K-ras4B and prenylated proteins lacking "second signals" associate dynamically with cellular membranes

Abstract

We have used fluorescence microscopy and the technique of rapamycin-regulated protein heterodimerization to examine the dynamics of the subcellular localizations of fluorescent proteins fused to lipid-modified protein sequences and to wild-type and mutated forms of full-length K-ras4B. Singly prenylated or myristoylated fluorescent protein derivatives lacking a "second signal" to direct them to specific subcellular destinations, but incorporating a rapamycin-dependent heterodimerization module, rapidly translocate to mitochondria upon rapamycin addition to bind to a mitochondrial outer membrane protein incorporating a complementary heterodimerization module. Under the same conditions analogous constructs anchored to the plasma membrane by multiply lipid-modified sequences, or by a transmembrane helix, show very slow or no transfer to mitochondria, respectively. Interestingly, however, fluorescent protein constructs incorporating either full-length K-ras4B or its plasma membrane-targeting sequence alone undergo rapamycin-induced transfer from the plasma membrane to mitochondria on a time scale of minutes, demonstrating the rapidly reversible nature of K-ras4B binding to the plasma membrane. The dynamic nature of the plasma membrane targeting of K-ras4B could contribute to K-ras4B function by facilitating redistribution of the protein between subcellular compartments under particular conditions.

Figure 1.

Structures of fluorescent protein constructs created for this study. The construct FRB₂-CFP-tK(S181A) was generated by mutating to alanine the serine residue shown in boldface in the structure of FRB₂-CFP-tK.

Figure 2.

Confocal microscopic images showing the subcellular distributions of prenylated and myristoylated fluorescent proteins in CV-1 cells. (A–C) Images of (A) FRB₂-CFP-tH′(F), (B) FRB₂-CFP-tH′(GG), and (C) myr-FRB₂-CFP. (D and E) Images of FRB₂-CFP-tH′(GG) and YFP-tH′(GG), respectively; (F) merged image (green is YFP-tH′(GG)). (G and H) Images of Arf6(1–7)-CFP and YFP-tH′(GG), respectively; (I) merged image (green is YFP-tH′(GG)).

Figure 3.

Rapamycin-induced redistribution to mitochondria of farnesylated FRB₂-CFP-tH′(F) in cells transiently expressing the protein together with mitochondrially anchored mitoRFP-FKBP₃. Limited resolution in some images is due to the use of short exposure times to minimize potential photobleaching. (A and B) Fluorescence images of mitoRFP-FKBP₃ and of a mitochondrial matrix-targeted CFP derivative, pOCT(1–37), respectively, in a COS-1 cell transiently expressing the two proteins. (C and D) Fluorescence images of FRB₂-CFP-tH′(F) before (C) and 5 min after (D) addition of rapamycin (0.5 μM). (E and F) Fluorescence images of mitoRFP-FKBP₃ before (E) and 5 min after (F) addition of rapamycin to the same cell as shown in C and D.

Figure 4.

Rapamycin-induced redistribution of lipid-modified, FRB-domain-incorporating fluorescent proteins. (A–C) Subcellular distribution of FRB₂-CFP-tH′(GG) at 0 (A) and 5 min after (B) addition of rapamycin to a CV-1 cell coexpressing the protein together with mitoRFP-FKBP₃, whose distribution 5 min after rapamycin addition is shown in C. (D–F) Subcellular distribution of myrFRB₂-CFP at 0 (D) and 5 min after (E) addition of rapamycin to CV-1 cells coexpressing the protein together with mitoRFP-FKBP₃, whose distribution 5 min after rapamycin addition is shown in F. (G and H) Subcellular distribution of FRB₂-CFP-tH′(GG) at 0 (G) and 5 min after (H) addition of rapamycin to a CV-1 cell not coexpressing the mitoRFP-FKBP₃ construct.

Figure 5.

Quantitative analysis, carried out as described in the text, of the kinetics of rapamycin- or rapalog-induced translocation to mitochondria of protein constructs coexpressed with mitoRFP-FKBP₃ in CV-1 cells. (A) Time courses of redistribution of FRB₂-CFP-tH′(GG) upon addition of 0.5 μM rapamycin (•) or 1 μM rapalog AP21967 (○). The complete time courses determined for rapamycin-treated cells, and the rapid (post-lag) phases of those determined for AP29167-treated cells, were fit as illustrated to an equation of the form I = A + B·exp(–kt), where I is the mean pixel intensity determined for a fixed region of the cell image, A and B are adjustable scaling parameters, and k is the rate constant for the redistribution process. Values of k determined in this manner are summarized in Table 1. (B–D) Time courses of rapamycin-induced redistribution of (B) FRB₂-CFP-tH, (C) lck(1–10)-FRB₂-CFP, and (D) LAT(1–35)-FRB₂-CFP in CV-1 cells under conditions similar to those described for A. Data shown in B and C are fit with a biexponential decay curve and those in D to a monoexponential decay curve (in the latter case yielding a rate constant << 10^–4 min^–1, indicating negligible transfer from the plasma membrane). The values of k summarized in Table 1 for translocation of FRB₂-CFP-tH, and lck(1–10)-FRB₂-CFP correspond to the rate constants estimated for the slow phase of fluorescence redistribution. Because of the nature of the image analysis and the deliberate use of overly conservative image-background corrections in data quantification, the magnitudes of the intensity decays shown (expressed as a percentage of the initial intensity) should not be directly equated to the percentages of translocation of the protein constructs, which will typically be somewhat larger and can be better assessed from original images like those shown in Figures 3, 4, and 6.

Figure 6.

Effects of rapamycin on the subcellular distribution (in CV-1 cells coexpressing mitoRFP-FKBP₃) of fluorescent FRB domain-containing protein constructs anchored to membranes via multiply lipid-modified or transmembrane sequences. (A–C) Cell expressing FRB₂-CFP-tH imaged before (A) and 5 min (B) or 60 min (C) after addition of rapamycin (0.5 μM). (D–I) Cells expressing lck(1–10)-FRB₂-CFP (D–F) and LAT(1–35)-FRB₂-CFP (G–I), imaged at the same times as indicated for A–C.

Figure 7.

Rapamycin- or rapalog-induced redistribution to mitochondria of fluorescent FRB-domain-containing constructs incorporating the K-ras4B-targeting sequence or fulllength K-ras4B in CV-1 cells coexpressing the mitoRFP-FKBP₃ construct. (A–F) Subcellular distribution of FRB₂-CFP-tK at 0 (A), 1 (B), 5 (C), 20 (D), and 60 min (E) after addition of rapamycin; (F) distribution of mitoRFP-FKBP₃ in the same cells 60 min after rapamycin addition. The distribution of mitoRFP-FKBP₃ at 0 min (unpublished data) was very similar. (G–L) Subcellular distribution of FRB₂-CFP-Kras at 0 (G), 2 (H), 5 (I), 10 (J), and 45 min (K) after rapamycin addition; (L) distribution of mitoRFP-FKBP₃ 45 min after rapamycin addition. (M–O) Distribution of FRB₂-CFP-Kras at 0 (M), 10 (N), and 45 min (O) after addition of rapalog AP29167.

Figure 8.

Kinetic analysis of rapamycin-induced intracellular redistribution of FRB₂-CFP-tK and FRB₂-CFP-Kras upon addition of rapamycin or rapalog AP29167 to CV-1 cells coexpressing the mitoRFP-FKBP₃ protein. (A) Time courses of redistribution of FRB₂-CFP-tK upon addition of 0.5 μM rapamycin (•)or1 μM rapalog (○). (B) Redistribution of FRB₂-CFP-Kras upon addition of rapamycin (•) or rapalog (○).