Search

Poliketidi i neribosomalno sintetizirani peptidi su vrlo važne kemijske supstancije za farmaceutsku industriju i agroindustriju. Njihov biosintetski put obuhvaća spajanje jednostavnih građevnih jedinica u složene kemijske strukture katalitičkim djelovanjem enzimskih kompleksa poliketid sintaza ili sintetaza neribosomalnih peptida. U posljednjem desetljeću u znanstvenoj javnosti postoji osobito zanimanje za oblikovanje novih supstancija u proizvodnji novih lijekova manipulacijom genskih nakupina tih enzimskih kompleksa u uvjetima in vitro, postupcima kombinatorne biosinteze. Međutim, značajna je prepreka napretku na tom području što većina promjena u uvjetima in vitro ne dovodi do sinteze produkta ili su mu prinosi vrlo mali. Jedno od mogućih rješenja toga problema bilo bi oblikovanje novih genskih nakupina homolognom rekombinacijom u uvjetima in vivo jer bi se tako omogućilo spajanje identičnih sekvencija i smanjile poteškoće zbog pojave nefunkcionalnih čvorišta te opće nedovoljne identičnosti različitih modula. Osim toga, homolognom bi se rekombinacijom povećala učestalost rekombinacije, te potaknula kombinatorna raznolikost rekombinanata. U tijeku je razvoj integralnih računalnih programskih paketa, CompGen i ClustScan, za modeliranje tih procesa u uvjetima in silico. Okosnica je programskog paketa CompGen specifično strukturirana baza podataka koja povezuje biosintetski put sinteze sa sekvencijama DNA genskih nakupina. Povezanost sekvencija DNA s biosintetskim putem omogućava njezinu povezanost sa strukturom produkta. Jedna je od funkcija računalnoga programskog paketa, temeljena na toj povezanosti, sposobnost oblikovanja virtualnih rekombinanata između genskih nakupina. To se obavlja pomoću modela rekombinacije da bi se in silico predvidjele sekvencije DNA u kojima dolazi do homologne rekombinacije. CompGen iz tih podataka predviđa kemijsku strukturu nove supstancije i sprema je u virtualnu bazu podataka kemijskih struktura radi daljnjega molekularnog modeliranja. Računalni programski paket omogućava i analizu tzv. 'obrnutom genetikom'. Naime, ako se pretpostavi poželjna kemijska struktura, program može predvidjeti kako bi trebala izgledati genska nakupina poliketid sintaza ili sintetaza neribosomalnih peptida, koja bi sintetizirala takvu strukturu na temelju građevnih jedinica genskih nakupina u bazi podataka. U cjelini, CompGen će omogućiti oblikovanje baze podataka novih prirodnih kemijskih entiteta u uvjetima in silico, koja se zatim može upotrijebiti za istraživanja na području tehnologije računalnoga dizajna novih lijekova. Drugi će integralni generički računalni programski paket, ClustScan, moći prepoznati i anotirati nove genske nakupine iz sekvencija cjelovitih mikrobnih genoma ili genskih nakupina u metagenomima mikroorganizama što žive u tlu ili u simbiozi s morskim organizmima.

Many polyketide compounds are used in the agro-industry including antiparasitics (e.g. avermectin), coccidiostatics (e.g. monensin), animal growth promoters (e.g. virginiamycin) and insecticides (e.g. spinosin). Traditional screening programmes suffer from the disadvantage that subsequent purification and characterisation of compounds is slow and labour intensive. Combinatorial chemistry looked promising, but failed to produce bioactive compounds. The modular polyketide biosynthesis clusters (PKS) in Streptomyces synthesise polyketides in a step-wise manner and the extender unit incorporated at each step is specified by its own module in the large PKS enzyme. Combinatorial biosynthesis, which combines modules of different specificity using recombinant DNA technology should produce a large number of polyketide products. Although such approaches can produce the predicted products, the yield is usually extremely low restricting their practical application. It seemed likely that the low yield resulted from unsuitable junctions between modules. Homologous recombination, which should be important in the natural evolution of new PKS clusters, favours closely related sequences and should result in better junctions and higher yields. It was, therefore, decided to model homologous recombination between known PKS clusters in silico. The program CompGen predicted the number of recombinants that would be produced and using a specially constructed database to link genetic structure to biosynthesis, it was possible to predict the novel chemical structures of the linear polyketide backbone produced by recombinants. The problems of predicting the cyclisation of these structures will be discussed. A second program ClustScan was developed to annotate new clusters and, thus, expand the number of substrates for CompGen.

Parasites are a major problem both for meat production and in fish farming. Avermectin and its derivatives have proved very successful antiparasitics, but there is a constant demand for new compounds. The most important traditional path to new compounds is by screening for secondary metabolites. The diminishing returns from this approach have stimulated the development of alternative approaches in recent years. Avermectin, which is produced by Streptomyces avermitilis, is a polyketide that is synthesised by a modular polyketide synthase (PKS). Such PKSs function on a "building block" basis so that it is possible to obtain derivatives in a rational way by carrying out defined genetic manipulation on individual modules. Unfortunately, this usually results in dramatic decreases of yield making industrial production impossible. We have suggested an alternative approach using homologous recombination between different clusters. This will select recombinants between closely related modules without unnatural junctions and may, thus, result in higher yields. We have developed a generic computer program package CompGen that contains programs to predict recombination sites. The package also predicts the primary linear polyketide chain products of recombinant clusters. The package was used to examine potential recombinants between the avermectin gene cluster and other PKS clusters. The compounds were classified in terms of their similarity to avermectin to identify candidates for novel antiparasitics.

Our approach to the in silico drug design and discovery links the DNA sequences encoding modular enzymes that synthesise natural products of clinical importance, polyketides and non-ribosomal peptides, to their in silico medicinal chemistry. This can be done because there is a direct correlation of the gene cluster DNA sequence, by virtue of amino acid sequences of modules and catalytically active domains of these multi-functional and multi-modular proteins, to the chemical structure of their biosynthetic product. Therefore, polyketide or peptide chemical structures can be predicted from gene-cluster DNA sequences. It is well known that more than 50% of drugs - that are today in clinical use has - as their biologically active principls - polyketide or peptide compounds. To them we should add antiparasitics, coccidiostatics, animal growth promotants and natural insecticides used in food and agro-industries. Polyketides and peptides were, therefore, very important for the discovery of present drugs and will continue to be important for the novel drug discovery. Since 1940s the pharmaceutical companies and academic institutions have screened culturable microorganisms from different habitats for their natural products. They discovered a significant number of biologically active polyketide and peptide compounds with important pharmacological properties, some of which has found their way to the market. However, during the last 30 years or so, culturable microorganisms with novel biologically activities become exhausted and the discovery of novel drugs dropped down considerably. The major question now is: are there any alternative strategies to exploit the chemical potential of small molecule natural products? The answer might lie in recent development of rapid and relatively in-expensive DNA sequencing technologies, which have delivered more than nine hundred sequenced bacterial genomes, with more than thirteen hundred genome sequencing projects in progress. Polyketide and peptide gene clusters were analysed within this genomes. The most prolific producers of these natural products are Actinomyces species. It appeared that genome of every Actinobacteria has more than one gene cluster for the synthesis of these secondary metabolites. For instance, the sequenced genome of Streptomyces avermitilis has, along the gene clusters for the polyketide antiparasitic avermectin and the polyketide antibiotic oligomycin, another nine polyketide and eight peptide gene clusters. Today, there are a large number of sequenced polyketide, peptide and hybrid gene clusters of unknown products in public databases and their number grows exponentially with every new bacterial genome sequenced. There is, therefore, significant chemical diversity potential in these sequenced genomes and gene clusters. Even more important, these developments have also opened up new opportunities allowing DNA sequencing of unculturable microorganisms. Today they permit the DNA sequencing of metagenomes, in which a large number of unculturable microorganisms may be present, that can not be grown in the laboratory and whose biosynthetic potential can not be assessed by the traditional methods. Computational tools are now needed to rapidly mine such massive datasets to hunt for gene clusters encoding entirely novel compounds which could be the medicines of the future. This is where we hope our approach will become important. We are developing two generic computer program packages ClustScan and CompGen. ClustScan and CompGen are acronyms of "Cluster Scanner and Compound Generator". The currents status of their development will be described at this meeting.

Poliketidi su važna skupina prirodnih spojeva koji se upotrebljavaju u farmaceutskoj industriji. U javnim bazama podataka danas postoji veliki broj sekvenciranih poliketidnih (PKS) genskih nakupina nepoznatih produkata i njihov broj raste eksponencijalno sa svakim novo sekvenciranim genomom aktinomiceta. Mi smo razvili programski paket ClustScan za analizu sekvencija DNA i anotaciju modularnih PKS genskih nakupina. Budući da se predviđanja temelje na prepoznavanju bitnih aminokiselina, a ne na sličnosti sekvencija, program uspješno predviđa potpuno nove genske nakupine. ClustScan ispisuje predviđene kemijske strukture u obliku izomeričkih zapisa SMILES (engl. "Simplified Molecular Input Line Entry System") i dopušta daljnje analize kemijskim računalnim programima. Razvili smo i programski paket CompGen koji generira potencijalne rekombinante između PKS genskih nakupina. Pohranjene su u posebnoj bazi podataka koja povezuje sekvencije DNA s biosintetskim redoslijedom i tako omogućava predviđanje kemijskih struktura novih spojeva. CompGen pokazuje da je malen broj mjesta za homolognu rekombinaciju između PKS genskih nakupina i zbog toga generira odabranu skupinu kemijskih spojeva koja imitira one u prirodi. ClustScan i CompGen omogućuju potragu za genskim nakupina što sadržavaju genetičku uputu za nove kemijske spojeve u genomima i metagenomima koji bi mogli postati lijekovima budućnosti. Funkcionalnost će programa biti ilustrirana prikladnim primjerima.

The program package ClustScan is designed for rapid, semi-automatic, annotation of DNA sequences encoding modular biosynthetic enzymes including polyketide synthases (PKS), non-ribosomal peptide synthetases (NRPS) and hybrid (PKS/NRPS) enzymes. Potential protein coding regions and protein domains are identified by programs running on a server and the results are presented to the user in a graphical interface with a Java client program. The user can edit the results and assemble biosynthetic clusters. The versatility of this program package has been demonstrated by annotating biochemical pathways in microbial, invertebrate animal and metagenomic datasets. The novelty of ClustScan over existing computer programs is the automatic prediction of substrate activity and specificity, as well as steriochemical outcomes of biosynthetic pathways, allowing linear and cyclic chemical structures to be deduced in silico. The open architecture of ClustScan allows easy integration with other programs, facilitating further analyses of results, which would be of use to a broad range of researchers in the chemical and biological sciences. There are two main criteria for the usefulness of ClustScan. The accuracy of prediction and the speed and convenience of annotating large DNA data sets. The accuracy of prediction will be demonstrated using the well-known gene cluster encoding the erythromycin biosynthetic pathway. The speed and convenience of annotating large DNA data sets will also be demonstrated using genomic and metagenomic data sets. The modular biosynthetic clusters in an actinomycete genome sequence can be annotated by a single person in about 3 hours. The ClustScan program package will be demonstrated online during the meeting on reguest. To download ClustScan and access services, one should go to the Web site http://reg.bioserv.pbf.hr/. STARCEVIC, Antonio, ZUCKO, Jurica, SIMUNKOVIC, Jurica, LONG, F. Paul, CULLUM, John, HRANUELI, Daslav. ClustScan: Semi-automatic annotation of modular gene clusters and prediction of chemical structures. Nucleic Acids Res., 2008 (submitted). STARCEVIC, Antonio, AKTHAR, Shamima, DUNLAP, C. Walter, SHICK, J. Malcolm, HRANUELI, Daslav, CULLUM, John, LONG, F. Paul. Enzymes of the shikimic acid pathway encoded in the genome of a basal metazoan, Nematostella vectensis, have microbial origins. Proc. Natl. Acad. Sci. U.S.A., 105, 2533-2537, 2008. ZUCKO, Jurica, SKUNCA, Nives, CURK, Tomaz, ZUPAN, Blaz, LONG, F. Paul, CULLUM, John, KESSIN, Richard, HRANUELI, Daslav. Polyketide synthase genes and the natural products potential of Dictyostelium discoideum. Bioinformatics, 23, 2543– 2549, 2007.

Traditional screening has been immensely successful in identifying valuable secondary metabolites. Polyketides are a very important class of natural products and modular polyketide synthases (PKS) have a building block structure that makes them an inviting target for genetic manipulation. In vitro genetics has been successful in achieving desired compounds, but the yield is often extremely low. The natural evolution of modular systems must involve homologous recombination. We have been modeling recombination in silico in the expectation that homologous recombination will produce more compatible junctions than in vitro manipulations and, thus, better yields. Molecular genetic methods are available in Streptomyces and related genera, which would allow the laboratory realization of the in silico constructs. The program suite CompGen generates potential recombinants from PKS clusters. The clusters are held in a special database, which links DNA sequence to the biosynthetic order and, thus, allows prediction of the chemical structure of novel products. CompGen shows that the number of good sites for homologous recombination between clusters is usually low and, thus, produces a strongly selected set of product molecules, which are likely to mimic those that are available in nature. 16 well annotated clusters were used to test the program and 216 recombination sites were predicted in the total of 120 possible cluster pairs. A second program suite ClustScan was developed to analyze DNA sequences and annotate modular PKS clusters. The specificity of the modules is predicted from published knowledge about PKS proteins and new analyses are done to identify amino acid residues involved in specificity. This predicts the structure of the products including the stereochemistry. As the prediction is based on identification of critical amino acid residues rather than similarity with known PKS clusters, it is particularly useful for novel clusters that are not closely related to known clusters. ClustScan can output the predicted chemical structure as an isomeric SMILES (Simplified Molecular Input Line Entry System) allowing analysis by further chemistry-based programs. ClustScan was initially developed to build the database for CompGen, but is useful in its own right for annotating the ever growing flood of PKS sequences in the DNA databases. The modular biosynthetic clusters in an actinomycete genome sequence can be annotated by a single person in about 3 hours. The ClustScan and CompGen program packages will be demonstrated online during the meeting. To download ClustScan and access services, one should go to the Web site http://reg.bioserv.pbf.hr/.

The ongoing genomic revolution, together with the application of computational tools, has provided the opportunity to mine whole genome and mixed-population metagenome sequences revealing numerous examples of 'cryptic' or 'orphan' biosynthetic gene clusters. The discovery of these secondary metabolic pathways, for which the encoded natural product is unknown, represents a treasure trove for gaining deeper insight into the evolution of chemical diversity and biological function. Several tools have been developed to assist these processes which, in the main, use local sequence alignment or motif searching to annotate based on finding the closest related sequences in a database. The reliability of such annotation depends, therefore, not only on the size and taxonomic diversity of the database against which a query sequence is compared, but also on the search criteria used to make the comparison. We have used a computational systems biology approach that integrates global sequence alignment and family profiling methods as a first step to overcome database composition bias. In addition, we have also developed an integrated set of computer programs to find modular gene clusters in substantial DNA data sets, which we call the ClustScan (Cluster Scanner) program package. The novelty of ClustScan over existing computer programs is the automatic knowledge-based prediction of substrate specificity and steriochemical outcomes of biosynthetic pathways, allowing linear and cyclic chemical structures to be deduced in silico. The versatility of our approaches to extract data on biosynthetic processes using genome sequences from marine invertebrate-microbial symbiotic assemblages and free-living terrestrial microorganisms will be discussed.

Secondary metabolites play an important role in agricultural production. They include antiparasitics as well as coccidiostatics and animal growth promoters. Polyketides are one of the most important chemical classes. Many products are synthesised by modular polyketide synthetases (PKS) in bacteria. In this case, successive biosynthesis steps are carried out in successive modules in the PKS. Each module incorporates a module-specific extender unit. The variety of choice for extender unit allows the synthesis of a large number of different polyketides. The extender units can differ in number of carbon atoms (C2, C3 or larger) and the degree of reduction (keto group, hydroxyl group, C=C double bond or fully reduced) as well as different chiral forms. Comparing the DNA sequences of modules from well-characterised clusters has given information about protein sequence differences associated with the incorporation of different extender units. Large scale DNA sequencing is resulting in ever increasing numbers of sequenced PKS clusters, whose product is unknown. We have developed the program ClustScan, which helps the semi-automatic annotation of unknown modular PKS clusters including prediction of the extender units incorporated by each module. We will show the results of such analyses and discuss the limits of prediction and the problem of selecting clusters, which may produce products of agricultural interest for further characterisation.

Modular polyketide synthases (PKSs) are multifunctional enzyme complexes that catalyse the synthesis of a structurally diverse group of secondary metabolites – the polyketides. Many polyketides are produced by Streptomyces and related filamentous bacteria ; some are of commercial significance including the antibiotics erythromycin A and rifamycin, the immunosuppressive macrolide rapamycin and the antitumour compound mithramycin. Since many of these compounds are high value biopharmaceuticals, there has been intense interest in studying the chemistry and molecular biology of biosynthesis on these megasynthases [1]. The modular PKS leading to the formation of the erythromycin polyketide backbone (6-deoxyerythronolide B synthase or DEBS) in Saccharopolyspora erythraea is one of the best studied [2], especially regarding stereochemistry. Polyketide biosynthesis proceeds via a Claisen condensation between simple carboxylic acids with inversion of configuration to give a (2R) methyl centre [3]. However, the methyl stereochemistry present in the erythromycin polyketide backbone, namely (2R), (4R), (6S), (8R), (10R) and (12S), implies that two epimerisation steps occur. We used a bioinformatics approach to identify a putative epimerase function in the relevant ketoreduction (KR) domains. For profile analysis the current version of HMMER [http://hmmer.wustl.edu/] and current release of the Pfam database [http://www.sanger.ac.uk/Software/Pfam/] were used. All three DEBS protein sequences were extracted from GenBank and searched against the entire Pfam database locally. The Secondary Structure Matching server at EBI [http://www.ebi.ac.uk/msdsrv/ssm/] was used as a tool for overlapping known crystal structures. Multiple alignments of A- and B-type KR sequences were constructed using ClustalW [http://www.ebi.ac.uk]. Family profiling was first undertaken to predict the most likely location for epimerisation within the domains of the DEBS proteins. Sequence comparisons were then used to identify particular amino-acids that could be involved in this catalysis. Family profiling within DEBS confirmed the sequence positions of all the known catalytic domains. Interestingly, however, when the NAD dependent epimerase/dehydratase family within Pfam was used as a search template, putative epimerase motifs were identified within the boundaries of all the KR domains. Therefore, we postulate that C2 epimerization and C3 ketoreduction are coordinated processes and that the KR domains within modular polyketide synthases should be referred to as isomeroreductases to reflect their bifunctional nature. [1] J. Staunton, K.J. Weissman, Nat. Prod. Rep. 18 (2001) 380-416. [2] D. Hranueli, J. Cullum, B. Basrak, P. Goldstein, P.F. Long, Curr. Med. Chem. 12 (2005) 1697-1704. [3] A. Siskos, A. Baerga-Ortiz, S. Bali, V. Stein, H. Mamdani, D. Spiteller, B. Popovic, J. Spencer, J. Staunton, K. Weissman, Chem. Biol. 12 (2005) 1145-1153.

In tropical marine environments where nutrient concentrations are often poor, sessile invertebrates compensate for a lack of nutrients by adopting a symbiotic relationship with microorganisms, especially endosymbiotic dinoflagellates and unicellular algae. In these photoautotrophic symbioses, organic carbon is released to the animal host for nutrition by the microorganism that benefits from recycling of waste nutrients and carbon dioxide from the host animal. The requirements of algal symbiosis necessitate exposure to high levels of solar radiation that can lead to potential photo-oxidative toxicity, exacerbated by the release of photosynthetic oxygen to cause hyperoxic saturation of host tissues. One mechanism symbioses have evolved in defence to chronic UV exposure is the synthesis of protective sunscreen compounds that absorb environmentally relevant wavelengths of UV radiation. These compounds, termed mycosporine-like amino acids (MAAs), are a family of some 20+ related natural products possessing a cyclohexenone or cyclohexenimine chromophore conjugated with the nitrogen constituent of an amino acid having UV– absorption maxima in the range 310-360 nm with high molar absorptivity (e = 20, 000-50, 000). Specific inhibitors of shikimic acid biosynthesis block the accumulation of MAAs in coral tissues suggesting that MAAs biosynthesis is derived from a branch point of the shikimate pathway. In aposymbiotic animals and some symbiotic examples, accumulation of MAAs is via a dietary route since animals cannot synthesise de novo essential products of the shikimate pathway. In all other assemblages the MAAs are, therefore, believed to be derived from the symbiotic phototrophic partner. Enzymes forming the putative branch point in the shikimate pathway leading to the formation of MAAs have never been isolated and the genes encoding these enzymes have not been described. We report on our progress to identify the genes responsible for MAA biosynthesis that may enable us in the future, using recombinant DNA technology, to biosynthesize new classes of sunscreens for the prevention of UV-associated diseases such as skin cancer. We anticipate that genetic modification could also be used to incorporate the MAA shikimate branch-point pathway in higher plants to produce new strains of UV-tolerant plants for enhanced crop production in the third-world tropics.

The social amoeba (cellular slime mould) Dictyostelium discoideum is an important model organism for studies in development, cell communication and, more recently, evolution [1]. Yet, one aspect of the life of these organisms has been ignored: like other soil micro-organisms they should produce a complex chemical repertoire. The genome sequence was recently completed showing around 12, 500 genes in a relatively small genome of 34 Mb. The organism is exceptionally rich in polyketide synthases (PKS) encoded by more than 40 recognizable genes spread on all six chromosomes [2, http://dictybase.org/]. We know of no other organisms with as many PKS genes [3]. Therefore, we have started with the analysis of the potential PKS genes using bioinformatics methods. The putative PKS genes were initially identified using BLAST [http://www.ncbi.nlm.nih.gov/blast/], with the erythromycin KS4 domain [3] as a query against the translated D. discoideum genome sequence. In order to recover potentially new PKS genes the dicyBase version 2.5 was used. A knowledge based resource for analysis of non-ribosomal peptide synthetases and polyketide synthases [http://www.nii.res.in/nrps-pks.html] was used to retrieve PKS domain sequences for building custom profiles. The current version of HMMER [http://hmmer.wustl.edu/] and the current release of the Pfam database [http://www.sanger.ac.uk/Software/Pfam/] were used for profile analysis. All PKS protein sequences were extracted from dictyBase and searched against the entire Pfam database locally. From the 45 PKS genes identified, 29 occur in 12 potential gene-clusters (containing 2-5 genes each), which might correspond to the biosynthetic pathways of different natural products. The genes show a similar domain structure to Type I PKS genes except that they contain two additional conserved amino acid stretches (of 450-1500 and 300-600 amino acids) of unknown function that does not show any homology to any known protein. No loading domains could be identified, which suggests there might be a novel initiation mechanism. Sixteen of the genes resemble Type I iterative genes, but no Type II genes were found. In this poster we will present the use of bioinformatics approaches to analyse two conserved amino acid sequences of unknown function with detailed sequence annotation using Artemis [http://www.sanger.ac.uk/Software/ Artemis/]. Detailed annotation of all genes showed that they encoded β -ketoacylsynthase (KS), acyltransferase (AT), dehydratase (DH), methyltransferase (MT), enoylreductase (ER), ketoreductase (KR) and acyl carrier protein (ACP) domains expected in a PKS [3]. [1] R.H. Kessin, Dictyostelium - evolution, cell biology, and the development of multicellularity, Cambridge Univ. Press, Cambridge, 2001, pp. 1-294. [2] L. Eichinger, J.A. Pachebat, G. Glockner, M.A. Rajandream, R., Sucgang et al., Nature 435 (2005) 43-57. [3] D. Hranueli, J. Cullum, B. Basrak, P. Goldstein, P.F. Long, Curr. Med. Chem. 12 (2005) 1697-1704.

Modular polyketides and nonribosomal peptides are very important compounds for the pharmaceutical and agricultural industries. Their biosynthesis involves assembly of simple chemical building blocks into complex chemical structures by the catalytic activity of polyketide synthases (PKS) or nonribosomal peptide synthetases (NRPS) [1]. There has been a lot of interest in the last few years in generating new compounds for the production of novel drugs by manipulating the programming of such clusters in vitro (e.g. the idea of combinatorial biosynthesis) [2]. However, an important barrier to the progress is the fact that most changes made by in vitro methods result in very low yields or no detectable product. A possible solution to the yield problem would be the generation of novel clusters by homologous recombination in vivo [3], because this would favour more closely related sequences and should reduce problems caused by incompatible junctions. An integral generic program package is under development to model this process in silico. The heart of the program is a specially structured database that connects the biosynthetic order of synthase/synthetase enzymes to the sequences of the component polypeptides. The additional linkage to the gene sequences allows the integration of DNA sequence with product structure. The database contains sequences of the well-characterised PKS and NRPS clusters to act as building blocks for the production of novel products. It is easy to add custom sequences to the database. One function of the program is the ability to generate virtual recombinants between clusters. This can be done using a recombination model to predict sites for homologous recombination or by user defined recombination sites (e.g. to model in vitro genetic manipulation such as module replacement). The program predicts the chemical structure of the resulting products and allows them to be inserted into a virtual compound database for molecular modelling studies. An optional 'reverse Genetics' module analyses a given chemical structure to see if it could be produced by a PKS or NRPS synthesis cluster and suggests the DNA sequence of a suitable cluster based on building blocks derived from clusters contained in the data base. There are also component programs to recognise and annotate new gene clusters from microbial genome sequencing projects or in metagenomes of soil and/or marine microorganisms [4]. The integral generic program package allows in silico generation of databases of novel chemical compounds that can be used for in silico screening using CDD technology. [1] D. Hranueli, J. Cullum, B. Basrak, P. Goldstein, P.F. Long, Curr. Med. Chem. 12 (2005) 1697-1704. [2] K.J. Weissman, P.F. Leadlay, Nat. Rev. Microbiol. 3 (2005) 925-936. [3] C. Rayssiguier, D.S. Thaler, M. Radman, Nature 342 (1989) 396-401. [4] W.C. Dunlap, M. Jaspars, D. Hranueli, C.N. Battershill, N. Perić-Concha, J. Zucko, S.H. Wright, P.F. Long, Curr. Med. Chem., 13 (2006) 697-710.

Natural products from symbiotic or commensal associations between marine invertebrate and microbial organisms show exceptional promise as pharmaceuticals in many therapeutic areas. An economic and sustainable global market supply due to difficulty of synthesis is cited as the main obstacle for exploitation of these otherwise exciting marine bioactive compounds [1]. Different strategies have been evoked to overcome this impediment as long-term harvesting of wild stocks from the environment is considered unsound, and other modes of production based on biosynthesis, such as aquaculture, have not yet been proven as reliable [2]. One option is to clone the genes encoding the biosynthetic expression of a lead metabolite into a Surrogate host suitable for industrial-scale fermentation. To facilitate this goal we are developing a universal system to clone and express genes responsible for biosynthesis of natural products from both eukaryotic and prokaryotic partners of marine symbioses. The ability to harness the complete meta-transcriptome of entire biosynthetic pathways is particularly valuable where the biogenesis of a target natural product occurring within a complex symbiotic association is unclear.

Polyketides are important chemical substances for both pharmaceutical and agro-industries. Today, polyketide antibiotics, antifungals, antiparasitics, animal growth promotants, natural insecticides, immunosuppressives, cytostatics and anticholesterolemics are in commercial use. Polyketides are synthesized by Type I modular and Type II iterative polyketide synthases (PKS's) in bacteria, Type I iterative PKS's in fungi and by chalcone synthases in plants (The nature of polyketide antibiotic synthesis will be explained at the poster with an animated cartoon). As soil organisms that exist in a world of bacteria, fungi, nematodes, and other potential predators, the Dictyostelia should also have a rich chemical repertoire. The genome of Dictyostelium discoideum is exceptionally rich in polyketide synthase genes (PKS) encoded by more then 40 recognizable genes spread on all six chromosomes. At least one of the PKSs is dedicated to the synthesis of Differentiation Inducing Factor, as shown by Rob Kay and his collaborators. We know of no other organisms with as many PKS genes. We are using bioinformatics and gene expression approaches to confirm existing and to detect novel PKS genes. This sequence and expression information was used to ask how the PKS genes of Dictyostelium evolved. Similarity searches revealed all forty three known putative PKS genes present in the dictyBase. These were analyzed further using an approach based on hidden Markov models implemented within the HMMER program. Additional analyses showed that 3 dictyBase pairs of PKS genes (6 genes) could be combined into 3 single genes. In addition, 3 novel genes have been identified. Therefore, our analysis also revealed 43 PKS genes in total. PKS proteins are highly modular. Detailed annotation of all genes showed that they encoded β -ketoacylsynthase (KS), acyltransferase (AT), dehydratase (DH), methyltransferase (MT), enoylreductase (ER), ketoreductase (KR) and acyl carrier protein (ACP) domains expected in a polyketide synthase. In about half of the genes, there is an additional domain translated downstream of the ACP. In two genes, these domains show homology to chalcone synthases, which creates a novel structure. Thirty-two of the PKS genes occur in thirteen potential gene-clusters (containing 2-5 genes each). At present we do not know whether any of the putative clustered PKS genes represent an individual biosynthetic pathway. Twelve single genes were found and they resemble Type I iterative genes. We have constructed a phylogenetic tree by neighbor joining of the KS domain sequences. This showed that the D. discoideum sequences formed a separate but very diverse group with distances between extreme members nearly as great as the distance to bacterial and fungal sequences. Many of the retrieved PKS genes are transcribed. We found expressed sequence tags for many PKS genes in cDNA libraries and we have recovered information on transcription of others by examining gene array experiments. Real time PCR techniques using primers to the 3' ends of the genes were performed with 10 selected PKS genes to obtain their time course of expression. All ten genes tested showed different expression patterns during growth and development.

Modular polyketides and nonribosomal peptides are very important compounds for the pharmaceutical and agricultural industries. Their biosynthesis involves assembly of simple chemical building blocks into complex chemical structures by the catalytic activity of polyketide synthases (PKS) or nonribosomal peptide synthetases (NRPS). There has been a lot of interest in the last few years in generating new compounds for the production of novel drugs by manipulating the programming of such clusters in vitro (e.g. the idea of combinatorial biosynthesis). However, an important barrier to the progress is the fact that most changes made by in vitro methods result in very low yields or no detectable product. A possible solution to the yield problem would be the generation of novel clusters by homologous recombination system in vivo, because this would favour more closely related sequences and should reduce problems caused by incompatible junctions. An integral generic program package was developed to model this process in silico. This utilises a specially structured database, which links genetic structure to biosynthetic order, and contains a suite of computer programs to model the processes of the homologous recombination and predict the products of the recombinant clusters. There are also component programs to recognise and annotate new gene clusters from microbial genome sequencing projects or in gene-clusters present of the metagenomes of soil and/or marine microorganisms. The integral generic program package allows in silico generation of databases of novel chemical compounds that can be used for in silico screening using CDD technology.

The social amoeba (cellular slime mould) Dictyostelium discoideum is an important model organism for studies in development, cell communication and recently evolution. Yet, one aspect of the life of these organisms has been ignored: like other soil micro-organisms they should produce a complex chemical repertoire. The genome sequence was recently completed showing around 12, 500 genes in a relatively small genome of 34 Mb. We have been analysing the potential polyketide synthase (PKS) clusters using bioinformatics methods. Similarity searches revealed 43 potential PKS genes and these were analysed further using an approach based on HMMER profiles. Twenty nine of the PKS genes occur in 12 potential gene-clusters (containing 2-5 genes each), which presumably correspond to the biosynthesis pathways of different natural products. Most of the genes are represented in EST libraries or microarray signals indicating that the pathways are active. RT-PCR studies suggest that the genes tested are transcribed and also that they have different patterns of regulation during the elaborate development of the organism. The genes show a similar domain structure to bacterial Type I PKS genes except that they contain a conserved additional 800 amino acid sequence of unknown function, which does not show any homology to any known protein. However, no loading domains could be identified, which suggests there might be a novel initiation mechanism. Twelve of the genes resemble Type I iterative genes, but no Type II iterative genes were found.

Many phenotypic characteristics that are associated with differentiation and secondary metabolism are genetically unstable, including aerial mycelium formation, pigmentation, sporulation, resistance to antibiotics, and/or antibiotic production. Genetic instability in Streptomyces is a very common phenomenon with spontaneous mutants occurring at frequencies higher than 10-3. S. rimosus, the oxytetracycline (OTC) producer, is not an exception. However, higher frequencies (up to 100%) of instability at the genome level are induced by genetic modification of the otc genes of S rimosus strains, causing the frequency of occurrence of gene rearrangements to be close to 100%. The studies on the genetics of OTC biosynthesis in S. rimosus have been severely hampered by genetic rearrangements. Therefore, we have considered various approaches to avoid or at least reduce this problem. The transfer of entire otc cluster into the genetically amenable host S. lividans reduced the gene rearrangement almost entirely. Unfortunately, the yield of the OTC produced was reduced up to 100-fold. The compound(s) produced by the genetically engineered S. lividans strain produced only a trace amount of the desired products thus making the compound purification and structure determination work extremely difficult. Despite the observed gene instability induced by genetic manipulations, the over-producing S. rimosus strains seemed to be the best system to study OTC biosynthesis. In our attempt to reduce gene instability, we have employed various gene disruption and replacement strategies using streptomycete replicative vectors, E. coli non-replicative vectors ("suicide disruption") and linearised double and single strand constructs. The frequency of plasmid DNA transformation (integration) and the DNA rearrangement patterns varied according to the method applied. This work has allowed us to asses the best approach to reduce the gene instability in order to isolate the desired mutants.

Mathematical models were applied to define the behaviour kinetics distinction among microbial strains. In the first series of experiments the growth kinetics of microbial colonies of several S. rimosus mutant strains cultivated on agar plates were compared. Later, the interest was focused on the chosen two strains, in order to express mathematically their differences with respect to their colony growth and antibiotic biosynthesis kinetics. Finally, the behaviour of selected three S. rimosus derivative strains at different culture conditions was subjected to the study, with an aim to define strain distinction parameters. Mathematical models based on the three-dimensional growth concept and describing the microorganism growth, substrate uptake and antibiotic biosynthesis kinetics were developed. The computer simulation was applied to verify the applicability of mathematical models. The excellent agreement of computer simulation with experimental data confirmed the hypothesis that the kinetics parameters can be successfully applied to define the behaviour distinction among different S. rimosus strains. In the case of selected three strains, S. rimosus R6-500, S. rimosus MV9R-1 and MV9R-2, it was established that they can be distinguished by their growth kinetics parameters, their substrate uptake kinetics parameters and their antibiotic biosynthesis kinetics parameters.The strain S. rimosus R6-500 showed to be superior with respect to all kinetics parameters, the strain S. rimosus MV9R-2 showed to be slightly inferior to it, whereas the strain S. rimosus MV9R-1 showed to be inferior with respect to the both mentioned strains, especially because it showed the pronounced active biomass reduction rate at all investigated culture conditions. Based on these and the corresponding previous results one can conclude that appropriate mathematical models can be recommended to be used in defining parameters of microbial behaviour distinction among different microbial strains of S. rimosus species.

Recent DNA sequencing projects (e.g. genome sequences) are yielding a large number of new potential PKS and NRPS modular clusters whose functionality and products are unknown. A program has been written to scan the EMBL database using BLAST to identify potential clusters. This program is run regularly and used to identify new clusters that have been added to the EMBL database. The new clusters are downloaded and screened for potential modules using further programs. The output is then screened by eye to decide if a complete cluster is likely to be present. Further programs analyse these potential clusters to identify domains. By using models for the structure of typical modules it is possible in most cases to assemble the domains into modules. Operator intervention is necessary if atypical module structures occur. The final program enters the newly annotated clusters into the database. Clusters are only entered after a human decision that the annotation is appropriate. This system will allow the rapid addition of novel sequences that have not been adequately annotated previously. There will be a computer demonstration of programs.

The DNA sequences of many modular biosynthesis clusters are known. Most clusters contain several synthetase genes and the order in which the polypeptides are used in the biosynthesis pathway cannot yet be deduced from the DNA sequence alone. A database of modular clusters was developed that included a description linking the order of the genes to the biosynthetic order. The database was developed in MySQL allowing the easy implementation of interfaces for different purposes. The database is designed to have complete information about all the well-characterised clusters as well as containing the many potential clusters deduced from DNA sequencing, where the product (if any) is not known. This database will be useful in a variety of studies including the in silico prediction of the products of homologous recombination between biosynthesis clusters. There will be a computer demonstration of the database.

Determining function of a protein from its primary structure is among the most important problems in molecular biology. The usual approach – inferring function from homology – requires detailed description of functional units, considerable amount of data and is computationally rather demanding. Instead, we describe here a simple method for dividing a family of proteins into two functional subfamilies that is based on an underlying model of evolution. More precisely, we consider the alignment of a protein family and choose the sites in the alignment that strongly prefer bimodal evolutionary model over the unimodal one. In order to obtain two functional groups, we combine the information from these sites by an unsupervised learning algorithm. The method is demonstrated by analysing families of AT and KR domains from type I polyketide synthases.

Polyketides (PK) and non-ribosomal peptides (NRP) are large families of biologically active compounds (e.g. the immunosupressants rapamycin and cyclosporin, the antibiotics erythromycin and penicillin). The complex enzymes that produce these substances - PK synthases and NRP synthetases - are divided into modules, which, in turn, consist of clearly defined functional units, called domains. Moreover, knowing the function of each domain enables one to determine the resulting compound. Genome sequencing projects of several PKS and NRPS hosts have revealed a large number of PKS and NRPS gene-clusters with unknown domain functions. It is of great interest to predict the function of these domains and scan the resulting compounds for biological activity or combinatorial biosynthesis potential. The prediction is obtained by considering a probabilistic model of the multiple alignment of domains in question, and combining it with a hidden Markov model (HMM)-based unsupervised learning method. This yields a division of the family of domains into several functional subfamilies. For example, we were able to predict two major functional groups for substrate recognition among acyltransferases, one of them showing further subdivision into two subfamilies. We can also recognise two major functional groups among ketoreductases with putative different stereochemistry during the ketoreductase reaction. Further applications of the method will be discussed.

Modular clusters for the biosynthesis of polyketides (PKS) and non-ribosomal peptide synthesis (NRPS) are common and include many commercially important products. Each step of the biosynthesis is encoded by a different module, which are usually present on several polypeptides. It is likely that the observed diversity of clusters has evolved by processes including homologous recombination and the generation of novel clusters by homologous recombination is a potential route to novel chemical structures. A BLAST algorithm was used to model recombination by selection of sites of local similarity. This showed many potential recombination sites in the PKS clusters, which were mostly present in the ketosynthase domains of the modules. A range of BLAST parameters was used and it was shown that the prediction of the modules that recombine was not very sensitive to changes in stringency. The modules could be grouped into several discrete classes on the basis of their predicted recombination partners. A program was also written to predict the linear backbone structure of the novel polyketides produced by the recombinant clusters. In contrast to the PKS clusters, the NRPS clusters did not show many potential recombination sites.

DNA sequences of microorganisms that are important in food production have been accumulating rapidly. This includes genome sequences of food grade organisms such as brewers' and bakers' yeast Saccharomyces cerevisiae or Lactococcus lactis used in dairy industry. Secondary products produced by Streptomyces and related genera are also important in food production in fighting infections in animals, fish and plants (e.g. tetracycline from S. aureofaciens). As well as complete genome sequences there are also many sequences of secondary metabolite biosynthesis clusters. The chromosomes of Streptomycetes are linear molecules of about 8.5 Mb in size, which makes them considerably larger than those in most other bacteria. Analysis of genomic sequences suggests that there are many clusters coding for secondary metabolites, which have not yet been detected. Bioinformatics can help in accessing the biodiversity in this group of organisms. The most important group of chemically diverse and biologically active Streptomyces compounds are the complex polyketides synthesized by Type I polyketide synthases. These multi-functional enzymes with a modular organisation share considerable DNA homology allowing prediction of polyketide structure from the DNA sequence and genetic manipulation of individual modules. The major approach that has been used up till now is targeted manipulation, e.g. disruption and replacement of certain modules involved in the biosynthetic pathways. Our work in S. rimosus showed that there is frequent recombination between the chromosome end and the linear plasmid present in the host cells. This suggested a general strategy for obtaining recombinants between two polyketide biosynthesis clusters. We are developing computer programs to model recombination between modular polyketide gene-clusters. These programs should be useful tools for predicting novel polyketide structures in silico that might then be produced by an appropriate genetics manipulation in the laboratory.

x

Most chromosomal and linear plasmid ends in Streptomyces are similar and share terminal proteins. In S. rochei the different copies of terminal protein genes (tpg) that differ slightly in sequence seem to be used interchangeably and the pSLA2 ends can be replicated in S. lividans 66 and S. coelicolor A3(2) that have slight differences in the terminal proteins from S. rochei. It seems likely that the two ends of an element may interact with each other &#8211 ; ; the circular genetic maps of the chromosome may be due to this. One would expect that the linear plasmid DNA is negatively supercoiled in the cell like other DNA and this would be most easily achieved by the end proteins interacting. This may prevent stable existence of a plasmid with two totally different ends. There are at least three candidates for such elements in S. rimosus: pPZG102 does not have long IRs [1] and elements (pPZG103 and chromosome of MV25) formed by single cross overs between pPZG101 and the chromosome does not have much homology at the ends [2]. The end fragments of pPZG101, pPZG102 and the chromosome were cloned and sequenced. They only differ by a few bp from each other and fit the consensus for the other ends. Thus, there is no reason that they need different Tpgs. It is striking that the changes in the early palindromes are such that pairing would be conserved. This suggests that the secondary structure is important. The AseI patterns of the chromosomes in R6-65 and R6-501 are indistinguishable as are the macrorestriction maps of one half of pPZG101 and one end of pPZG102. Surprisingly, in both cases the terminal fragments showed small differences in size. Sequencing showed that these differences were not in the terminal 200 bp, but they are within 0.9 kb in the plasmid and 4.7 kb in the chromosome. This indicates that there is genetic lability associated with the ends of linear elements. DNA sequencing will show the nature of the changes. Ideas for explaining lability: maybe DNA topology &#8211 ; ; the ends could often be in a non-supercoiled state during replication ; maybe the end protein can sometimes prime DNA synthesis past the single stranded region resulting in strand displacement and ssDNA regions. [1] Gravius, B., Glocker, D., Pigac, J., Pandža, K., Hranueli, D. and Cullum, J. 1994. Microbiol. 140: 2271-2277. [2] Pandza, S., Biuković, G., Paravić, A., Dadbin, A., Cullum, J. and Hranueli, D. 1998. Mol. Microbiol. 28: 1165-1176.

The explosive growth of DNA sequence information in the last few years has made it vital for all scientists working in bacterial genetics to acquire a basic knowledge of bioinformatics. As relatively modest amounts of sequence data will be generated in a typical laboratory and little specialised support is available, much of the analysis depends on publicly available sites on the world wide web. It is often important to detect the protein coding regions in a sequence and this typically uses a program based on statistical differences between coding and non-coding sequences such as GeneMark. Another important question is whether similar genes have been analysed before so that inferences about possible function can be drawn. This is done by similarity search programs such as BLAST. The critical interpretation of the results from such programs needs a certain knowledge of their mode of action and also input of biological knowledge. These problems will be illustrated by some examples which include: searching sequence databases for polyketide synthetases, analysis of cofactor genes interesting for vitamin C production in Gluconobacter oxydans and comparison of genome organisation between Streptomyces species.

New technologic, economic and social challenges

Historically, the Streptomycetes were classified as fungi, with which they share a common habitat and superficial appearance. Subsequently, they were classified correctly as bacteria - based on their histology, biochemical composition and genetics. Therefore, it came as a surprise to discover that members of the genus have a single linear chromosome, in contrast to most other bacterial species that have circular chromosomes. They thus share the common feature of linear chromosomes with the eukaryotes, although the exact details of chromosomal architecture may be different. Our experiments with two distantly-related species of Streptomyces showed that sequences from the terminal regions could be retained at low copy number in apparent deletion strains. In the case of S. lividans, it was shown that there was a free chromosome end present in the low copy number sequences. In S. rimosus the Class II mutants showed apparent loss of one chromosome end. The new junction bands for AseI and XbaI were weak in comparison to neighbouring bands. An explanation for this would be if two copies of the chromosome had fused in inverted orientation, as this would produce junction fragments of the observed sizes. The junction bands would be weak, because they are only present in one copy, whereas all other bands would be double bands and they would be distant from the two-replication origins. Circularisation of the chromosome in S. lividans and recombination in S. rimosus of two chromosomes to form a single fusion chromosome with large segments in inverted orientation would each produce a second smaller recombination product carrying both chromosome ends, but no known replication origin. As S. lividans and S. rimosus are among the most distantly related strains in the genus Streptomyces, it seems likely that these phenomena are general. The existence of a relatively stable diploid structure would make Streptomycetes a possible link in evolution between unicellular prokaryotes (Eubacteria) and mycelial lower eukaryotes, so understanding more about the biology of their genomes is important.

Molecular biology has allowed genes to be cloned, sequenced and analysed. In recent years, advances in the sequencing hardware and robotics have resulted in an explosion in the rate of data acquired. This has changed the approach of molecular biologists from traditional molecular genetics to functional Genomics implying sequencing of entire genomes. The new science of Bioinformatics is required to process and analyse the data coming from sequencing projects. The first genome to be sequenced was the genome of the human pathogen Haemophilus influenzae. Genomes of model organisms followed, including Escherichia coli, brewers yeast, the worm, the fruit fly and the first plant. From the applied point of view sequencing of industrial microorganisms and agricultural plants is important for their further breeding. Sequencing of human and veterinary pathogens is also vital for the validation of new biological targets for the generation of novel antiinfectives. However, the most important of all is the sequencing of the human genome consisting of three billion nucleotides, whose first draft is now completed. This breakthrough will undoubtedly bring human beings the highest reward. Now that most of the 30, 000 or so genes that make up the human genome have been decoded, the biggest issue for genomics today is no longer genes. In a sense, at this very moment, they are also becoming passé. What is interesting now is what we will do with the genes. They are turned on or off at different times according to the tissue they are in and their role in the body. Therefore, the attention has to be paid to the transcriptome - to the mRNAs being produced by a cell at any given time - and the proteome, all the proteins being made according to the instructions in those mRNAs. The new scientific field Proteomics is emerging. It seems that Genomics and Proteomics, coupled with Bioinformatics, will change the world. Many discovery steps that are nowdays done in the wet labs will be done in silico, although conformation in the laboratory will remain essential. Hence, all will depend on brainware, not hardware or software.

Poliketidi prokariotskih i eukariotskih mikroorganizama najveća su skupina kemijskih supstancija među sekundarnim metabolitima. Mnoge od tih supstancija imaju antibakterijska, antifungalna, antitumorska, antihelmintička, imunosupresivna, hipolipemička ili druga svojstva. Taj je raznovrsni spektar bioloških aktivnosti neposredan odraz raznolikosti kemijskih struktura poliketida. Iako su poliketidi vrlo različitih kemijskih struktura, oni u osnovi imaju jedinstven biosintetski put. Biosintetski put započinje uzastopnom kondenzacijom jednostavnih građevnih jedinica (karboksilnih kiselina) i dovodi do izgradnje ugljikovih lanaca određenih dužina. Ravnolančani kondenzati zatim podliježu ciklizaciji, aromatizaciji i postkondenzacijskim izmjenama. Do sada je prikupljen veliki broj podataka koji upućuje na veliku sličnost između biosinteze ravnolančanih masnih kiselina (katalitičkim djelovanjem 'sintaza masnih kiselina') i biosinteze poliketida (katalitičkim djelovanjem 'poliketid sintaza'). Saznanja molekularne genetike o biosintezi poliketida u aktinomiceta usmjerio je zanimanje znanstvenika prema istraživanjima kojima se otkriva na koji su način genetički programirane različite poliketid-sintaze, te prema smišljenom oblikovanju novih (tzv. 'hibridnih') poliketidnih supstancija metodama genetičkog inženjerstva biosintetskih puteva. Do sada je klonirano i sekvencirano više od 50 genskih nakupina čiji su produkti odgovorni za biosintezu poliketida. Sekvenciranje genskih nakupina pokazalo je da među njima postoji bitna podudarnost na razini DNA te da, prema tome, vjerojatno potječu od zajedničkog pretka. Ti rezultati upućuju na mogućnost primjene kombinatorne biosinteze u kreiranju novih poliketidnih struktura. Biosinteza novih hibridnih poliketida, od kojih se očekuju nove biološki aktivne supstancije (tj. novi terapeutici), može se postići genetičkom manipulacijom nakupina gena što sudjeluju u sintezi prije kondenzacije, a i nakupina gena čiji produkti sudjeluju u postkondenzacijskim izmjenama poliketidnih struktura. U ovom su pregledu prikazane najnovije svjetske spoznaje i dostignuća s područja kombinatorne biosinteze poliketida.

Polyketides produced by actinomycetes are of great industrial importance (e.g. avermectin from Streptomyces avermitilis, oxytetracycline from Streptomyces rimosus, rifamycin from Amycolatopsis mediterranei). Some polyketide synthetases are of the Type I modular form (e.g. avermectin, rifamycin) in which each extension step is mediated by its own module. In Type II polyketide synthetases (e.g. oxytetracycline) the enzyme is used iteratively. In both cases, genetic manipulation of the biosynthetic clusters can produce novel polyketides. Because of the large size of the clusters (typically 30-150 kb) this is usually achieved by introducing constructs made by in vitro genetics into the clusters by homologous recombination. Many industrially important strains give problems for genetic manipulation including low transformation frequencies, the possession of methylation-dependent restriction systems, the failure of common cloning vectors to replicate in the strain, resistance to antibiotics commonly used in cloning vectors and the occurrence of genetic instability. Instead of protoplast transformation it is possible to use electroporation or conjugation from E. coli to introduce DNA into some species. One strategy to select the required recombination events is the use of counterselectable markers and linear plasmid vectors should be useful in future.

The strain S. rimosus R6-501 has a linear chromosome of 8 Mb and a linear plasmid pPZG101 of 387 kb. The ends of linear DNA molecules in Streptomyces usually possess covalently bound proteins on the 5' ends. A procedure was developed to isolate protein-bound DNA fragments in quantities sufficient for visualisation by agarose gel electrophoresis. When this was carried out with S. rimosus R6-501, in addition to the end fragments from the chromosome and pPZG101, a further unexpected fragment was detected and cloned. The additional fragment was due to the presence of a 4 kb linear element, which had not been detected previously, because it runs out of pulsed-field gels if standard programs designed for larger DNA fragments are used. An isolation method based on binding to glass beads allowed the preparative isolation of the element with its terminal protein intact. This is the first example of such isolation of intact linear plasmid from Streptomyces. This is important for analysing the replication of the ends and the development of novel cloning vectors.

OtcC encodes a hydroxylase that acts at the 6-position of the tetracyclic nucleus, after it has beenformed. OtcC is located at the edge of the otc cluster, distant (and transcribed separately) from the genes for biosynthesis of the polyketide backbone. Disruption of otcC resulted, surprisingly, in a series of incorrectly-folded shorter polyketides (we had expected the tetracyclic structure without a 6-hydroxyl, which has been characterized previously in mutants studied in the 1960's). When a catalytically-inactive mutant version of otcC was used to replace this disrupted copy, the ability to make tetracyclic structures was restored. We hypothesize that OtcC is required as a structural component at the quaternary level to ensure fidelity of the polyketide synthase complex in generating the tetracyclic nucleus.

Streptomyces species and related genera synthesize a large number of secondary metabolites, many of which are biologically-active. Amongst them, polyketides is the largest class. Polyketides are a structurally-diverse family of natural products with a broad range of biological activities. The formation of polyketides is very similar to biosynthesis of long chain fatty acids - both in the enzymatic reactions that take place and the enzyme proteins that are involved. During the last decade many polyketide gene-clusters have been cloned and sequenced. DNA sequencing has shown that the clusters have substantial homology suggesting that they originated from a common ancestor. This similarity has resulted in development of combinatorial biology techniques to create novel chemical entities. Two approaches have been used: targeted manipulation, e.g. disruption and, often, replacement of certain genes involved in the biosynthetic pathway, and the random approach, e.g. 'gene shuffling'. A targeted approach has been used to generate several novel scaffolds by manipulation of the S. rimosus oxytetracycline gene-cluster. Genes encoding ketosynthase, a and b, ketoreductase, cyclase/aromatase and C-6 hydroxylase were disrupted to construct four recombinant strains. Thin-layer-chromatography and high-liquid-pressure-chromatography of extracts from fermentation broths of all four recombinants showed that they produce about 20 potentially novel metabolites of which 12 have been characterised chemically. In this review, data on the disruption and replacement of the otcC gene will be described in more detail.

Genetic instability is very common in Streptomyces species. The phenotypic and genetic bases of instability are reviewed and the question of whether there is a common mechanism for all forms of instability is discussed. The phenomenon of "reversion" of deletion mutations is considered in the context of a model that distinguishes the genetic content of mycelium and spores.

Streptomyces species and related genera synthesize a large number of secondary metabolites many of which are biologically active. Among them polyketides are the largest class. Polyketides are a structurally diverse family of natural products with a broad range of biological activities and pharmacological properties. Polyketide antibiotics, antifungals, antiparasitics, animal growth promotants, natural insecticides, cytostatics, immunosuppressants and anticholesterolemics are in com-mercial use. The evidences accumulated up till now have indicated a substantial analogy between the formation of long chain fatty acids and biosynthesis of polyketides carried out by Type I and Type II polyketide synthases (Hopwood, Chem. Rev., 97, 2465, 1997). During the last decade many polyketide gene-clusters have been cloned and sequenced. DNA sequencing showed considerable DNA homology hence they originated from the same ancestor. This allowed the use of combinatorial biology in creation of novel chemical entities. Combinatorial biology aims to create biodiversity by engineering the way the natural molecules are formed to generate "natural" substances that Nature never imagined. In principal, two approaches have been used: targeted manipulation (e.g. disruption and replacement of certain genes involved in the biosynthetic pathway ; McDaniel et al., Proc. Natl. Acad. Sci. USA, 96, 1846, 1999) and random approach (e.g. gene shuffling technology ; Del Cardayre et al., WO 00/4190). In PLIVA, the targeted approach that generated several novel scaffolds by the manipu-lation of the S. rimosus oxytetreacycline gene-cluster has been used (Petković et al., J. Biol. Chem., 274, 32829, 1999 ; Perić et al., P-990153A). In this talk the latest data from this field will be presented.

Streptomyces rimosus synthesise 19-carbon polyketide antibiotic oxytetracyline (OTC) by a 'Type II' polyketide synthase (PKS) complex. OTC biosynthetic gene-cluster occupies a contiguous 34 kb segment on the S. rimosus chromosome. One open reading frame, designated otcC, was showen to encode a hydroxylase involved in hydroxylation of C-6 of the OTC backbone. The disruption of otcC gene by the gentamicin (gmr) resistance cassette resulted in the strain producing polyketides of shorter chain length (17 C atoms). Therefore, disruption leads in 'derailment' of the ability of the OTC PKS to form a chain of the usual length. The otcC gene modified by in vitro site-directed mutagenesis was used to replace the disrupted gene. Residues that were predicted to be essential for C-6 hydroxylase specificity of OtcC were altered. Structure determination of the product made by this new strain revealed that it confers OTC production. This raises questions concerning either the specificity of the mutation or of the specificity of OtcC itself.

In spite of the introduction of novel technologies in pharmaceutical research (e.g. genomics, high throughput screening, combinatorial and parallel synthesis) the success of these programs is lagging behind the high expectations. This may be partially due to the lack of novel compound classes to be screened for interaction with various targets. Therefore, the interest in natural compound pools has recently increased again. The significant advances in our understanding of the biosynthetic machinery of complex compound classes such as polyketides has led to the introduction of combinatorial genetics, allowing the modification of natural pathways. Targeted manipulation, e.g. of the polyketide synthetases type I (reduced complex polyketides, such as macrolides) and type II (polyaromatic polyketides, such as tetracyclines and anthracyclines), has led to the generation of novel compounds with significant structural diversity as compared to the natural parental compound. In principal the approaches can be divided into targeted and random. The targeted approach allows precise tailoring of known scaffold. However, it requires substantial workload and most often the compounds turn out to be weakly active or inactive for the desired purpose. On the other hand, random approaches such as gene shuffling technology, although aimed for broad derivatisation, are so far limited for the use on a gene level. The present technologies can't be applied on gene clusters or complex pathways. To overcome this bottleneck we are currently evaluating the random approach, which enables the broadest recombination based on the inhibition of mismatch repair system (MRS). MRS consists of various proteins that correct mismatches occurring during transcription and functions as a guardian of genomic integrity. Besides, it prevents recombination of diverged sequences. Inactivation of MRS enables recombination of genes, gene clusters or even the whole genomes that diverge more than 30-50% (isn't it too much?). In this way, it is possible to screen in vivo the generation of novel compounds by combining various pathways. We are now adapting this system for broad modification of secondary metabolite biosynthesis to identify new chemical entities for their application in life sciences rapidly and efficiently.

Petri-dishes containing agarised media layers are frequently applied in cultivating microorganisms for different research purposes. Determination of microbial specific activities and screening of microorganisms are very often the main goals of studies. However, a disadvantage of the use of common agarised media layers in Petri-dishes is that such layers do not enable the simple use of both layer sides. This disadvantage can be overcome by inserting the net of inert metal (e.g. stainless steel) into agarised medium layer, because in such a way the layer becomes fortified and therefore susceptible for manipulation. The convenience of fortified agarised media layers was investigated comparing antibiotic productivities of different derivatives of Streptomyces rimosus R6-500 strain. Sterile small circular stainless steel wire nets with wire supports at their brims were used to prepare the fortified sterile agarised media layers. After the fortified layers have been prepared, cultures of strains chosen to be studied were used to start the growth of their colonies. The transfer of cultures to inoculate the upper layer surface parts was performed by means of a sharp needle, glass rod or a pipette, depending on the design of the experiment. After the layers in Petri-dishes have been inoculated, their incubation at 28°C and/or 37°C for a given time was performed. Then the fortified layers were reversed by means of corresponding tweezers, in order to inoculate the opposite layer side with the test microorganism. After seeding the mentioned layer surface with the test microorganism, the dishes were incubated at the same temperature for 20 hours. Then all incubated agarised media layers were examined to evaluate the formed inhibition zones. The series of experiments was performed and the experimental results were analysed. The results showed that the differences between particular strains could be well defined. Because of their advantages the fortified media layers can be recommended for the application in screening procedures of streptomycetes. In experiments performed we observed that the investigated strains could be classified as follows: S.rimosus R6-500 > S.rimosus R6-ZGL 1 > S.rimosus R6-MV25W, if the potency index is accepted as the classification criterion.

Recent advances in the molecular genetics of Streptomyces have increased our understanding of polyketide antibiotic biosynthesis, to the point where recombinant DNA approaches to generate novel structures are possible. Our understanding of how antibiotic pathways are regulated and integrated into central metabolism also provides the opportunity for strain manipulation to enhance productivity.

Transposon mutagenesis is a very useful technique in bacterial genetics. It was planned to construct a transposon mutagenesis system for the oxytetracycline producer Streptomyces rimosus R6. The transposon Tn5424 (Irnich and Cullum, Biotech. Lett.,16:437-442,1994) had shown useful properties in Streptomyces lividans 66 so a new transposon delivery vector carrying an erythromycin resistance gene (pPZG200) was constructed for use in S. rimosus R6. Surprisingly all transposition events analysed proved to carry DNA amplifications in the chromosome which varied in size from 1.5 kb to 9.0 kb. The amplifications were characterised using restriction analysis and hybridisation. In order to identify the chromosomal regions involved in the DNA amplification a cosmid gene bank of S. rimosus R6 containing 1765 clones was constructed in the vector Supercos 1. The results of the analyses of the amplifications and the cosmids will be discussed.

S. rimosus has a linear chromosome of 8 Mb in size. In the "Zagreb strain" S. rimosus R6-65 there is a 320 kb linear plasmid pPZG102 which does not have any extensive terminal inverted repeats (Pandza et al., Mol. Microbiol., 28:1165, 1998). In a later derivative R6-500 in the strain development programme there is a 390 kb linear plasmid pPZG101 which has been derived from pPZG102 by the generation of long terminal repeats of about 180 kb and loss of about 110 kb of the pPZG102 sequences. Single cross overs between pPZG101 and the chromosome can generate chromosomes or plasmid primes carrying one chromosome end and one plasmid end. In order to study the structure of the plasmids in more detail it was necessary to characterise the ends. The ends of the linear chromosomes and linear plasmids in Streptomyces carry covalently bound proteins at the 5' ends. The ends can be isolated from DNA preparations using protein-binding materials such as glass milk (Huang et al., Mol Microbiol., 28:905, 1998). The isolation procedure was optimised to allow visualisation of terminal restriction fragments in ethidium-bromide stained agarose gels. In R6-65 three XhoI fragments of 4.7, 2.3 and 0.9 kb were observed. Southern hybridisation experiments identified the 4.7 kb fragment as the chromosome end. The other two fragments appear to correspond to the two ends of pPZG102. The lack of cross-hybridisation with the chromosome terminal fragment showed that despite the ability for stable replication of molecules carrying one chromosome and one plasmid end there is little conserved DNA homology between the ends.