Cross-linking tBid affects its translocation to the mitochondria

Bid was cross-linked at two different sets of positions shown in Figures 2a and b to test which conformational changes are necessary for translocation. These locations are not expected to affect their caspase-8 cleavage activity because the loop containing Leu56 is still exposed. Before translocation, the intracellular diffusion of the uncross-linked variants and cross-linked variants of Bid are very similar, Figure 2c. The auto-correlation function for each of the Bid variants could be fit with two diffusive behaviors that correspond to a fast and slow time constant. The majority (67–71%) of the diffusive behavior exhibited the fast time constant (7.3 × 10⁻⁴ to 1.6 × 10⁻³s). The rest of the diffusive behavior (29–33%) exhibited the slow time constant (3.3 × 10⁻² to 1.1 × 10⁻¹ s). These relative populations are the result of fitting without correction to differences in quantum yields for the two time constants. As a result, they are only 'apparent populations'. In addition, it is not clear whether to fit the slower time constant to a 3-D or 2-D diffusion model because multiple scenarios could account for the different models of diffusion at a time constant of this magnitude, such as oligomers, association with an organelle or diffusion within a membrane environment. However, these apparent populations are reported to demonstrate the heterogeneity of the diffusion of Bid in a cellular environment.

Open in a separate window

Figure 2

Intramolecular cross-linking affects translocation behavior. (a) Bid was cross-linked between positions Bid-88 and Bid-158. (b) Bid was also cross-linked at positions Bid-136 and Bid-180. Cylinders represent helices in the Bid structure. They are color coded as in Figure 1. (c) The intracellular diffusion of both cross-linked (Bid-88-158-BM in green, Bid-136-180-BM in pink) species is similar to the uncross-linked variant (Bid-88-158 in blue, Bid-136-180 in red) before translocation as determined by FCS. (d) After translocation, the cross-linker between positions Bid-88 and Bid-158 prevented translocation, but not the cross-linker between positions Bid-136 and Bid-180 in Bid. The same color code was used as in 2C

After translocation, the intracellular diffusion of these variants differs significantly from that of the uncross-linked variant (Figure 2d). The uncross-linked variants of Bid diffuse at a significantly slower rate, which is consistent with the translocation of Bid to the MOM or the formation of large oligomers. However, cross-linking positions Bid-88 and Bid-158 abrogated translocation. The diffusion of Bid cross-linked at positions 88 and 158 was almost identical before and after the addition of STS. In contrast, cross-linking Bid at positions Bid-136 and Bid-180 did not affect translocation. The diffusive behavior of this variant, before and after the addition of STS, was the same with or without the cross-linker. By cross-linking positions 136 and 180, the helix that contains the BH3 helix is free to associate with the OMM. Thus, these results demonstrate that in order to associate with the OMM, Bid must adopt an extended conformation and have the BH3 domain exposed.

Intermolecular contacts between tBid molecules at the OMM

FRET between separate molecules of tBid at the OMM were also measured. These experiments were carried out as described before.²² Figure 3 contains images that show the translocation of Bid from the cytosol to the mitochondria, as well as a per-pixel scatterplot that utilizes correlative trends to determine if there is FRET between two sites within tBid, as analyzed previously.²² In both experiments, the intracellular distribution of Bid is diffuse and punctate before and after translocation to the mitochondria, respectively. In Figure 3a, before translocation, there is no FRET between positions Bid-158 and Bid-118 of tBid, as evidenced by the similar correlative slope in the presence and absence of the FRET acceptor. After translocation, however, in Figure 3b, the correlative slope in the presence of the acceptor is lower, which indicates FRET between positions Bid-158 and Bid-118 of tBid after translocation. In Figures 3c and d, contact between positions Bid-118 and Bid-88 of tBid were investigated. As indicated by their similar slopes in the scatter plots, there is no FRET between these sites and presumably no contact between positions Bid-118 and Bid-88.

Open in a separate window

Figure 3

Intermolecular contacts between tBid molecules. (a) Before translocation, preSTS, there is no FRET between Bid-158 and Bid-118. (b) After translocation, postSTS, FRET is observed between the two sites as indicated by a lower correlative slope when the FRET Acceptor is present, red dots, compared with when it is absent, black dots. (c) There is no FRET between Bid-118 and Bid-88 before translocation. (d) A similar correlative trend in the presence (red dots) and absence (black dots) of the acceptor, indicates that there is no FRET between Bid-118 and Bid-88 after translocation

Intermolecular FRET efficiencies between other positions are shown in Table 2. Most of the contacts between tBid molecules are between the N-terminal of one molecule, positions Bid-88 and Bid-118, and the C-terminal region of another molecule, positions Bid-158 and Bid-180. The absence of FRET between symmetric regions in tBid indicates that an extended conformation is required to satisfy these distance restraints. This is in agreement with the picture formulated from intramolecular FRET efficiencies. One of the observed contacts is between positions Bid-118 and Bid-180, which explains the residual FRET efficiency after translocation in the intramolecular FRET experiments. The magnitude of some of these FRET efficiencies is similar to those observed in the intermolecular contacts between Bax molecules reported previously.²² The FRET efficiencies between positions Bid-118, Bid-140, and Bid-158 are particularly interesting. The observed intermolecular FRET efficiencies between positions Bid-118 and Bid-158 and between Bid-140 and Bid-158 in Table 2 are consistent with intramolecular distances in Bid. Positions Bid-118, Bid-140, and Bid-158 are within helices α4, α5, and α6, respectively. According to the NMR structure of Bid, the side-chains of these helices are closely packed. It is very likely that the same interactions that pack these helices in soluble, monomeric Bid are maintained in the intermolecular contacts of tBid when it is associated with the OMM in an extended conformation.

Labeling of mutants with fluorescent dyes for FRET measurements

Bid was site-specifically labeled at engineered cysteines using maleimide chemistry. AlexaFluor546 and AlexaFluor488 (Invitrogen) were used as FRET donors and Dabcyl (Anaspec) was used as the FRET acceptor. For the conjugation reaction of each probe to Bid, the conditions were 20mM Tris-HCl at pH 8.0 with 1mM TCEP, in the dark for 1h at 25^°C. The protein concentration was 100μM. To react the FRET donors to Bid, the ratio of protein to dye was 2:1 for proteins with two cysteines and 1:1 for single-cysteine variants. To conjugate the FRET acceptor, Dabcyl, to Bid a 7:1 excess of dye to protein was used. Unreacted and reacted species with one fluorescent dye molecule were separated with a mono-Q column as described previously.²⁷ The internal reference, AlexaFluor633, was conjugated to the above dye-labeled Bid proteins using recently published protocol at Lys144 or Lys146 (Fragment from residues 142–148 had an additional mass corresponding to the AlexaFluor633 tag: Exp. 1801.90 Da, Theor. 1801.90 Da).²⁷ Liquid chromatography mass spectrometry (LC-MS) of the labeled protein variants and their fragments, following pepsin digestion, confirmed the stoichiometry and location of dyes, respectively, as previously described.²⁷

Confocal imaging for FRET measurements and activity assays

The confocal microscopy and microinjection setup to study Bid was described earlier.^{22, 27} All imaging experiments were carried out using serially immortalized mouse embryonic fibroblast (MEF), which was generously provided by Richard Youle, (NINDS, NIH).⁴⁸ Additional experiments were performed on a Zeiss 510 system but with the same excitation and emission filter sets that were used in the previous studies. The total concentration of conjugated protein that was microinjected for FRET measurements ranged from 200 to 400nM. Equimolar amounts of 200nM of different Bid variants were co-injected for intermolecular FRET measurements, whereas equimolar amounts of 200nM of Bid and Bax variants were co-injected for interprotein FRET. To ensure that Bax:Bax contacts were intact for Bid:Bax experiments, an additional 200nM of unlabeled Bid was added and co-injected with 200nM of fluorescent Bax labeled at position 78 and 200nM of Bax labeled at position 62. Previously we estimated that there was a dilution factor of 1:5–1:10 of the injected protein concentration relative to the final concentration in the cell.

FRET efficiency depends on a distance separating two fluorophores. The choice of pair of fluorophores depends on their characteristic Föster radius (R₀), which corresponds to a distance between the fluorophores with 50% FRET efficiency. To measure intramolecular FRET between two cysteines, the F' sample consisted of a Bid variant with two cysteines with only AlexaFluor546 or AlexaFluor488 labeled on one of the cysteines and the internal reference, AlexaFluor633. Then, a F_Q' sample was prepared for this same variant that was labeled with the either AlexaFluor546 or AlexaFluor488 as well as Dabcyl on the other cysteine in addition to the internal reference, AlexaFluor633. FRET efficiency, E' was calculated as described previously.²⁷ To measure intermolecular FRET between two cysteines, the F' sample was labeled with AlexaFluor546 and the internal reference, AlexaFluor633. The F_Q' sample consisted of co-microinjecting the F' sample with another single-cysteine variant of Bid labeled with Dabcyl. FRET efficiency, E' was calculated as described previously.²⁷ The locations of the cysteines used in these experiments are indicated in the Tables.

Translocation of Bid to the mitochondria was initiated by the addition of 3.6μM staurosporin (STS) into the cell culture medium. This method of initiation was chosen to ensure consistency with our previous studies of Bax.^{22, 27} To verify that the translocation of Bid using STS was consistent with inducing translocation of Bid with TNFα,^{50, 51} a colocalization assay with Mitotracker Green FM was performed using 100μg/ml CHX and 1μg/ml TNFα with cells in PBS. Colocalization of Bid and Mitotracker Green FM, as well as similar Bid translocation times, was observed using both methods of apoptosis initiation (Supplementary Figure S1). The caspase inhibitor, Q-VD-OPh (Biovision, Milpitas, CA, USA) was utilized to examine the effects of caspase-8 induced fragmentation. For these experiments, cellular media was incubated for 15min of 10μM Q-VD-OPh before microinjection and STS-induced apoptosis. Experiments utilized 10–15 cells to determine the FRET efficiency between two locations. Quantification of fluorescence intensities and further analysis of confocal images were described previously.^{22, 27}

This work was supported by the Intramural Research Program of the NIH, and NHLBI. We thank Duck-Yeon Lee of the NHLBI Biochemical Facility for assistance in mass spectrometry and Xufeng Wu of the NHLBI Light Microscopy Core for assistance in obtaining confocal images. We also would like to thank Richard Youle and Ashley C Barnes for helpful comments and suggestions on the manuscript.