Abstract

BCG, only vaccine available to prevent tuberculosis, was established in the early 20th century by prolonged passaging of a virulent clinical strain of Mycobacterium bovis. BCG Tokyo-172, originally distributed within Japan in 1924, is one of the currently used reference substrains for the vaccine. Recently, this substrain was reported to contain two spontaneously arising, heterogeneous subpopulations (Types I and II). The proportions of the subpopulations changed over time in both distributed seed lots and commercial lots. To maintain the homogeneity of live vaccines, such variations and subpopulational mutations in lots should be restrained and monitored. We incorporated deep sequencing techniques to validate such heterogeneity in lots of the BCG Tokyo-172 substrain without cloning. By bioinformatics analysis, we not only detected the two subpopulations but also detected two intrinsic variations within these populations. The intrinsic variants could be isolated from respective lots as colonies cultured on plate media, suggesting analyses incorporating deep sequencing techniques are powerful, valid tools to detect mutations in live bacterial vaccine lots. Our data showed that spontaneous mutations in BCG vaccines could be easily monitored by deep sequencing without direct isolation of variants, revealing the complex heterogeneity of BCG Tokyo-172 and its daughter lots currently in use.