Skip to main content

Thank you for visiting You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

A novel Bcl-x splice product, Bcl-xAK, triggers apoptosis in human melanoma cells without BH3 domain


Pro- and antiapoptotic proteins of the large Bcl-2 family are critical regulators of apoptosis via the mitochondrial pathway. Whereas antiapoptotic proteins of the family share all four Bcl-2 homology domains (BH1–BH4), proapoptotic members may lack some of these domains, but all so far described proapoptotic Bcl-2 proteins enclose BH3. The bcl-x gene gives rise to several alternative splice products resulting in proteins with distinct functions as the antiapoptotic Bcl-xL and proapoptotic Bcl-xS. Here, we describe a novel Bcl-x splice product of 138 amino acids termed Bcl-xAK (Atypical Killer), which encloses the Bcl-2 homology domains BH2 and BH4 as well as the transmembrane domain, but lacks BH1 and BH3. Weak endogenous expression of Bcl-xAK was seen in melanoma and other tumor cells. Interestingly, its overexpression by applying a tetracycline-inducible expression system resulted in significant induction of apoptosis in melanoma cells, which occurred in synergism with drug-induced apoptosis. After exogenous overexpression, Bcl-xAK was localized both in mitochondrial and in cytosolic cell fractions. By these findings, a completely new class of Bcl-2-related proteins is introduced, which promotes apoptosis independently from the BH3 domain and implies additional, new mechanisms for apoptosis regulation in melanoma cells.


Programmed cell death/apoptosis is a biological process essential for normal development and maintenance of tissue homeostasis (Petros et al., 2004). Excess of apoptosis contributes to a wide variety of pathologic conditions including AIDS as well as cardiovascular, infectious and neurodegenerative diseases (Osford et al., 2004). On the other hand, apoptosis resistance is a common feature of human malignancies, contributing both to cancer development as well as to resistance to conventional therapies such as radiation and cytotoxic drugs (Johnstone et al., 2002).

Bcl-2-related proteins are strongly engaged in apoptosis control, and are characterized by one to four structural motifs termed Bcl-2 homology domains (BH1–BH4). Antiapoptotic Bcl-2 proteins including Bcl-2, Mcl-1, Bcl-w and Bcl-xL share all four domains. In contrast, proapoptotic Bcl-2 homologs are missing one or several BH domains and are subdivided into multidomain proteins (including Bax, Bak, Bok/Mtd and Bcl-xS) as well as BH3-only proteins (including Puma, Noxa, Bid, Bad, Nbk/Bik) (Borner, 2003; Daniel et al., 2003). A characteristic feature of Bcl-2 proteins is the formation of homodimers and heterodimers with other members of the family. For example, Bcl-2 can form heterodimers with Bax, Bad or other proapoptotic molecules, and may thus block their proapoptotic activity (Gross et al., 1999). Inefficient apoptotic signalling in melanoma cells has been attributed to high expression of the antiapoptotic Bcl-2 proteins (Raisova et al., 2001; Hossini et al., 2003a; Hussein et al., 2003; Oppermann et al., 2005).

Of the bcl-x gene, several alternative splice products had been described giving rise to functionally distinct proteins such as Bcl-xL and Bcl-xS (Shiraiwa et al., 1996; Grillot et al., 1997; Ban et al., 1998). As Bcl-2, the four BH domain-containing Bcl-xL can heterodimerize with proapoptotic Bcl-2 proteins and thus inhibit apoptosis (Diaz et al., 1997), whereas the proapoptotic Bcl-xS containing only BH3 and BH4 antagonizes survival proteins such as Bcl-2 and Bcl-xL (Yang and Korsmeyer, 1996; Kelekar and Thompson, 1998).

In the present investigation, we identified Bcl-xAK (Atypical Killer), a novel splice form of the bcl-x gene, which encloses BH4 and BH2 and lacks BH3 and BH1, but nevertheless exerts proapoptotic functions in human melanoma cells. This is the first report on a Bcl-2 protein, which induces apoptosis without a functional BH3 domain.


Identification of a novel bcl-x splice variant missing BH1 and BH3

Western blot analyses for Bcl-xS and Bcl-xL had often revealed an extra band of lower molecular weight than Bcl-xS. For identifying this unknown protein, reverse transcriptase–polymerase chain reaction (RT–PCR) from total RNA of melanoma cells was performed with primers covering the whole open reading frame of Bcl-xL. After subcloning the RT–PCR products, along with clones for Bcl-xS and Bcl-xL, a third type of clones was found. The cDNA was characterized by DNA sequence analysis and revealed an open reading frame of 138 amino acids. It is derived from the bcl-x gene and shares identical amino and carboxy termini with Bcl-xL including BH4 and BH2 but excludes the central portion carrying BH3 and BH1. It also contains the interdomain region between BH4 and BH3 of 41 amino acids as well as the transmembrane anchor (Figure 1). This new splice variant was termed Bcl-xAK (Atypical Killer, see below).

Figure 1

Bcl-xAK lacks characteristic domains of Bcl-xL and Bcl-xS. (a) Amino-acid sequence alignment of Bcl-xL, Bcl-xS and Bcl-xAK. The amino-acid sequence of Bcl-xAK was deduced from the sequence of the cloned cDNA. Bcl-2 homology domains (BH1–BH4) and the transmembrane region (TM) are indicated. Amino acids missing in Bcl-xS or in Bcl-xAK are indicated by hyphens. (b) Schematic comparative listing of conserved domains.

Endogenous expression of Bcl-xAK in melanoma cells

Endogenous expression of Bcl-xAK protein was investigated by Western blotting in nine human melanoma cell lines (A-375, SK-Mel-13, Mel-2a, SK-Mel-23, M-221, Mel-HO, MeWo, M-186, M-5), in normal human melanocytes (NHM), in a cervical tumor cell line (HeLa) and in a neuroectodermal tumor cell line (PFSK-1). Weak but significant expression of Bcl-xAK with a size of 18 kDa was seen in eight of nine melanoma cell lines, in both normal human melanocyte cultures and both nonmelanoma cell lines, investigated. In contrast, expression of Bcl-xS protein with a size of 20 kDa was considerably weaker in all cell lines, studied, and appeared only as a faint protein band, although a high amount of protein extracts had been applied in Western analysis (200 μg). As compared to Bcl-xL, endogenous expression of Bcl-xAK and of Bcl-xS was much weaker (Figure 2a).

Figure 2

Significant endogenous expression of Bcl-xAK in melanoma cells. (a) Expression of Bcl-xAK, Bcl-xS and of Bcl-xL as determined by Western blot analysis in nine human melanoma cell lines (A-375, SK-Mel-13, Mel-2a, SK-Mel-23, M-221, Mel-HO, MeWo, M-186, M-5), two independent normal human melanocyte cultures (NHM-1, NHM-2) and two nonmelanoma cell lines (HeLa, PFSK-1). Equal protein amounts (200 μg) were loaded in each lane and consistent blotting was confirmed by ponceau staining. As positive controls for Bcl-xAK and Bcl-xS, extracts of SKM13-Tet-On cells stably transfected with either Bcl-xAK (SKM13-Bcl-xAK) or with Bcl-xS (SKM13-Bcl-xS) were used (each 10 μg). As Bcl-xL was much stronger expressed than Bcl-xAK and Bcl-xS, its corresponding part of the membrane was incubated separately with antibody. Exposure times were 30 min (Bcl-xL) and overnight (Bcl-xAK). (b) Length comparison of a nonspecified protein band (ns, 22 kDa) and Bcl-xS. Cell extracts of three melanoma cell lines (M-221, MeWo and M-186, each 200 μg) were analysed side by side with extracts of SK-Mel-13 stably transfected with Bcl-xS (SKM13-Bcl-xS; 10 μg). Western blot analysis clearly proved that this band is distinct from Bcl-xS. (c) Products of nested RT–PCR representing Bcl-xL, Bcl-xS and Bcl-xAK mRNA expression in SK-Mel-28, A-375 and M-5 melanoma cells as well as in PFSK-1 and HeLa nonmelanoma cells were analysed by agarose gel electrophoresis. As a positive control, SKM13-Tet-On melanoma cells stably transfected with the tetracycline-inducible Bcl-xAK construct were used (SKM13-Bcl-xAK).

In four of the melanoma cell lines (M-221, MeWo, M-186 and Mel-2a), an extra nonspecified protein band (n.s.) was seen of similar molecular weight as Bcl-xS (22 kDa). To rule out the possibility that this extra band corresponded to Bcl-xS, protein extracts of three cell lines positive for this band were loaded side by side with the positive control for Bcl-xS (SK-Mel-13 melanoma cells stably transfected with Bcl-xS). Figure 2b clearly shows, that the 22 kDa band was distinct from Bcl-xS in size.

Also, the respective mRNA was detected by nested RT–PCR in a selection of cell lines positive in Western analysis (Figure 2c).

Exogenous expression of the new splice variant in melanoma cells

To study the function of Bcl-xAK, the cDNA of the new splice variant was subcloned into the expression plasmid pTRE-1 downstream of the tetracycline-inducible promoter. After stable transfection into the tetracycline-regulatable cell line SKM13-Tet-On and clonal selection, cell clones were screened by Western blot analysis. Several cell clones showed expression of an 18 kDa protein after doxycycline induction, thus proving that the new cDNA indeed codes for a stable protein which can be immunologically detected. Induction with doxycycline for 48 h resulted in a strong increase of Bcl-xAK protein in transfected cells (Figure 3a).

Figure 3

Induction of apoptosis by Bcl-xAK in melanoma cells is accompanied by nuclear chromatin condensation and DNA fragmentation. (a) DNA fragmentation triggered by overexpression of Bcl-xAK is compared to the effect of Bcl-xS overexpression. Relative apoptotic rates are given as compared to untreated control cells (−) set to 1. Means and standard deviations of triple values of a representative experiment are shown. The experiment has been repeated at least five times, resulting in highly comparable results. Corresponding protein extracts as determined by Western blot analysis are shown on the right-hand side. Equal protein amounts (50 μg) were loaded. (b) Chromatin condensation and nuclear fragmentation in SKM13-Bcl-xAK cells were visualized by bisbenzimide (Hoechst-33258) staining, 48 h after doxycycline induction. Four typical apoptotic nuclei are indicated by arrows. (c) Examples of typical apoptotic nuclei after Bcl-xAK expression are visualized after higher magnification. (d) Quantification after bisbenzimide staining revealed a high percentage of apoptotic nuclei after Dox induction in Bcl-xAK cells as compared to mock-transfected cells. Values represent the mean of three wells counted±s.d. The experiment has been repeated once giving comparable result. (e) DNA strand breaks were visualized using TUNEL staining, at 48 h after induction with doxycycline. Positive cell examples are indicated. (f) Quantification after TUNEL staining revealed an increased number of apoptotic cells after Dox induction in Bcl-xAK cells as compared to mock-transfected cells. Mean values±s.d. of two independent experiments each consisting of triplicate values are shown.

Bcl-xAK triggers apoptosis in melanoma cells and reveals additive proapoptotic effects with chemotherapeutics

Doxycycline induction of Bcl-xAK in SKM13-Bcl-xAK resulted after 48 h in significantly increased apoptotic rates as compared to untreated control cells (mean factor 1.7±0.1; P<0.001; Figure 3a). In contrast, cytotoxicity levels determined in parallel by release of LDH were unaffected after 48 h (data not shown).

DNA fragmentation rates in the Bcl-xAK clone were, however, weaker than in an isogenic cell clone stably transfected with Bcl-xS cDNA (SKM13-Bcl-xS), which had been described previously (Hossini et al., 2003a) (Figure 3a).

Typical morphological hallmarks of apoptosis (chromatin condensation and nuclear fragmentation) were also observed in SKM13-Bcl-xAK cells after doxycycline induction (Figure 3b). High magnification clearly identified apoptotic nuclei (Figure 3c). Quantification revealed 48 h after Bcl-xAK induction 14±3% apoptotic cells as compared to only 4±2% found in noninduced cells (Figure 3d). Again as compared with Bcl-xS apoptosis induced in SKM13-Bcl-xS cells was stronger: 40±7% cells with condensed chromatin or fragmented nuclei (Hossini et al., 2003a).

Labeling of free 3′-OH DNA ends by TUNEL technique also proved the proapoptotic effect of Bcl-xAK. After Bcl-xAK induction, 5±0.5% of cells were TUNEL-positive as compared to only 1.5±0.9% in noninduced cultures (Figure 3e and f). And finally, FACS analysis (Nicoletti assay) showed a similar result: 5% hypodiploid cells after Bcl-xAK induction as compared to 1.8% in noninduced cells (data not shown). Thus, Bcl-xAK clearly induced apoptotic not necrotic cell death in melanoma cells.

The proapoptotic effects of chemotherapeutics (etoposide, pamidronate, doxorubucine) and of the agonistic CD95 antibody CH-11 were not reduced, thus ruling out an antiapoptotic activity of Bcl-xAK, but were even enhanced upon doxycycline-mediated induction of Bcl-xAK. On the other hand, doxycycline was without effect on sensitivity of mock-transfected isogenic cells (Figure 4).

Figure 4

Enhancement of chemotherapeutic effects by Bcl-xAK. Induction of apoptosis as determined by DNA fragmentation in SKM13-Bcl-xAK. Cells were pretreated for 24 h with doxycyline (Dox) for induction of Bcl-xAK (+) or were not induced (−). After preincubation, cells were in addition treated with etoposide (VP-16), doxorubucine (Doxo) or pamidronate (Pam) for additional 24 h or were treated with an agonistic CD95 antibody (CH-11) for 16 h, whereas other cells did not receive any extra treatment (no). Relative apoptotic rates are given as compared to untreated cells, which were set to 100%. Means and standard deviations of triple values of a representative experiment are shown. The experiment was repeated three times giving comparable results.

Mitochondrial and cytosolic localization but no caspase-3 activation by Bcl-xAK

To understand the possible impact of the domain structure on subcellular distribution of Bcl-x proteins, localization of Bcl-xAK was investigated in mitochondrial and cytosolic fractions after doxycycline induction. For reasons of comparison, an isogenic cell line stably transfected with Bcl-xS (SKM13-Bcl-xS) was examined in parallel. For demonstrating purity of cytosolic fractions and especially to exclude any mitochondrial contamination, voltage-dependent anion channel protein (VDAC-1) was shown to coincide exclusively with the mitochondrial fractions. Bcl-xAK was found both in the mitochondrial and in the cytosolic cell fractions, in clear contrast to Bcl-xS, which was exclusively detected in the mitochondrial/membranous fraction (Figure 5a).

Figure 5

Localization of Bcl-xAK in mitochondria and in the cytosol – no activation of caspase-3. (a) Western blot analysis after separation of mitochondrial and cytosolic cell fractions is shown for Tet-On cells stably transfected with Bcl-xAK (SKM13-Bcl-xAK) and with Bcl-xS (SKM13-Bcl-xS), respectively. Cells were treated for 48 h with doxycyline for induction (+) or were left without (−). Equal protein amounts of each fraction were separated by SDS–PAGE, and consistent blotting was confirmed by ponceau staining. Purity of the cytosolic fractions was monitored by the absence of immunostaining for voltage-dependent anion channel protein (VDAC-1). (b) Proteins were extracted from SKM13-Tet-On melanoma cells stably transfected with pTRE-Bcl-xAK (SKM13-Bcl-xAK) and with pTRE-CD95L (SKM13-CD95L), respectively. Bcl-xAK and CD95L expression was induced by doxycycline (+Dox) for 12, 24 and 48 h as indicated below. Equal amounts of proteins (40 μg/lane) were loaded and equal transfer of proteins was confirmed by ponceau red staining. Cleavage products described as catalytically active forms of caspase-3 (p17, p20) were increased in SKM13-CD95L cells after doxycycline induction but not in SKM13-Bcl-xAK. As loading control, β-actin Western blots are shown.

To further understand the possible mechanism of Bcl-xAK-induced apoptosis in melanoma cells, engagement of the main effector caspase-3 was investigated. The SKM13-Bcl-xAK clone as well an isogenic SK-Mel-13 cell clone stably transfected with CD95L (SKM13-CD95L) used as a positive control were triggered with doxycycline for 12, 24, and 48 h before protein extraction. Whereas Western blot analysis revealed for SKM13-CD95L cells some increasement in processing of caspase-3 to its p17 and p20 cleavage products after doxycycline induction, no caspase-3 processing was seen for SKM13-Bcl-xAK cells (Figure 5b).


Pro- and antiapoptotic proteins of the Bcl-2 family are critical regulators of apoptosis exerting their activities at membranes of mitochondria, endoplasmatic reticulum and nuclear envelope (Zamzami et al., 1998; Daniel et al., 2003). In addition to their high number, complexity is further increased by several alternative splice products as described for bax (Bargou et al., 1995; Jin et al., 2001), bid (Renshaw et al., 2004), bcl-G (Guo et al., 2001) and bcl-x genes (Boise et al., 1993). Whereas the long splice variant of the bcl-x gene, Bcl-xL, exerts well-described antiapoptotic functions (Ogata and Takahashi, 2003), expression of the short splice form, Bcl-xS, was shown to induce apoptosis in neuroblastoma, Kaposi's sarcoma, breast cancer, prostate cancer, colon carcinoma and in melanoma cells (Clarke et al., 1995; Lindenboim et al., 2000; Mercatante et al., 2001; Hossini et al., 2003a). An additional bcl-x splice variant containing all four BH domains, Bcl-xβ, has been identified and reported as proapoptotic in rat promyeloid cells (Shiraiwa et al., 1996) but exerted antiapoptotic activities in mouse neuronal cells (Ban et al., 1998). Its physiological role in human cells remains to be proven. Bcl-xAK, the novel splice variant reported here is clearly distinguished from these previously described variants.

Antiapoptotic Bcl-2 proteins display sequence conservation of all four conserved Bcl-2 homology domains (BH1–BH4), thereby indicating the functional significance of these domains for inhibition of apoptosis. The proapoptotic Bcl-2 proteins subdivide in multidomain proteins, such as Bax, Bak and Bok, and BH3-only proteins. However, all so far described proapoptotic Bcl-2 proteins contain BH3 indicating a functional significance of this domain for apoptosis induction (Petros et al., 2004). The precise mechanism, however, how the different proapoptotic Bcl-2 proteins promote cell death is still under discussion and seems to be highly varying for the individual proteins. According to a present model, multidomain proteins such as Bax and Bak, which are monomeric and cytosolic in their inactive form, undergo upon induction conformational changes and translocate to mitochondria, where they oligomerize and form pores that allow release of proapoptotic mitochondrial factors (Eskes et al., 2000; Gillissen et al., 2003).

BH3-only proteins such as Bid and Bik/Nbk appear to function as sensors and triggers for induction of apoptosis. They are activated at the transcriptional or at the post-translational level, translocate to the mitochondrial membrane and bind to antiapoptotic Bcl-2 proteins thereby blocking their activity (Gross et al., 1999). Others may also directly influence the activity of multidomain proapoptotic proteins (Eskes et al., 2000; Gillissen et al., 2003).

Bcl-xAK is characterized by a new composition of BH2, BH4 and transmembrane domain. In a most recent report, even another splice product of bcl-x containing BH4 and BH2 (Bcl-xES) has been reported as antiapoptotic when overexpressed in human B lymphoma cells (Schmitt et al., 2004). Also this protein is clearly distinguished from Bcl-xAK as it lacks 41 AA of the BH4–BH3 interdomain region present in Bcl-xAK, and it includes a stretch of 20 amino acids in front of the BH2 domain missing in Bcl-xAK (Figure 6).

Figure 6

Variant inclusion of alpha-helices in bcl-x splice variants. An amino-acid sequence alignment of described splice products of bcl-x is shown. Positions of the 7 described α-helices (Hlx) are indicated above, and positions of the BH domains are indicated below the sequences. Missing amino acids in splice products are indicated by hyphens.

As shown previously for Bcl-xL, both BH1 and BH2 seem to be required for antiapoptotic function and heterodimerization with Bax (Yin et al., 1994). X-ray analysis further revealed that BH1, BH2 and BH3 can form a hydrophobic pocket into which the BH3 domains of proapoptotic proteins may bind (Muchmore et al., 1996). A mechanism of this kind, however, has to be ruled out for Bcl-xAK as well as for Bcl-xES. Three-dimensional structure analysis of Bcl-xL had revealed two central hydrophobic α-helices (helices 5 and 6) suggested as to be directly involved in ion channel formation, which are surrounded by five additional, amphipathic helices (Muchmore et al., 1996; Schendel et al., 1997; Losonczi et al., 2000). Significant roles of helices 5 and 6 were shown for the function of antiapoptotic Bcl-2 proteins (Matsuyama et al., 1998), whereas, in proapoptotic proteins, this pore structure has been hypothesized as important for enabling the release of cytochrome c (Vander Heiden and Thompson, 1999). The new splice variant Bcl-xAK includes only helices 1 and 7 but lacks helices 2–6 (Figure 6) in full agreement with the lack of antiapoptotic function as demonstrated for chemotherapeutics and agonistic CD95 stimulation. Furthermore, the lack of any antiapoptotic activity of Bcl-xAK in contrast to Bcl-xES (Schmitt et al., 2004) may be related to the fact that helix 6 is contained in Bcl-xES but not in Bcl-xAK.

Analysis of DNA fragmentation, DNA strand breaks (TUNEL), cell cycle analysis (Nicoletti) as well as of nuclear condensation and fragmentation (Hoechst) applied here are all generally accepted techniques for identification of apoptotic cells (Nicoletti et al., 1991; Hipp and Bauer, 1997; Vander Heiden and Thompson, 1999; Leech et al., 2000; Kim et al., 2001). Cytotoxicity, on the other hand, was not measured within the time of observation (48 h) clearly indicating that induction of apoptosis was a primary effect. Thus, Bcl-xAK represents the first proapoptotic Bcl-2-related protein that lacks a functional BH3 domain.

The proapoptotic potential of Bcl-xAK was weaker than that of Bcl-xS in the two stably transfected cell clones analysed. This may be seen in relation to the weaker overexpression of Bcl-xAK after doxycycline induction in the isolated cell clone, or may also reflect the relatively higher basal expression level of Bcl-xAK in melanoma cells as compared to Bcl-xS. Furthermore, enhanced apoptosis was seen after combination of Bcl-xAK overexpression with anticancer drugs or death receptor ligation, possibly indicating involvement in controlling DNA damage-induced cell death. Comparable additive effects with chemotherapeutics had previously been shown also for Bcl-xS expression (Hossini et al., 2003a).

Mitochondria are of critical importance in melanoma cells for induction of apoptosis by death ligands or chemotherapeutics (Raisova et al., 2001; Eberle et al., 2003; Hussein et al., 2003). Bcl-xAK was found both in the mitochondrial as well as in the cytosolic fraction, in clear contrast to Bcl-xS that was localized exclusively in the mitochondrial fraction. Several proapoptotic proteins of the Bcl-2 family, such as Bax, Bad and Bid are also found in the cytosol and translocate after apoptotic stimulation (Reed, 1998; Li and Yuan, 1999; Hussein et al., 2003).

The lack of a BH3 domain and partial cytosolic localization strongly suggest that Bcl-xAK utilizes other mechanisms for induction of apoptosis than Bcl-xS. Several connections between Bcl-2-related proteins and signaling cascades have been identified: for example, the proapoptotic Bcl-2 protein Bad is phosphorylated and inactivated by protein kinase B as well as by MEK (Cardone et al., 1997; del Peso et al., 1997). Also, the proapoptotic BH3-only protein Bim was shown to be phosphorylated by MAP kinases resulting in its degradation in epithelial cells (Marani et al., 2004). On the other hand, Bcl-2 can be activated by the Ras-Raf-initiated pathways in neuroblastoma cells (Schwarz et al., 2002). Activated Ras, integrin, vitronectin, HGF, NF-κB and MITF signaling cascades are further linked to changes in the expression of Bcl-2-related proteins (Bours et al., 2000; Grad et al., 2000; McGill et al., 2002). It is therefore conceivable that the new Bcl-xAK interferes with these signaling cascades thus promoting apoptosis.

In melanoma cells, there are strong indications for caspase-independent pathways triggered by proapoptotic Bcl-2 proteins, as we have recently shown for Nbk/Bik (Oppermann et al., 2005) as well as for Bcl-xS (Hossini et al., 2003b). In agreement, we saw here no activation of the main effector caspase-3 after Bcl-xAK overexpression, whereas caspase-3 was activated in isogenic cells after CD95L/FasL expression. Alternative, caspase-independent but so far unspecified pathways seem to be activated in melanoma cells by proapoptotic Bcl-2 proteins, and their elucidation will draw a new picture of apoptosis regulation in melanoma.

Overall, Bcl-xAK introduces a completely new group of Bcl-2-related proteins; it neither shares typical characteristics of antiapoptotic nor such of typical proapoptotic molecules. Nevertheless, it evidently triggered apoptosis in human melanoma cells and even enhanced apoptosis induction by anticancer drugs. A new, BH3-independent mechanism for apoptosis control has to be assumed, which might yield interesting insights into intracellular signaling cascades. With respect to therapeutic approaches, targeting these new pathways might provide additional strategies for overcoming apoptosis resistance of human melanoma cells.

Materials and methods

Cell culture

NHM were isolated from human foreskins after trypsin digestion, and were cultured under serum-free conditions as previously described (Eberle et al., 1999). Eight human melanoma cell lines, A-375 (Giard et al., 1973); Mel-2a (Bruggen et al., 1981); M-5 (Liao et al., 1975); Mel-HO (Holzmann et al., 1988); MeWo (Bean et al., 1975); SK-Mel-13; SK-Mel-23; SK-Mel-28 (Carey et al., 1976) and two cell populations established from histologically confirmed metastatic melanomas of two different patients (M-186, M-221) were investigated. In addition, the neuroectodermal tumor cell line PFSK-1 (Fults et al., 1992) and the cervical carcinoma cell line HeLa (Scherer et al., 1953) were investigated. For control experiments, SKM13-Tet-On melanoma cell clones stably transfected with a tetracycline-inducible Bcl-xS construct (SKM13-Bcl-xS) and with tetracycline-inducible CD95L/FasL (SKM13-CD95L), respectively, were applied, which we had established and described previously (Eberle et al., 2003; Hossini et al., 2003a). Melanoma cells were maintained in DMEM (4.5 g/l glucose; Gibco, Invitrogen, Karlsruhe, Germany) supplemented with 10% fetal calf serum and antibiotics (Biochrom, Berlin, Germany).

For induction of apoptosis, cells were treated for 10 h with 0.5 μg/ml of the agonistic anti-CD95 antibody CH-11 (Immunotech, Marseille, France) or were treated for 24 h with 100 μg/ml pamidronate (3-amino-1-hydroxy-propylidene-1,1-bisphosphonate; Novartis Pharmaceuticals, Basel, Switzerland), 200 nM doxorubucine (Alexis, Grünberg, Germany) or 2 μ M etoposide (VP-16; Sigma, Taufkirchen, Germany).

RT–PCR cloning and subcloning

Along with clones for Bcl-xS and Bcl-xL, the full-length cDNA for Bcl-xAK was obtained by RT–PCR performed from total RNA of the melanoma cell line M-5 using primers corresponding to the untranslated regions at the 5′ end (IndexTermTTGGACAATGGACTGGTTGA) and the 3′ end (IndexTermGTAGAGTGGATGGTCAGTG) of bcl-x mRNA. The resulting cDNA fragment of bcl-xAK (491 bp) was subcloned into plasmid TOPO-TA (Invitrogen, Groningen, Netherlands), and identity was confirmed by restriction and sequence analysis. The cDNA fragment was further subcloned into the EcoRI site of plasmid pTRE-1 (Clontech, Heidelberg, Germany) downstream of a tetracycline-regulatable promotor resulting in pTRE-Bcl-xAK.

Cellular transfection

For inducible gene expression, the tetracycline-regulatable gene expression system (Tet-On; Clontech) (Gossen and Bujard, 1992) was applied in melanoma cells. The tetracycline-regulatable cell line SKM13-Tet-On established by stable transfection of plasmid pTet-On (Clontech) into the melanoma cell line SK-Mel-13 has been described previously (Eberle et al., 2002). Stable transfections of melanoma cells were performed at a cellular confluence of 50% using 0.5% Pfx-2 (Invitrogen, Groningen, Netherlands) and 1.5 μg/ml plasmid DNA as described previously (Fecker et al., 2002). After transfection of SKM13-Tet-On with pTRE-Bcl-xAK, stable cell clones were selected with hygromycin (100 μg/ml; Life Technologies) and geneticin (400 μg/ml; Boehringer, Mannheim, Germany). Individual cell clones were obtained by limited dilution in microtiter plates and were continuously cultured under antibiotic pressure. Promoter induction was achieved by addition of doxycycline (2 μg/ml; ICN, Aurora, OH, USA) to the growth medium for 48 h.

Quantification of apoptosis and cytotoxicity

Stably transfected cell clones were seeded in six-well plates and were induced after 24 h at a confluence of 50% with doxycycline for another 48 h. Apoptosis was subsequently quantified by using a cell death detection ELISA (Roche Diagnostics, Mannheim, Germany), which detects mono- and oligonucleosomes formed in apoptotic cells. Relative apoptotic rates were calculated as the ratio of the OD determined from doxycycline-induced and noninduced cultures. Each assay consisted of triple values, and at least three independent experiments were performed. Cytotoxicity was determined in parallel by a LDH-Assay (Lactate Dehydrogenase; Roche Diagnostics) as described previously (Fecker et al., 2002). Statistical significance was determined by unpaired Student's t-test.

For determination of chromatin condensation, cells were harvested by trypsinization, fixed in 4% formaldehyde (methanol-free) for 30 min at 4°C and washed once with phosphate-buffered saline (PBS). Cytospins were incubated with Hoechst-33258 dye (1 μg/ml; Sigma) for 20 min at room temperature after which cells were washed again with PBS. Finally, cells were mounted (MoBiTec, Göttingen, Germany) and examined by fluorescence microscopy. Apoptotic cells were identified by condensed and fragmented nuclei. For quantification, a minimum of 450 cells were counted, and the ratio of apoptotic cells was calculated as percent of total cells counted.

For analysing DNA strand breaks in individual cells, free 3′-OH termini were labeled by terminal deoxynucleotidyl transferase (TUNEL technique) using a commercially available kit (Roche Diagnostics). After induction with doxycycline for 2 days, adherent cells were fixed in 4% paraformaldehyde (1 h) and were permeabilized in a buffer containing 0.1% Triton-X-100 and 0.1% sodium citrate (2 min, 4°C). After washing two times with PBS, the TUNEL assay reagents (containing terminal deoxynucleotidyl transferase and fluorescein-labeled dUTP) were added. Incubation at 37°C was for 60 min, followed by three times washing with PBS. Cells were counterstained with 1 μg/ml bisbenzimide in PBS for 30 min at room temperature, followed by two times washing steps with PBS. Apoptotic cells were identified as bright green-stained nuclei. For quantification, a minimum of 500 cells were counted in each analysis, and the ratio of apoptotic cells was calculated as percent of total cell counts.

Cell cycle analysis was performed by detection of hypodiploid nuclei in flow cytometric analyses (Nicoletti assay). Melanoma cells were induced with doxycycline for 96 h, harvested by trypsinization, stained with propidium iodide (200 μg/ml) and analysed by FACS (Nicoletti et al., 1991).

Expression analyses and subcellular fractionation

Total RNA was extracted by the RNeasy Mini Kit (Qiagen, Hilden, Germany). Expression of mRNA was analysed by Northern blotting and by nested RT–PCR. Primers used for RT–PCR enabled simultaneous amplification and distinction of the three reading frames for Bcl-xL, Bcl-xS and Bcl-xAK. Primer sequences: 5′ end, outer: IndexTermTTGGACAATGGACTGGTTGA; 3′ end, outer: IndexTermGTAGAGTGGATGGTCAGTG; 5′ end, inner: IndexTermATCCTGGCACCTGGCAGACA; 3′ end, inner: IndexTermGCTCTCGGCTGCTGCATTGT.

Protein extracts were generated in lysis buffer containing 50 mM Tris-HCl, pH 7.6, 150 mM NaCl, 1% Nonidet P-40, 10% glycerin, 0.2% SDS, 10 μg/ml aprotinin (Bayer, Leverkusen, Germany), 1 mM PMSF, 10 μg/ml leupeptin and 5 μg/ml pepstatin. Extracts were homogenized and cleared by centrifugation at 10 000 g for 10 min. For separation of cytosolic and mitochondrial cellular fractions, a mitochondria/cytosol fractionation kit (Alexis, Grünberg, Komma; Germany) was applied.

For immunodetection in Western blot analysis, an anti-Bcl-x polyclonal antibody (M-125, sc-1690; Santa Cruz, Heidelberg, Germany; 1:200) was used, which detected Bcl-xL, Bcl-xS and Bcl-xAK. For detection of the cleavage products of caspase-3, an anti-cleaved-caspase-3 polyclonal antibody was used (#9661, Cell Signaling, Frankfurt, Germany; 1:1000).

Cytosolic and mitochondrial fractions were controlled by immunodetection of voltage-dependent anion channel protein (VDAC-1, 31HL, mouse, Ab-3; Calbiochem, Schwalbach, Germany; 1:1000). Following incubation with a peroxidase-labeled anti-rabbit secondary antibody (Dako, Hamburg, Germany; 1:5000), antigen–antibody complexes were visualized by ECL Western blotting reagents on Hyperfilm ECL (Amersham Pharmacia Biotech, Freiburg, Germany).



normal human melanocytes


reverse transcriptase–polymerase chain reaction


lactate dehydrogenase

BH domain:

Bcl-2 homology domain


  1. Ban J, Eckhart L, Weninger W, Mildner M, Tschachler E . (1998). Biochem Biophys Res Commun 248: 147–152.

  2. Bargou RC, Daniel PT, Mapara MY, Bommert K, Wagener C, Kallinich B et al. (1995). Int J Cancer 60: 854–859.

  3. Bean MA, Bloom BR, Herberman RB, Old LJ, Oettgen HF, Klein G et al. (1975). Cancer Res 35: 2902–2913.

  4. Boise LH, Gonzalez-Garcia M, Postema CE, Ding L, Lindsten T, Turka LA et al. (1993). Cell 74: 597–608.

  5. Borner C . (2003). Mol Immunol 39: 615–647.

  6. Bours V, Bentires-Alj M, Hellin AC, Viatour P, Robe P, Delhalle S et al. (2000). Biochem Pharmacol 60: 1085–1089.

  7. Bruggen J, Fogh J, Sorg C . (1981). J Cancer Res Clin Oncol 102: 141–152.

  8. Cardone MH, Salvesen GS, Widmann C, Johnson G, Frisch SM . (1997). Cell 90: 315–323.

  9. Carey TE, Takahashi T, Resnick LA, Oettgen HF, Old LJ . (1976). Proc Natl Acad Sci USA 73: 3278–3282.

  10. Clarke MF, Apel IJ, Benedict MA, Eipers PG, Sumantran V, Gonzalez-Garcia M et al. (1995). Proc Natl Acad Sci USA 92: 11024–11028.

  11. Daniel PT, Schulze-Osthoff K, Belka C, Guner D . (2003). Essays Biochem 39: 73–88.

  12. del Peso L, Gonzalez-Garcia M, Page C, Herrera R, Nunez G . (1997). Science 278: 687–689.

  13. Diaz JL, Oltersdorf T, Horne W, McConnell M, Wilson G, Weeks S et al. (1997). J Biol Chem 272: 11350–11355.

  14. Eberle J, Fecker LF, Bittner JU, Orfanos CE, Geilen CC . (2002). Br J Cancer 86: 1957–1962.

  15. Eberle J, Fecker LF, Hossini AM, Wieder T, Daniel PT, Orfanos CE et al. (2003). Oncogene 22: 9131–9141.

  16. Eberle J, Weitmann S, Thieck O, Pech H, Paul M, Orfanos CE . (1999). J Invest Dermatol 112: 925–932.

  17. Eskes R, Desagher S, Antonsson B, Martinou JC . (2000). Mol Cell Biol 20: 929–935.

  18. Fecker LF, Eberle J, Orfanos CE, Geilen CC . (2002). J Invest Dermatol 118: 1019–1025.

  19. Fults D, Pedone CA, Morse HG, Rose JW, McKay RD . (1992). J Neuropathol Exp Neurol 51: 272–280.

  20. Giard DJ, Aaronson SA, Todaro GJ, Arnstein P, Kersey JH, Dosik H et al. (1973). J Natl Cancer Inst 51: 1417–1423.

  21. Gillissen B, Essmann F, Graupner V, Starck L, Radetzki S, Dorken B et al. (2003). EMBO J 22: 3580–3590.

  22. Gossen M, Bujard H . (1992). Proc Natl Acad Sci USA 89: 5547–5551.

  23. Grad JM, Zeng XR, Boise LH . (2000). Curr Opin Oncol 12: 543–549.

  24. Grillot DA, Gonzalez-Garcia M, Ekhterae D, Duan L, Inohara N, Ohta S et al. (1997). J Immunol 158: 4750–4757.

  25. Gross A, McDonnell JM, Korsmeyer SJ . (1999). Genes Dev 13: 1899–1911.

  26. Guo B, Godzik A, Reed JC . (2001). J Biol Chem 276: 2780–2785.

  27. Hipp ML, Bauer G . (1997). Oncogene 15: 791–797.

  28. Holzmann B, Lehmann JM, Ziegler-Heitbrock HW, Funke I, Riethmuller G, Johnson JP . (1988). Int J Cancer 41: 542–547.

  29. Hossini AM, Eberle J, Fecker LF, Orfanos CE, Geilen CC . (2003a). FEBS Lett 553: 250–256.

  30. Hossini AM, Eberle J, Fecker LF, Orfanos CE, Geilen CC . (2003b). Arch Dermatol Res 294: 499, abstract 153.

  31. Hussein MR, Haemel AK, Wood GS . (2003). J Pathol 199: 275–288.

  32. Jin KL, Graham SH, Mao XO, He X, Nagayama T, Simon RP et al. (2001). J Neurochem 77: 1508–1519.

  33. Johnstone RW, Ruefli AA, Lowe SW . (2002). Cell 108: 153–164.

  34. Kelekar A, Thompson CB . (1998). Trends Cell Biol 8: 324–330.

  35. Kim MR, Lee JY, Park MT, Chun YJ, Jang YJ, Kang CM et al. (2001). FEBS Lett 505: 179–184.

  36. Leech SH, Olie RA, Gautschi O, Simoes-Wust AP, Tschopp S, Haner R et al. (2000). Int J Cancer 86: 570–576.

  37. Li H, Yuan J . (1999). Curr Opin Cell Biol 11: 261–266.

  38. Liao SK, Dent PB, McCulloch PB . (1975). J Natl Cancer Inst 54: 1037–1044.

  39. Lindenboim L, Yuan J, Stein R . (2000). Oncogene 19: 1783–1793.

  40. Losonczi JA, Olejniczak ET, Betz SF, Harlan JE, Mack J, Fesik SW . (2000). Biochemistry 39: 11024–11033.

  41. Marani M, Hancock D, Lopes R, Tenev T, Downward J, Lemoine NR . (2004). Oncogene 23: 2431–2441.

  42. Matsuyama S, Schendel SL, Xie Z, Reed JC . (1998). J Biol Chem 273: 30995–31001.

  43. McGill GG, Horstmann M, Widlund HR, Du J, Motyckova G, Nishimura EK et al. (2002). Cell 109: 707–718.

  44. Mercatante DR, Bortner CD, Cidlowski JA, Kole R . (2001). J Biol Chem 276: 16411–16417.

  45. Muchmore SW, Sattler M, Liang H, Meadows RP, Harlan JE, Yoon HS et al. (1996). Nature 381: 335–341.

  46. Nicoletti I, Migliorati G, Pagliacci MC, Grignani F, Riccardi C . (1991). J Immunol Methods 139: 271–279.

  47. Ogata Y, Takahashi M . (2003). Drug News Perspect 16: 446–452.

  48. Oppermann M, Geilen CC, Fecker LF, Gillissen B, Daniel PT, Eberle J . (2005). Oncogene, Epub ahead of print, PMID. 16007125.

  49. Osford SM, Dallman CL, Johnson PW, Ganesan A, Packham G . (2004). Curr Med Chem 11: 1031–1039.

  50. Petros AM, Olejniczak ET, Fesik SW . (2004). Biochim Biophys Acta 1644: 83–94.

  51. Raisova M, Hossini AM, Eberle J, Riebeling C, Wieder T, Sturm I et al. (2001). J Invest Dermatol 117: 333–340.

  52. Reed JC . (1998). Oncogene 17: 3225–3236.

  53. Renshaw SA, Dempsey CE, Barnes FA, Bagstaff SM, Dower SK, Bingle CD et al. (2004). J Biol Chem 279: 2846–2855.

  54. Schendel SL, Xie Z, Montal MO, Matsuyama S, Montal M, Reed JC . (1997). Proc Natl Acad Sci USA 94: 5113–5118.

  55. Scherer WF, Syverton JT, Gey GO . (1953). J Exp Med 97: 695–710.

  56. Schmitt E, Paquet C, Beauchemin M, Bertrand R . (2004). Oncogene 23: 3915–3931.

  57. Schwarz CS, Seyfried J, Evert BO, Klockgether T, Wullner U . (2002). Neuroreport 13: 2439–2442.

  58. Shiraiwa N, Inohara N, Okada S, Yuzaki M, Shoji S, Ohta S . (1996). J Biol Chem 271: 13258–13265.

  59. Vander Heiden MG, Thompson CB . (1999). Nat Cell Biol 1: E209–E216.

  60. Yang E, Korsmeyer SJ . (1996). Blood 88: 386–401.

  61. Yin XM, Oltvai ZN, Korsmeyer SJ . (1994). Nature 369: 321–323.

  62. Zamzami N, Brenner C, Marzo I, Susin SA, Kroemer G . (1998). Oncogene 16: 2265–2282.

Download references


The study was supported by grants from the Deutsche Krebshilfe/Mildred-Scheel-Stiftung (10-1434-Eb2), the Deutsche Forschungsgemeinschaft (SFB 366, TP B8) and the Sonnenfeld Stiftung, Berlin.

Author information



Corresponding author

Correspondence to J Eberle.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Hossini, A., Geilen, C., Fecker, L. et al. A novel Bcl-x splice product, Bcl-xAK, triggers apoptosis in human melanoma cells without BH3 domain. Oncogene 25, 2160–2169 (2006).

Download citation


  • Bcl-x
  • apoptosis
  • alternative splicing
  • melanoma
  • chemotherapeutics

Further reading


Quick links