Membrane interactions of apoptotic inhibitor Bcl-xL: What can be learned using fluorescence spectroscopy

Highlights

• •Apoptotic inhibitor Bcl-xL adopts various conformations in membranes.
• •Release of BH4-domain is linked to non-canonical inhibition of apoptosis.
• •Lipid composition and divalent cations modulate membrane targeting and refolding of Bcl-xL.
• •Fluorescence tools play a critical role in studies of protein-membrane interactions.

Abstract

Permeabilization of the mitochondrial outer membrane–a point of no return in apoptotic regulation–is tightly controlled by proteins of the Bcl-2 family. Apoptotic inhibitor Bcl-xL is an important member of this family, responsible for blocking the permeabilization, and is also a promising target for anti-cancer drugs. Bcl-xL exists in the following conformations, each believed to play a role in the inhibition of apoptosis: (i) a soluble folded conformation, (ii) a membrane-anchored (by its C-terminal α8 helix) form, which retains the same fold as in solution and (iii) refolded membrane-inserted conformations, for which no structural data are available. In this review, we present the summary of the application of various methods of fluorescence spectroscopy for studying membrane interaction of Bcl-xL, and specifically the formation of the refolded inserted conformation. We discuss the application of environment-sensitive probes, Förster resonance energy transfer, fluorescence correlation spectroscopy, and fluorescent quenching for structural, thermodynamic, and functional characterization of protein-lipid interactions, which can benefit studies of other members of Bcl-2 (e.g., Bax, BAK, Bid). The conformational switching between various conformations of Bcl-xL depends on the presence of divalent cations, pH and lipid composition. This insertion-refolding transition also results in the release of the BH4 regulatory domain from the folded structure of Bcl-xL, which is relevant to the lipid-regulated conversion between canonical and non-canonical modes of apoptotic inhibition.

1. Introduction

Bcl-2-family proteins play central roles in apoptotic regulation. Alterations to their function contribute to cancer pathogenesis and chemotherapy resistance, making them important targets for drug discovery [1], [2], [3], [4]. These proteins regulate and execute MOMP (mitochondrial outer membrane permeabilization), recognized as the point of no return in intrinsic apoptosis, via a series of protein–protein and protein–membrane interactions [5], [6], [7]. The functions of individual family members are largely determined by the specific combination of the four conserved motifs known as Bcl-2 homology (BH) regions [8]. The pro-apoptotic pore-forming proteins (e.g., Bax, Bak, Bok) contain BH regions 1–3. The anti-apoptotic proteins (e.g., BCL-2, Bcl-w, Bcl-xL) contain all four BH regions. The BH3-only effector proteins (Bid, Bim, Bad, Puma, etc.) have a single BH3 region. During intrinsic apoptosis, BH3-only effectors cause Bax/Bak to oligomerize within and permeabilize the MOM, whereas the anti-apoptotic proteins inhibit this process at multiple steps [9,10]. The prevailing Embedded Together model [6,11] proposes that many of the functional interactions between Bcl-2 proteins occur only in membranes, though the exact conformational rearrangements upon membrane insertion remain largely unknown and their deciphering will require new approaches, as described in this review.

Bcl-xL is a 26 kDa anti-apoptotic regulator whose function is inhibiting MOMP [8,11]. Bcl-xL can exist in multiple soluble, membrane-anchored and membrane-inserted conformations (Fig. 1), each of them potentially playing a role in suppressing apoptosis. In solution, Bcl-xL consists of a globular structure [12], in which two long central helices, α5-α6, are surrounded by shorter amphiphilic helices. The soluble domain apparently retains its fold when anchored to the bilayer via a C-terminal TM anchor segment [13]. This anchored conformation is different from the inserted ones, for which no high-resolution structures are available. It has been shown that Bcl-xL is able to prevent the formation of high-order oligomers by Bax on the membrane [14], presumably by making a heterodimeric complex. Based on co-crystallization of regulatory domains with soluble conformations of partner proteins, two inhibition models have been suggested: (1) a canonical mode, in which anti-apoptotic protein captures the BH3 domain of Bax to prevent its homodimerization [15] and (2) a novel, non-canonical mode, in which the BH4 domain of anti-apoptotic protein engages Bax to prevent its activation [16]. In this review, we demonstrate how various fluorescence techniques can be used to gain important structural, thermodynamic and functional insights into the conformational switching by lipids, pH and divalent cations, resulting in the membrane insertion and refolding of Bcl-xL.

Fig. 1.

Simplified schematic representation of apoptotic regulation by Bcl-2 family of proteins. (a) Pro-apoptotic Bax and anti-apoptotic Bcl-xL are typical representatives of the Bcl-2 family of apoptotic regulators that bind to mitochondrial outer membrane (MOM) and control its permeabilization (MOMP). (b) Schematic illustration of the two structural motifs suggested for inhibitory action of Bcl-xL on Bax. The canonical model proposes short-range interactions in which the folded Bcl-xL traps the BH3 domain (green cylinder) of Bax. The non-canonical model suggests that Bax is engaged remotely by a the BH4 domain (red cylinder) released from the refolded Bcl-xL (while still connected by a long unfolded loop to the rest of the Bcl-xL structure. The two models of inhibition are not mutually exclusive and can be implemented by the various conformations of the Bcl-xL. (c) Schematic representation of the three main conformations of Bcl-xL: (top) soluble, (left) anchored to the membrane via helix a8 (yellow cylinder), and (right) membrane-inserted with the released domain BH4 (helix a1, red cylinder). The changes in lipid composition can lead to membrane insertion of either soluble or anchored Bcl-xL, resulting into presumable switching between canonical and non-canonical models of apoptotic inhibition.

2. Fluorescence tools for studying protein-membrane interactions

One of the main reasons for the lack in our understanding of the physiological action of Bcl-2 proteins, including Bcl-xL, is the shortage of appropriate experimental tools for studying the protein-membrane interactions. High-resolution methods are difficult to apply because the refolding on membrane interface and subsequent trans-bilayer insertion produces multiple intermediate states. However, structural, kinetic and thermodynamic informations can be acquired by using various fluorescence tools, as summarized in Fig. 2 [17,18]. In particular, fluorescence-based approaches can be utilized to monitor various aspects of membrane protein interactions, including protein-membrane association, formation of kinetic intermediates, the pathways of membrane insertion, specific conformational rearrangements, structure and membrane-topology of embedded states (Fig. 2). Moreover, fluorescence tools make it possible to elucidate the role of pH and lipid composition on membrane interactions of Bcl-xL.

Fig. 2.

Summary of fluorescence tools used for studying interactions of membrane-inserting proteins (e.g., Bcl-xL) with large unilamellar lipid vesicles (LUV), modeling the biological membranes. (a) Protein binding to a bilayer can be monitored by FRET occurring between a donor-dye-labeled Bcl-xL and acceptor-labeled lipids doped into a bilayer. (b) The site-selective labeling of Bcl-xL with environment-sensitive NBD fluorescent dye allows monitoring the bilayer insertion by the “turn-on” fluorescence effect. (c) Depth-dependent fluorescence quenching of NBD-labeled Bcl-xL by a series of lipid-attached spin quenchers with varying immersion depths makes it possible to estimate the membrane topology. (d) Studying the bilayer topology of NBD-labeled Bcl-xL by quenching with the water-soluble dithionite, added from the outside of the vesicles. (e) Bcl-xL refolding upon the membrane association can be studied by intramolecular FRET between donor dye Alexa488 (attached to a specifically introduced single Cys-residue) and an acceptor mCherry fluorescent protein (engineered into the sequence at the N-terminus).

Förster resonance energy transfer (FRET) is a common strategy to monitor binding interaction between two fluorescent-labeled particles [18], [19], [20], [21], [22], [23]. The method is based on a change in signal that can be experimentally monitored because a donor probe transfers energy to an acceptor probe when they become within the Förster's distance (radius) [24,25] (Fig. 2a and e). A typical membrane-binding experiment consists on site-selective-labeling of the protein and a lipid membrane doped with the dye, so that the membrane association of the protein causes FRET between the donor-acceptor probes (Fig. 2a) [17,24,26]. FRET technique can also be applied to monitor Bcl-xL protein refolding (Fig. 2e). More details on applications of fluorescence spectroscopy in membrane protein studies can be found in the recent reviews [18,[27], [28], [29], [30], [31]].

It should also be noted that beside the above-mentioned in vitro fluorescence protocols, the Bcl-2 family proteins have been studied in vivo by using confocal single-molecule fluorescence microscopy and imaging [32], [33], [34].

3. Lipid modulation of pH-dependent membrane insertion of Bcl-xL

3.1. pH-Dependent of Bcl-xL membrane insertion and BH4 domain release

Membrane association of Bcl-xL is pH-dependent [25], which could be monitored by FRET-based experiments (Fig. 2). In our earlier study, we used the FRET pair composed of the donor AlexaFluor-488 dye, linked to Bcl-xL as a thiol-reactive probe attached to the cysteine residue R102C. A FRET acceptor is a lipid-attached dye Rhodamine-PE [20,35] incorporated into a lipid membrane. Full details of this experiment are given elsewhere [25]. Briefly, acidification of the solution containing premixed AlexaFluor488-labeled Bcl-xL and Rhodamine-PE-doped LUVs resulted in a decrease in steady-state emission intensity of the donor dye, which was accompanied by a shortening of its fluorescence decays (not shown). These emission changes are indicative of protein binding to the membrane. Moreover, adjusting pH back to reference pH 8.0 caused the recovery of the initial spectral features. In summary, these FRET measurements suggest that (i) there is FRET between the donor/acceptor pair, which occurs under acidic conditions and (ii) FRET vanishes when reversing the pH conditions. These observations are consistent with a population of dye-labeled Bcl-xL binding/partitioning to the labeled lipid bilayer in a pH-dependent and reversible manner [25].

A convenient method to monitor the partitioning of Bcl-xL protein to a lipid bilayer is the use of fluorescence environment-sensitive probes, such as NBD (derivatives of 7-nitro-2–1,3-benzoxadiazol-4-amine dye) [25,36]. Environment-sensitive NBD probes change their fluorescent behavior depending on the polarity of its immediate environment [24,[37], [38], [39]]. To be specific, when a probe moves from a polar to a non-polar milieu, e.g. from aqueous to membranous environment, the emission spectrum increases in intensity and the position of the emission band maximum shifts towards shorter wavelengths (blue-shift) [40,[41], [42]]. Typically, these changes are also associated with longer fluorescence decay of a probe in the non-polar milieu. These fluorescence changes can be used to monitor the partitioning of NBD-labeled-Bcl-xL between the aqueous solution and a lipid bilayer, as shown in Fig. 3a [36]. Fig. 3a (gray) shows that compared to Bcl-xL in solution, the addition of LUVs containing 50% of cardiolipin (1TOCL:2POPC) at pH 8.0 does not result in significant spectroscopic changes (Fig. 3a, blue). However, solution acidification leads to large pH-dependent increases in the NBD intensity that is saturated at pH 4.5 (Fig. 3a, orange). The intensity increase is accompanied by a blue shift of the NBD emission maxima from 535.1 nm at pH 8.0 to 525.9 nm at pH 4.5, respectively. Both these fluorescence changes are associated with the transition of an NBD probe to hydrophobic environments and indicate the pH-dependent membrane interaction of the hydrophobic α6 helix of Bcl-xL.

Fig. 3.

Protonation-dependent membrane insertion and refolding of Bcl-xL. (a) Representative spectra of Bcl-xL ΔTM N175C—NBD inserting into cardiolipin containing LUV (1TOCL:2POPC). (b) The NMR structure of Bcl-xL is shown by backbone conformation in gray with the following color highlights: hydrophobic helix α6 in blue, BH4 helix α1 (BH4 domain) in red, the loop between α1 and α2 helices in green. The A488 FRET donor site attached at the D189C mutant is shown in yellow. An mCherry fluorescence protein (magenta), conjugated to a N-terminus, is used as a FRET acceptor. (c) Changes in fluorescence emission spectra of donor- and acceptor-labeled Bcl-xL ΔTM observed upon solution acidification in the presence of 1TOCL:2POPC LUV. The release of BH4 domain was monitored by the loss of FRET between Alexa 488 donor dye, attached to single-Cys-mutant D189C, and mCherry acceptor moiety, fused to the N-terminus next to the BH4 domain. (d) Comparison of protonation-dependent insertion and refolding of Bcl-xL. A solid line shows that both datasets can be accurately described by a single global fitting by with the apparent pKa of 5.8.

The membrane insertion-related refolding of Bcl-xL is accompanied by the release of the N-terminal BH4 (α1) regulatory domain from the folded structure. This domain is believed to be relevant to the so-called non-canonical mode of apoptotic inhibition, originally suggested for a close structural homologue of Bcl-xL, Bcl2 [16]. Our recent studies have revealed that alteration in lipid composition can serve as a regulatory factor for preparing Bcl-xL for the switching between the canonical and non-canonical modes of inhibition [36]. Moreover, it was previously shown that the ability of Bcl-xL to refold and release the BH4 helix in its membrane-inserted state is not to affected by the deletion of the C-terminal helix [36].

We demonstrate that a FRET technique can also be used for studying Bcl-xL refolding and the release of its N-terminal BH4 (α1) regulatory domain [43]. This technique requires the use both donor and acceptor dyes, linked to the same protein [20,28,[44], [45], [46]]. To monitor the release of the N-terminal BH4 domain of Bcl-xL we conjugated the fluorescent protein mCherry to the N-terminus of Bcl-xL ΔTM (Fig. 3b). This construct was then used to perform FRET measurements between AlexaFluor-488 labeled Bcl-xL ΔTM (a donor dye) and N-terminal mCherry (an acceptor dye) [36,43]. The fluorescent probe AlexaFluor488 was introduced using the single Cys-D189C mutant, so that in the folded structure the donor dye locates at distances <20 Å from the acceptor fluorophore (high FRET regime). Therefore, both fluorophores form an efficient FRET pair (R₀ ∼61 Å) when Bcl-xL is in solution and its BH4 domain is packed within the main body of the protein (Fig. 3b).

Fig. 3c (blue) shows the presence of both AlexaFluor488 and mCherry bands in the emission spectra when only the donor A488 is excited in the presence of 1TOCL:2POPC LUV, pH 8.0. However, solution acidification in the presence of membranes leads to complete loss of FRET, as seen by the large recovery in the intensity of the donor A488 band and the complete elimination of the long-wavelength band of the acceptor mCherry at pH 5 (Fig. 3c, red). These spectroscopic changes are governed by the increase in spatial distance between the donor AlexaFluor488 fluorophore and the mCherry-labeled N-terminal BH4 domain (Fig. 3b). These FRET results suggest that the pH-dependent insertion of Bcl-xL into membranes correlates with the release of its N-terminal BH4 domain (Fig. 3a and c).

Combined use of the different fluorescence tools allowed us to demonstrate that both membrane insertion and refolding of Bcl-xL are modulated by pH and lipid composition (Fig. 3d). The quantitative analysis of the percent change observed in the NBD intensity signals and the FRET band ratio showed that the relative changes in the insertion of Bcl-xL ΔTM into membranes completely overlapped with the release of its N-terminal BH4 domain. These findings suggest that both events are correlated at equilibrium (Fig. 3d) [36]. Changes in lipid composition have been demonstrated to modulate the pK_a of the transition and bring it at the threshold of physiological pH [36].

3.2. Determination of interfacial vs transmembrane topology by depth-dependent fluorescence quenching by lipid-attached quenchers

Determination of the insertion topology of a certain site on a membrane protein (i.e., its position and immersion depth with respect to the two leaflets of a lipid bilayer) is an important step in establishing proteins structural organization in the lipid bilayer [31,47,48]. However, broad distributions of the transverse positions of both quenchers and fluorophores in membrane often complicate accurate determination of transverse location of an intrinsic (e.g., Trp) or extrinsic (e.g., NBD) fluorescence probe [18,49]. Therefore, depth-dependent fluorescence quenching with a series of lipid-attached paramagnetic Tempo and Doxyl quenchers has become an important spectroscopic tool for calculating of a precise depth of a fluorescent moiety [31,[48], [49], [50], [51], [52]]. These spin moieties are capable of quenching most organic fluorophores, including NBD, rhodamine, bimane and antracene [51,[53], [54], [55]]. To increase the accuracy of estimation of membrane depth of a probe, we have calibrated the method by determining transverse positions of the quencher groups in a series of six spin-labeled lipids immersed into the model POPC using atomistic molecular dynamics (MD) simulations [56]. It has been suggested that measurements of topology are especially important in the determination of membrane insertion pathways for a variety of spontaneously inserting non-constitutive proteins, such as bacterial toxins [17,57,58], colicins [59,60], some annexins [61,62] and numerous apoptotic regulators of the Bcl-2 family [7,8,11,25,63].

The qualitative analysis of the depth-dependent fluorescence quenching methodology is self-evident and assumes that fluorescence quenching is the highest if a dye molecule and a quencher are located at the same depth within the bilayer [50]. Extracting the quantitate information about membrane penetration, however, is more challenging and requires the application of the Distribution Analysis (DA) method [31,48]. First, fluorescence quenching profiles (QPs) are generated from fluorescence measurements with a series of lipid-attached quenchers (nitroxide probes or bromine atoms) of a known depth (h). The DA methodology approximates the transverse QPs of a fluorophore with a symmetrical twin Gaussian function (Eq. (1)), which has three fitting parameters: h_m is the center (mean) of the quenching profile, σ is the width of the distribution, and S is the area of the quenching profile. These fitting parameters correspond to the most probable depth of probe penetration, its fluctuations in the transverse position, and its overall accessibility to quenchers (i.e., fluorescence quenching efficiency), respectively [49]. In order to account for possible trans-leaflet quenching of deeply penetrating fluorophores, the mirror-image G(-h) component is added to main Gaussian component G(h) [49].

$$\text{QP}\left(h\right)=G\left(h\right)+G\left(-h\right)=\frac{\mathrm{S}}{\sigma \sqrt{2\pi }}\exp \left[-\frac{{\left(h-h_{\mathrm{m}}\right)}^2}{2{\sigma }^2}\right]+\frac{\mathrm{S}}{\sigma \sqrt{2\pi }}\exp \left[-\frac{{\left(h+h_{\mathrm{m}}\right)}^2}{2{\sigma }^2}\right]$$ (1)

Recently, we have demonstrated the use of a combination of MD simulations and depth-dependent fluorescence quenching to calibrate the methodology for extracting quantitative information on membrane penetration of the fluorescent-labeled T-domain of diphtheria toxin [64]. Because the T-domain has similar fold in soluble conformation to that of many Bcl-2 proteins (including Bcl-xL), it has been suggested that their topology in membrane-inserted states might be similar as well [65]. To investigate this issue, we have conducted a series of depth-dependent fluorescence quenching measurements for the helices of the central hairpin of Bcl-xL and T-domain and applied the methodology of the Distribution Analysis to obtained quenching profiles (Fig. 4).

Fig. 4.

Illustration of the difference in membrane topology of helices of the central hairpin of Bcl-xL (left panels, helix α6 [36]) and diphtheria toxin translocation domain (right panels, helix α9 [66]). Depth-dependent fluorescence quenching profiles of NBD probe selectively attached to a single cysteine residue at position 161 (a) or 169 of Bcl-xL (b), and at position 378 (c) or 364 of the translocation domain of diphtheria toxin (d). Symbols correspond to the values of the differential quenching profile observed with NBD quenching observed with a series of spin-label quenchers attached at various depth (see Fig. 2c). The transverse position of the maximum of the differential quenching profile, h_m, corresponds to the most probable distance from the bilayer center to the probe. (e) Schematic representation of the membrane topology of α-helix 6 of Bcl-xL and α-helix 9 of the T-domain, based on measurements of several single-Cys-mutants identified in the figure (corresponding h_m parameters are given in parenthesis). The multitude of quenching data indicates that helix a6 of Bcl-xL is deeply penetrated into the bilayer, but doesn't adapt transmembrane topology, similar to a9 of the translocation domain of the toxin.

To ensure complete insertion in the presence or absence of 30 mol% spin-lipid quenchers, steady-state and lifetime fluorescence measurements were carried out in membranes composed of high cardiolipin content of 3TOCL:2POPC at pH 4.5. Five different spin-labeled lipid quenchers of an NBD probe were used with either TEMPO or Doxyl paramagnetic moieties attached at progressively deeper positions along the lipid molecules from its head-group up to a lipid tail. The differential fluorescence quenching profiles (QPs) for the depth-dependent measurements in membranes composed of 3TOCL:2POPC are summarized in Fig. 4a and b. Differential QPs for each tested mutant were obtained by subtracting the dynamic quenching component from steady-state quenching measurements, as described in detail in [31,66]. Finally, the most probable immersion depth (h_m) of each labeled residue were estimated by fitting the data to Eq. (1).

The fluorescence quenching profiles of residues V161C and W169C, located in the middle of the α6 helix in membrane inserted Bcl-xL ΔTM, suggest that these residues are as far as 10–11 Å from the bilayer center (Fig. 4a and b), so that the helix does not insert in a transmembrane orientation. Comparison with the membrane topology of the α-helix 9 of the T-domain of diphtheria toxin is given for the two limiting cases of shallow (P378C—NBD) and deep penetration of the probe (I364C—NBD) in Fig. 4c, d. The analysis of differential QPs of seven residues located along the α9 helix of the T-domain (Fig. 4e) confirming the transmembrane topology [66]. Our results clearly indicate that the two proteins do not share the same topology and while the α9 of the T-domain is in the transmembrane orientation [66], the corresponding a6 helix of Bcl-xL is interfacial [36]. These structural results are also corroborated by other differences in the insertion pathways and the nature of molecular mechanisms of conformational switching [25].

4. Membrane targeting and refolding modulated by lipid composition and divalent cations

4.1. Lipid modulation of Bcl-xL membrane insertion and BH4 domain release

Numerous studies have shown that protein–lipid interactions depend on the membrane composition [43,[67], [68], [69]]. For example, Bax interacts with neutral membranes only at high concentration. Neutral membranes are mainly composed of neutral lipids POPC and POPE, which mimic the outer mitochondrial membrane compositions [67,70]. Membrane interaction of Bax causes its insertion, following by membrane disordering. However, the presence of anionic phospholipids, such as cardiolipin (CL), promotes strong and preferential electrostatic interactions with Bax, targeting it onto the negatively charged membrane surface [70].

It has been shown that the anti-apoptotic Bcl-xL and pro-apoptotic Bcl-2 protein BAX have an inherent ability to interact with membranes in the absence of their canonical activators (BH3-only effector proteins), which, however, requires the presence of cellular concentrations of Mg²⁺/Ca²⁺. Under these conditions, BAX and Bcl-xL are selectively targeted to membranes, refolded, and activated in the presence of anionic lipids, especially the mitochondrial-specific lipid cardiolipin [71].

The Mg²⁺/Ca²⁺-induced membrane interactions of Bcl-xL was studied using an environment-sensitive NBD probe attached to Cys-residue W169C. In the absence of LUV at pH 8, Bcl-xL W169C—NBD reveals a position of emission maximum of 529 nm (Fig. 5a, gray). This band remains mostly unaffected after the addition of 1TOCL:2POPC LUV (Fig. 5a, black). However, the addition of 2 mM Mg²⁺ leads to a twofold intensity increase accompanied by a blue shift to 525 nm (Fig. 5a, red). These spectral changes indicate the transition of NBD-labeled Bcl-xL α6 into the bilayer. These effects were further enhanced by the protonation of Bcl-xL (Fig. 5a, brown) [71].

Fig. 5.

Membrane insertion of Bcl-xL modulated by divalent cations. The insertion of Bcl-xL ΔTM was monitored using fluorescence of NBD probe attached to W169C mutant. (a) Representative NBD spectrum before and after the addition of Mg²⁺ at 25 °C in the presence of 1TOCL:2POPC LUV. (b) Relative membrane insertion of Bcl-xL quantified using the increases in NBD fluorescence intensity at 510 nm. The membrane-inserted populations were proportional to [Mg²⁺]/[Ca²⁺] [71].

The pH dependence of fluorescence changes of Bcl-xL W169C—NBD indicates a crucial role of Mg²⁺ in its membrane insertion and suggests some correlation between the presence of Mg²⁺ and protein protonation [71]. However, these data shows also that the presence of Mg²⁺does not induce the membrane insertion of the entire Bcl-xL population at pH 8 (Fig. 5a). The coupling between divalent cations and solution pH was further characterized by performing pH titrations in the presence of variable concentrations of Mg²⁺ and Ca²⁺ (Fig. 5b). In the absence of divalent cations, Bcl-xL insertion is strongly pH dependent (Fig. 5b, black). In the presence of Mg²⁺ and Ca²⁺, the Bcl-xL inserted fraction was increased at neutral pH. At basic pH, the pH dependence converges to some plateau without changes in the apparent pK_a of Bcl-xL (Fig. 5b). It should also be noted that both Mg²⁺ and Ca²⁺ have the same effect on insertion of Bcl-xL [71].

4.2. Kinetics and thermodynamics of Bcl-xL membrane insertion and refolding

In the soluble conformation of Bcl-xL, the N-terminal BH4 helix (α1) is packed against the hydrophobic α6 helix and is connected through a long loop to the rest of the protein fold. Bilayer insertion of Bcl-xL destabilizes its helix packing that subsequently result in its refolding in the membrane, releasing the regulatory BH4 domain (Fig. 6). The latter domain has been suggested to play an important role in a non-canonical anti-apoptotic mechanism [72], [73], [74]. Therefore, establishing the membrane topology of the inserted form of Bcl-xL is crucial to understanding of apoptotic repression.

Fig. 6.

Determination of the conformation of the C-terminal helix α8 during membrane insertion of the Bcl-xL. (a) The anchoring via α8 helix was measured by the protection of the NBD fluorophore from the soluble quencher dithionite. The NBD probe was attached in the middle of α8 at G222C of Bcl-xL inserted in 1TOCL:2POPC LUV (pH 7.5, 37 °C). For soluble Bcl-xL, adition of sodium dithionite results in a rapid and complete decrease in fluorescence signal (gray, black). In contrast, when lipid nanodiscs are formed in the presence of Bcl-xL, little quenching is observed (green), consistent with the transmembrane orientation of α8. The insertion of full-length Bcl-xL (Bcl-xL FL) into vesicles in the presence of divalent cations results in intermediate signal quenching (Mg²⁺, blue; Ca²⁺, red), suggesting a distribution of interfacial and transmembrane topologies (b) Pore-inhibiting action of Bcl-xL in the absence of BH3 effectors (1TOCL:2POPC, pH 7.5, 37 °C). The ANTS/DPX leakage assay reveals that BH3-independent BAX mediated leakage by Bcl-xL FL and Bcl-xL is inhibited in the absence of BH3-effectors in the presence of 1 mM Mg²⁺. Strong inhibition occurring in the absence of helix α8, indicates that prior anchoring is not required of action on the membrane-inserted and refolded Bcl-xL.

To monitor the kinetics of the membrane targeting of Bcl-xL FL and membrane topology of an anchored conformation, the intensity of NBD-labeled Bcl-xL FL at its C-terminal tail (G222C) was measured followed by quenching by irreversible soluble quencher dithionite. Sodium dithionite is membrane-impermeable and quenches water-assessable NBD-labeled sites on membrane proteins assessable on the outer lipid leaflet that allow discriminating between interfacial and transmembrane membrane topologies (Fig. 6a) [75], [76], [77], [78].

In order to determine the range of protection of the NBD probe attached to C-terminal residue G222C from dithionite quencher, two samples can be compared: (1) soluble Bcl-xL in the absence of membranes, and (2) tail-anchored Bcl-xL. It should be noted that the membrane-anchored conformation of the protein does not spontaneously anchor itself upon mixing with LUVs and is usually prepared in vitro by the formation of nanodiscs around the anchoring helix [13,79,80]. Therefore, tail-anchored Bcl-xL was prepared so that lipid nanodiscs were formed around the C-terminal helix. Fig. 6a(gray) shows that in the absence of membranes, the addition of dithionite to the soluble form of Bcl-xL led to a rapid decrease of the probe intensity in due to complete exposure of the NBD fluorophore to the quencher. In contrast, the anchoring of the C-terminal helix into lipid nanodiscs prevented signal quenching due to the protection of the NBD probe by the bilayer (Fig. 6a, green). The addition of 1TOCL:2POPC LUV in the absence of divalent cations resulted in no protection (Fig. 6a, black); however, the addition of either 2 mM Mg²⁺ or Ca²⁺ (Fig. 6a, blue and red) recovers an intermediate level of the NBD protection.

To determine whether the Bcl-xL conformation that releases its BH4 helix is active, either Bcl-xL FL (Fig. 6b, blue) or Bcl-xL (Fig. 6b, red) were incubated with 1TOCL:2POPC LUV containing the fluorophore/quencher pair 8-aminonapthalene-1,3,6 trisulfonic acid/p-xylene-bis-pyridinium bromide (ANTS/DPX) in the presence 1 mM Mg²⁺. An ANTS/DPX fluorescence de-quenching-based LUV leakage assay allows measuring membrane permeabilization by pore-forming proteins [81]. Disruption of liposome integrity and subsequent leakage result in measurable fluorescence emitted by ANTS dye. Fig. 6b shows that under these conditions, the addition of BAX resulted in only 10% leakage and it was independent the used Bcl-xL construct. At the same time, in the absence of Bcl-xL/Bcl-xL FL up to 85% membrane leakage was observed (Fig. 6b, gray). The large decrease in BAX-induced leakage in the presence of Bcl-xL demonstrates that the BH3-independent membrane-inserted form of Bcl-xL in high cardiolipin membranes is active and is able to inhibit BAX activity. Furthermore, some similarity between the inhibitory effects of both Bcl-xL variants demonstrates that the C-terminal helix does not play a crucial role in the BH3-independent antiapoptotic activity of Bcl-xL [71].

Fluorescence technique allows discriminating between the kinetics of Bcl-xL membrane insertion and refolding by combining FRET measurements and the NBD “turn-on” fluorescence effect (Fig. 7). FRET measurements were performed in 1TOCL:2POPC LUV in the presence of 1 mM Mg²⁺ or Ca²⁺, which is the most favorable refolding conditions according to the steady-state data shown in Fig. 6a. We found that both divalent cations revealed similar kinetic traces for membrane insertion (Fig. 7a, triangles) and refolding (Fig. 7a, circles). This behavior provides further evidence for the equivalency of Mg²⁺ and Ca²⁺ for the BH3-independent membrane interactions of Bcl-xL observed earlier by the steady-state measurements. Importantly, these results also showed that Bcl-xL insertion and refolding in the bilayer do not occur simultaneously, so that Bcl-xL refolding is delayed relative to its membrane insertion. This complex kinetic behavior further suggests that Bcl-xL can be associated with the bilayer without incurring extensive refolding and the release of BH4 helix, trapped by a relatively high-energy barrier [71].

Fig. 7.

Comparison of the kinetics of membrane binding and refolding of Bcl-xL. (a) Time dependence of Bcl-xL membrane insertion (triangles) and BH4 release (circles) in the presence of 1 mM Mg²⁺ (blue) or 1 mM Ca²⁺ (red). (b) Scheme of the joint use of FRET measurements and the NBD “turn-on” fluorescence effect to trace the membrane insertion and refolding of Bcl-xL modulated by caldiolipin content [71].

The activation energy required for Bcl-xL refolding was estimated by measuring the rates of refolding as a function of the temperature in a range from 25 to 42 °C in the presence of 1 mM Mg²⁺ or Ca²⁺. After plotting these temperature dependence as an Arrhenius plot, ln(k) vs. 1/T, where T is an absolute temperature, the activation energy was found to be E_a=30±5 kcal/mol [71].

As demonstrated by our FRET studies (Fig. 3), acidic pH induced membrane insertion of Bcl-xL is accompanied by the release of its N-terminal BH4 helix (Fig. 3b and c) [36,43]. The same FRET methodology and the construct can be utilized to study the kinetics of Bcl-xL insertion and refolding in the presence of Mg²⁺ or Ca²⁺, as summarized in Fig. 8 [71]. The BH3-independent release of the BH4 helix was further characterized by single-molecule FRET (smFRET) using the same AlexaFluor488-labeled mCherry-Bcl-xL construct [43,71]. In the presence of 1TOCL:2POPC LUV at pH 7.5 but in the absence of divalent cations, AlexaFluor488-Bcl-xL- mCherry construct is characterized by a smFRET efficiency distribution centered at 0.4 (Fig. 8b, gray). This value is consistent with the ensemble-measured FRET efficiency acquired by conventional steady-state and lifetime measurements [43]. Addition of 2 mM Mg²⁺ decrease the smFRET efficiency up to 0.05 (Fig. 8b, purple), while at 0.5 mM Mg²⁺ a distribution was centered in between those determined at 0 and 2 mM Mg²⁺ (Fig. 8b, orange). Multiple intermediate smFRET efficiencies were observed at various concentrations of Mg²⁺ (Fig. 6c, green), indicating that the BH3-independent release of the BH4 helix involves multiple Bcl-xL conformations with different degrees of compactness [71]. The presence of multiple Bcl-xL conformations in the bilayer is consistent with the protonation-dependent refolding of Bcl-xL [43]. These smFRET efficiencies (Fig. 8c, green) also agree well with those of ensemble FRET measurements (Fig. 8c, pink).

Fig. 8.

Evidence for multiple folding intermediates of membrane-inserted Bcl-xL. (a) Scheme of multiple refolding intermediate of Bcl-xL monitored by intramolecular FRET between two fluorophores. (b) FRET-measured refolding of membrane-inserted Bcl-xL in 1TOCL:2POPC LUV at pH 7.5 and 25 °C. FRET was monitored between an AlexaFluor488 donor fluorophore attached at Bcl-xL D189C and mCherry acceptor conjugated at the N terminus of Bcl-xL. Representative smFRET distributions of Bcl-xL measured using fluorescence correlation spectroscopy (FCS) at different concentrations of Mg²⁺: 0 mM Mg²⁺ (black) soluble Bcl-xL, 2 mM Mg²⁺ (purple) membrane-inserted Bcl-xL with BH4 released, and 0.5 mM Mg²⁺ (orange) membrane-inserted Bcl-xL with intermediate compactness. (c) Comparison of FRET efficiencies calculated for ensemble measurements (pink) and single-molecule FCS regime (green). (d) Increase in cardiolipin content modulates lipid-dependent Bcl-xL refolding in membranes in the presence of varying concentrations of Mg²⁺ (blue) or Ca²⁺ (red) [71].

Finally, to monitor the impact of cardiolipin on membrane targeting of Bcl-xL, a varying fraction of cardiolipin was incorporated into the model POPC bilayer. In 1TOCL:2POPC LUV, the presence of cytosolic 1 mM Mg²⁺ led to 60% of the Bcl-xL population to release its BH4 helix, while only 10% refolding was observed in 1TOCL:6POPC LUV and 0% in POPC LUV. Therefore, the release of the BH4 helix become significantly higher in membranes composed of larger contents of cardiolipin (Fig. 8d). These findings suggest that alteration in mitochondrial lipid composition in the presence of divalent cations affects the propensity for membrane bound Bcl-xL and enables switching among different conformational states [71].

Our FRET-based studies of protein refolding suggest that the BH3-independent insertion leads to multiple membrane-inserted conformations of Bcl-xL in the bilayer, and that cardiolipin enables switching among these structures. Two of these conformations can be readily identified upon membrane insertion: (i) at high cardiolipin content: membrane-inserted Bcl-xL with a released BH4 helix and (ii) at low cardiolipin content: membrane-inserted Bcl-xL with a non-released BH4 helix. Structural insights into the dependence of the conformational dynamics of Bcl-xL on protonation states and membrane composition have recently been studied by molecular dynamics simulations [82]. The coupled effect of protonation and cardiolipin content on the membrane-anchored conformations of Bcl-xL was observed, suggesting that the presence of cardiolipin produces more favorable protonation-dependent insertion free energy.

5. Conclusions and perspectives

The field of Bcl-2-controlled apoptosis is in a peculiar state, marked by apparent contradictions [83]. On one hand, a number of high-resolution structures have become available, but almost exclusively of soluble conformations, often with deleted membrane anchors and unfolded loops [63]. The structural information for membrane-inserted conformations is scarce and is often modeled solely from solution structures. On the other hand, the general consensus from cellular studies is that the action of Bcl-2 proteins is intimately associated with their membrane interactions that produce active conformations [6,7,11]. The nature of the modulation of regulatory interactions within the Bcl-2 family by the membrane remains elusive.

In this review, we summarize recent achievements in studying the membrane targeting, insertion and refolding of Bcl-xL using multiple fluorescence tools (more detailed applications of fluorescence methodology to other membrane-inserting proteins are presented in [28,30]. The role of the enrichment of the mitochondrial outer membrane with anionic lipids, primarily cardiolipin is summarized in Fig. 9. We demonstrate that the lipid-dependent activation is coupled to the presence of physiologically relevant concentrations of divalent cations, such as Mg²⁺ and Ca²⁺ [71]. (These interactions should not be confused with the role of Ca²⁺ as a second messenger, which requires a highly specific action of Ca²⁺ over that of Mg²⁺, which is clearly not the case here. This lack of selectivity has important implications for the Bcl-2-bilayer interactions, given that cytoplasmic concentrations of Mg²⁺ are in millimolar range. Consequently, it should be emphasized that the regulatory changes are not expected to be driven by the changes in the concentration of divalent cations, but through the changes in lipid composition [71]). Membrane insertion of Bcl-xL results in the release of the BH4 regulatory domain from the folded structure via a series of intermediate steps [36,43], ultimately leading to inhibition of pore formation by BAX (Fig. 9). The use of a series of spectroscopic measurements provides a powerful technique to rationale mechanistic aspects of recent cellular studies that demonstrate the mitochondrial localization and activity of Bcl-2 proteins in cells lacking BH3-only effectors. In addition, our data support the existence of the conformation of Bcl-xL capable of acting via a non-canonical mode of inhibition (Fig. 1b). This confirms previous suggestions that the BH4 domain as a valid drug target for cancer therapies for activation of apoptosis by “inhibiting the inhibitor” (i.e., blocking the anti-apoptotic action of Bcl-xL and allowing BAX and other pro-apoptotic factors to permeabilize MOM).

Fig. 9.

Schematic summary of lipid modulation the BH3-independent membrane targeting and activation of Bcl-xL and BAX.

Future advances in studies of various cellular processes involving conformational switching between soluble and membrane-bound protein conformations require not only advanced experimental tools, but also development of new computational approaches. One such approach, involving microsecond-scale molecular dynamics (MD) simulations, has been already applied to the anchored conformation of Bcl-xL and revealed lipid-dependent modulation of its structure and dynamics [82]. The applications of atomistic MD simulations to refolded and membrane inserted conformations, however, are complicated by the lack of structural information needed to set up the initial conformation. Recently, some encouraging progress has been made in application of low-resolution methods, such as fluorescence depth-dependent quenching, in overcoming this limitation [66]. Another aspect of computational methods that can abate studies of protein-membrane interactions is the advancement of sequence-based prediction tools. One of the best examples of the latter is a web-based tool MPEx [84] (https://blanco.biomol.uci.edu/mpex), which has been recently updated to predict binding in the presence of physiologically important anionic lipids [85]. Another advancement that is critically required for accurate predictions, is the ability to account for the action of divalent cations which affect all aspects of thermodynamics of protein-membrane interactions [86]. Recently, divalent cations have been implicated in other systems undergoing membrane insertion, noticeably a cancer-targeting peptide pHLIP [69,87], providing possible mechanistic insights into conformational switching of many systems, including Bcl-xL.

Data availability

No data was used for the research described in the article.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

• No data was used for the research described in the article.