The length of ribosomal binding site spacer sequence controls the production yield for intracellular and secreted proteins by Bacillus subtilis

Abstract

Background: Bacillus subtilis is widely used for the industrial production of recombinant proteins, mainly due to its high secretion capacity, but higher production yields can be achieved only if bottlenecks are removed. To this end, a crucial process is translation initiation which takes place at the ribosome binding site enclosing the Shine Dalgarno sequence, the start codon of the target gene and a short spacer sequence in between. Here, we have studied the effects of varying spacer sequence lengths in vivo on the production yield of different intra- and extracellular proteins. Results: The shuttle vector pBSMul1 containing the strong constitutive promoter P HpaII and the optimal Shine Dalgarno sequence TAAGGAGG was used as a template to construct a series of vectors with spacer lengths varying from 4 to 12 adenosines. For the intracellular proteins GFPmut3 and β-glucuronidase, an increase of spacer lengths from 4 to 7-9 nucleotides resulted in a gradual increase of product yields up to 27-fold reaching a plateau for even longer spacers. The production of secreted proteins was tested with cutinase Cut and swollenin EXLX1 which were N-terminally fused to one of the Sec-dependent signal peptides SPPel, SPEpr or SPBsn. Again, longer spacer sequences resulted in up to tenfold increased yields of extracellular proteins. Fusions with signal peptides SPPel or SPBsn revealed the highest production yields with spacers of 7-10nt length. Remarkably, fusions with SPEpr resulted in a twofold lower production yield with 6 or 7nt spacers reaching a maximum with 10-12nt spacers. This pattern was observed for both secreted proteins fused to SPEpr indicating a dominant role also of the nucleotide sequence encoding the respective signal peptide for translation initiation. This conclusion was corroborated by RT qPCR revealing only slightly different amounts of transcript. Also, the effect of a putative alternative translation initiation site could be ruled out. Conclusion: Our results confirm the importance of the 5′ end sequence of a target gene for translation initiation. Optimizing production yields thus may require screenings for optimal spacer sequence lengths. In case of secreted proteins, the 5′ sequence encoding the signal peptide for Sec-depended secretion should also be considered.