The Beilstein workshops address contemporary issues in the chemical and related sciences by employing an interdisciplinary approach. Scientists from a wide range of areas – often outside chemistry – are invited to present aspects of their work for discussion with the aim of not only to advance science, but also, to enhance interdisciplinary communication.

Molecular interactions are of central importance to chemistry and biology; they control molecular events and states. The macrostructure and function of biomolecular compounds – proteins, nucleic acids, carbohydrates and lipids – are governed by molecular interactions, as are synthetic chemistry and catalysis of chemical reactions.

Understanding the evolution of biopolymers is required to rationalise the directed and undirected design of functional molecules. Large scale experiments or detailed computational studies are often impractical. Therefore, simple model systems, such as RNA secondary structure and lattice proteins have to be adapted to study general statistical and topological features of genotype (sequence) to phenotype (structure) maps.

Cellular processes require the interaction of many biomacromolecules such as proteins, RNA, carbohydrates etc. within and across several cellular compartments. Determining the collective network of such interactions is an important aspect of understanding the role and regulation of the individual members of such interacting networks.

Molecular Interactions bring chemistry to life in living organisms, but chemistry is a science that scientists can bring to life. Complex syntheses of natural products, elegantly controlled chemical reactions, the understanding of how proteins fold or DNA replicates, design of new pharmaceuticals or the docking of ligands in targets are all good examples of this. Central to all this work is not only the structure of the molecules in question but also a well founded understanding of how these molecules interact.

The rapid progress in structural and molecular biology over the past fifteen years has allowed chemists to access the structures of enzymes, of their complexes and of mutants. This wealth of structural information has led to a surge in the interest in enzymes as elegant chemical catalysts in such a way that enzymology became to be a distinguished field with important contributions to medicine and basic science.

We would like to thank particularly the authors who provided us with written versions of the papers that they presented. Special thanks go to all those involved with the preparation and organization of the workshop, to the chairmen who piloted us successfully through the sessions and to the speakers and participants for their contribution in making this workshop a success

Frankfurt/Main, July 2007

Martin G. Hicks
Carsten Kettner

In the past twenty years structural biology has come to play a major role in lead optimization and target identification in the process of drug discovery. Only recently, however, has the development of highthroughput methods of structure determination provided a powerful approach to the screening of fragment binding so that structural biology can now contribute directly to lead discovery. Most targets for the new approaches continue to be enzymes, channels or receptors which tend to be “tractable” with deep and well defined cavities that bind a range of ligands. Here we discuss the challenges and opportunities in moving fragment-based approaches to target less-tractable multiprotein systems.

Understanding the effects that non-synonymous single nucleotide polymorphisms have on the structures of the gene products, the proteins, is important in identifying the origins of complex diseases. A method based on amino acid substitutions observed within homologous protein families with known 3D structures was used to predict changes in stability caused by mutations. In the task of predicting only the sign of stability change, our method performs comparably or better to other published methods with an accuracy of 71%. The method was applied to a set of disease associated and non-disease associated mutations and was shown to distinguish the two sets in terms of protein stability. Our method may therefore have application in correlating SNPs with diseases caused by protein instability.

The definition of life has excited little interest among molecular biologists during the past half-century, and the enormous development in biology during that time has been largely based on an analytical approach in which all biological entities are studied in terms of their components, the process being extended to greater and greater detail without limit. The benefits of this reductionism are so obvious that they need no discussion, but there have been costs as well, and future advances, for example for creating artificial life or for taking biotechnology beyond the level of tinkering, will need more serious attention to be given to the question of what makes a living organism alive. According to Robert Rosen's theory of (M,R)-systems (metabolismreplacement systems), the central idea missing from molecular biology is that of metabolic circularity, most evident from the obvious but commonly ignored fact that proteins are not given from outside but are products of metabolism, and thus metabolites. Life can be embodied in a mathematical formalism that treats metabolism as a function able to act on an instance of itself to produce a new instance of itself.

Chemical genomics aims to systematically explore the interactions between small molecules and biological systems at different levels of organization ranging from individual macromolecules to whole organisms. By analogy to the progression of creating genetic maps over the past century, which now provide nucleotide-level resolution of entire genomes, chemical genomics allows the annotation of 'chemomes', the full set of biologically relevant chemicals capable of interaction with a particular biological system. This article aims to discuss recent progress made toward the goal of mapping multidimensional chemical and biological descriptor spaces. The focus is on the complementary nature of these efforts and the importance of recognizing the distinction between computed versus observed descriptors. Recent examples of identifying small molecules the molecular interactions of which give rise to novel phenotypes relevant to human disease and our understanding of complex biological pathways will be described. To further advance the field, information being derived from computational studies of molecular structure and observational studies of molecular function must be integrated into global models of biological activity that are both explanatory and predictive. Unlike the complementarity principle in physics, which describes the impossibility of simultaneously observing both the wave and particle nature of light and electrons, it is possible to simultaneously observe and model chemical and biological space. In doing so a fuller description of the interaction of small molecules with biological systems arises than if either of the two spaces is considered separately.

Recent developments in molecular biology offer new approaches for improving our understanding of enzyme-ligand interactions. The complexity of enzyme catalysis, consisting of ligand recognition and exquisite discrimination followed by rapid catalytic turnover, is now being tapped by modifying ligand specificity. While various approaches to modify specificity have been developed, we apply a 'semi-rational design' approach, whereby residues in proximity to the bound ligand are mutated. Mutations are either random (20 possible amino acids) or semi-random (a subset of amino acids is encoded), and several positions are mutated simultaneously to allow the occurrence of complementary, or compensatory, mutations.

By conducting combinatorial mutagenesis specifically directed toward the active-site area of enzymes involved in drug resistances, we are gaining insights into the nature of the enzyme-ligand interactions underlying these resistance mechanisms. We screen libraries of mutated enzymes for resistance toward their target drug(s), selecting a variety of modified active-site environments that we characterize for binding and reactivity. Here, we present modification of the specificity of a β-lactamase and of a dihydrofolate reductase. Insights into the molecular nature of these modified enzyme-ligand interactions will provide new information for design of more advanced drug generations. Our work contributes to demonstrate the generality of the combinatorial active-site mutation strategy for modifying enzyme specificity.

DNA-directed chemical ligations enable the highly sequence specific analysis of mutations in DNA. The full diagnostic potential of DNAdirected chemistry can be harvested when DNA-analogues that provide new opportunities such as improved affinity and selectivity in DNA-binding and/or ease and accuracy of detection are employed. It is shown that peptide nucleic acid (PNA) conjugates, non-ionic biostable DNA analogues, can be ligated by using native chemical ligation. This reaction proceeds as rapidly and more selective than T4-ligase mediated oligonucleotide ligations. The selectivity is higher than 3000-fold in discriminating matched from single mismatched DNA. This high selectivity is the result of a particular ligation architecture which involves an unpaired DNA-base opposite to the ligation site. It is suggested that the high sequence specificity of this so-called abasic ligation architecture facilitates the analysis of early cancer onset. As an example it is shown that as little as 0.2% of single-base mutant in presence of 99.8% wild-type DNA can be detected by massspectrometric analysis of the PNA-native chemical ligation. The PNAligation chemistry can also be applied to double stranded DNA-templates produced by PCR. In this case, auxiliary PNA needs to be added in order to help binding of the rather short PNA ligation probes to DNA. One drawback of using chemical methods for ligation of oligonucleotides and analogues is product inhibition. Usually the products of ligation bind to the template with higher affinity than the probes before ligation. This prevents catalytic turnover, which, however, is desired if targets are present at low concentration. An approach to overcome product inhibition in PNA ligation is presented. The approach is based on native chemical reactions involving isocysteine rather than cysteine. Ligations at isocysteine succeed through the intermediacy of a ligated thioester that rearranges via chain-extension. This step increases the flexibility of the final product, thereby reducing its affinity to the DNA template. To facilitate real-time monitoring of product formation a fluorescence resonance energy transfer (FRET)-based detection method was established. By using the FRET technique single base mutations can be detected within minutes and with perfect sequence selectivity at optimized conditions.

Noncanonical amino acids as building blocks for the biosynthesis of tailor-made proteins represent a nearly infinite supply for the introduction of unusual functions, molecular scaffolds exerting conformational constraints or pharmacologically active entities into proteins. Exploitation of this supply for biotechnological or medical application is owed to the flexibility of the cellular systems involved in the incorporation of noncanonical amino acids into proteins. The broad substrate specificity of cellular amino acid transport systems allows for transmembrane passage of many noncanonical amino acid analogues. Subsequently, the intracellular amount of noncanonical amino acids can be tuned to levels high enough for efficient activation and tRNA charging by aminoacyl-tRNA synthetases (AARS). As a variety of noncanonical amino acids with different chemical properties are incorporated into polypeptide sequences, they are obviously metabolically stable. The indiscriminateness of the AARS towards many noncanonical substrates, i. e. their “catalytic promiscuity”, is the central principle for expanding the scope of ribosomal protein synthesis. tRNAs charged with noncanonical amino acids mediate their efficient translation into nascent polypeptide chains by codon reassignment owing to the adaptability of the ribosome. Our contribution specifically highlights all theses principles as sine qua non for protein translation reprogramming with an expanded genetic code.

Using beta-hairpin model systems, we have investigated the role of aromatic interactions in providing selectivity to protein folding and protein-protein interactions. In particular, we have explored the role of cation-π and amide-π interactions and their role in controlling proteinprotein interactions with post-translationally modified proteins. Specific modifications that have been studied are Lysine methylation and acylation and Arginine methylation. The molecular recognition of these modified amino acids has particular relevance to understanding the “histone code”, which has been proposed to control chromatin structure and gene expression.

The use and characteristics of local properties designed to describe intermolecular interactions projected onto molecular surfaces and based on semiempirical molecular orbital theory are described. After a discussion of the local properties themselves and their relationship to intermolecular interactions and chemical reactivity, two applications are described. The first, surface-integral models for physical properties, involve integrating a functional of the local properties over the molecular surface. In the second example, we discuss a possible approach to determining the potential specificity of biological interactions based on Shannon's theory of communication.

Stereochemistry is information, and stereoselective reactions are the means by which that information may be communicated within and between molecules. The control of remote stereogenic centres can be achieved by stereochemical relay, and the use of thermodynamic control over conformational preference is turning out to be a very powerful method for long-range transmission of stereochemical information.

The design of focused compound libraries aims at enriching bioactive molecules that contain different scaffold structures. Pharmacophorebased similarity searching has been shown to provide a means to achieve this goal. We have developed such a method (LIQUID) that is grounded on the representation of potential pharmacophore points by trivariate Gaussian densities. This “fuzzy” pharmacophore technique is described and discussed in detail, together with a retrospective virtual screening application. LIQUID succeeded in retrieving activity-enriched sets of compound with diverse backbone architecture.

Particle-based computer simulation is a powerful tool to study the behaviour of membranes at molecular resolution. Atomic-level models have been employed for decades now, and have given an understanding of many membrane phenomena. However, these studies are computationally very expensive, for an enormous amount of calculation has to be performed to model the interactions between all atoms in the system. This problem can be tackled by adopting simplified, “coarse-grain” descriptions, in which the number of interacting particles is significantly reduced. In this review, we summarize and discuss the most representative work reported in the literature concerning coarse-grain computer models of lipid bilayers. Every model is analysed in terms of the force-field employed, parameterization procedure, and predictive power in relation to the corresponding experimental observables. We also highlight general advantages and drawbacks of the coarse-grain approach with respect to the traditional atomic-level methodology.

The present impact of supramolecular chemistry in biology is not as large as it could be. The affinities of most water-soluble supramolecular receptors are many orders of magnitude lower than those of their biological counterparts, preventing their application in biological systems. We believe that the superiority of biological hosts is due to noncovalent interactions within their structures that enhance binding affinity. We have recently discovered that the synthetic receptors for anions that we have developed exhibit enhanced affinities as a result of similar intra-receptor interactions. This effect has as yet unexplored potential as a tool for pushing binding affinities of synthetic receptors into the desirable nanomolar affinity range. Another area where an expanding role of supramolecular chemistry is expected is that of complex systems. We have investigated the behaviour of dynamic combinatorial libraries of hosts in response to the introduction of guest molecules. These investigations have improved our understanding of thermodynamically controlled molecular networks, which is relevant for the use of dynamic combinatorial libraries as a method to discover new receptors, but also provides a new entry into the emerging field of systems chemistry.

Gene therapy for HIV-1 infection has been proposed as an alternative to antiretroviral drug regimens due to emerging drug resistance and toxicity that raises concerns about HAART as a long-term therapy. VIRxSYS has developed an HIV-based lentiviral vector platform for delivery of genetic therapies. For the first clinical application of the gene delivery technology, VIRxSYS created VRX496, a lentiviral vector expressing a 937-base long antisense against the HIV envelope gene. Along with the lentivector, a packaging vector, VIRPAC, was created based on a single plasmid approach for transient production. Many safety features were incorporated into VRX496 and VIRPAC. VRX496 pre-clinical efficacy was demonstrated in vitro by achieving high transduction efficiencies with stable gene transfer into human primary CD4+ T lymphocytes and by showing selective resistance to CD4 down regulation with over 4 logs (99.99%) of HIV replication inhibition in challenge assays using various X4, R5 or dual tropism strains of HIV. In December 2002, the US Food and Drug Administration approved the first ever Phase I clinical trial of lentiviral vectors in humans, testing the safety and tolerability of a single infusion of autologous HIV infected CD4+ T Cells transduced with VRX496. No adverse events due to the product were observed. Although the purpose of the Phase I clinical trial was to establish the safety of the therapy, and the number of patients in the Phase I clinical trial is too small to make any conclusions with respect to efficacy, potential effects of VRX496 were observed in patients during the monitoring of their circulating CD4 counts and HIV viral load.

We analyse the evolution of haemagglutinin from human influenza H3 using a model that allows for variations in selective pressure, both at different locations in the protein as well as during the course of evolution. It has been observed that, in contrast to the steady rate of sequence change, the antigenic properties of the haemagglutinin changes in a punctuated manner between well-defined clusters. We find that the changes in antigenic properties correspond to increased rate of change in selective pressure, as if these antigenic clusters correspond to different interactions between the virus and the immune system. Conversely, despite a large increase in glycosylation during the past 40 years, these changes in glycosylation do not generally seem to be correlated either with changes in antigenic properties or with significantly more rapid changes in selective pressure.

Chemical glycomics uses synthetic chemistry to procure defined carbohydrate molecules to study the glycans involved in many functions in the living cell. Based on an automated synthesis platform, a host of synthetic tools including carbohydrate microarrays has been developed. These tools have been employed to dissect carbohydrate-copolymer interactions. Basic research in the glycomics arena is beginning to impact on drug discovery, especially the development of carbohydrate- based vaccines. The development of vaccine candidates to protect from malaria and leishmaniasis infections is discussed.

Molecular interactions, a notion which is familiar to any chemist, many a physicist, and most biologists, probably since the early days of Emil Fischer, has been discussed enthusiastically ever since [1]. But is it worth having a workshop on this topic still today and does it really create any new momentum “bringing it to life”? The answer, being far from simple, I would like to split it into two parts. Firstly, I would like to reflect on the role of molecular interactions in the contributions to the workshop, and secondly, I would like to comment on the efforts taken in the context of the workshop, to bring chemistry to life. The order in which the individual contributions are mentioned is not the order of the workshop programme, but follows my personal view of the topics.