New technologies let researchers think outside the square

Just one in-house technology is not enough, if you're a company that's always looking for new ways to boost your drug discovery and manufacturing capabilities, delegates to this week's AusBiotech conference were told.

Dr Michael Breuer, from German manufacturing giant BASF, described how metabolic engineering and isolation of novel enzymes from biodiverse environmental samples were allowing the company to improve its enzyme-based vitamin and small molecule drug manufacturing capabilities.

"It's definitely not enough to have just one technology in-house," he said, explaining that scientists at BASF were using a number of functional genomic and proteomic techniques to improve biotechnology processes.

Breuer's was one of several addresses at the conference that described a variety of innovative methods for integrating genomic data with function and phenotype.

One of BASF's more innovative techniques, he said, was to use biological processes to discriminate between and separate chiral products, molecules that are chemically identical but structurally mirror images of each other.

Breuer described a technique that uses enzymes to modify one form of the chiral molecule, also known as an enantiomer, while leaving the other unchanged. This allows the two forms to be separated.

He also explained how BASF screened biodiverse environmental samples for novel enzymes, in a process that used DNA isolated from a sample to prepare a library that is then screened for biocatalytic activity. In one screen of 50,000 clones, 45 novel lipases, 25 phosphatases, two proteases and 25 amylases were discovered. Breuer noted that all were novel, and originated from diverse taxonomic groups.

New mouse models

Also on the subject of functional genomics, Prof Chris Goodnow, from the Australian Phenomic Facility at the John Curtin School of Medical Research, described his technique of developing new mouse models for disease and drug discovery using the mammalian phenome.

"For every gene in our genome, there is a correlating gene in the mouse genome spelled almost the same way," he said.

He said genome-wide mutations were induced in mice at a frequency of one in 2000 bases, and then the animals were bred for homozygosity and to separate useful mutations. In this way, large numbers of mice were created that can be screened for useful phenotypes.

"We start with mutations, look at the phenotype and then go to the genes," Goodnow explained. "In this way you get clusters of hits that allow functional analysis of pathways."

The platform is being commercialised by Phenomix, a company resulting from a joint venture between Goodnow's team at ANU, Baylor College of Medicine in Texas and the Genomics Institute of the Novartis Research Foundation in San Diego.

The company has already identified three strains of mice with useful phenotypes. 'Munchkin' is a dwarf strain, 'Blobby' is an obese strain with type II diabetes and 'Hannibal' contains a mutation in a gene associated with lymphoma.

While Goodnow is focusing on mice as a model for human disease, Prof Simon Easteal, co-director of the Centre for BioInformation Science at ANU and the director of the Genetic Epidemiology unit at Tasmania's Menzies Research Institute, is using the rich source of information available in small human populations to look at complex diseases.

He noted that the Tasmanian population was excellent for genetic population studies, as it has good records extending back six to nine generations. The database currently has about 500,000 records, which are being used to examine a variety of diseases including multiple sclerosis, glaucoma and other multigenic diseases.

But Easteal said that the challenges of developing computational processes to analyse the data were very substantial.

Prof Limsoon Wong, from the Laboratories for Information Technology in Singapore, described some of the informatics approaches he had developed to facilitate searching and modelling of bioinformatics data. Among the software developed by his team is the Dragon Promoter Finder, which is used to predict transcription start sites in eukaryotic DNA.

"We need technology to extract data on protein-protein interactions," he said.