that explains where we came from, what we are, and where the physical world will allow us to go. Chemical language has great esthetic beauty and links the physical sciences to the biological sciences.

A major obstacle in strengthening the linkage of chemistry to biology is a cultural gulf that has long separated these disciplines. In this regard, Louis Pasteur stands out in history and can teach us much by his example. Trained and distinguished as a chemist, he then felt driven to pursue fundamental biologic questions, such as spontaneous generation, and practical issues such as diseases of wine and people and vaccination. In his brilliant solutions to these questions also founded the disciplines of microbiology and immunology. Yet, in his preoccupation with biologic and clinical issues, he strayed from his chemical heritage and neglected the chemical basis of the phenomena he had done so much to clarify.

Today in science, we are faced with a gap far more serious than that between chemistry and biology. It is the lack of economic support needed to attack the major problems in biology. In science, as with microbial cultures, we have left the growth phase and are now in the stationary phase. Because resources will be severely limited, we will need Pasteurian skills to convince the citizenry to support the basic research that, as in the past, will be the most cost-effective investment for the health and welfare of our society.

Supramolecular chemistry has relied on more or less preorganized molecular receptors for effecting molecular recognition, catalysis, and transport processes. A step beyond consists in the design of systems undergoing molecular self-organization, i.e., systems capable of spontaneously generating a well-defined supramolecular architecture by self-assembling from their components in a given set of conditions.

The molecular information necessary for the process to take place must be stored in the components and acts through selective molecular interactions. Thus, these programmed supramolecular systems operate via molecular recognition.

Self-assembly processes may lead to the formation of chiral supramolecular architects from either achiral or chiral components. Three cases will be discussed:

1) the generation of chiral supermolecules from achiral components;

2) the formation of helical liquid crystalline polymers from complementary units derived from tartaric acids, amounting to the macroscopic expression of molecular asymmetry in supramolecular helicity;

3) the assembly of double-stranded and triple-stranded helicates, helical metal complexes generated from ligand strands and specific metal ions; homoduplex, heteroduplex, and circular duplex arrangements will be described.

The achiral world of the title is the world of racemases, which use chiral amino acids to construct a chiral active site that is nonetheless capable of interconverting both enantiomers of chiral substrates. The crystal structures of mandelate racemase from Pseudomonas putida and alanine racemase from Bacillus stearothermophilus show a common strategy for solving this catalytic problem: both enzymes construct a "functionally symmetric" reactive subsite. The catalytic mechanism in each case appears to be a two-base mechanism. Although the bases are nonidentical, their pKa values are adjusted by environmental perturbation to be close to each other and to that of the carbon acid substrate. Binding of either enantiomer in the correct orientation for catalysis is achieved by anchoring a nonreactive portion of the substrate to a cofactor, magnesium in the case of mandelate racemase and pyridoxal phosphate in the case of alanine racemase. The cofactor also serves to stabilize a carbanion intermediate in both reaction pathways. Mandelate racemase further employs general acid catalysis to reduce the carbon acid pKa, while alanine racemase uses electron delocalization to direct the PLP chemistry along the pathway of proton transfer. Site-directed mutagenesis and direct crystallographic observation of the enzyme-substrate complex are being used to confirm these mechanistic conclusions.

Molecular assemblies provide one means of evaluating the intermolecular forces and the chemical information involved in molecular recognition. Multiple copies of a molecule often give rise to ordered superstructures with functions that are unique to their assembled states. Nature provides many examples of assembly of temporary cavities for the reversible encapsulation and transport of smaller guest species. Synthetic structures of low molecular weight can also show these properties and permit exploration of the effects of size, shape and solvation on the encapsulation of smaller molecules. We will discuss these issues in the context of complementary and self-complementary small molecules.

The human rhinoviruses are the leading cause of the ubiquitous, mild, and self-limiting infections generally referred to as the common cold. Considerable research effort has been expended in the search for well-tolerated antiviral agents capable of preventing and treating the common cold. Although no antirhinovirus drug is yet commercially available, considerable progress has been made in the discovery and development of novel, viral-specific inhibitors of rhinovirus replication that inhibit early steps in the virus life cycle: attachment to the cellular receptor and uncoating of the viral RNA. Molecules directed at these targets currently possess the greatest potential for generating a safe and efficacious treatment for the rhinovirus common cold.

Among the problems encountered in the development of antiviral drugs is the rapid selection for drug-resistant mutants. Thus, the analysis of the mechanism of resistance is of prime interest. Human rhinovirus 14 has been grown in the presence of the antiviral compounds WIN 52035 and WIN 52084. The selected resistant mutants are either displacement or two types of compensation single-site mutations. The former are changes of a small residue to a large residue within the drug binding pocket, thus inhibiting drug binding. The compensation mutants either decrease the binding affinity of the drug to the virus or increase the binding affinity of the cellular receptor to the virus. As the binding sites of the drugs and of the receptor are overlapping, the compensation mutants interfere with the delicate equilibrium between a natural pocket factor displaced by the drugs and binding of receptor.

Insights into signaling pathways and other cellular processes have resulted from studies of cell permeable, organic molecules identified from natural sources and designed and synthesized in the laboratory. This lecture will present results of studies using such molecules to understand and control intracellular signaling pathways. These low-molecular-weight ligands cause either a conditional loss of function following binding to the products of wild-type alleles or a gain of function following binding to the products of rationally designed conditional alleles. Examples are seen in studies of immunophilin-natural product complexes that led to the identification of calcineurin as a mediator of T cell receptor signaling and of FRAP as a mediator of signaling that links mitogenic pathways to the cell cycle machinery. A family of cell permeable ligands that induce intracellular proteins to associate, developed in collaboration with Gerald Crabtree, has been used to regulate transcription and signal transduction (including pathways emanating from the T cell receptor and the apoptosis-inducing Fas antigen), and other cellular processes such as intracellular protein degradation and translocation.

The humoral immune system is perhaps the best single example in which large diverse libraries of molecules are generated and screened for a specific biological function. One of the first examples in which the chemical potential of large libraries of this sort was exploited was the use of transition state theory to select catalysts from among the large population of antibody molecules. Since these early experiments which involved relatively simple transformations, the reactions catalyzed by antibodies have increased in complexity and degree of difficulty. At the same time the strategies for generating antibody catalysts have become increasingly sophisticated. Recent structural and mechanistic studies are showing that the chemical notions used to generate catalytic antibodies are indeed reflected in their active site structures. Today, there is interest in screening libraries not only of antibodies but also Fab fragments, RNAs, peptides, synthetic organic molecules, and even solid state materials, for interesting new properties and functions. This lecture will survey recent advances in the field and lessons that have been learned from the study of antibody catalysis.

The nucleolus of all eucaryotic cells is inhabited by multiple small nucleolar RNPs (snoRNPs). Some of these are transcribed (by RNA polymerase II or III) from their own independent transcription units, while many more are encoded within the introns of protein-coding genes. U3, U8, U13, and MRP belong to the first class, while U14-U30 belong to the latter. Many, but not all, snoRNAs bind the autoantigen fibrillarin.

Recently, considerable progress has been made in assigning functions to snoRNPs in vertebrate cells. U3, the most abundant snoRNP, acts at an early processing site near the 5' end of the pre-rRNA transcript and also influences that maturation of 18S rRNA. By exploiting the Xenopus oocyte, where a small RNA can be targeted for destruction by injection of a complementary oligonucleotide, we have recently defined roles for U8, U22, and U14. U22 and U14 are required for maturation of 18S rRNA at both ends, whereas U8 is essential for 5.8S and 28S rRNA processing. Thus, direct parallels can be drawn between the mode of rRNA processing in bacteria and eucaryotes.

In vitro studies suggest that most intron-encoded snoRNAs arise by exonucleolytic trimming after excision (and debranching) of their host introns. However, many questions regarding this novel class of small RNAs remain. Why are all intron-encoded RNAs destined for the nucleolus? Why have only some snoRNAs been selected to be intron-encoded? What is the mobility mechanism whereby these snoRNAs "move" from intron to intron among highly expressed genes of vertebrate organisms? What is the function of the abundant, polyadenylated, but apparently noncoding, transcript of the host gene for human U22? Does the extensive complementarity exhibited by many intron-encoded snoRNAs to conserved regions of 18S and 28S rRNAs mean that they function as chaperones in ribosome assembly? The state of progress on each of these issues will be discussed.

The mode of binding of antigenic peptides to the class I and class II major histocompatibility (MHC) glycoproteins has been determined in atomic detail by x-ray crystallography as well as the interaction between MHC class II molecules with bacterial protein superantigens. Structural data on an intermediate in the in vivo assembly of MHC class II-peptide complexes and the MHC Invariant chain protein argue that an unstructured domain of the Invariant chain protein may bind to MHC class II molecules, stabilizing its structure like a chaperon, until the assembly reaches the endosomal compartment containing peptides derived from extracellular antigens. These observations will be reviewed in the context of the intracellular transport involved in antigen presentation.

For a quarter century ray diffraction in single crystals was unique in its ability to solve three-dimensional structures of proteins and nucleic acids at atomic resolution. The situation changed in 1984 with the completion of a protein structure determination by nuclear magnetic resonance (NMR) spectroscopy in solution (1), and today NMR is a second, widely used method for biomacromolecular structure determination (2). NMR is most powerful for studies of relatively small systems with molecular weights up to about 30,000, but these structures can be obtained in near-physiological solution. Diffraction techniques and NMR have widely different time scales, which afford different insights into internal molecular mobility as well as different views of protein or nucleic acid molecular surfaces and hydration. In addition to the information on the average three-dimensional structure, NMR provides information on a wide array of short-lived transient conformational states, and of the transition frequencies between these states. Combining information from crystallography and NMR can therefore yield more detailed insights into the structural basis of protein and nucleic acid functions, and thus provide a more reliable platform for rational drug design and the engineering of novel protein functions.