Model-Driven Minimization of the B. Subtilis Genome

Bacillus subtilis is a gram positive, sporulating bacteria often utilized in industry as a producer of high quality enzymes and proteins [1]. The flexible growth conditions, rapid growth rate, and natural competence of B. subtilis make it an ideal candidate for scientific study and industrial use. Yet many aspects of B. subtilis are still poorly understood, and the complexity of this organism makes experimental analysis and reengineering difficult. To enhance our understanding of B. subtilis and develop a more malleable and productive strain for use in a wide variety of industrial applications, we are systematically deleting all regions of the B. subtilis chromosome that are dispensable for growth on rich defined media [2]. We have applied bioinformatics techniques [3] and experimental knowledge [4] to design a set of 157 intervals of the B. subtilis genome that are predicted to be dispensable. These intervals (which range from 2 to 130 Kb in size) include all but ~832 (20%) of the genes in the B. subtilis genome. The intervals were individually deleted at 37°C in rich defined medium. 135 intervals were found to be dispensable, with the remaining 21 being essential for viability. To identify the coessential genes responsible for the 21 nonviable deletions, these intervals were iteratively subdivided into 140 smaller intervals, with new coessential and essential genes in B. subtilis ultimately being identified. All 297 experimentally implemented deletions were simulated in silico using the iBsu1103 metabolic model of B. subtilis [5]. Initially, the model incorrectly predicted the outcome of 60 (20%) interval deletions. The model was subsequently adjusted using a combination of computational analysis and experimental study to correct nearly all erroneous predictions. In several cases, the model assisted in the interval splitting process by predicting coessential gene sets. The model also predicted additional nutrients that could be added to the rich defined media to restore viability for some unviable intervals. Phenotypes for the 135 successfully implemented deletions were also studied experimentally and computationally predicting and measuring growth in minimal media, rich defined media, and LB media. Insights gained from the initial 297 interval deletions and the subsequent combination of interval deletions into a 1.5 MB knockout mutant will be discussed. 1.) Zweers JC, Barak I, Becher D et al (2008) Towards the development of Bacillus subtilis as a cell factory for membrane proteins and protein complexes. Microb Cell Fact 7:10 2.) Fabret C, Ehrlich SD, and Noirot P (2002) A new mutation delivery system for genome-scale approaches in Bacillus subtilis. Mol Microbiol 46(1):25-36 3.) Overbeek R, Disz T, and Stevens R (2004) The SEED: A peer-to-peer environment for genome annotation. Communications of the ACM 47(11):46-51 4.) Kobayashi K, Ehrlich SD, Albertini A et al (2003) Essential Bacillus subtilis genes. Proc Natl Acad Sci U S A 100(8):4678-83 5.) Henry CS, Zinner J, Cohoon M et al (2009) iBsu1103: a new genome scale metabolic model of B. subtilis based on SEED annotations. Genome Biology 10:R69