Co-Expression of a Chimeric Protease Inhibitor Secreted by a Tumor-Targeted Salmonella Protects Therapeutic Proteins from Proteolytic Degradation

Abstract

Sunflower trypsin inhibitor (SFTI) is a 14-amino-acid bicyclic peptide that contains a single internal disulfide bond. We initially constructed chimeras of SFTI with N-terminal secretion signals from the Escherichia coli OmpA and Pseudomonas aeruginosa ToxA, but only detected small amounts of protease inhibition resulting from these constructs. A substantially higher degree of protease inhibition was detected from a C-terminal SFTI fusion with E. coli YebF, which radiated more than a centimeter from an individual colony of E. coli using a culture-based inhibitor assay. Inhibitory activity was further improved in YebF-SFTI fusions by the addition of a trypsin cleavage signal immediately upstream of SFTI, and resulted in production of a 14-amino-acid, disulfide-bonded SFTI free in the culture supernatant. To assess the potential of the secreted SFTI to protect the ability of a cytotoxic protein to kill tumor cells, we utilized a tumor-selective form of the Pseudomonas ToxA (OTG-PE38K) alone and expressed as a polycistronic construct with YebF-SFTI in the tumor-targeted Salmonella VNP20009. When we assessed the ability of toxin-containing culture supernatants to kill MDA-MB-468 breast cancer cells, the untreated OTG-PE38K was able to eliminate all detectable tumor cells, while pretreatment with trypsin resulted in the complete loss of anticancer cytotoxicity. However, when OTG-PE38K was co-expressed with YebF-SFTI, cytotoxicity was completely retained in the presence of trypsin. These data demonstrate SFTI chimeras are secreted in a functional form and that co-expression of protease inhibitors with therapeutic proteins by tumor-targeted bacteria has the potential to enhance the activity of therapeutic proteins by suppressing their degradation within a proteolytic environment.

Keywords: OTG-PE38K; Protease inhibitors; VNP20009; YebF; sunflower trypsin inhibitor (SFTI); tumor-targeted Salmonella.

Fig. 1.. Comparison of N-terminal hydrophobic secretion signal fusions with sunflower trypsin inhibitor (SFTI).

SFTI +/− the colicin E3 (ColE3) lysis protein was expressed as fusions with N-terminal hydrophobic secretion signals and as a polycistronic construct with a separate ribosomal binding site (RBS) in the pTet99a plasmid, expressed in E. coli strain JW2197–1. Left panel showing the DNA constructs. (1) The empty vector; no insert cloned into the pTet99a expression site (pDB221). (2) The SFTI DNA sequence fused to the signal sequence (SS) peptide derived from OmpA; OmpA-SFTI (pDQ71). (3) SFTI fused with the ToxA SS; ToxA-SFTI (pDQ70). (4) OmpA-SFTI and the ColE3 lysis protein; OmpA-SFTI E3 lysis (pDQ75). 5) ToxA-SFTI and the ColE3 lysis protein; ToxA-SFTI E3 lysis (pDQ76). Right panel showing the protease inhibition assay (A) Colonies of E. coli harboring 1–5 from panel A growing on a casein petri plate with Tet induction. (B) The protease inhibition staining pattern of 1–5.

Fig. 2.. Detection of secreted protease inhibitory activity by YebF chimeric with SFTI in E. coli strain JW2197–1 and Salmonella strain VNP20009.

(A) The DNA constructs showing (1) the pTet99a empty cloning vector (pDB221), (2) the YebF construct (pDQ84) and (3) the YebF-SFTI fusion construct engineered without (YebF-SFTI (pDQ221)) and (4) with the trypsin cleavage sequence LK (YebF-^LKSFTI (pDQ44)). (B) A petri plate protease inhibition assay of induced E. coli colonies carrying the pTet99a plasmid with (1) pTet99a empty vector, (2) YebF, and (3) YebF-SFTI. (C) A petri plate protease inhibition assay of induced Salmonella strain VNP20009 colonies carrying (1) pTet99a empty vector, (2) YebF, and (3) YebF-SFTI. (D) A petri plate protease inhibition assay of induced Salmonella strain VNP20009 colonies carrying (1) pTet99a empty vector, (3) YebF-SFTI and (4) YebF-^LKSFTI.

Fig. 3.. Protease inhibition of E. coli JW2197–1 and VNP20009 strains assessed by the BApNA hydrolysis inhibition assay.

(A) A standard curve constructed using the synthetic SFTI peptide. (B) The inhibitory activity of (1) E. coli YebF, (2) E. coli YebF-SFTI, (3) E. coli YebF^LKSFTI, (4) VNP20009 YebF, (5) VNP20009 YebF-SFTI, and (6) VNP20009 YebF-^LKSFTI, each of which had a p < 0.0001 (****).

Fig. 4.. Protease inhibition of E. coli JW2197–1 strains assessed by reverse zymography, and liquid chromatography and mass spectrometry analysis of trypsin-affinity purified peptides produced by E. coli strain JW2197–1.

Top panel: Protease inhibition was assessed by non-denaturing polyacrylamide gel electrophoresis containing the protein substrate gelatin of (1) the SFTI peptide, (2) YebF-SFTI, and (3) YebF-^LKSFTI, followed by incubation with trypsin and staining with Coomassie brilliant blue. Bottom panel: Purified peptides were fractionated online using reversed phase chromatography followed by ESI-MS determination of molecular weights (enclosed boxes). (A) The pTet99a empty vector strain. (B) The synthetic SFTI peptide. (C) The peptides produced in the absence of an LK trypsin cleavage site through induction of YebF-SFTI. (D) The peptides produced with the LK trypsin cleavage site through induction of YebF-^LKSFTI.

Fig. 5.. Protection of a therapeutic protein from trypsin degradation by Salmonella VNP20009.

In vitro cytotoxicity assays were performed against MDA-MB-468 tumor cells and displayed as the percentage of surviving cells compared with untreated following treatment with cell culture supernatants from Salmonella strain VNP20009 containing various DNA constructs. (A) The arabinose inducible chimeric Pseudomonas exotoxin A OTG-PE38K (upper construct) and the OTG-PE38K RBS YebF-^LKSFTI (lower construct). (B) The percentage of surviving cancer cells treated with (1) YebF, (2) YebF-^LKSFTI, (3) the OTG-PE38K chimeric Pseudomonas ToxA, (4) the OTG-PE38K pretreated with trypsin and (5) the OTG-PE38K RBS YebF-^LKSFTI pretreated with trypsin (bar for small negative value not drawn to scale). Statistically significant comparisons shown as **** (p < 0.001).