4. New Analytical and Diagnostic Approaches 4.1 Bioinformatics and Proteomics S4.1-1 Strategies to cover microbial proteomes P. R. Jungblut Core Facility Protein Analysis, Max Planck Institute for Infection Biology, Berlin, Germany. E-mail: jungblut@mpiib-berlin.mpg.de The aim of our studies is to get access to as much as possible of the protein species present in the proteome under analysis. Already at the low complexity level of proteasomes it becomes clear that proteins are massively modified. There are about 70 spots representing different protein forms from which only some of them were identified at the protein species level. The mycobac- terial proteome of a cell extract of late growth phase state revealed by 2-DE/MS 1800 spots with 379 different proteins iden- tified (10% of the predicted genes). Isotope coded affinity tag (ICAT)-LC/MS resulted in 619 (15% of predicted genes) identi- fied proteins. Because of the overlap of 158 proteins showing the high complementarity of 2-DE/MS and ICAT-LC/MS the prote- ome is covered at present by 21%. For Helicobacter pylori we revealed with a pre-fractionation strategy about 520 different proteins (33% of predicted proteome). One of the proteins deter- mined as vaccine candidate was successful as vaccine in animal models and is now in a clinical study. A proteome 2-DE database was established (http://www.mpiib-berlin.mpg.de/2D-PAGE), which contains data of more than 10 different bacterial species. Proteomics does not end with the identification of a protein. At present identification at the protein species level, which needs 100% sequence coverage is very time-consuming and affords fur- ther technology development. Today, the enormous information content of a 2-DE gel is only accessible to a small extent. S4.1-2 Targeted peptide-centric protemics: a versatile tool for quantitative proteomics J. Vandekerckhove, L. Martens, J. Pinxteren, P. Van Damme, X. Hanoulle, B. Ghesquie ` re, A. Staes, J. Van Damme, E. Timmerman, M. Goethals, H. Demol, K. Hugelier and K. Gevaert Department of Biochemistry and Medical Protein Research, VIB and Ghent University, Ghent, Belgium. E-mail: joel.vandekerckhove@ugent.be We have developed a peptide-centric proteome method, based on the principle of diagonal chromatography. The procedure has been called COmbined FRActional DIagonal Chromatography (COFRADIC) and can sort for any kind of peptide or subset of peptides, provided it can be selectively altered by chemical or enzymatic reaction. We can sort for either methionine or cysteine peptides. In addition, the protein N-terminal peptides can be iso- lated from highly complex mixtures, such as tryptic digests of total cell or tissue lysates. The latter approach holds interesting possibilities: (a) the complexity of a proteome is reduced to a one protein-one peptide ratio, allowing high coverage proteomics of complex samples; (b) the N-termini of proteins each harbor highly characteristic sequences opening the possibility to analyze, two proteomes at the same time such as is the case in xeno-pro- teomics; and (c) protein cleavage or processing yields novel N-termini which can be specifically isolated allowing global ana- lysis of protein degradation events. Examples are shown of: (a) high-coverage proteomics of human multipotent progenitor cells; (b) mouse plasma proteomics after human hepatocyte grafting; (c) global analysis of apoptotic cells; (d) glycoproteomics in human interstitial fluids; and (e) analysis of oxidative stress in CSF. We have further adapted our COFRADIC procedure to identify different types of ATP-binding to proteins in a total cellular con- text. S4.1-3 Learning from evolution to predict protein structure and function A. Tramontano, D. Cozzetto, A. Giorgetti, D. Raimondo Department of Biochemical Sciences, University ‘La Sapienza’, Rome, Italy. E-mail: anna.tramontano@uniroma1.it Since decades, the problem of deciphering the code that relates the amino acid sequence of a protein and its native three-dimen- sional structure has been the subject of innumerable investiga- tions and, in spite of the many frustrations caused by its elusiveness, the interest in the problem is not fading away, on the contrary. What stands in our way, notwithstanding all these efforts, is the complexity of protein structures. In a protein struc- ture thousands of atoms are held together by weak forces and give rise to a conformation which is only marginally stable. It is very unlikely that we can use the laws of physics to compute the native functional structure of a protein in the foreseeable future. However, we have at our disposal the experimentally solved structures of a reasonable number of proteins, a few thousands as of today. Among them, there are many examples of proteins that descend from a common ancestral protein. We know that these proteins are functional, or at least not deleterious, as they have been accepted in the population, therefore they are expected to have a stable native structure. Because of the requirement that the evo- lutionary changes preserve function, we also expect their struc- ture to be similar. This has allowed us to devise methods to construct models of the structure of their evolutionary neigh- bours. I will describe how we can evaluate and estimate the quality of a protein model and analyze how this impacts on its usefulness. Abstracts 24 4.2 Clinical Proteomics S4.2-1 Proteomic strategies to the analysis of breast apocrine carcinomas J. E. Celis, I. Gromova, P. Gromov, J. M. A. Moreira, T. Cabezo ´ n, E. Friis and F. Rank Danish Centre for Translational Breast Cancer Research, Copen- hagen, Denmark. E-mail: jec@cancer.dk Breast apocrine carcinomas represent about 0.5% of all invasive breast cancers, and despite their distinct morphological features, there are at present no standard molecular criteria available for their diagnosis. Recent proteome expression profiling studies of breast apocrine macrocysts, normal breast tissue and breast tumours have identified specific apocrine biomarkers (15-PGDH and HMG-CoA reductase) present in early and advanced apocrine lesions. These biomarkers in combination with proteins found to be characteristically upregulated in pure apocrine carcinomas (psoriasin, S100A9, p53) provide a protein signature distinctive for benign apocrine metaplasias. These studies have also presented compelling evidence for a direct link, through the expression of 15-PGDH, between early apocrine lesions and pure apocrine carci- nomas. Moreover, specific antibodies against components of the signature have identified precursor lesions in the linear progression to apocrine carcinoma. Finally, the identification of proteins that characterize the early stages of mammary apocrine differentiation such as 15-PGDH, HMG-CoA reductase and COX-2, has opened a window of opportunity for chemoprevention. S4.2-2 Uniprot and beyond: integrated database infrastructure for the life sciences R. Apweiler EMBL-EBI, Sequence Database Group, Wellcome Trust Genome Campus, Hinxton, UK. E-mail: apweiler@ebi.ac.uk The increasing sophistication of electronic mechanisms to store, manipulate and communicate information has transformed the way we work. In particular, the opportunities and pitfalls for sci- ence opened up by information technology are profound. Nowhere has this been more apparent than in molecular biology. It is possible to collect, analyse, communicate and share huge amounts of information rapidly and accurately. Molecular bio- logy, driven by the need to deal with large volumes of informa- tion, was quick to build large collections of shared scientific information. Nowadays, access to and skill in exploiting elec- tronic information repositories is crucial to biological research. It is on such resources (especially the protein sequence databases), their history, their usage, their advantages, their pitfalls and how scientific progress leads to new challenges for bioinformatics and the development of resources for biological information that my talk will concentrate. S4.2-3 Nanobiotechnology in clinical proteomics A. I. Archakov and Y. D. Ivanov Institute of Biomedical Chemistry RAMS, Pogodinskaya, Moscow, Russia. E-mail: archak @ibmc.msk.ru A certain concentration barrier exists in proteomics: the protein molecules residing in the concentration range below 10 )12 M can- not be identified. The atomic force microscopy (AFM) nanotech- nological method appears to be one of the most perspective diagnostic tools and enables not to measure protein concentration but merely to count single protein molecules and their complexes. Biospecific fishing method enables to extend the concentration limits of proteomics, as it enables to fetch the protein from the solution and therefore to increase its concentration on the surface of the biochip with immobilized partner proteins. Based on com- bination of AFM and biospecific fishing method we have created the AFM biochips for registration of hepatitis B and C and ovar- ian cancer biomarkers in patient serum. The described approach attains the concentrations of about 10 )15 M; however, to reveal the lower concentrations (e.g. 10 )18 M) it is necessary to convert the reversible complex formation into irreversible process. Several other nanotechnological approaches can be useful for proteomic research, among them are nanowire electrical detectors and optical biosensors. The optical biosensor method enables to register in real time the intermolecular interaction reactions (including Ag/Ab) without labels and calculate the kinetic rate constants and affinities for the complexes. Optical principle of detection can be easily implemented in the widely available CD- ROM readers, which convert the biospecific Ag/Ab binding directly into the informational signal. Nanowire electrical detec- tor is designed as nanowire arrays modified with antibodies. These detectors show discrete conductance changes with biospe- cific binding of viruses to nanowire. 4.3 DNA and Protein Microarrays S4.3-1 Identification and analysis of signaling networks regulating the development of the mammalian cerebral cortex G. Eichele Max Planck Institute for Experimental Endocrinology, Hannover, Germany Determining the spatiotemporal distribution of gene products at cellular resolution on a transcriptome-wide scale is crucial for functional genomics but presents a daunting task. To tackle this challenge, we have developed a platform that consists of high- throughput detection of gene products and a web database – gene- paint.org – from which annotated gene expression patterns can be retrieved and viewed. This platform was used to construct a mid- gestation mouse gene expression atlas containing a balanced repre- sentation of 10% of the transcriptome. As proof of concept, we analyze the Small eye mouse mutant that lacks the transcription factor Pax6, indispensable for the proper development of many organs including the cerebral cortex. From genepaint.org we retrieved 300 genes regionally expressed in developing cortex, 87 of which we then found to be differentially expressed in Small eye cortex, mostly in proliferative zones and mediating cell division, migration, and fate determination, processes known to be abnor- mal in Small eye. Band-shift assays suggest that many of the dif- ferentially expressed genes are direct Pax6 transcriptional targets. We discuss the emerging regulatory network of genetic cascades that contribute to mammalian brain development. Abstracts 25 S4.3-2 High-density protein and antibody arrays D. J. Cahill Conway Institute, University College Dublin, Dublin, Ireland. E-mail: dolores.cahill@ucd.ie The ability to systematically interrogate thousands of human pro- teins simultaneously will facilitate large-scale biological approa- ches. We have generated a set of over 10 000 different human recombinant proteins. Results of screening antibodies and pep- tides against these thousands of proteins, including validation of the results obtained, will be shown. Recent results of profiling the antibody repertoire in sera from subjects with autoimmune disease versus controls will also be described. Recent improve- ments to optimize the protein and antibody chip format will be presented. S4.3-3 Microarray-based comparative genomic hybridization (array-CGH) H. Fiegler The Wellcome Trust Sanger Institute, The Wellcome Trust Genome Campus, Hinxton, Cambridge, UK. E-mail: heike@sanger.ac.uk Comparative genomic hybridization has become a widespread method for a genome-wide analysis of DNA copy number changes in tumours. However, the detection of amplifications and deletions is limited when performing CGH to metaphase chromosomes. Replacing metaphase chromosomes with mapped sequences arrayed onto glass slides overcomes these limitations. As a result, resolution is drastically improved and limited only by the insert size and the density of the mapped sequences used. We have initially constructed an array with a resolution of approximately 1 Mb across the human genome. We have now increased this resolution by using overlapping tiling path clones for array construction. Array-based CGH (array-CGH) has initially been used for the analysis of copy number changes in tumours. We have also applied this technology to detect microduplications and microde- letions in patients with constitutional rearrangements. In an ini- tial study using the 1 Mb array we detected copy number changes in 24% of the patients studied. However, five of the 12 copy number changes identified were found to be inherited from one of the phenotypically normal parents. This suggested that these imbalances might not be linked to the phenotype of the patient but represent copy number variations in the normal pop- ulation. In a newly established research programme we are now screening more than 1000 normal individuals for copy number variations using the whole genome tiling path array. 4.4 Pharmacogenomics and Toxicogenomics S4.4-1 Pharmacogenomics and pharmacoproteomics in cardiovascular diseases and risks G. Siest Universite ´ Inserm U.525 Equipe IV, Henri Poincare ´ Nancy 1, Nancy, France. E-mail: gerard.siest@pharma.uhp-nancy.fr We have recently proposed a five-step pharmacogenomic approach for a new cardiovascular drug. After making a sum- mary of the most well-known genes involved in drug metabolism (in particular the new transporter ones) we will take as example of drugs the statins. These drugs are responding differently in human in function on multiple polymorphisms and we will give examples with common ApoE polymorphisms influencing the hypolipaemic effects of statins. These drugs have also pleiotropic effects and are decreasing inflammation including some plasma markers. This is an example of the need to separate the clinical diseases phenotypes and to propose specific metabolic pathways. However, given the current existing pharmacogenomic tools and the development of the pharmacoproteomics field, a complete strategy needs the integration of gene products, including some peptides measured in laboratory medicine. These biomarkers could be the metabolites resulting from drug-metabolising enzymes activities, clusters of peptides related to an inflammatory process, specific proteins or the metabolites resulting from envi- ronmental effects. With the extraordinary amount of information which could now be given by the proteins separated from plasma by proteomic devices, we now have to add all these data to the genomic and transcriptomic ones. Biological systems are then the only way to put all this information in a comprehensive and clin- ically usable set of data. S4.4-2 Pharmacogenomics: towards individualized drug therapy M. Eichelbaum Dr Margarete-Fischer-Bosch-Institut, fu ¨ r Klinische Pharmakologie, Stuttgart, Germany. E-mail: michel.eichelbaum@ikp-stuttgart.de With the complete sequence of the human genome available, individualized medicine may soon become reality. Genomic infor- mation may allow more accurate prediction of an individual’s drug response and selection of the appropriate drug dosage to achieve the optimal therapeutic response, avoid therapeutic fail- ure and minimize side effects and toxicity. Although many genes encoding proteins involved in the metabo- lism, transport and mechanism of action of medications are known to exhibit polymorphism in humans, use of this knowledge in routine clinical practice is limited. Excepting a few examples of drug-metabolizing enzymes, the contribution of genetic polymor- phisms to individual differences in drug effects and toxicity is not well understood. Moreover, most studies have focused on the effect of a single polymorphism on drug response. This approach neglects the fact that drug response phenotype, like most disease phenotypes, is a complex polygenic trait also determined by non- genetic factors (three). The extent to which genetic factors con- tribute to the drug response/toxicity phenotype depends on the extent of the candidate gene’s influence on drug disposition and effects. Even if a gene has a large effect on a drug’s pharmacoki- netics or pharmacodynamics, the presence of, a single nucleotide polymorphism (SNP) in that gene will not provide an unequivocal answer but, rather, will indicate the likelihood that an individual patient will show an altered drug response. Abstracts 26 S4.4-3 Ecogenetics and pharmacogenetics of coronaropathies P. F. Pignatti Department of Mother and Child, Biology and Genetics, University of Verona, Verona, Italy. E-mail: pierfranco.pignatti@univr.it A few examples from our work on the interactions between gen- etic and environmental factors to determine a modification of (a) disease risk or (b) response to treatment will be discussed. Given the interactions between MTHFR gene and pholatemia in determining hyperomocysteinemia (Girelli et al. 1998), threshold values of pholatemia relative to genotype will be indicated (Gi- relli et al. 2003). A correlation between APOCIII genotypes and coronaropathies (Olivieri et al. 2002) or metabolic syndrome (Olivieri et al. 2003) will be the second example, including the response to dietary PUFA (Olivieri et al. 2005). The third ecogenetic example will be the correlation between PON genotypes and miocardial infarction risk in coronaropatics relative to smoking habit (Martinelli et al. 2004), or metabolic syndrome (Martinelli et al. 2005). Pharmacogenetic response to clopidogrel platelet antiaggregating therapy in patients who underwent coronary stenting, and the role of several genes involved in the drug’s mechanism of action or metabolism will last be analyzed: GPIIIa (Angiolillo et al. 2004a), GPIa (Angiolillo et al. 2004b, 2005a), and P2Y12 recep- tor (Angiolillo et al. 2005b), or CYP3A4 (Angiolillo et al. sub- mitted), respectively. 4.5 Glycobiology and Glycomics S4.5-1 Carbohydrate-based vaccines J. F. G. Vliegenthart Department Bioorganic Chemistry, Bijvoet Center, Utrecht University, Utrecht, the Netherlands. E-mail: j.f.g.vliegenthart@chem.uu.nl In recent decades, significant progress has been made in the struc- ture determination and chemo (enzymatic) synthesis of glycan chains that are part of cell surface molecules. This development has rendered feasible the preparation of new generations of car- bohydrate-based vaccines. They are urgently needed to prevent bacterial infections, because a gradual increase in the resistance to antibiotics has been observed leading to a serious health thread. Recently, the application of a human vaccine against Hemophilus influenzae type B has found approval. This vaccine consists of a fully synthetic glycan that has been coupled to a suitable carrier protein. Obviously, the problems arising from an infection by (tropical) parasites has focused the attention on their cell surfaces glycans, which are being explored as immuno-epitopes for prepar- ing glycoconjugate vaccines. Also the possible application is investigated of vaccines directed against cancer-related cell surface epitopes. These vaccines are primarily aimed to prevent the recur- rence of cancer after treatment and/or to avoid the formation of metastasis. The development in this area will be discussed. S4.5-2 Sialic acid-binding proteins, siglecs, as modulators of immune responses P. R. Crocker Cell Biology and Immunology, University of Dundee, Dundee, UK. E-mail: p.r.crocker@dundee.ac.uk Siglecs are transmembrane sialic acid-binding receptors of the immunoglobulin superfamily expressed in the haemopoietic, immune and nervous systems where they play roles in cellular adhesion and signalling. Sialoadhesin (Sn) is a prototypic siglec that mediates adhesive functions of macrophages. We have recently generated Sn-deficient mice, which exhibit subtle differ- ences in subsets of T and B cells, but no defect in steady-state haemopoiesis. Immune responses are significantly altered in Sn-deficient mice indicating that this receptor is required for opti- mal responses of T and B cell to foreign antigens. The CD33- related siglecs are a distinct subset expressed predominantly by cells of the innate immune system. They typically contain two conserved tyrosine-based signalling motifs, comprising a mem- brane proximal immune receptor tyrosine-based inhibitory motif (ITIM) and an ITIM-like motif that can recruit and activate the tyrosine phosphatases, SHP-1 and SHP-2. We have studied the inhibitory signalling functions of CD33-related siglecs and shown that even in the absence of tyrosine phosphorylation, Siglec-5 is able to deliver inhibitory signalling functions. Together with other published data, these findings suggest that glycan-depend- ent interactions of siglecs leads to modulation of leucocyte activa- tion and may play an important role in preventing inappropriate tissue damage during immune and inflammatory responses. S4.5-3 Oligosaccharide microarrays to unravel carbohydrate recognition in the immune system T. Feizi The Glycosciences Laboratory, Imperial College London, Northwick Park and St Mark’s Hospital Campus, Harrow, UK. E-mail: t.feizi@imperial.ac.uk The diverse oligosaccharides of glycoproteins, glycolipids, prote- oglycans and polysaccharides are potentially a vast source of information, harbouring ‘glyco codes’ that are waiting to be deci- phered in various contexts of biological and medical importance. Towards developing a knowledge base of biological systems that operate through oligosaccharide recognition, we have developed a carbohydrate microarray platform. The approach is based on the neoglycolipid technology that we introduced in 1985 to address the need for a microscale technology to present oligosac- charides in a clustered form for studies of carbohydrate–protein interactions. This is a powerful means of generating probes: from N- and O-glycans, glycolipids, glycosaminoglycans, polysaccha- rides and diverse chemically synthesized oligosaccharides. A key development has been to combine carbohydrate–protein interac- tion studies with mass spectrometry in situ to determine mono- saccharide composition, sequence and branching pattern, as well exotic substituents at low picomole levels. Our state-of-the-art microarray system, with almost 300 sequence-defined oligosac- charide probes, is continually expanding towards thousands, and has provision for generating ‘designer’ microarrays from desired sources. I shall describe ligand assignments to classic receptors of (C-type lectins and siglecs) and a newly described receptor Dec- tin-1 of the immune system (Nat Biotechnol 2002; 20: 1011; J Biol Chem, ahead of print, December 2005). Abstracts 27 4.6 New Proteomic Assays for Cancer Biomarkers S4.6-1 Clinical utility of biomarkers in cancer M. Mu ¨ ller Institute of Laboratory Diagnostics, Kaiser Franz Joseph Hospital, Vienna, Austria. E-mail: mathias.mueller@wienkav.at Biomarkers in cancer (approximately 25% of the mortality rate in the population) are characteristic for the malignant phenotype of a tumor cell. The tumor markers belong to various groups of substances: enzymes, hormones, oncofetal antigens, carbohydrate markers, blood group carbohydrates, proteins, and genetic mark- ers. The concentration of the circulating marker depends upon its expression, its production, its release, its catabolism, its clearance as well as upon the blood supply of the tumor. They can be iden- tified in intact cells by immuno-histopathology, flowcytometry, cancer genetics, and proteomics technologies. In blood they are commonly measured by immunochemical techniques. In addition to established markers for malignancy several new biomarkers based on genomics, proteomics, and bioinformatics research are now evolving. There are promising results on cal- reticulin for bladder cancer, on mammaglobin, on autoantibod- ies, on the combination of the tumor-suppressor protein p53, the growth factor C-erbB2, and the cell adhesion protein E-cadherin for breast cancer and on surface proteins for colon carcinoma. The ideal tumour marker for differentiation between normal and neoplastic conditions offers a high degree of organ specificity and at the same time correlates with the prognosis and stage of the malignancy. These goals have not yet been reached in most cases. However, multivariate combination of markers results in signifi- cant improvements of the diagnostic validities. S4.6-2 Endogenous proteinase inhibitors as future targets in the treatment of cancer A. S. Schroll 1 , U. Lademann 1 , I. J. Christensen 2 , J. Foekens 3 , H. J. Nielsen 2 ,S.Wu ¨ rtz 1 , L. Davidsen 1 and N. Bru ¨ nner 1 1 Veterinary Pathobiology, KVL, Frederiksberg, Denmark, 2 Department of Surgical Gastroenterology, Hvidovre Hospital, Copenhagen, Denmark, 3 Rotterdam University, Rotterdam, the Netherlands. E-mail: nbr@kvl.dk Proteinases and their natural occurring inhibitors are involved in cancer progression. As such, they may be used as biomarkers or represent targets for new treatment interventions. We have devel- oped and validated immunoassays for the detection and quantita- tion of Tissue Inhibitor of Metalloproteinases-1 (TIMP-1) in plasma and tumor tissue. Analyses of plasma samples collected retrospectively have shown that patients with colorectal cancer (CRC) have elevated plasma TIMP-1 levels as compared to non- CRC individuals. We could show that plasma TIMP-1 was eleva- ted in early as well as late stage disease and that plasma TIMP-1 detected the disease with high sensitivity and high specificity (Holten-Andersen et al. Clin Cancer Res 2002). Based on these results, we have now performed a large prospective study to val- idate TIMP-1 as a marker for early detection of CRC. The study includes plasma samples from 5330 individuals who have under- gone endoscopy. At the same time, we have shown in experimen- tal models that TIMP-1 protects against chemotherapy-induced apoptosis. By inhibiting TIMP-1 expression, cancer cells become significantly more sensitive to chemotherapy. In support of an apoptosis-inhibiting function of TIMP-1 in cancer, we have shown that in patients with metastatic breast cancer and high tumor tissue TIMP-1 levels, the objective response to conven- tional chemotherapy is 0% as opposed to 45% in patients with low TIMP-1 tumor tissue levels. S4.6-3 SH2-profiling: a new approach for the global characterization of the tyrosine phosphorylation state K. Dierck 1 , K. Machida 2 , B. J. Mayer 2 and P. Nollau 1 1 Department of Clinical Chemistry, Center of Clinical Pathology, University Medical Center Hamburg-Eppendorf, Hamburg, Germany, 2 Department of Genetics and Developmental Biology, Raymond and Beverly Sackler Laboratory of Genetics and Molecular Medicine, University of Connecticut Health Center, Farmington, CT, USA. E-mail: nollau@uke.uni-hamburg.de Deciphering complex signaling networks is of great importance for the understanding of many important biologic functions such as cell proliferation, differentiation or apoptosis. Dysregulation of signal transduction plays a critical role in the pathogenesis of many malignant and other diseases. In this context, the phos- phorylation of tyrosine residues is a key event and it is of great interest to gain detailed insights into the state and regulation of the phosphoproteome. Current techniques for the comprehensive characterization of the cellular state of tyrosine phosphorylation are limited. Applying src-homology regions 2 (SH2)-domains, we developed a novel strategy termed SH2-profiling for the analysis of phosphotyrosine-dependent signaling networks. SH2-domains are modular-binding domains present in many important signa- ling proteins specifically recognizing tyrosine phosphorylated pro- teins and thus serving as ideal tools for the comprehensive analysis of the tyrosine phosphorylation state. Our current SH2-profiling data demonstrate that tumor- and patient-specific phosphorylation profiles are detectable with high sensitivity and specificity in small amounts of complex protein mixtures allowing to uncover aberrant signaling pathways with the potential to clas- sify tumors and to develop rational therapeutic concepts based on differential signaling states. Abstracts 28 4.7 Imaging and Noninvasive Techniques: Fluorescence Techniques S4.7-1 Quantum dot ligands reveal a new mechanism for retrograde transport of activated growth factor receptor T. M. Jovin, D. S. Lidke and D. J. Arndt-Jovin Department of Molecular Biology, Max Planck Inst. f. Biophys. Chem., Goettingen, Germany. E-mail: tjovin@gwdg.de Biotinylated epidermal growth factor (EGF) attached to strept- avidin-conjugated Quantum Dots (QDs) binds to its cell surface receptor (EGFR, ErbB1, HER1). Wide field and laser scanning microscopy have revealed the ligand–receptor complexes diffuse freely on the cell surface, including filopodia, cellular extensions with an F-actin bundle core. Interaction of at least two com- plexes leads to activation of the receptor tyrosine kinase and initiation of retrograde transport from the filopodia to the cell body. According to a FRET assay based on QDs as donors, endocytosis of the complexes occurs at the base of the filopodia. We have measured the rates of free diffusion and retrograde transport as a function of cell type, temperature, presence of var- ious inhibitors, and mutations or manipulation of the expression levels of putative protein components of the transport mechan- ism. Diffusion constants and transport rates were determined with single molecule sensitivity by tracking individual QD–recep- tor complexes, exploiting the high sensitivity provided by an elec- tron multiplying CCD camera. The QDs, the associated receptors (visualized as a fusion protein with eGFP), and F-actin are trans- ported with the same rate on filopodia, implying a coupling of receptor movement to treadmilling of the F-actin bundles. We conclude that filopodia act as non-linear biosensors for growth factors located far from the cell body (Lidke et al. J Cell Biol 2005; 170: 619–626). S4.7-2 Functional optical imaging of the neocortex in action A. Grinvald Department of Neurobiology, The Weizmann Institute of Science, Rehovot, Israel, E-mail: amiram.grinvald@weizmann.ac.il During the last few decades, neuroscientists have benefited from the emergence of many powerful functional imaging techniques that cover broad spatial and temporal scales. We can now image single molecules controlling cell differentiation, growth and death; single cells and their neurites processing electrical inputs and sending outputs; neuronal circuits performing neural compu- tations in vitro; and the intact brain. At present, imaging based on voltage-sensitive dyes (VSDI) offers the highest spatial and temporal resolution for imaging neocortical functions in the liv- ing brain, and has paved the way for a new era in the functional imaging of cortical dynamics. It has facilitated the exploration of fundamental mechanisms that underlie neocortical development, function and plasticity at the fundamental level of the cortical column. To follow neuronal computations at this fundamental level requires a spatial resolution of at least 200 lm. Further- more, because communication occurs over milliseconds, an ima- ging approach with millisecond time resolution is required. Without an imaging technique that can reveal the spatio-tem- poral patterns of the subthreshold input, perception and other higher brain functions cannot be fully understood. Finally, to investigate higher brain functions, it is essential to be able to image the neocortical function of behaving mammals, partic- ularly monkeys. In the present study, we review the technique of VSDI and describe some of its applications. S4.7-3 Visualization and sensing of intermolecular interactions with two-color fluorescent probes A. Demchenko Laboratory of Spectroscopy, Research Institute for Genetic Engineering and Biotechnology, Gebze-Kocaeli, Turkey. E-mail: dem@gmbae.tubitak.gov.tr The key issue in biosensing is the transformation of response to sensor–analyte interaction into measurable signal. We developed a new generic technology for fluorescence response that may find a broad range of applications – from homogeneous assays in solutions to microarrays, microfluidic devices, nanosensors, and whole-cell imaging. In sensing event, the novel functional dyes change their color due to interplay of intensities between blue- green and orange-red well-separated emission bands. Due to two- channel ratiometric detection the signal is self-calibrating and allowing easy quantitative assays. These dyes belong to the fam- ily of 3-hydroxychromones (3HCs) exhibiting an excited-state intramolecular proton transfer (ESIPT) reaction. By proper chemical substitutions we endowed 3HCs the ability to respond in a wavelength-ratiometric manner to all major types of non-covalent interactions that can be used in sensing – to polar- ity, hydrogen bonding ability and to electrostatic fields. Their derivatives for covalent labeling of proteins and peptides were synthesized. We present the data showing how with incorpor- ation of new dyes the biomembrane modes such as micelles and phospholipid vesicles can be transformed into nanosensor devices. In cellular research a fast response to electrostatic mem- brane potential can be achieved and the membrane changes occurring on apoptosis detected. Abstracts 29 4.8 Imaging and Noninvasive Techniques: MRI S4.8-1 Use of intracortical recordings and fMRI to study mass action N. K. Logothetis Department of Physiology of Cognitive Processes, Max-Planck-Institute for Biological Cybernetics, Tuebingen, Germany. E-mail: conchy.moya@tuebingen.mpg.de The combination of fMRI with electrophysiology, histology, and neurochemistry promises great insights into a level of neural organization that could have never been studied with either tech- nique alone. My talk will focus on (a) spatially resolved fMRI/ MRSI; (b) the study of in vivo connectivity using fMRI and elec- trical microstimulation, or manganese-enhanced MRI; (c) com- bined physiology and MRI for examining the electrical activity occurring during increases and decreases of BOLD activation; and (d) molecular imaging based on smart agents. MRSI optimi- zation enabled sufficiently high spectral dispersion and spatiotem- poral resolution to obtain isolated glutamate maps in the primate brain. Ongoing research attempts the differentiation of brain structures in the millimeter range and/or detection of small con- centration differences in the same structure (activated versus non- activated cortex). Finally, I will report our first attempt to develop chelates for Gd and endow them with the appropriate coordinating groups that reversibly block some of the Gd’s free coordination sites with changes in pH or [Ca++]. S4.8-2 Functional and molecular assessment of breast cancer using high field MRI and MRS M. Garwood Radiology/CMRR, University of Minnesota, Minneapolis, MN, USA. E-mail: gar@cmrr.umn.edu The techniques of magnetic resonance spectroscopy (MRS) and imaging (MRI) can be used to assess molecular, functional, and anatomical features of breast lesions in patients non-invasively. MRI scanning of the breast following intravenous injection of a contrast-enhancing agent such as Gd-DTPA is useful for deter- mining the presence or absence of a breast lesion, and provides information about lesion morphology, blood flow, and vascular properties. In the same examination, in vivo proton (1H) MRS can be used to obtain molecular information about the lesion(s). Resonances from choline-containing compounds are commonly observed in 1H spectra of malignant lesions, but not in benign or normal tissues. The primary constituents are those with a tri- methylamine moiety, including free choline, phosphocholine, and glycerophosphocholine. Used in combination, these MR tech- niques are capable of providing information to make an accurate non-invasive diagnosis and to identify a clinical response early in the course of cancer therapy. Current methods and technical issues in performing MRS and MRI of breast at high magnetic field (4 Tesla) will also be discussed. S4.8-3 MRI-guided gene therapy E. Atalar 1 and X. Yang 2 1 Electrical and Electronics Engineering, Bilkent University, Ankara, Turkey, 2 Department of Radiology, Johns Hopkins University, Baltimore, MD, USA. E-mail: ergin@ee.bilkent.edu.tr Gene therapy is designed to target the causes rather than the symptoms of diseases. Our current knowledge about the biodis- tribution and pharmacokinetics of gene therapy relies mainly on results of in vitro laboratory examinations of tissues obtained at biopsy or autopsy. In many cases, actual distribution of gene materials is unknown. Thus, it is essential to develop non-inva- sive methods to monitor and guide gene therapy. To date, different imaging modalities, such as ultrasound, nuclear medicine, and optical imaging, have been applied in monitoring of gene therapy. Magnetic resonance imaging (MRI) has several prominent advantages comparing to other imaging techniques, including the abilities to generate high-contrast and high-resolu- tion images, to obtain multiple diagnostic evaluations of organ functions and morphology, and to provide multiple image planes with no risk of ionizing radiation. Recent efforts have focused on using MR technology to monitor gene delivery, to enhance gene transfection/transduction, and to track gene expression. This review summarizes the current status of MRI-guided gene ther- apy. Abstracts 30 . Diagnostic Approaches 4.1 Bioinformatics and Proteomics S4.1-1 Strategies to cover microbial proteomes P. R. Jungblut Core Facility Protein Analysis, Max. also determined by non- genetic factors (three). The extent to which genetic factors con- tribute to the drug response/toxicity phenotype depends on the extent