Figure 1.
Schematic depiction of HIF-1α-Activated Protein Switches (Haps).
These protein switches are composed of the prodrug-converting enzyme yeast cytosine deaminase (yCD), which converts the non-toxic prodrug 5-fluorocytosine (5FC) to the highly toxic chemotherapeutic 5-fluorouracil (5FU), and the HIF-1α-binding CH1 domain of the human p300 protein. The cancer marker, HIF-1α, modulates the enzymatic activity of the yCD-domain of the switch, either through an allosteric mechanism or through stabilizing the protein such that it accumulates at higher levels in the cell. In normal cells HIF-1α is absent and the yCD domain is inactive. In the presence of HIF-1α – ideally only in cancer cells – the yCD domain is active and able to produce the chemotherapeutic.
Figure 2.
Strategies for creating improved HIF-1α-activate protein switches.
(A) Sequences of parental switches Haps3 and Haps59. Numbers above indicate the amino acid residues in yCD. Letters below indicate the 1-letter abbreviation amino acid sequence of the corresponding linker. (B) Schematic of the linker libraries created in Haps59. Lines indicate which combinations of N- and C-terminal linkers were combined to comprise the 18 libraries. (C) Schematic of random mutagenesis libraries of Haps59 and Haps3. A * indicates a representative point mutation. (D) Schematic of the construction of libraries of random circular permutations of CH1 that are randomly inserted into the yCD domain of FLAG-tagged yCD. Multiplex PCR was used to create all possible circular permutations of the CH1 domain. Multiplex inverse PCR was used to create all possible insertion sites within yCD. These two populations of DNA were ligated to create the library.
Table 1.
List of linker library members with switching abilities similar to Haps59.
Figure 3.
Characterization of Ehaps22 and Ehaps08 in E. coli.
(A) Diagram of Ehaps22 and Ehaps08 sequences. Ehaps22 was selected from the linker library containing 2-amino acid N- and C-terminal linkers. Ehaps08 was selected from the random mutagenesis library of Haps59. Numbers above indicate the amino acid residues in yCD. Letters below indicate the 1-letter abbreviation of the amino acid sequence of the corresponding linker or mutation. (B) Representative 5FC dot toxicity assay on cells expressing yCD, Haps59, Ehaps22, or Ehaps08. Serial dilutions of equal density log phase cultures containing either pGA (−HIF-1α, i.e. not expressing HIF-1α) or pGA-HIF (+ HIF-1α, i.e. expressing HIF-1α) were spotted on minimal media plates containing increasing concentrations of 5FC. (C) Quantification of dot toxicity assay. The highest dilution at which growth was observed for each culture on each plate is plotted against the concentration of 5FC in each plate. The number of replicates of each plate is listed above. Error bars represent max and min for N = 2 plates and standard deviation for N = 4. See Fig S1, S2, S3, and S4 for replicate plates.
Figure 4.
Characterization of RICP12 and RICP95 in E. coli.
(A) Diagrams of RICP12 and RICP95 sequences identified from the random circular permutation and random insertion library. Numbers above indicate the amino acid residues of the corresponding wild-type proteins. Letters below indicate the 1-letter abbreviation of the amino acid linker sequence used in the circular permutation of the CH1 domain. (B) Dot toxicity assay of RICP12 and RICP95.
Figure 5.
Characterization of the effects of Haps59 expression in Flp-In 293 cells and E. coli.
(A) Western blot with anti-HIF-1α antibodies showing that the addition of CoCl2 causes the accumulation of HIF-1α in Flp-In 293 cells. See Figure S5 for full blot. (B) Toxicity of 5FC to stable isogenic pools of Flp-In 293 cells containing EV, yCD, and Haps59. Cells were grown in the absence (i.e. −HIF-1α) or presence (i.e. + HIF-1α) of 50 µM CoCl2, and increasing concentrations of 5FC. Percent survival was calculated by measuring the total DNA of surviving cells in each well and normalizing between wells containing no 5FC (i.e. 100% survival) and wells containing no cells (i.e. 0% survival). Error bars represent the standard deviation across three replicates. (C) Dot toxicity assay indicates that CITED2 but not E1A activates Hap59 in E. coli. Equal density dilutions of log phase cultures of strains expressing Haps59 and either the p300 binding fragments of HIF-1α, CITED2, or E1A were plated on increasing concentrations of 5FC.