Secretion stress caused by overexpression of heterologous α-amylase in Bacillus subtilis

Transcriptome analysis was used to investigate the global stress response of the Gram-positive bacterium Bacillus subtilis caused by overproduction of the well-secreted AmyQ α-amylase from Bacillus amyloliquefaciens. Analyses of the control and overproducing strains were carried out at the end of exponential growth and in stationary phase, when protein secretion from B. subtilis is optimal. Among the genes that showed increased expression were htrA and htrB, which are part of the CssRS regulon that responds to high-level protein secretion and heat stress. The analysis of the transcriptome profiles of a cssS mutant compared to the wild-type, under identical secretion stress conditions, revealed several genes with altered transcription in a CssRS-dependent manner, for example citM, ylxF, yloA, ykoJ and several genes of the flgB operon. However, a high affinity CssR-binding was only observed for htrA and htrB, and possibly for citM. In addition, the DNA macroarray approach reveal that several genes of the sporulation pathway are downregulated by AmyQ overexpression, and a group of motility-specific (σD-dependent) transcripts were clearly upregulated. Subsequent flow cytometric analyses demonstrate that upon overproduction of AmyQ as well as a non-secretable variant of the α-amylase, the process of sporulation is severely inhibited. The same experiments were implemented to investigate the expression levels of the hag promoter, a well-established reporter for σD-dependent gene expression. This approach confirmed the observations based on our DNA macroarray analyses and led us to conclude that expression levels of several genes involved in motility are maintained at high levels under all conditions of α-amylase overproduction. Keywords: secretion stress response Secretion stress was applied by overproducing the well-secreted AmyQ α-amylase (pKTH10 vector) from B. amyloliquefaciens. Besides examining secretion stress in wild-type cells, we compared transcriptome profiles of a cssS mutant strain under conditions of high-level AmyQ production. Samples for transcriptome analyses were collected at the late exponential growth stage (one hour before the transition point) and 3 hours upon entry in the stationary growth phase. Three independent cultures of each strain were used and cells were sampled for macroarray experiments. Duplicate spots were averaged in Array-Pro software (Media Cybernetics, Inc.) and the signal was normalized after background subtraction by calculation of the percentage of total signal per gene using Microsoft Excel.