VALIDATION OF A METHOD FOR AUTOMATED NUCLEOTIDE-SEQUENCE ANALYSIS ANDESTIMATION OF THE LIMITS OF DETECTION OF VARIANT SEQUENCES IN A HETEROGENIC DNA SAMPLE

The assessment of the genetic stability of an expression-system used f or the production of recombinant drugs and the verification of the nuc leotide sequence of expression plasmids makes it necessary to estimate a limit of detection for variant sequences. Therefore, a series of ex periments was performed in which defined mixtures of two different pla smid-DNAs were subjected to nucleotide sequence analysis using an auto matic sequencer (ABI 377) and PCR-based sequencing reactions. A plasmi d pGR201 (Fig. 1a), encoding recombinant non-glycosylated human prouro kinase (Saruplase), was spiked in varying ratios (90:10; 80:20; 70:30; 60:40) with a related plasmid, pSJ41 (Fig. 1a), differing in a single base-substitution and a cluster of multiple mutations, respectively. Each of these mixtures was sequenced six times in a 50 bp-region cover ing tile variant sequences in order to estimate a limit of detection f or point mutations and clusters of multiple mutations. The results are of general interest since we analysed enough point-mutations as well as deletion-mutations to allow a statistical assessment. It was shown that a single point mutation resulting in a base-substitution will be detected by sequencing in one direction if at least 20 % of the plasmi ds within the DNA sample are mutated, while the sequencing method is s ensitive enough to detect a cluster of multiple mutations when at leas t 10 % of the sequenced plasmid DNA molecules carry the altered sequen ce. In contrast to other publications the sequencing-procedure present ed in this paper results in electropherograms with small background si gnals although only a very simple purification of the sequencing react ion products was used in order to present a method that is not time-co nsuming.