Volume 382, Issue 9

After finishing the sequence of the human genome, a functional understanding of genome dynamics is the next major step on the agenda of the biosciences. New approaches, such as microarray techniques, and new methods of bioinformatics provide powerful tools aiming in this direction. In the last few years, important parts of genome organization and dynamics in a number of model organisms have been determined. However, an integrated view of gene regulation on a genomic scale is still lacking. Here, genome function is discussed from a complex dynamical systems perspective: which dynamical properties can a large genomic system exhibit in principle, given the local mechanisms governing the small subsystems that we know today? Models of artificial genetic networks are used to explore dynamical principles and possible emergent dynamical phenomena in networks of genetic switches. One observes evolution of robustness and dynamical selforganization in large networks of artificial regulators that are based on the dynamic mechanism of transcriptional regulators as observed in biological gene regulation. Possible biological observables and ways of experimental testing of global phenomena in genome function and dynamics are discussed. Models of artificial genetic networks provide a tool to address questions in genome dynamics and their evolution and allow simulation studies in evolutionary genomics.

Theoretical concepts and experiments dealing with the evolution of molecules in vitro reached a state that allows for direct applications to the design of biomolecules with predefined properties. RNA evolution in vitro represents a basis for the development of a new and comprehensive model of evolution, focusing on the phenotype and its fitness relevant properties. Relations between genotypes and phenotypes are described by mappings from genotype space onto a space of phenotypes, which are manytoone and thus give ample room for neutrality as expressed by the existence of extended neutral networks in genotype space. The RNA model reduces genotypephenotype relations to mappings from sequences into secondary structures of minimal free energies and allows for derivation of otherwise inaccessible quantitative results. Continuity and discontinuity in evolution are defined through a new notion of accessibility in phenotype space that provides a basis for straightforward interpretation of computer simulations on RNA optimization; furthermore, it reveals the constructive role of random genomic drift in the search for phenotypes of higher fitness. The effects of population size on the course of evolutionary optimization can be predicted quantitatively by means of a simple stochastic model based on a birthanddeath process with immigration.

The (?[alpha)8-barrel is the most versatile and most frequently encountered fold among enzymes. It is an interesting question how the contemporary (?[alpha)8-barrels are evolutionarily related and by which mechanisms they evolved from more simple precursors. Comprehensive comparisons of amino acid sequences and threedimensional structures suggest that a large fraction of the known (?[alpha)8-barrels have divergently evolved from a common ancestor. The mutational interconversion of enzymatic activities of several (?[alpha)8-barrels further supports their common evolutionary origin. Moreover, the high structural similarity between the N and Cterminal (?[alpha)4 units of two (?[alpha)8-barrel enzymes from histidine biosynthesis indicates that the contemporary proteins evolved by tandem duplication and fusion of the gene of an ancestral halfbarrel precursor. In support of this hypothesis, recombinantly produced halfbarrels were shown to be folded, dimeric proteins.

RNA molecules with catalytic properties have been isolated by in vitro selection from combinatorial libraries. A broad range of chemical reactions can be catalyzed, and nucleic acids can accelerate bond formation between small organic substrates. The catalytic performance of nucleic acids can be enhanced by incorporation of additional functional groups. This minireview focuses on carboncarbon bond formation accelerated by in vitro selected ribozymes.

Methods for automation of nucleic acid selections are being developed. The selection of aptamers has been successfully automated using a Biomek 2000 workstation. Several binding species with nanomolar affinities were isolated from diverse populations. Automation of a deoxyribozyme ligase selection is in progress. The process requires eleven times more robotic manipulations than an aptamer selection. The random sequence pool contained a 5 iodine residue and the ligation substrate contained a 3 phosphorothioate. Initially, a manual deoxyribozyme ligase selection was performed. Thirteen rounds of selection yielded ligators with a 400-fold increase in activity over the initial pool. Several difficulties were encountered during the automation of DNA catalyst selection, including effectively washing beadbound DNA, pipetting 50% glycerol solutions, purifying single strand DNA, and monitoring the progress of the selection as it is performed. Nonetheless, automated selection experiments for deoxyribozyme ligases were carried out starting from either a naive pool or round eight of the manually selected pool. In both instances, the first round of selection revealed an increase in ligase activity. However, this activity was lost in subsequent rounds. A possible cause could be mispriming during the unmonitored PCR reactions. Potential solutions include pool redesign, fewer PCR cycles, and integration of a fluorescence microtiter plate reader to allow robotic observation of the selections as they progress.

Anticalins comprise a novel class of receptor proteins with predetermined ligand specificities which were engineered using the lipocalin fold. Attractive features of these artificial ligandbinding proteins include their small size and monomeric nature, being composed of a single polypeptide chain. Here we report the construction of a functional fusion protein from two independent anticalins, a socalled duocalin. The gene for the fusion protein was assembled from nucleotide sequences encoding an anticalin with fluorescein specificity on the one hand and an anticalin with digoxigenin specificity on the other. Both engineered lipocalins were previously selected from a random library prepared on the basis of the bilinbinding protein, a natural lipocalin abundant in insects. The corresponding fusion protein was expressed in a secretable form in E. coli cells and isolated from the periplasmic fraction using the Streptag method. The major fraction of the purified protein appeared to possess the proper pattern of altogether four disulphide bonds. The ligandbinding behaviour of the fusion protein was investigated both by solid phase ELISA and in fluorescence titration experiments. Our results demonstrate that the novel fusion protein has retained both ligand specificities. Up to now, dimerized ligandbinding proteins were mostly derived from recombinant antibody fragments. Compared with those constructs the duocalins, either with bispecific or with bivalent target recognition properties, should provide useful reagents for various purposes in biotechnology.

A fully stochastic chemical modelling technique is derived which describes the influence of spatial separation and discrete population size on the evolutionary stability of coupled amplification in biopolymers. The model is analytically tractable for an dimensional space (simplex geometry), which also provides insight into evolution in normal Euclidean space. The results are compared with stochastic simulations describing the coevolution of combinatorial families of molecular sequences both in the simplex geometry and in lower (one, two and three) space dimensions. They demonstrate analytically the generic limits which exploitation place on coevolving multicomponent amplification systems. In particular, there is an optimal diffusion (or migration) coefficient for cooperative amplification and minimal and maximal threshold values for stable cooperation. Over a bounded range of diffusion rates, the model also exhibits stable limit cycles. Furthermore, the cooperatively coupled system has a maximum tolerable error rate at intermediate rates of diffusion. A tractable model is thereby established which demonstrates that spatial effects can stabilize catalytic biological information. The analytic behaviour in dimensional simplex space is seen to provide a reasonable guide to the spatial dependence of the error threshold in physical space. Nanoscale possibilities for the evolution of catalysis on the basis of the model are outlined. We denote the modelling technique by PRESS, Probability Reduced Evolution of Spatiallydiscrete Species.

A new computational approach for the efficient docking of flexible ligands in a rigid protein is presented. It exploits the binding modes of functional groups determined by an exhaustive search with solvation. The search in ligand conformational space is performed by a genetic algorithm whose scoring function approximates steric effects and intermolecular hydrogen bonds. Ligand conformations generated by the genetic algorithm are docked in the protein binding site by optimizing the fit of their fragments to optimal positions of chemically related functional groups. We show that the use of optimal binding modes of molecular fragments allows to dock known inhibitors with about ten rotatable bonds in the active site of the uncomplexed and complexed conformations of thrombin and HIV-1 protease.

Adenoviral DNA was examined within capsids by dimethyl sulfate footprinting. ProteinDNA interactions were visualized through ligationmediated PCR (LMPCR). Signals for protein binding were found adjacent to both inverted terminal repeats (ITR). There were no indications of close protein binding at several other loci of the viral genome. Therefore, adenovirus type 5 seems to contain sequence or locusspecific DNA binding proteins within the virion.

Drosophila enhancer of rudimentary [e(r)] interacts genetically with the rudimentary gene, which encodes a protein possessing the first three enzymatic activities of the pyrimidine biosynthesis pathway. A regulatory or enzymatic activity of e(r) in pyrimidine biosynthesis and the cell cycle has been suggested, but nothing is known about its molecular function. The factor is evolutionarily highly conserved since homologues exist in plants and mammals. We cloned the Xenopus enhancer of rudimentary homologue (XERH) as an interaction partner of DCoH/PCD (dimerisation cofactor of HNF1/pterin-4 carbinolamine dehydratase) in the yeast twohybrid assay. DCoH/PCD is a multifunctional factor originally identified as a positive cofactor of the HNF1 homeobox transcription factors. XERH is a 104 amino acid protein that is identical to its mammalian homologues. The mRNA is expressed maternally, enriched in ectodermal derivatives during development and ubiquitously detectable in the adult. Fused to the DNA binding region of the GAL4 transcription factor domain, XERH represses the activity of a GAL4 responsive reporter in HeLa, but not in NIH3T3 cells. Furthermore, the DCoH/PCD coactivation of a HNF1 responsive reporter is inhibited by XERH. We propose that XERH is a cell typespecific transcriptional repressor, probably interfering with HNF1-dependent gene regulation via DCoH/PCD.

The human chemokine receptors CXCR5 and CXCR1 activate signaling pathways via pertussis toxinsensitive as well as insensitive G proteins. CXCR5 induces Ca2+signaling and chemotaxis independently of inhibitory G proteins, whereas the same signaling pathways are entirely dependent on inhibitory G proteins for CXCR1. In contrast, activation of the MAP kinase cascade via ERK1/2 is a pertussis toxinsensitive signaling event for both receptors. Using chimeric CXCR1/CXCR5 receptors we investigated structural requirements for the activation of signal transduction pathways by CXCR5. Individual or multiple intracellular domains of CXCR1 were exchanged for the corresponding sequences of CXCR5, leading to receptors resembling CXCR5 at the cytoplasmic surface to a varying extent. Replacing the second intracellular domain of CXCR1 had a major influence on signaling mediated by inhibitory G proteins, whereas the exchange of the third or carboxyterminal intracellular domain had only minor effects on signal transduction. Activation of the MAP kinase cascade via ERK1/2 and chemotaxis are largely reduced in chimeras comprising the second intracellular domain of CXCR5, although coupling to inibitory G proteins is retained in all chimeric receptors. In summary, these data characterize the contribution of the intracellular domains of CXCR5 to receptor signaling, thereby disclosing unique structural requirements that modulate G protein coupling by the receptor.

RgpB, a cysteine proteinase produced by Porphyromonas gingivalis, exhibits proteolytic activity selectively directed against peptide bonds containing an arginine residue in the P1 position. Here we show that this enzyme can be used for very efficient and specific protein cleavage. RgpB is highly active even at high concentrations of denaturing agents, including urea (up to 6 M) and SDS (0.1%), both of them being commonly used for solubilization of insoluble proteins and peptides. Moreover, RgpB is able to digest polypeptide chains in buffers supplemented with 1% Triton X-100, 1% octyl or decylpyranoside, detergents employed for the enzymatic digestion of proteins transferred onto nitrocellulose membranes. These features render RgpB a suitable tool for use in protein chemistry.

Chemokines help control normal leukocyte trafficking as well as their infiltration into tissues during acute and chronic inflammation. Matrix metalloproteinases (MMPs) help support the extravasation and infiltration of leukocytes through limited proteolysis of basement membranes and matrix material. The effect of the chemokines RANTES/CCL5, MCP-1/CCL and SDF-1/CXCL12 on secretion of the matrix metalloproteinase B and its endogenous inhibitor TIMP-1 was studied. RANTES/CCL5 and SDF-1/CXCL12 were found to induce MMP-9 secretion in primary human monocytes while TIMP-1 secretion was not affected. RANTES/CCL5 effects were mediated through CCR1 because the CCR1 antagonist BX471 was found to effectively block RANTES/CCL5-induced MMP-9 secretion.