An improved reversibly dimerizing mutant of the FK506-binding protein FKBP

Abstract

FK506-binding protein (FKBP) is a monomer that binds to FK506, rapamycin, and related ligands. The F36M substitution, in which Phe36 in the ligand-binding pocket is changed to Met, leads to formation of antiparallel FKBP dimers, which can be dissociated into monomers by ligand binding. This FKBP(M) mutant has been employed in the mammalian secretory pathway to generate aggregates that can be dissolved by ligand addition to create cargo waves. However, when testing this approach in yeast, we found that dissolution of FKBP(M) aggregates was inefficient. An improved reversibly dimerizing FKBP formed aggregates that dissolved more readily. This FKBP(L,V) mutant carries the F36L mutation, which increases the affinity of ligand binding, and the I90V mutation, which accelerates ligand-induced dissociation of the dimers. The FKBP(L,V) mutant expands the utility of reversibly dimerizing FKBP.

Keywords: F36M; FK506-binding protein; FKBP; dimerization; regulatable secretory cargo.

Figure 1.

Strategy for generating fluorescent aggregates using dimerizing FKBP. Tetrameric E2-Crimson (red) is fused to a dimerizing FKBP mutant (gold) to generate cross-linked aggregates. Addition of ligand (blue) disrupts FKBP dimerization, thereby yielding soluble tetramers.

Figure 2.

Influence of the fluorescent protein tag on the formation of aggregates containing 4 tandem copies of FKBP(M). Yeast cells were transformed with integrating vectors to express either EGFP-FKBP(M)x4 or mEGFP-FKBP(M)x4. Green fluorescence and differential interference contrast (DIC) images of logarithmically growing cells were captured by confocal microscopy, using the same parameters for both strains. Representative flourescence and merged images are shown. To enhance weaker signals, the gamma value of the fluorescence images was adjusted to 2.0 using Adobe Photoshop. Scale bar, 2 μm.

Figure 3.

Enhanced aggregate dissolution in drug-sensitive yeast strains. An integrating vector encoding E2-Crimson-FKBP(M) was transformed into isogenic wild-type and pdr5Δ and pdr1Δ pdr3Δ strains. Logarithmically growing cultures were treated with 250 μM SLF for 0 to 120 min as indicated, and were imaged by confocal microscopy to capture red fluorescence and DIC images, using the same parameters in all cases. Representative merged images are shown. Scale bar, 2 μm.

Figure 4.

Aggregate dissolution at lower ligand concentrations with the F36L mutation. Integrating vectors encoding E2-Crimson-FKBP(M) or E2-Crimson-FKBP(L) were transformed into a pdr1Δ pdr3Δ strain. Logarithmically growing cultures were treated with the indicated concentrations of SLF for 30 min, and were imaged by confocal microscopy to capture red fluorescence and DIC images, using the same parameters in all cases. For each strain and condition, approximately 30 – 100 cells were examined, and the percentage of the fluorescence signal that was punctate was quantified as described in Materials and Methods. Representative merged images are shown together with quantitation of the image data. Scale bar, 2 μm. Error bars represent s.e.m.

Figure 5.

Faster aggregate dissolution with the I90V mutation. (A) Diagram of a cross-section through the center of the flow chamber. Media enters and exits the chamber through Tygon tubing attached to stainless steel tubes embedded in a slab of hard PMMA plastic. Those tubes connect to a groove in the plastic. The PMMA slab is placed on a concanavalin A-coated coverslip to which yeast cells have been attached. This assembly sits in a well in a metal piece on a microscope stage. To hold the assembly in place, an additional clear plastic slab, not shown in the diagram, is placed on top and attached with screws to the metal piece. A tapered opening in the metal piece allows an oil objective to be positioned close to the bottom of the coverslip. Medium is pulled through the flow chamber using a negative pressure pump. Hydrophobic compounds such as SLF flow directly past the cells. (B) Selected frames from Videos 1 and 2, which show SLF-induced dissolution of E2-Crimson-FKBP(L) aggregates and E2-Crimson-FKBP(L,V) aggregates, respectively. SLF reached the cells approximately 20 s after the beginning of a movie. Scale bar, 2 μm. (C) Quantitation of aggregate dissolution from Videos 1 and 2. The radius of each fluorescent aggregate was measured at each time point, beginning at the time point when SLF was estimated to reach the cells. The inset shows the mean time needed for complete dissolution of aggregates. For this quantitation, a total of 43 E2-Crimson-FKBP(L) aggregates and 36 E2-Crimson-FKBP(L,V) aggregates were examined from 3 movies for each construct, and the times needed for complete dissolution of the aggregates were recorded. Error bars represent s.e.m. The mean dissolution times were 159 ± 15 s for E2-Crimson-FKBP(L) and 40 ± 4 s for E2-Crimson-FKBP(L,V).