Failed transfection studies lead to discovery of metabolic co-operation Discussions arising from chance meetings at coffee time can sometimes bring together people and areas of research that would otherwise not interact. In 1962, a paper by Szybalska and Szybalski (1) generated such a discussion between Bobby Bu ยจrk, John Subak-Sharpe, and myself in the Institute of Virology, Glasgow, and, perhaps unusually, led to a series of experiments. The experiments failed in their intended aim but resulted in the accidental discovery of metabolic co-operation, the first evidence for the intercellular exchange of metabolites through what are now thought to be gap junctions.
The Szybalskis' paper provided the first detailed description of DNA-mediated transformation of animal cells. Using DNA from wild-type cells to transform HPRT -cells, they measured transformation frequencies by survival in HAT medium (containing hypoxanthine, aminopterin, and thymidine, which selects for HPRT ฯฉ cells). Transformation frequency was unexpectedly high, reaching a maximum of 4 ฯซ 10 -4 . Our discussion centred on this value, which was much higher than that typically found in bacterial systems, even though the number of copies of any gene per milligram of bacterial DNA is about a 1,000-fold higher. It seemed surprising that such a potentially powerful system for studying mammalian genetics had not been exploited.
John Subak-Sharpe had recently isolated HPRT -variants of transformed BHK 21 hamster cells, (2) and I had been using autoradiography to examine stimulation of host cell DNA synthesis by polyomavirus. We realised that we could use autoradiography to provide a faster assay for transformation frequency than the clonal selection used by the Szybalskis. After treatment of HPRT -cells with wild-type DNA, addition of 3 H hypoxanthine and autoradiography should allow black (covered with silver grains) HPRT ฯฉ cells to be identified against a background of unlabelled HPRT -cells. It should, we thought, be possible to detect even one black cell per dish. However, despite several attempts, no black, HPRT ฯฉ cells could be found. Either transformation was less frequent and