Information Technology Reference
In-Depth Information
peptide synthases, abbreviated as PKSs and NRPSs respectively, catalyse condensations of
carboxylic acid and amino acid derivatives into polyketide and peptide structures like
precursors of Erythromycin or Penicillin, respectively. DNA sequencing of PKS and NRPS
gene-clusters showed that they are multi-functional enzymes with modular organisation.
Each module is responsible for a single cycle of polyketide and peptide chain extension and
contains catalytic domains for necessary ketoreduction, dehydratation and enoylreduction
as well as epimerisation, N-methylation and reduction activities. The last modules of both
multi-functional enzymes contain thioesterase domains responsible for the release of linear
chains from enzymes and their cyclisation. Therefore, there is a one-to-one correlation
between the product structure and the active domains in modular PKSs and NRPSs that
generate linear polyketide and peptide chains. This allows the prediction of polyketide and
peptide backbone structures from DNA sequences. PKSs and NRPSs share considerable
DNA homology between them so they must have originated from the same ancestors.
Therefore, the creation of directed changes in the backbone structures by genetic
manipulation of modules is possible. Major approaches that have been used up until now
are targeted manipulations, which are the disruptions, deletions or replacements of certain
catalytic domains or whole modules in the existing gene-clusters [see references in: 34 and
36].
There are a number of small biotechnology companies that use these approaches.
The approaches of two most important ones, Biotica Technology Ltd.
(http://www.biotica.co. uk/) and Kosan Biosciences Inc. (http://www.kosan.com/) are
illustrated to show what they do. With the introduction of restriction enzyme sites,
scientists at Kosan designed genetic cassettes that allowed relatively easy manipulation of
individual active domains by deletion, or insertions of active domains from other modular
PKSs. In complex polyketides like macrolides, Biotica and Kosan inactivated specific
enzyme active domains or inserted active domains from other clusters to generate novel
macrolides with differently reduced oxo groups that possess different stereochemistry. They
also deleted whole modules from multi-functional PKSs or used chemo-biosynthesis to
generate novel macrolides with smaller or larger polyketide backbones. Moreover, recently
the new structural class of polyketides (having 2,4-dioxa-adamantane ring-system) has been
isolated from an engineered
Streptomyces
strain, thus supporting the claim that
combinatorial biology is capable of producing novel chemotypes. However, all these
approaches are labour and time consuming, allowing the development of relatively small
libraries of novel polyketides and peptides [see references in: 34].
While studying the genome topology and genetic stability of
S. rimosus
, an
industrial producer of the antibiotic oxytetracycline, the frequent interaction between its
chromosome and linear plasmid present in the host cells was noticed in a number of strains
that had not been in any way selected. Genetic elements like plasmid pPZG103 and the
chromosome of the strain MV25W are formed by single crossover between the plasmid
pPZG101 and the chromosome end [37]. This suggested a general strategy for obtaining
recombinants between two polyketide biosynthesis clusters. The polyketide gene-cluster
could be cloned into a linear plasmid vector in between selectable and counter-selectable
gene-cassettes [38], introduced in host cells carrying the second gene-cluster cloned within
similar gene-cassettes near the end of the chromosome and selected for a single cross over
between them (22).
Bioinformatics can also help in accessing the biodiversity in this group of
organisms. Until 2000 only 19 polyketide gene-clusters for modular PKSs were cloned and
sequenced from a number of different species [34]. However, as it was already mentioned,
genome sequencing of
S. coelicolor
and
S. avermitilis
suggests that there are more than 40
presumed gene-clusters in these two species that code for secondary metabolites, which
have not yet been analysed [29, 30]. Bioinformatics tools can be used to annotate as yet
