Biosensor-Based Optimization of Cutinase Secretion by Corynebacterium glutamicum

The industrial microbe Corynebacterium glutamicum is gaining substantial importance as a platform host for recombinant protein secretion. We recently developed a fluorescence-based (eYFP) C. glutamicum reporter strain for the quantification of Sec-dependent protein secretion by monitoring the secretion-related stress response and now demonstrate its applicability in optimizing the secretion of the heterologous enzyme cutinase from Fusarium solani pisi. To drive secretion, either the poor-performing PelSP or the potent NprESP Sec signal peptide from Bacillus subtilis was used. To enable easy detection and quantification of the secreted cutinase we implemented the split green fluorescent protein (GFP) assay, which relies on the GFP11-tag fused to the C-terminus of the cutinase, which can complement a truncated GFP thereby reconstituting its fluorescence. The reporter strain was transformed with different mutant libraries created by error-prone PCR, which covered the region of the signal peptide and the N-terminus of the cutinase. Fluorescence-activated cell sorting (FACS) was performed to isolate cells that show increased fluorescence in response to increased protein secretion stress. Five PelSP variants were identified that showed a 4- to 6-fold increase in the amount and activity of the secreted cutinase (up to 4,100 U/L), whereas two improved NprESP variants were identified that showed a ∼35% increase in secretion, achieving ∼5,500 U/L. Most of the isolated variants carried mutations in the h-region of the signal peptide that increased its overall hydrophobicity. Using site-directed mutagenesis it was shown that the combined mutations F11I and P16S within the hydrophobic core of the PelSP are sufficient to boost cutinase secretion in batch cultivations to the same level as achieved by the NprESP. Screening of a PelSP mutant library in addition resulted in the identification of a cutinase variant with an increased specific activity, which was attributed to the mutation A85V located within the substrate-binding region. Taken together the biosensor-based optimization approach resulted in a substantial improvement of cutinase secretion by C. glutamicum, and therefore represents a valuable tool that can be applied to any secretory protein of interest.

A secretion biosensor for monitoring Sec-dependent protein export in Corynebacterium glutamicum

BACKGROUND

In recent years, the industrial workhorse Corynebacterium glutamicum has gained increasing interest as a host organism for the secretory production of heterologous proteins. Generally, the yield of a target protein in the culture supernatant depends on a multitude of interdependent biological and bioprocess parameters which have to be optimized. So far, the monitoring of such optimization processes depends on the availability of a direct assay for the respective target protein that can be handled also in high throughput approaches. Since simple assays, such as standard enzymatic activity assays, are not always at hand, the availability of a general protein secretion biosensor is highly desirable.

RESULTS

High level secretion of proteins via the Sec protein export pathway leads to secretion stress, a phenomenon that is thought to be caused by the accumulation of incompletely or misfolded proteins at the membrane-cell envelope interface. We have analyzed the transcriptional responses of C. glutamicum to the secretory production of two different heterologous proteins and found that, in both cases, the expression of the gene encoding a homologue of the extracytosolic HtrA protease was highly upregulated. Based on this finding, a C. glutamicum Sec secretion biosensor strain was constructed in which the htrA gene on the chromosome was replaced by the eyfp gene. The fluorescence of the resulting reporter strain responded to the secretion of different heterologous proteins (cutinase from Fusarium solani pisi and alkaline phosphatase PhoA from Escherichia coli) in a dose-dependent manner. In addition, three differently efficient signal peptides for the secretory production of the cutinase could be differentiated by the biosensor signal. Furthermore, we have shown that an efficient signal peptide can be separated from a poor signal peptide by using the biosensor signal of the respective cells in fluorescence activated cell sorting experiments.

CONCLUSIONS

We have succeeded in the construction of a C. glutamicum biosensor strain that allows for the monitoring of Sec-dependent secretion of heterologous proteins in a dose-dependent manner, independent of a direct assay for the desired target protein.