Cancer cell-specific internalizing ligands from phage displayed beta-lactamase-peptide fusion libraries

Abstract

The present study was focused on identifying cancer cell-specific internalizing ligands using a new kind of phage display library in which the linear or cysteine-constrained random peptides were at amino-terminus fusion to catalytically active P99 beta-lactamase molecules. The size and quality of libraries were comparable to other reported phage display systems. Several cancer cell-specific binding and internalizing beta-lactamase-peptide fusion ligands were isolated by selecting these libraries on the live BT-474 human breast cancer cells. The identified ligands shared several significant motifs, which showed their selectivity and possible binding to some common cancer cell targets. The beta-lactamase fusion made the whole process of clone screening efficient and simple. The ligands selected from such libraries do not require peptide synthesis and modifications, and can be used directly for applications that require ligand tracking. In addition, the selected beta-lactamase-peptide ligands have a potential for their direct use in targeted enzyme prodrug therapy. The cancer-specific peptides can also be adopted for other kinds of targeted delivery protocols requiring cell-specific affinity reagents. This is first report on the selection of cell-internalized enzyme conjugates using phage display technology, which opens the possibility for new fusion libraries with other relevant enzymes.

Fig. 1

(A) Schematic presentation of phagemid vector design. The vector expressed random peptide libraries at the amino-terminus of P99 β-lactamase, linked through Gly-Gly-Gly-Ser linker. The C-terminal end of P99 β-lactamase had FLAG and 6xHis tags, which were connected to pIII protein with the protease-cleavable linker (PAGLSEGSTIEGRGAHE). pIII signal peptide was used for periplasmic translocation of β-lactamase-peptide fusion protein. The vector has T7, SP6 and lac promoters. CAT, chloramphenicol acetyltransferase; Amber stop, amber stop codon (tag); pIII signal peptide, oligonucleotide sequence that encode the peptide; P lac, lac promoter; gIII, pIII protein gene; f1 Ori, f1 origin of replications from filamentous bacteriophage; (B) This vector in suppressor host TG1 E. coli, with the help of KM13 helper phage, generated phage particles displaying β-lactamase-random peptide as an N-terminal fusion protein to phage coat pIII. In non-suppressor strain Top10 E. coli, which recognizes the amber codon, the same vector expressed free β-lactamase fusion proteins with a random peptide library at the N-terminal end and FLAG and 6xHis tags at the C-terminal end.

Fig. 2

Enrichment of cell-internalized clones at subsequent rounds of biopanning. The phage enrichments were evaluated by comparing the phage recovery ratios (output/input phage) in different rounds of the biopanning. The phage were quantified as cfu using biological titration method. Each bar represents the ratio ± SE of three determinations.

Fig. 3

Binding of β-lactamase-peptide fusion proteins to live human breast cancer cells. Soluble fusions from 62 randomly selected clones following the third round of panning were screened for their binding to live BT-474 cells. The lysates used for the screening were normalized for their β-lactamase activity. The human non-cancer MCF 10A epithelial cells and mouse fibroblast 3T3 cell-line were used as negative control cells. The screening was based on assaying the cell-bound β-lactamase activities. A pool of peptide fusion proteins from the unselected library clones and wild-type β-lactamase (Wt-BLA; without fusion peptide) were used as negative fusion peptide controls. The bars represent β-lactamase activity as change in relative fluorescence unit (RFU)/min ± SE of three determinations. *P ≤ 0.05, significance of difference in the binding of fusion proteins to BT-474 cells in comparison to MCF 10A or 3T3 cells were determined using Student's t-test.

Fig. 4

Microscopic imaging of the human breast cancer cells for internalization of β-lactamase-peptide fusion proteins. (A) The figure shows the results of fluorescence imaging of the internalized fusions using anti-6xHis antibody-AlexaFluor 488 conjugate as a probe. The β-lactamase-peptide proteins from clones 201 (a), 218 (b), 221 (c), 225 (d), 228 (e), 232 (f) and 235 (g) showed the fluorescence staining only in BT-474 cells (first row) not in MCF 10A cells (third row). The negative clone control (h), which comprised a pool of β-lactamase fusion proteins from the unselected library clones did not show any staining in either BT-474 cells or MCF 10A cells. The images in the second and fourth rows are corresponding DAPI nuclear staining in BT-474 and MCF 10A cells, respectively. The magnified confocal images of some of the positively stained cells for 6xHis are presented in the fifth row. The figures i (clone 228), j (clone 232) and k (clone 235) represent cytoplasmic staining of internalized β-lactamase-peptide fusion proteins. (B) The cell internalization of β-lactamase-peptide fusion proteins was also visualized by the enzymatic cleavage of β-lactamase substrate CCF4 loaded within the cells. The results of this study are represented by the fusion proteins from clones 232 (l,m) and 235 (n,o) showing positive blue fluorescence in BT-474 cells (l,n) and negative green in MCF 10A cells (m,o). The cells exposed to a pool of fusion proteins from unselected library (negative clone control) exhibit green fluorescence in both BT-474 cells (p) and MCF 10A cells (q), indicating no internalization. The size bars measure 20 µm in length.