Research Paper 913

Regulation of apoptosis by BH3 domains in a cell-free system

Sabina C. Cosulich*, Vivienne Worrall†, Philip J. Hedge† , Stephen Green†
and Paul R. Clarke*

Background: The Bcl-2 family of proteins plays a key role in the regulation of Addresses: *Zeneca Laboratory of Molecular and
apoptosis. Some family members prevent apoptosis induced by a variety of Cellular Biology, School of Biological Sciences,
G38 Stopford Building, University of Manchester,
stimuli, whereas others promote apoptosis. Competitive dimerisation between Oxford Road, Manchester M13 9PT, UK. †Cancer,
family members is thought to regulate their function. Homologous domains Metabolism and Endocrine Research Department,
within individual proteins are necessary for interactions with other family Zeneca Pharmaceuticals, Alderley Park,
members and for activity, although the specific mechanisms might differ Macclesfield, Cheshire SK10 4TG, UK.
between the pro-apoptotic and anti-apoptotic proteins.
Correspondence: Paul R. Clarke
E-mail: paul.clarke@man.ac.uk
Results: Using a cell-free system based on extracts of Xenopus eggs, we have
investigated the role of the Bcl-2 homology domain 3 (BH3) from different Received: 13 August 1997
Revised: 17 September 1997
members of the Bcl-2 family. BH3 domains from the pro-apoptotic proteins Bax
Accepted: 2 October 1997
and Bak, but not the BH3 domain of the anti-apoptotic protein Bcl-2, induced
apoptosis in this system, as determined by the rapid activation of specific Published: 31 October 1997
apoptotic proteases (caspases) and by DNA fragmentation. The apoptosis-
Current Biology 1997, 7:913–920
inducing activity of the BH3 domains requires both membrane and cytosolic http://biomednet.com/elecref/0960982200700913
fractions of cytoplasm, involves the release of cytochrome c from mitochondria
and is antagonistic to Bcl-2 function. Short peptides, corresponding to the © Current Biology Ltd ISSN 0960-9822
minimal sequence of BH3 domains required to bind anti-apoptotic Bcl-2 family
proteins, also trigger apoptosis in this system.

Conclusions: The BH3 domains of pro-apoptotic proteins are sufficient to

trigger cytochrome c release, caspase activation and apoptosis. These results
support a model in which pro-apoptotic proteins, such as Bax and Bak, bind to
Bcl-2 via their BH3 domains, inactivating the normal ability of Bcl-2 to suppress
apoptosis. The ability of synthetic peptides to reproduce the effect of
pro-apoptotic BH3 domains suggests that such peptides may provide the basis
for engineering reagents to control the initiation of apoptosis.

Background whereas others promote it [4]. Bcl-2 acts in a biochemical

Programmed cell death (PCD) plays an essential role in
the development and tissue homeostasis of metazoan
organisms [1]. Inappropriate PCD may disrupt these
processes, leading to disease, whereas resistance to the
induction of PCD contibutes to cancer [2]. Most cells
undergoing PCD exhibit similar and dramatic morphologi-
cal changes, including plasma membrane blebbing, cyto-
plasmic reorganisation, chromatin condensation and DNA
fragmentation. These morphological changes and their
underlying molecular mechanisms constitute the process
known as apoptosis [3].
The Bcl-2 oncoprotein delays or prevents apoptosis trig-
gered by a variety of stimuli and permits cell survival [4],
indicating that it acts at a common control point in the cell
death/survival pathway. There is remarkable conservation
of Bcl-2 function across species [5,6], suggesting that Bcl-2
controls an evolutionarily conserved mechanism. Bcl-2 is
one of a family of related proteins found in mammalian
cells, some of which function as suppressors of apoptosis
pathway upstream of members of the caspase family of
cysteine proteases, regulating their activation [7–9]. Cas-
pases cleave substrates thought to mediate the dramatic
restructuring of the cell during apoptosis [10]. Caspases
are themselves activated by proteolytic cleavage of their
inactive precursors; one mechanism by which caspases are
activated involves the release of cytochrome c from
mitochondria in response to death-inducing signals. Bcl-2
and Bcl-xL block cytochrome c release [11–14]. A homo-
logue of the Caenorhabditis elegans CED-4 protein interacts
with Bcl-2 family proteins, caspase precursors, cytochrome
c and additional factors which participate in caspase
activation [15,16].
Although proteins of the Bcl-2 family can have opposing
biological effects, they share regions of sequence similar-
ity known as Bcl-2 homology (BH) domains [9,17,18]. It
has been proposed that these proteins are regulated
through their ability to interact and form homodimers or
heterodimers [19]. It has been unclear, however, whether
914 Current Biology, Vol 7 No 12

the anti-apoptotic or pro-apoptotic proteins, or both, are
the effector molecules that would be inhibited by associa-
tion with an opposing partner. Deletion analysis has
shown that the BH1, BH2 and BH3 domains are all
required for anti-apoptotic activity of death-inhibiting
members of the family when they are transfected into
cells and for heterodimerisation with death-promoting
members of the family [18,20]. In contrast, the BH3
domain of Bax and of Bak is is the only region which,
when deleted, abolishes both heterodimerisation with
Bcl-2 and Bcl-xL and the ability to promote cell death
[17,21]. Moreover, the BH3 domain of Bak is sufficient to
cause cell death when transfected into cells [17]. Other
pro-apoptotic proteins have been described, such as Hrk,
Bid and Nbk/Bik, which share homology with other Bcl-2
family members only in the BH3 domain [22–25]. Togeth-
er, these results suggest that BH3 domains are critical for
the induction of apoptosis by pro-apoptotic Bcl-2 family
members, although the mechanism is still uncertain.
Recently, we and others have characterised a cell-free
system based on extracts of Xenopus eggs that reproduce
morphological and biochemical events characteristic of
apoptosis [8,26–29]. In this system, spontaneous activation
of caspases occurs after a lag period of several hours [8].
The mechanism of caspase activation is unclear, but it
involves the release of cytochrome c from mitochondria
that are present in the extracts [13]. Cytochrome c release,
caspase activation and apoptosis are all inhibited by
addition of Bcl-2 [8,13,26] or Bcl-xL [27]. Xenopus egg
extracts, therefore, provide a useful biochemical system
with which to study the regulation of apoptosis by the
Bcl-2 family of proteins.
In this study, we have used Xenopus egg extracts to study
the function of BH3 domains from Bcl-2 family members.
BH3 domains from Bax and Bak were found to induce the
release of cytochrome c from mitochondria, triggering
caspase activation and apoptosis in this system, whereas
the BH3 domain from Bcl-2 was found to be ineffectual.
Pro-apoptotic BH3 domains bind to Bcl-2 and antagonise
its function, possibly by directly neutralising its anti-apop-
totic activity. In addition, peptides derived from pro-apop-
totic BH3 domains were found to be sufficient for caspase
activation; this observation may provide the basis for
engineering reagents to control the initiation of apoptosis.
Results
In order to elucidate further the molecular basis of the
activity of Bcl-2 family members, in particular, the pro-
apoptotic proteins, we have investigated the role of BH3
domains in the regulation of apoptosis. We produced a glu-
tathione-S-transferase (GST) fusion protein containing the
BH3 domain from Bak (amino acids 71–123) and tested its
ability to control apoptosis in Xenopus egg extracts. Human
HeLa cell nuclei were added to the extracts to provide
markers of nuclear events. In the absence of the GST–Bak
BH3 fusion protein, DNA cleavage into oligonucleosomal

Figure 1

(a) (b) Bak (c)

Bak BH3 +
Control Bak BH3 Bcl-2 BH3 Control DEVD Control Bak BH3
Incubation Incubation Incubation
time (h) 1 2 3 4 1 2 3 4 3 4 M time (h) 1 2 3 4 1 2 3 4 1 234 time (h) 1 2 3 4 1 2 3 4

127

73

2036
1636 (d) 20
2036
Fluorescence (arbitrary units)

1018
1636
15 Control
507 1018
GST–Bak BH3
396
10 GST–Bax BH3
507
GST–Bcl-2 BH3
5 GST alone

0
0 1 2 3 4 5
Incubation time (h)
Current Biology

BH3 domains trigger apoptosis in Xenopus egg extracts. Egg extracts
were incubated with (a–c) or without (d) the addition of HeLa nuclei
and the additions described. At the times shown, samples were
removed for analysis by agarose gel electrophoresis (a,b), SDS–PAGE
(8% acrylamide) and western blotting with anti-PARP antibodies (c) or
determination of caspase activity by assaying cleavage of the
fluorogenic substrate, Ac-DEVD–AMC (d). Additions were: (a) PBS
(control), GST–Bak BH3 or Bcl-2, (b) PBS (control), GST–Bak BH3
or GST–Bak BH3 plus 100 nM Ac-DEVD-CHO, (c) PBS (control) or
GST–Bak BH3, (d) GST–Bak BH3, GST–Bax BH3, GST–Bcl-2 BH3,
GST alone or PBS (control). All additions were made to a final
concentration of 0.1 mg/ml. The migration positions of DNA (a,b) and
protein (c) standards are indicated on the right with sizes in base pairs
and molecular mass (kDa), respectively.
 Research Paper Regulation of apoptosis by BH3 domains Cosulich et al.
fragments (Figure 1a) and chromatin condensation (data
not shown) typical of apoptosis occurred between 3 and
4 hours of incubation. Addition of GST–Bak BH3 to the
extracts to a final concentration of 0.1 mg/ml accelerated
greatly the timing and amount of DNA fragmentation
(Figure 1a), whereas Bcl-2 protein completely blocked
fragmentation, as has been shown previously [8,26].
Control incubations, GST alone and unrelated GST fusion
proteins had no effect, whereas GST–Bak BH3 was still
effective at triggering caspase activation at a final concen-
tration of 0.01 mg/ml (data not shown). A similar accelera-
tion of DNA fragmentation was obtained using GST
fusion proteins containing the BH3 domain of Bax (amino
acids 50–78 and 50–105; data not shown).
Previously, we have shown that DNA fragmentation
occurring spontaneously after prolonged incubation of
egg extracts is dependent on the activation of one or more
caspases of the class potently inhibited by the tetra-
peptide derivative Ac-DEVD-CHO [8]. Similarly, the
rapid induction of DNA fragmentation in response to
GST–Bak BH3 (Figure 1b) or GST–Bax BH3 (data not
shown) was completely blocked by Ac-DEVD-CHO. Fur-
thermore, addition of these BH3 domains to egg extracts
resulted in the rapid activation of caspases, as determined
by assaying cleavage of the substrate poly (ADP-ribose)
polymerase (PARP; Figure 1c) or by cleavage of a fluoro-
genic substrate, Ac-DEVD–AMC (Figure 1d). Upon
incubation at 23°C, control extracts exhibited sponta-
neous caspase activation after a 2–3 hour lag period (Fig-
ure 1c,d). However, in the presence of GST–Bak BH3 or
GST–Bax BH3 proteins, activation of caspases occurred
within 30 minutes of the start of the incubation
(Figure 1c,d). Rapid activation of caspases was also
observed when a 29 amino-acid peptide derived from the
Bax BH3 domain (amino acids 50–78) was added to the
egg extracts (data not shown). In contrast to BH3 domains
from Bax and Bak, addition of GST–Bcl-2 BH3 (amino
acids 86–111) had no effect, caspase activation occurring
with the same kinetics as in control extracts (Figure 1d).
Likewise, no effect of GST–Bcl-2 BH3 was observed on
the timing or extent of DNA fragmentation, even at high
Table 1
Peptides derived from BH3 domains.
Protein Peptide
Bak 72 GQVGRQLAIIGDDINR 87
Bak mutant 72 GQVGRQAAIIGDDINR 87
Bax 57 KKLSECLKRIGDELDS 72
Bid 84 RNIARHLAQVGDSMDR 99
Bcl-2 91 PVVHLALRQAGDDFSR 106
Sequences are shown using the single-letter code for amino acids.
Numbers refer to the position of the terminal amino acids of the
peptide in the parent protein. The altered residue in the Bak mutant
peptide (Leu78→Ala) is underlined.
915
concentrations (1 mg/ml; data not shown), although intact
Bcl-2 completely blocks caspase activation (see below and
[8]) and DNA fragmentation (Figure 1a).
Previous studies have shown that a 16 amino-acid region
of the Bak BH3 domain (amino acids 72–87) is sufficient
for binding to Bcl-xL [30]. We therefore investigated
whether this region is also sufficient for pro-apoptotic
activity. Peptides of 16 amino acids derived from the BH3
regions of Bak, Bax or Bid (Table 1) caused rapid activa-
tion of caspases in Xenopus egg extracts (Figure 2). To test
the possible relationship between binding to Bcl-xL —
and, by extension, probably to other anti-apoptotic Bcl-2
family proteins — and ability to trigger caspase activation,
we used a peptide derived from Bak BH3 with a single
amino-acid substitution (Leu 78→Ala; Table 1). This sub-
stitution, which reduces the affinity of the Bak BH3
peptide for Bcl-xL by about 800-fold [30], abolished the
effect on caspase activity. In addition, a peptide derived
from the BH3 domain of Bcl-2 had no effect upon activa-
tion of caspases (Figure 2).
These results suggest that pro-apoptotic BH3 domains act
by binding endogenous Bcl-2 homologues in egg extracts.
Figure 2
20
Fluorescence (arbitrary units)
15
10
5
0
0 1 2 3 4 5
Incubation time (h)
Control Bax (55–72) Bak (72–87)
Bak
(72–87, L78A) Bid (84–99) Bcl-2 (91–106)
Current Biology
Peptides from pro-apoptotic BH3 domains trigger caspase activation
in Xenopus egg extracts. Peptides from the BH3 domains of Bax, Bak,
Bid, Bcl-2 or a mutant Bak peptide (L78A) were added to egg extracts
to a final concentration of 0.1 mg/ml. At the times shown, samples
were removed and analysed for cleavage of the fluorogenic substrate
Ac-DEVD–AMC.
916 Current Biology, Vol 7 No 12

Figure 3 Figure 4

(a) GST–Bak BH3 GST–Bax BH3 20

Bcl-2

Bcl-2
Bcl-x

Bcl-x
Bak

Bak
Bax

Bax

Fluorescence (arbitrary units)

42 42 15

32 32

10

GST–Bcl-2 BH3 GST beads

Bcl-2

Bcl-2
Bcl-x

Bcl-x
5
Bak

Bak
Bax

Bax
42 42

32 32 0
0 1 2 3 4 5
Incubation time (h)
HS HS + BH3 HS + HM
HS + HM HS + BH3
HM + BH3
(b) 20 Current Biology

BH3 domains require both cytosolic and membrane components to

Fluorescence (arbitrary units)

trigger caspase activation. Egg extracts were thawed and immediately

15 subjected to high speed centrifugation (100,000 × g). Membrane
fractions were washed in membrane buffer before incubation alone
(HM), with GST–Bak BH3 (HM + BH3), high-speed supernatant
fractions (HS) or high-speed supernatant and GST–Bak BH3 (HS +
10 HM + BH3). High-speed supernatant was also incubated alone (HS)
or with GST–Bak BH3 (HS + BH3). At the times shown, samples
were removed and caspase activity assayed using Ac-DEVD–AMC as
a substrate.
5

Although Bcl-2 homologues have been identified in

0
0 1 2 3
Incubation time (h)
Control BH3 (1x)
BH3 (2x)/Bcl-2 (1x)
Bcl-2 (2x)/BH3 (1x)
(a) BH3 domains from Bax and Bak can bind to pro-apoptotic and
anti-apoptotic Bcl-2 family members in vitro. GST–Bak BH3,
GST–Bax BH3 and GST–Bcl-2 BH3 were incubated with glutathione
beads and equal amounts of each [35S]-methionine-labelled protein in
reticulocyte lysates from in vitro transcription/translation reactions.
Labelled products bound to GST fusion proteins or beads alone were
analysed by SDS–PAGE (12% acrylamide) and autoradiography. The
migration positions of protein standards are shown on the right with
molecular masses in kDa. (b) BH3 domains interfere with protective
function of Bcl-2. Egg extracts were incubated with GST–Bak BH3
and/or Bcl-2. Concentrations used were 3 µM (1×) and 6 µM (2×).
Caspase activity was measured at the times shown using the substrate
Ac-DEVD–AMC.
Xenopus [31], they have not been characterised at the level
4 5 of protein expression. To test binding of the GST–BH3
fusion proteins to Bcl-2 family proteins, we compared the
Bcl-2 (1x) ability of each GST–BH3 fusion protein to precipitate in
vitro translated Bax, Bak, Bcl-2 or Bcl-xL. GST–Bak BH3
BH3 (1x)/ Bcl-2 (1x) was able to bind to all four Bcl-2 family proteins
(Figure 3a), but not to other proteins such as caspases in
Current Biology control precipitations (data not shown). GST–Bax BH3
bound to Bcl-2, Bcl-xL and Bax, but not Bak. In contrast,
GST–Bcl-2 BH3 was only able to bind to Bcl-xL in this
assay. GST alone failed to precipitate any of the tested
proteins (Figure 3a).
Pre-binding of Bcl-2 protein to GST–Bak BH3 fusion
protein or simultaneous addition of equimolar amounts of
the two proteins to the egg extracts neutralised the anti-
apoptotic effect of Bcl-2 (Figure 3b). Indeed, spontaneous
caspase activation occurred with very similar kinetics in
the presence or absence of a 1:1 ratio of Bcl-2:GST–Bak
BH3. When Bcl-2 was added at a two-fold excess over
GST–Bak BH3, inhibition of caspase activity was
 Research Paper Regulation of apoptosis by BH3 domains Cosulich et al.
Figure 5
BH3 domains trigger release of cytochrome c
(Cyt c) from mitochondria. (a) Egg extracts (a)
were incubated with PBS (control), GST–Bak Incubation
BH3 or Bcl-2 for the times shown and then time (h) 0 0.5
subjected to high-speed centrifugation
(100,000 × g). The high-speed supernatant
fractions were analysed by SDS–PAGE (15%
acrylamide) and western blotting using
antibodies to cytochrome c. The migration Cyt c
positions of protein standards are shown on
the right with molecular masses in kDa. Caspase activity 14 19 16 16 114 18 228 187 154 78 31
(arbitary units)
Caspase activity was assayed using Ac-
DEVD–AMC as a substrate. (b) The
experiment was carried out as in (a), except (b)
that both supernatant (HS) and heavy
membrane pellet (HM) fractions were Time (h)
analysed by western blotting using antibodies
to cytochrome c and cytochrome a. Note that
the relative amounts of cytochrome c in the
pellet and supernatant are not directly
comparable, but the pellet fraction represents
approximately 25% of volume of extract Cyt a
required to generate the supernatant fraction. Cyt c
The supernatant fractions are comparable to
each other.
observed, whereas a two-fold excess of GST–Bak BH3
triggered caspase activation (Figure 3b). These results
suggest that pro-apoptotic BH3 domains bind to and inac-
tivate anti-apoptotic Bcl-2 family proteins; presumably, an
excess of Bak BH3 interacts with Bcl-2 homologues in the
extracts to trigger apoptosis.
The spontaneous activation of caspases in Xenopus egg
extracts requires a heavy membrane fraction that includes
mitochondria [26]. We therefore tested whether intracel-
lular membranes are also required for the pro-apoptotic
activity of BH3 domains. Egg extracts were separated by
high-speed centrifugation into a heavy membrane fraction
containing mitochondria (HM) and a post-mitochondrial
high-speed supernatant (HS). The HS and HM fractions
did not show caspase activity when incubated separately
(Figure 4). However, when extracts were reconstituted by
mixing the two fractions, the spontaneous activation of
caspases usually observed after 3 hours incubation was
restored. Addition of GST–Bak BH3 to either the HS frac-
tion or the HM fraction alone failed to trigger caspase acti-
vation. However, GST–Bak BH3 accelerated caspase
activation in reconstituted extracts, indicating a require-
ment for both membrane and cytosolic components for
GST–Bak BH3 pro-apoptotic activity.
In Xenopus egg extracts, cytochrome c has been shown to be
slowly released from mitochondria causing the spontaneous
917
Control Bak BH3 Bcl-2
1 2 4 0 0.5 1 2 4 0 0.5 1 2 4
32
18
27 32 26 34
Control BH3
1 4 4
HS HM HS HM HS HM
Current Biology
activation of caspases after prolonged incubation [13]. One
possible mechanism for the rapid caspase activation by
BH3 domains might involve the triggering of cytochrome c
release from mitochondria. When egg extracts containing
mitochondria were incubated at 23°C and subjected to
high-speed centrifugation, spontaneous release of cyto-
chrome c and the coincident appearance of caspase activity
in the post-mitochondrial supernatant (HS) occurred
approximately 3–4 hours after the start of the incubation
(Figure 5a). Addition of Bcl-2 protein completely pre-
vented the release of cytochrome c and caspase activation,
as has been found previously [13]. In contrast, when
GST–Bak BH3 was added, cytochrome c appeared in the
supernatant within 0.5 hours of the start of incubation, and
accumulated to a high level after 4 hours. Caspase activity
in the supernatant was again coincident with the appear-
ance of cytochrome c (Figure 5a). In contrast to cytochrome
c, cytochrome a, which is located in the inner mitochondrial
membrane, did not appear in the supernatant fraction fol-
lowing either prolonged incubation of the extracts or addi-
tion of GST–Bak BH3. The majority of cytochrome c also
remained in the heavy membrane fraction (Figure 5b). We
confirmed that addition of purified cytochrome c to the
post-mitochondrial supernatant at levels similar to those
released from mitochondria resulted in caspase activation
(data not shown; see also [32]). Therefore, BH3 domains
seem to trigger capsase activation by acting at the control
point regulating the release of cytochrome c.
918 Current Biology, Vol 7 No 12
Discussion
In this study, we have carried out a functional analysis of
the BH3 domains of Bcl-2 family proteins using a cell-free
system based on Xenopus egg extracts. BH3 domains from
Bak and Bax were found to be potent inducers of apopto-
sis in the presence of a membrane fraction; they trigger
the activation of caspases and this process involves the
release of cytochrome c from mitochondria. Apoptosis-
inducing activity was reproduced by peptides of 16 amino
acids corresponding to the minimal region of the Bak, Bax
and Bid BH3 domains required for binding to Bcl-xL [30].
This result demonstrates that this region is also sufficient
for the induction of apoptosis. Although BH3 domains
from all Bcl-2 family members exhibit sequence similarity,
only BH3 domains from pro-apoptotic members of the
family were able to trigger caspase activation in this cell-
free system.
One or more critical functions of Bcl-2 family proteins are
thought to be controlled by the formation of dimers
between family members [19]. Studies using recombinant
proteins have suggested that both homodimers and het-
erodimers are disrupted by peptides derived from Bax
BH3 or Bak BH3 domains in binding assays in vitro [33]. It
seems likely, therefore, that BH3 domains from pro-apop-
totic proteins trigger apoptosis in Xenopus egg extracts by
binding Bcl-2 family proteins that may be present in the
extracts, disrupting anti-apototic dimers and mimicking
the formation of pro-apoptotic dimers. This conclusion is
supported strongly by our demonstration that a single
amino-acid change which effectively abolishes the binding
of a Bak BH3 peptide to Bcl-xL [30] completely abrogates
its ability to induce apoptosis. In addition, our experi-
ments titrating the activity of Bak BH3 against exogenous
Bcl-2 show that the interaction of a BH3 domain with
Bcl-2 blocks its ability to prevent apoptosis. Interestingly,
this interaction does not turn Bcl-2 into a pro-apoptotic
molecule. Excess pro-apoptotic BH3 domains may trigger
apoptosis by releasing Bax or Bak-like proteins from sup-
pression by endogenous Bcl-2-like proteins. It also
remains possible that BH3 domains bind to pro-apoptotic
proteins, perhaps activating their function. Together,
these results emphasise the view that the control of
caspase activation is critically dependent on the relative
amounts of pro-apoptotic and anti-apoptotic homologues
in the system.
The three-dimensional structure of BH3 peptides bound
to Bcl-xL suggests that, in order to form a homodimer or
heterodimer, a pro-apoptotic Bcl-2 family protein presents
its BH3 domain to the recipient protein through a confor-
mational change [30]. Given that pro-apoptotic BH3
domains bind in a pocket created by the BH1, BH2 and
BH3 domains, it seems likely that an intact recipient
protein is required. This conclusion is supported by our
finding that the BH3 domain from Bcl-2 or a peptide
derived from this domain has no activity in egg extracts. It
is also in agreement with previous studies showing that
Bcl-2 requires additional structures for its anti-apoptotic
function when transfected into cells, and for heterodimeri-
sation with Bax [18]. It seems that, while Bcl-2 acts as a
recipient protein for BH3 domains from pro-apoptotic
members of the family, it does not act as a BH3 donor
itself. Our results show that a pro-apoptotic protein, unlike
Bcl-2, does not need to contribute any structure to acti-
vate caspases and trigger apoptosis other than the minimal
region of the BH3 domain required for binding to the
recipient partner.
To trigger apoptosis in egg extracts, we find that BH3
domains require both membrane and cytosolic compo-
nents. This result is also consistent with an interaction
between BH3 domains and Bcl-2 family proteins in the
extracts, given that Bcl-2 interacts with mitochondria in
this system whereas other factors involved in caspase acti-
vation are cytosolic [13]. We have also demonstrated that
BH3 domains trigger the rapid release of cytochrome c
from mitochondria. Although the addition of cytochrome c
to a post-mitochondrial supernatant is sufficient to activate
soluble caspases in Xenopus egg extracts (this study and
[32]), it remains possible that additional factors released
from mitochondria [34] are also involved in caspase activa-
tion by BH3 domains. The mechanism by which Bcl-2
family members regulate the release of cytochrome c is
still unclear. Intriguingly, recent evidence suggests that
Bcl-xL is able to interact directly with cytochrome c, and
this interaction is disrupted by the related BH3-containing
protein Bcl-xS [12]. BH3 domains may also disrupt the
interaction of anti-apoptotic Bcl-2 family proteins with
other proteins, such as a vertebrate CED-4 homologue
[16] that regulates caspase activation (Figure 6). Interest-
ingly, Bax and Bak, as well as Bik/Nbk, a protein contain-
ing a BH3 domain, have been shown to prevent CED-4
binding to Bcl-xL [35]. One possibility is that the release
of cytochrome c from mitochondria is coincident with the
activation of an initiator caspase and dissociation of a
complex assembled by a Bcl-2 family protein on the outer
mitochondrial membrane. BH3 domains could also inter-
fere with the proposed function of Bcl-2 family proteins as
ion channels [20,36–38], but, as yet, it is unclear how this
function might control the release of large molecules such
as cytochrome c from mitochondria.
Whatever the precise mechanism of action of Bcl-2 family
proteins, the ability of BH3 domains to trigger apoptosis in
this cell-free system may have implications for the devel-
opment of reagents to control apoptosis in cells. Impor-
tantly, we have shown that short, synthetic peptides
comprising the core of the BH3 domain have apoptosis-
inducing activity. These small peptides are likely to be
independent of controls over full-length pro-apoptotic
proteins and may be mimicked by other small molecules
 Research Paper Regulation of apoptosis by BH3 domains Cosulich et al.
Figure 6
Model for the action of BH3 domains. A BH3
domain of a pro-apoptotic Bcl-2 family protein,
such as Bax or Bak, binds to an anti-apoptotic
family protein, such as Bcl-2 or Bcl-xL, at an
intracellular membrane site, in this case the
outer mitochodrial membrane. The BH3
heterodimeric complex triggers the release of
cytochrome c from mitochondria by an Bax
unknown mechanism. It may also disrupt
complexes with a CED-4 homologue (such as
Apaf-1 [16]). These events trigger the
proteolytic processing and activation of an
initiator caspase, which may be associated
with the complex, leading to the activation of
other caspases that bring about cytoplasmic
and nuclear changes by cleavage of specific
proteins. Caspase activity and many of the
characteristic biochemical and morphological
events of apoptosis are blocked by Ac-DEVD-
CHO.
that could be more easily introduced into cells. We would
anticipate that such reagents would not show a non-spe-
cific toxicity towards cells, as might be the case with intact
pro-apoptotic proteins that can form membrane pores.
Rather, they may selectively kill cells in which apoptosis
is actively suppressed by elevated levels of anti-apoptotic
Bcl-2 family proteins.
Conclusions
BH3 domains from pro-apoptotic Bcl-2 family members
and peptides derived from their sequences are potent
activators of caspases and apoptosis in a model cell-free
system. The mechanism of caspase activation by BH3
domains involves the release of cytochrome c from mito-
chondria. Our results provide support for a model in
which the role of pro-apoptotic BH3 domains is to disrupt
the function of anti-apoptotic Bcl-2 family members.
These results may open the way for the development of
novel reagents to control apoptosis in cells. Such reagents
may provide a useful strategy for intervention where
resistance to apoptosis, or its inappropriate initiation,
contributes to a disease.
Materials and methods
Xenopus egg extracts and DNA fragmentation analysis
Preparation of Xenopus egg extracts and DNA fragmentation analysis
was carried out as described previously [8]. Where stated, extracts
were fractionated into high-speed supernatant (HS) and heavy
membrane (HM) fractions by centrifugation at 100,000 × g in a
Beckman centrifuge using a TLA100 rotor. The HM fraction was
washed in three volumes of membrane buffer (50 mM KCl, 250 mM
sucrose, 2.5 mM MgCl2, 50 mM Hepes–NaOH, pH 8.0, 1 mM DTT,
1 µg/ml leupeptin, 1 µg/ml aprotinin) and resuspended in an equal
volume of buffer.
919
Mitochondrion
Bcl-2 Bax
Bcl-2
CED-4 Cytochrome c
CED-4
Caspase precursor
Active
caspase
AC-DEVD-CHO
APOPTOSIS
• Cleavage of specific proteins
• Cytoplasmic reorganisation
• Chromosome condensation
• DNA fragmentation
Current Biology
Recombinant proteins and peptides
GST fusion protein constructs were made by amplifying the appropriate
cDNA fragments using PCR and by cloning the fragments into pGEX
vectors (Pharmacia). GST–Bak BH3 included amino acids 71–123,
GST–Bax BH3 constructs included amino acids 50–78 and 50–105,
GST–Bcl-2 BH3 included amino acids 86–111. Protein expression in
Escherichia coli BL21 DE3 and purification on glutathione beads (Phar-
macia) were carried out using standard techniques. The following pep-
tides were obtained from Genosys or Zeneca Pharmaceuticals: Bak
(amino acids 72–87), Bak mutant (as Bak, with alanine at position 78),
Bax (amino acids 57–72), Bid (amino acids 84–99), Bcl-2 (amino acids
91–106) as described previously [30]. Bcl-2 (amino acids 1–218)
tagged at the carboxyl terminus with six histidines was constructed
using standard PCR techniques and cloned into vector ptb375
(Zeneca) for expression in E. coli BL21 DE3. Unless stated otherwise,
recombinant proteins and peptides dissolved in a 10-fold dilution of
phosphate-buffered saline (PBS) were added to egg extracts to 1/10 of
the final volume to give a concentration of 0.1 mg/ml.
SDS–PAGE and western blot analysis
For analysis of poly (ADP–ribose) polymerase (PARP) cleavage, 10 µl
samples were removed from the incubation with HeLa nuclei (see
above) and spun in a microcentrifuge at 4000 r.p.m. The top 5 µl
supernatant was then discarded and the remaining volume dissolved in
SDS–PAGE gel sample buffer. Western analysis was performed using
a PARP monoclonal antibody (gift from K. Caldecott), horseradish per-
oxidase-conjugated anti-mouse immunoglobulin and enhanced chemi-
luminescence (ECL) detection reagents (Amersham). For analysis of
cytochrome c release from mitochondria, extracts were subjected to
centrifugation at 100,000 × g in a Beckman centrifuge using a TLA100
rotor. Supernatants were then removed and added to SDS–PAGE
sample buffer. Western analysis was performed using a cytochrome c
monoclonal antibody (gift from R. Jemmerson) and standard ECL tech-
niques. The monoclonal antibody was found to also react with another
protein of > 50 kDa, as reported previously [11].
Caspase activation assay
Extracts were incubated in the presence of peptides (0.1 mg/ml final
concentration) or GST fusion proteins (also 0.1 mg/ml final
920 Current Biology, Vol 7 No 12
concentration). At the times indicated, 5 µl samples were removed and
incubated for exactly 10 min with 50 µM Ac-DEVD–AMC (Cal-
biochem). After incubation, samples were diluted to 2 ml in PBS. Cleav-
age of Ac-DEVD–AMC was detected with a fluorimeter using an
excitation wavelength of 380 nm and an emission wavelength of
460 nm.
In vitro transcription/translation of proteins and binding assays
Human cDNA clones of human Bcl-2, Bcl-xL (encoding amino acids
1–211), Bax (encoding amino acids 1–171) and Bak in pcDNA3 (Invit-
rogen) were translated using a coupled transcription/translation system
(Promega) in the presence of T7 polymerase and [35S]-methionine
according to the manufacturer’s instructions. Binding of GST–BH3
fusion proteins to glutathione beads (Pharmacia) was carried out in
binding buffer (20 mM Tris pH 8.0, 50 mM NaCl, 0.1 % Triton, 10 %
glycerol) according to the manufacturer’s instructions. Briefly, 20 µl
beads were bound to 5 µg of each GST fusion protein at 4°C for
30 min. Beads were washed three times in binding buffer, before addi-
tion of equal amounts of [35S]-methionine-labelled reticulocyte lysates
from in vitro translation reactions and incubation at 4°C for 60 min.
Beads were washed three times in binding buffer before addition of
20 µl GSB (50 mM Tris pH 6.8, 100 mM DTT, 2% SDS, 0.1% bro-
mophenol blue) and analysis by SDS–PAGE. Gels were dried and
analysed by autoradiography.
Acknowledgements
We thank K. Caldecott and R. Jemmerson for generous gifts of antibodies,
C. Streuli for comments on the manuscript and J.A. Hickman and P.
Workman for discussion and support.
References
1. Raff M: Social controls on cell survival and cell death. Nature
1992, 356:397-399.
2. Thompson CB: Apoptosis in the pathogenesis and treatment of
disease. Science 1995, 267:1456-1462.
3. Kerr JFR, Wyllie AH, Currie AR: Apoptosis: a basic biological
phenomenon with wide-ranging implications in tissue kinetics. Br
J Cancer 1972, 26:239-257.
4. Reed JC: Double identity for proteins of the Bcl-2 family. Nature
1997, 387:773-776.
5. Hengartner MO, Horvitz HR: C. elegans cell survival gene ced-9
encodes a functional homolog of the mammalian proto-oncogene
bcl-2. Cell 1994, 76:665-676.
6. Vaux DL, Weissman IL, Kim SK: Prevention of programmed cell
death in Caenorhabditis elegans by human bcl-2. Science 1992,
258:1955-1957.
7. Chinnaiyan AM, Orth K, O’Rouke K, Duan H, Poirier GG, Dixit VM:
Molecular ordering of the cell death pathway. J Biol Chem 1996,
271:4573-4576.
8. Cosulich SC, Green S, Clarke PR: Bcl-2 regulates activation of
apoptotic proteases in a cell-free system. Curr Biol 1996,
6:997-1005.
9. Rao L, White W: Bcl-2 and ICE family of apoptotic regulators:
making a connection. Curr Opin Genet Dev 1997, 7:52-58.
10. Martin SJ, Green DR: Protease activation during apoptosis: death
by a thousand cuts? Cell 1995, 82:349-352.
11. Liu X, Kim CN, Yang J, Jemmerson R, Wang X: Induction of
apoptotic program in cell free extracts: requirement for dATP and
cytochrome c. Cell 1996, 86:147-157.
12. Kharbanda S, Pandey P, Schofield L, Israels S, Roncinske R, Yoshida
K, et al.: Role for Bcl-xL as an inhibitor of cytochrome c
accumulation in DNA damage-induced apoptosis. Proc Natl Acad
Sci USA 1997, 94:6939-6942.
13. Kluck RM, Bossy-Wetzel E, Green DR, Newmeyer DD: The release
of cytochrome c from mitochondria: a primary site for Bcl-2
regulation of apoptosis. Science 1997, 275:1132-1136.
14. Yang J, Liu X, Bhalla K, Kim CN, Ibrado AM, Cai J, et al.: Prevention
of apoptosis by Bcl-2: release of cytochrome c from
mitochondria. Science 1997, 275:1129-1132.
15. Vaux DL: CED-4 — the third horseman of apoptosis. Cell 1997,
90:389-390.
16. Zou H, Henzel WJ, Liu X, Lutschg A, Wang X: Apaf-1, a human
protein homologous to C. elegans CED-4, participates in
cytochrome c-dependent activation of caspase-3. Cell 1997,
90:405-413.
17. Chittenden T, Flemington C, Houghton AB, Ebb RG, Gallo GJ,
Elangovan B, et al.: A conserved domain in Bak, distinct from BH1
and BH2, mediates cell death and protein binding functions.
EMBO J 1995, 14:5589-5596.
18. Yin X-M, Oltvai ZN, Korsmeyer SJ: BH1 and BH2 domains of Bcl-2
are required for inhibition of apoptosis and heterodimerisation
with Bax. Nature 1994, 369:321-323.
19. Oltvai ZN, Korsmeyer SJ: Checkpoints of dueling dimers foil death
wishes. Cell 1994, 79:189-192.
20. Muchmore SW, Sattler M, Liang H, Meadows M, Harlan JE, Yoon HS,
et al.: X-ray and NMR structure of human Bcl-XL, an inhibitor of
programmed cell death. Nature 1996, 381:335-341.
21. Zha HB, Aimesempe C, Sato T, Reed JC: Proapoptotic protein Bax
heterodimerizes with Bcl-2 and homodimerizes with Bax via a
novel domain (BH3) distinct from BH1 and BH2. J Biol Chem
1996, 271:7440-7444.
22. Wang K, Yin X-M, Chao D, Milliman CL, Korsmyer SJ: BID: a novel
BH3 domain-only death agonist. Genes Dev 1996, 10:2859-2869.
23. Inohara N, Ding LY, Chen S, Nunez G: harakiri, a novel regulator of
cell death, encodes a protein that activates apoptosis and
interacts selectively with survival-promoting proteins Bcl-2 and
Bcl-XL. EMBO J 1997, 16:1686-1694.
24. Han J, Sabbatini P, White E: Induction of apoptosis by human
Nbk/Bik, a BH3-containing protein that interacts with E1B 19K.
Mol Cell Biol 1996, 16:5857-5864.
25. Boyd JM, Gallo GJ, Elangovan B, Houghton AB, Malstrom S, Avery
BJ, et al.: Bik, a novel death-inducing protein shares a distinct
sequence motif with Bcl-2 family proteins and interacts with viral
and cellular survival-promoting proteins. Oncogene 1995,
11:1921-1928.
26. Newmeyer DD, Farschon DM, Reed JC: Cell-free apoptosis in
Xenopus egg extracts - Inhibition by Bcl-2 and requirement for an
organelle fraction enriched in mitochondria. Cell 1994,
79:353-364.
27. Cosulich SC, Horiuchi H, Zerial M, Clarke PR, Woodman PG:
Cleavage of Rabaptin-5 blocks endosome fusion during
apoptosis. EMBO J 1997, 16:6182-6191.
28. Le Romancer M, Cosulich SC, Jackson SP, Clarke PR: Cleavage and
inactivation of DNA-dependent protein kinase catalytic subunit
during apoptosis in Xenopus egg extracts. J Cell Sci 1996,
109:3121-3127.
29. Evans EK, Lu W, Strum SL, Mayer BJ, Kornbluth S: Crk is required
for apoptosis in Xenopus egg extracts. EMBO J 1997, 16:230-241.
30. Sattler M, Liang H, Nettesheim D, Meadows RP, Harlan JE, Eberstadt
M, et al.: Structure of Bcl-xL–Bak peptide complex: recognition
between regulators of apoptosis. Science 1997, 275:983-986.
31. Cruz-Reyes J, Tata JR: Cloning, characterization and expression of
two Xenopus bcl-2-like cell-survival genes. Gene 1995,
158:171-179.
32. Kluck RM, Martin SJ, Hoffman BM, Zhou JS, Green DR, Newmeyer
DD: Cytochrome c activation of CPP32-like proteolysis plays a
critical role in a Xenopus cell-free apoptosis system. EMBO J
1997, 16:4639-4649.
33. Diaz JL, Oltersdorf T, Horne W, McConnell M, Wilson G, Weeks S, et
al.: A common binding site mediates heterodimerization and
homodimerization of Bcl-2 family members. J Biol Chem 1997,
272:11350-11355.
34. Susin SA, Zamzami N, Castedo M, Hirsch T, Marchetti P, Macho A, et
al.: Bcl-2 inhibits the mitochondrial release of an apoptogenic
protease. J Exp Med 1996, 184:1331-1342.
35. Chinnaiyan AM, O’Rouke K, Lane B, Dixit VM: Interaction of CED-4
with CED-3 and CED-9: a molecular framework for cell death.
Science 1997, 275:1122-1128.
36. Schendel SL, Xie Z, Oblatt Montal M, Matsuyama S, Montal M, Reed
JC: Channel formation by antiapoptotic protein Bcl-2. Proc Natl
Acad Sci USA 1997, 94:5113-5118.
37. Minn AJ, Velez P, Schendel SL, Liang H, Muchmore SW, Fesik SW, et
al.: Bcl-XL forms ion channels in synthetic lipid membranes. Nature
1997, 385:353-357.
38. Antonsson B, Conti F, Ciavatta A, Montessuit S, Lewis S, Martinou I,
et al.: Inhibition of Bax channel-forming activity by Bcl-2. Science
1997, 277:370-372.