生物谷報(bào)道：Bcl-2家族是細(xì)胞凋亡是最關(guān)鍵性蛋白質(zhì)家族之一,，至少有14個(gè)成員，可分為抗凋亡因子和致凋亡因子兩大類,，這里介紹Bcl-w的作用的線粒體機(jī)制,。使大家對(duì)它有一個(gè)全面的認(rèn)識(shí)。

In cells, proteins often use membranes as partners to perform their biological function. This is especially true for members of the Bcl-2 family, which are crucial proapoptotic (Bax and Bak) and antiapoptotic (Bcl-2, Bcl-w and Bcl-xL) regulators [1]. These proteins are fully functional only when located on membranes of the endoplasmic reticulum (ER) or mitochondria. The survival factor Bcl-2 enters a tight partnership with intracellular membranes immediately following its synthesis ( Figure 1c). By contrast, the apoptotic factor Bax is retained in the cytosol because its membrane-targeting C-terminal ‘arm’ (domain) is folded back into a hydrophobic ‘pocket’ in the protein. In response to apoptotic stimuli, this arm can be released, allowing Bax to establish a tight ‘liaison’ with its partner membrane and perforate it (Figure 1d). Wilson-Annan and colleagues [2] have recently reported a novel type of membrane partnership for the survival factor Bcl-w. In surviving cells, Bcl-w has its C-terminal arm folded back, similar to Bax, but it is already loosely associated with the mitochondrial membrane and is functionally active ( Figure 1a). In response to apoptotic stimuli, the BH3-only protein Bim unleashes the C-terminal arm of Bcl-w, thereby leading to its inactivation and membrane insertion. The present article discusses this recent work in the context of mechanisms described for membrane targeting and/or insertion and protein–protein interactions of Bcl-2 family members.

 
 (32K)

Figure 1. Models of membrane targeting and/or insertion of the Bcl-2-family members Bcl-w (a), Bcl-xL (b), Bcl-2 (c) and Bax (d) in healthy and apoptotic cells. Although Bcl-2 has its C-terminal domain ‘arm’ exposed after synthesis and immediately inserts into ER and mitochondrial membranes (c), Bcl-w and Bax have their C-termini folded into a hydrophobic ‘pocket’ (a,d). Although the loose association of Bcl-w with the mitochondrial membrane has been shown by Wilson-Annan et al. [2], that of Bax is hypothetical as are the putative binding-proteins Y (probably humanin or 14–3-3 proteins) and Z (probably Bid). In healthy cells, depending on cell type and conditions, Bcl-xL could be cytosolic, loosely attached to mitochondria or inserted into the mitochondrial membrane (b); Bcl-2 and Bcl-xL most probably bind to proapoptotic effector proteins (X) at the hydrophobic pocket (b,c), whereas Bcl-w binds to these proteins at its C-terminal arm (a). In apoptotic cells, Bim might directly displace the proapoptotic effectors X in Bcl-xL and Bcl-2 (b,c), whereas in Bcl-w, it displaces the C-terminus after which the effector proteins are released (a). These effectors then activate Bax and/or Bak and trigger mitochondrial damage. Ch, chaperones.

1. Current model of the mode of action of Bcl-2-like survival factors
Following synthesis, Bcl-2-like survival factors are targeted to either ER or mitochondrial membranes, into which they integrate through their hydrophobic C-terminal arm [3, 4 and 5]. In this state, they probably sequester postulated proapoptotic effectors ( Figure 1c) required for the activation of Bax, Bak and/or caspases by a hydrophobic pocket that is formed from the four conserved domains (BH1–4) of these proteins [1]. In apoptotic cells, BH3-only proteins, which are sensor proteins containing a single Bcl-2 homology 3 (BH3) domain, such as Bim, are either transcriptionally induced or posttranslationally activated and displace the proapoptotic effectors at the hydrophobic pocket of Bcl-2-like proteins by competitive binding ( Figure 1c) [6]. The overexpression of Bcl-2-like proteins protects cells from apoptosis because they create interfaces sufficient to sequester both the death-triggering BH3-only proteins and the proapoptotic effectors [1].

The study by Wilson-Annan et al. [2] suggests that the current model describing the mechanistic action of Bcl-2-like proteins might not entirely apply to the survival factor Bcl-w. In contrast to Bcl-2 and Bcl-xL, which most probably have their hydrophobic C-termini exposed – three-dimensional structures of these proteins can only be obtained with deleted C-termini – and therefore need rapid membrane targeting and/or insertion [7 and 8], Bcl-w has its C-terminal arm folded back into its hydrophobic pocket and, similar to Bax, can be soluble [9, 10 and 11]. Consistent with this, Wilson-Annan et al. [2] found both endogenous and overexpressed Bcl-w in a cytosolic subfraction of healthy cell lysates. However, immunoelectron microscopy of the same cells revealed Bcl-w associated with mitochondria. Thus, Bcl-w is loosely associated with mitochondria ( Figure 1a) and readily dislodges after cell lysis, presumably because its C-terminal arm is inaccessible for membrane insertion.

2. From loose to tight: switch of Bcl-w activity
An obvious question is what happens to Bcl-w in apoptotic cells? Wilson-Annan et al. [2] nicely show that both endogenous and overexpressed Bcl-w become inserted into the mitochondrial membrane in response to various apoptotic stimuli, such as -radiation, staurosporine, UV, etoposide and cytokine deprivation, and bind to the BH3-only proteins BimEL (extra-long) and BimL (long) in coimmunoprecipitates. To determine whether Bim is sufficient to trigger membrane insertion of Bcl-w, the authors performed a series of elegant experiments. First, they used an optical biosensor (Box 1) to determine the binding affinity of BimL for Bcl-w. The affinity of this wild-type peptide (wtBH3) was high (22 nM), but BH3-mutants with one (L94ABH3) or four point-mutations (4EBH3) in the key binding residues to the hydrophobic pocket had reduced and abolished binding, respectively. Second, when these peptides were incubated with cell lysates, the wtBH3 caused Bcl-w to insert into membranes, whereas the mutants had reduced or no ability to do so. Third, when the wtBH3 peptides, or the 4EBH3 mutants, were tethered to the N-terminus of Bcl-w, by a flexible linker region to mimic the BH3-ligated conformer of Bcl-w, wtBH3–Bcl-w, but not 4EBH3–Bcl-w, chimera inserted into the membrane upon stable expression and was inaccessible to added BimEL in coimmunoprecipitation assays. The survival activity of the chimera was neutralized, that is, the stable expression of Bcl-w neither promoted apoptosis nor protected the cells from it. Importantly, it was the C-terminal arm of Bcl-w that mediated membrane insertion because a chimera with this domain excised did no longer associate with membranes. These data clearly show that in apoptotic cells the BH3 peptide of Bim occupies the hydrophobic pocket of Bcl-w, thereby displacing its C-terminal arm, which leads to membrane insertion. What remains unanswered is whether membrane insertion of Bcl-w is just a logical consequence of the displacement of the C-terminal arm or it actively contributes to the loss-of-survival activity of Bcl-w. This can be answered by testing the survival activity of a Bcl-w mutant containing a permanently exposed C-terminal arm such as that of Bcl-2 or TOM5, a tail-anchored component of the TOM import machinery on the mitochondrial outer membrane [12].

Box 1. Principles of the Biacore's surface plasmon resonance (SPR) biosensor

SPR-based biosensors monitor interactions between biomolecules by measuring their mass concentration close to the surface of a sensor chip. A major advantage of this technology is that molecules can be studied in their native state because there is no need to label them with fluorescent or radioactive tags.

The sensor surface is made specific by immobilizing one of the interaction partners (target) on a gold-coated glass slide modified with a carboxymethylated dextran, which forms a hydrophilic environment for attached biomolecules, thereby preserving them in a nondenatured state (Figure I). A sample containing the other partner(s) continuously flows over the sensor surface (microfluidic system), maintaining constant analyte concentrations at the surface of sensor chip. The gold layer in the sensor chip creates the physical conditions required for SPR. SPR arises when light is reflected from a conducting film (the gold layer) at the interface between two media – the interacting sample and the glass of the sensor chip – of different refractive indices. SPR causes a reduction in the intensity of reflected light at a specific angle of reflection [states ‘A’ and ‘B’ (Figure I)]. This angle varies with the refractive index that is close to the surface opposite from the reflected light. When molecules in the sample bind to the interactant attached to the sensor surface, the local concentration and therefore the refractive index at the surface changes and the SPR response is detected. Plotting the response against time during the course of an interaction provides a real-time quantitative measure (association and dissociation) of the progress of interaction (sensorgram). The response values of SPR are expressed in resonance units (RU); one RU represents the change of 0.0001 degree in the angle of the minimum intensity ( Figure I, upper plot). For most proteins, this is roughly equivalent to a change in concentration of ~1 pgmm−2 on the sensor surface. The calculations of association and/or dissociation rates, as well as binding affinities are described by Lackmann et al. [24].

 
 (30K)

Figure I. A diagram of the principles of the Biacore's surface plasmon resonance (SPR) biosensor. ‘A’ and ‘B’ are two different resonance signals in resonance units (RU) produced by the reflected light from a preceding sample–target interaction. For a quick-time animation of the SPR technology, see (http://www.biacore.com/technology/spr_technology.lasso) from where this figure was adapted. (BIACORE® is a registered trademark of Biacore AB. Copyright © Biacore 2003)

3. Sequestration of proapoptotic effectors by Bcl-w: an arm instead of a pocket
Wilson-Annan et al. [2] propose a model for the regulation of Bcl-w that is different from that of Bcl-2 ( Figure 1). In contrast to Bcl-2, which has its hydrophobic pocket accessible for binding to proapoptotic effectors in healthy cells, the pocket of Bcl-w is occupied by its C-terminal arm. Because Bcl-w is an active survival factor in this state, it must bind to the proapoptotic effectors by another domain. Wilson-Annan et al. [2] suggest that this domain is the backfolded C-terminal arm. Interestingly, some of these authors previously reported that although Bcl-w(ΔC5) (lacking the last 5 amino acid residues) still exerted survival activity, Bcl-w(ΔC10) lost this activity [10] raising the exciting possibility that proapoptotic effectors bind to amino acids 184–188 of Bcl-w. Upon membrane insertion of Bcl-w, which is provoked by a BH3-only protein, these effectors would be displaced from the C-terminal arm and trigger apoptosis. This model also nicely explains the antiapoptotic activity of overexpressed Bcl-w, which again is different from Bcl-2. Although overexpressed Bcl-2 sequesters both proapoptotic effectors and BH3-only proteins in its hydrophobic pocket, overexpressed Bcl-w sequesters these proteins in two different conformers. The membrane-inserted part of Bcl-w binds to BH3-only proteins, such as Bim, and the rest of this protein, which cannot insert and remains loosely attached to the membrane, binds to the proapoptotic effectors ( Figure 1). Indeed, Wilson-Annan et al. [2] show that only a third of overexpressed Bcl-w inserts into the membrane in response to apoptotic stimuli, whereas the rest remains loosely associated. Moreover, the two conformers of Bcl-w elute differently on gel filtration raising the possibility that binding proteins for each form could be isolated.

4. Concluding remarks
As with all good studies that propose new models, more questions arise. First, how Bcl-w can be specifically targeted to mitochondria when its C-terminal domain, a putative targeting sequence, is not unleashed (Figure 1a)? Does the back-folded C-terminus constitute a targeting signal that is accepted by a specific mitochondrial receptor, or is targeting mediated by another domain? Second, what happens to Bcl-w after membrane insertion? On the basis of their gel filtration data and the appearance of clusters detected by immunoelectron microscopy, Wilson-Annan et al. [2] propose that Bcl-w oligomerizes. Oligomer formation and clustering has been seen for Bax after insertion [13 and 14], but in contrast to Bcl-w, oligomerized Bax is cytotoxic [2]. Would oligomerization lead to another function of Bcl-w or do the large protein complexes found by gel filtration contain other, unidentified, protein partners? Finally, what can be learned from the regulation of Bcl-w about other Bcl-2-family members? Is it possible that, in healthy cells, Bax is not cytosolic [15 and 16] but loosely associated with mitochondria where it binds to inhibitory factors to its, possibly still backfolded, C-terminal arm ( Figure 1d)? Both the loose association of Bax in healthy cells [17] and its potential C-terminal-binding partners, such as humanin and 14–3-3 ( Figure 1d) have recently been suggested [18 and 19]. A BH3-only protein, such as Bid ( Figure 1d), could trigger C-terminal release and membrane insertion of Bax, as previously proposed [20]. And what about the survival factor Bcl-xL? Wilson-Annan et al. [2] propose that it might act similar to Bcl-w ( Figure 1b) (also see [10]). Indeed, in healthy cells, Bcl-xL exists in the cytosol and/or loosely attached to mitochondria, whereas in response to apoptotic stimuli, it translocates to and inserts into the mitochondrial membrane [21, 22 and 23]. However, this distribution might be cell-type specific, because in another study both endogenous and overexpressed Bcl-xL were found to be mostly membrane-inserted in healthy cells [5]. Moreover, in contrast to Bax and Bcl-w, Bcl-xL cannot be crystallized as an entire protein, indicating that it probably has its C-terminal arm exposed [7]. In this respect, a chaperone which binds to the C-terminal arm and/or hydrophobic pocket of Bcl-xL might have to accompany the survival factor to its ‘wedding’ with the mitochondrial membrane (Figure 1b). Thus, each Bcl-2 family member might have its own strategy to regulate anti or proapoptotic activity by membrane targeting and insertion. With more excellent studies, similar to that of Wilson-Annan et al.[2], the details of these strategies will probably be uncovered soon.

全文：Trends in Cell Biology,，Volume 14, Issue 1 , January 2004, Pages 8-12