Cannibalism by Sporulating Bacteria

Abstract

Spore formation by the bacterium Bacillus subtilis is an elaborate developmental process that is triggered by nutrient limitation. Here we report that cells that have entered
the pathway to sporulate produce and export a killing factor and a signaling protein that act cooperatively to block sister
cells from sporulating and to cause them to lyse. The sporulating cells feed on the nutrients thereby released, which allows
them to keep growing rather than to complete morphogenesis. We propose that sporulation is a stress-response pathway of last
resort and that B. subtilis delays a commitment to spore formation by cannibalizing its siblings.

Full text

DOI: 10.1126/science.1086462
, 510 (2003); 301Science
et al.José E. González-Pastor,
Cannibalism by Sporulating Bacteria
www.sciencemag.org (this information is current as of February 13, 2007 ):
The following resources related to this article are available online at
http://www.sciencemag.org/cgi/content/full/301/5632/510
version of this article at:
including high-resolution figures, can be found in the onlineUpdated information and services,
http://www.sciencemag.org/cgi/content/full/1086462/DC1
can be found at: Supporting Online Material
found at:
can berelated to this articleA list of selected additional articles on the Science Web sites
http://www.sciencemag.org/cgi/content/full/301/5632/510#related-content
http://www.sciencemag.org/cgi/content/full/301/5632/510#otherarticles
, 10 of which can be accessed for free: cites 17 articlesThis article
67 article(s) on the ISI Web of Science. cited byThis article has been
http://www.sciencemag.org/cgi/content/full/301/5632/510#otherarticles
29 articles hosted by HighWire Press; see: cited byThis article has been
http://www.sciencemag.org/cgi/collection/microbio
Microbiology
: subject collectionsThis article appears in the following
http://www.sciencemag.org/help/about/permissions.dtl
in whole or in part can be found at: this article
permission to reproduce of this article or about obtaining reprintsInformation about obtaining
registered trademark of AAAS.
c 2003 by the American Association for the Advancement of Science; all rights reserved. The title SCIENCE is a
CopyrightAmerican Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005.
Science (print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by the
on February 13, 2007 www.sciencemag.orgDownloaded from

because a single locus could retain higher
than expected levels of diversity. In particu-
lar, certain positions in the mammalian D
loop appear hypermutable. In our humpback
analysis, for example, some positions
changed four to eight times. Because these
positions contribute disproportionately to in-
traspecific diversity, they may inflate ␪. Re-
moving the 14 sites with more than three
intraspecific changes reduces diversity by
about 25 to 33%, suggesting that this may be
a source of error. Given the idiosyncratic
features of this single locus, data from more
loci are required before a fully accurate esti-
mate of historical populations is possible.
Unfortunately, no comparable nuclear data
sets are yet available (9). Yet even if we
assume that diversity is reduced by 50%, and
rely on the lowest 95% confidence limit, our
estimate of humpback populations would be
about 75,000. Populations would also be
halved if generation time estimates were dou-
bled, but errors of this magnitude are unlike-
ly. To bring our results completely in line
with historical humpback population sizes of
approximately 20,000 requires generation
times of more than 45 years plus a substitu-
tion rate nearly four times higher than esti-
mated (Fig. 2).
The genetic diversity of humpback,
minke, and fin whales is inconsistent with the
low historical population sizes currently as-
sumed (9). The discrepancy of these values
represents a crucial challenge. To reconcile
these results requires genetic analyses of ad-
ditional loci; more information about South
Atlantic populations; and reevaluation of the
time period, severity, and demographic im-
pacts of North Atlantic whaling.
Reconciling these numbers is crucial, be-
cause the possibility that vast cetacean pop-
ulations existed across deep ecological time
has fundamental implications not only for
their management but also for our perception
of the world’s oceans. In its Revised Man-
agement Procedure, the International Whal-
ing Commission (IWC) states, “catches
should not be allowed on stocks below 54%
of the estimated carrying capacity” (27 ). Ge-
netic data cannot be used alone to define
carrying capacity, because effective popula-
tion sizes are often orders of magnitude lower
than population censuses (5, 7, 9, 24 ), but
they can be useful in setting a lower limit to
these values. In light of our findings, current
populations of humpback or fin whales are
far from harvestable. Minke whales are closer
to genetically defined population limits, and
hunting decisions regarding them must be
based on other data.
References and Notes
1. R. Mather, Journal of Richard Mather (D. Clapp, Bos-
ton, MA, 1850).
2. M. Klinowska, Dolphins, Porpoises and Whales of the
World: The IUCN Red Data Book (IUCN, Cambridge,
UK, 1991).
3. P. T. Stevick et al., SC/53/NAH2, IWC (2001).
4. M. Soule´, in Molecular Evolution, F. J. Ayala, Ed.
(Sinauer, Sunderland, MA, 1976), pp. 60 –77.
5. R. Waples, Rep. Int. Whaling Comm. 15, 279 (1991).
6. R. Frankham, Conserv. Biol. 10, 1500 (1996).
7. J. E. Neigel, Conserv. Genet. 3, 167 (2002).
8. P. Beerli, J. Felsenstein, Proc. Natl. Acad. Sci. U.S.A.
98, 4563 (2001).
9. See supporting material on Science Online.
10. R. Kellog, Proc. U.S. Natl. Mus. 61, 1 (1922).
11. G. Pesole, C. Gissi, A. De Chirico, C. Saccone, J. Mol.
Evol. 48, 427 (1999).
12. A. P. Rooney, R. L. Honeycutt, J. N. Derr, Evolution 55,
1678 (2001).
13. J. Horwood, Biology and Exploitation of the Minke
Whale (CRC Press, Boca Raton, FL, 1990).
14. H. Kishino, H. Kato, F. Kasamatsu, Y. Fujise. Ann. Inst.
Stat. Math. 43, 435 (1991).
15. C. Lockyer, Rep. Int. Whaling Comm. 31, 389 (1981).
16. C. Lockyer, R. Gambell, S. G. Brown, Rep. Int. Whaling
Comm. 27, 427 (1977).
17. R. G. Chittleborough, Aust. J. Mar. Fresh. Res. 16,33
(1965).
18. For juvenile abundance, we conservatively assumed
that animals in year classes 1, 2, 3, 4, and 5 were
equally abundant.
19. Based on annual female survival of 96% (40) over a
reproductive life of 60 years.
20. C. Lockyer, Rep. Int. Whaling Comm. 6, 27 (1984).
21. L. Nunney, Evolution 47, 1329 (1993).
22. D. W. Weller, et al., J. Cetacean Res. Manage. 4,7
(2002).
23. R. P. Angliss, D. J. Rugh, D. E. Withrow, R. C. Hubbs,
Rep. Int. Whaling Comm. 45, 313 (1995).
24. J. Avise, R. M. Ball, J. A. Arnold, Mol. Biol. Evol. 5, 331
(1988).
25. J. Sigurjo´nsson, in Whales, Seals, Fish, and Man,A.S.
Blix, L. Walløe, O兾 . Ulltang, Eds. (Elsevier, Amsterdam,
1995), pp. 425– 441.
26. T. D. Smith et al., Mar. Mamm. Sci. 15, 1 (1999).
27. IWC documents are available at www.iwcoffice.org/
Estimate.htm.
28. J. B. C. Jackson et al., Science 293, 629 (2001).
29. J. K. Baum et al., Science 299, 389 (2003).
30. R. Myers, B. Worm, Nature 423, 280 (2003).
31. H. E. Winn, N. E Reichley, in Handbook of Marine
Mammals, vol. 3, S. H. Ridgeway, R. Harrison, Eds.
(Academic Press, New York, 1985), pp. 241–273.
32. D. E. Sergeant, Rep. Int. Whal. Commn. 27, 460
(1977).
33. R. M. Nowak, Walker’s Mammals of the World ( Johns
Hopkins Univ. Press, Baltimore, ed. 6, 1999).
34. H. C. Rosenbaum, et al., J. Hered. 93, 389 (2002).
35. J. Roman, S. R. Palumbi, data not shown.
36. D. Rand, Conserv. Biol. 10, 665 (1996).
37. C. S. Baker, G. M. Lento, F. Cipriano, S. R. Palumbi,
Proc. R. Soc. London Ser. B 267, 1191 (2000).
38. C. S. Baker et al., Proc. Natl. Acad. Sci. U.S.A. 90,
8239 (1993)
39. S. Wada, K. I. Numachi, Rep. Int. Whaling Comm. 13,
125 (1991).
40. J. Barlow, P. J. Clapham, Ecology 78, 535 (1997).
41. We thank J. Alipaz, C. Baker, P. Beerli, B. Bowen, R.
Brownell, P. Clapham, L. Farrell, J. Felsenstein, G.
Giribet, J. Jackson, M. Kuhner, D. Neafsy, D. Pauly, D.
Petrov, A. Platt, D. Rand, J. Robbins, T. Smith, M.
Turelli, J. Wakeley, and J. Wares for discussion and
suggestions. Supported by a Mia J. Tegner Memorial
Research Grant in Marine Environmental History and
Historical Marine Ecology from the Marine Conserva-
tion Biology Institute to J.R. and grants from NSF and
the Pew Charitable Trusts to S.R.P.
Supporting Online Material
www.sciencemag.org/cgi/content/full/301/5632/508/
DC1
Materials and Methods
Fig. S1
References
14 March 2003; accepted 17 June 2003
Cannibalism by
Sporulating Bacteria
Jose´ E. Gonza´lez-Pastor,* Errett C. Hobbs, Richard Losick†
Spore formation by the bacterium Bacillus subtilis is an elaborate developmental
process that is triggered by nutrient limitation. Here we report that cells that have
entered the pathway to sporulate produce and export a killing factor and a signaling
protein that act cooperatively to block sister cells from sporulating and to cause
them to lyse. The sporulating cells feed on the nutrients thereby released, which
allows them to keep growing rather than to complete morphogenesis. We propose
that sporulation is a stress-response pathway of last resort and that B. subtilis
delays a commitment to spore formation by cannibalizing its siblings.
Some microorganisms respond to nutritional
limitation by entering a resting state in which
they remain inactive for an extended time.
Bacillus subtilis produces a robust resting
cell, the endospore, that can remain dormant
for many years. Endospore formation is an
elaborate and energy intensive process that
requires several hours to complete (1–4 ). If
during this period nutrients were once again
to become plentiful, the sporulating cells
would be at a disadvantage relative to cells
able to resume growth rapidly. Thus, bacteria
could be expected to delay spore formation
until forced to do so by prolonged depletion
of nutrients. Here we present evidence that
cells that have entered the pathway to sporu-
late delay development by killing their sib-
lings and feeding on the nutrients thereby
released. Cannibalism is mediated by an ex-
tracellular killing factor and a novel intercel-
lular signaling protein that act cooperatively
to cause cell death and impede sporulation.
Department of Molecular and Cellular Biology, The
Biological Laboratories, Harvard University, 16 Divin-
ity Avenue, Cambridge, MA 02138, USA.
*Present address: Centro de Astrobiologı´a (CSIC-INTA),
Instituto Nacional de Te´cnica Aeroespacial, Carretera de
Ajalvir, km 4, 28850 Torrejo´n de Ardoz, Madrid, Spain.
†To whom correspondence should be addressed. E-
mail: losick@mcb.harvard.edu
R EPORTS
25 JULY 2003 VOL 301 SCIENCE www.sciencemag.org510
on February 13, 2007 www.sciencemag.orgDownloaded from

Entry into sporulation is governed by the
regulatory protein Spo0A (5). While building
mutants of genes under the control of Spo0A (6),
we discovered two operons (Fig. 1A) that are
strongly induced at the start of sporulation (fig.
S1) and in which mutations accelerated spore
formation (Fig. 1, B and C, and fig. S2). We refer
to these operons as skf for sporulation killing
factor and sdp for sporulation delaying protein.
Clues that the eight-gene skf operon di-
rects the production of an exported killing
factor came from the similarity of its gene
products to proteins involved in the produc-
tion of peptide antibiotics (7–9). The first
gene, skfA, encodes a small peptide, a char-
acteristic of operons involved in the produc-
tion of peptide antibiotics (7). The product of
the second gene, skfB, is similar to a B.
subtilis protein involved in the production of
an antilisterial peptide, subtilosin (10). Final-
ly, the product of skfD contains a domain
characteristic of the CAAX family of amino
terminal proteases (11, 12). The operon also
contains two genes, skfE and skfF, whose
products resemble an ATP– binding cassette
transport complex (ABC transporter) and
could be responsible for exporting the peptide
antibiotic and conferring resistance to it.
To investigate these possibilities, we asked
whether wild-type cells would kill cells of a
mutant of the skf operon that had been marked
with a lacZ fusion. Mutant and wild-type cells
were mixed in equal proportions and grown in
liquid sporulation medium. The ratio of mutant
to wild-type cells remained approximately con-
stant during growth but dropped dramatically
after the onset of sporulation (Fig. 2A). These
results indicate that the skf operon is involved in
the production of an extracellular killing factor
during sporulation. The operon must also con-
fer resistance to the factor, because the mutation
rendered cells sensitive to it. In keeping with
these ideas, cells engineered to express the
skf operon during growth in response to
IPTG (isopropyl ␤-
D-1-thiogalactopyranoside)
caused killing when spotted on a lawn of wild-
type or skf mutant cells and did so in a manner
that was dependent on the presence of the IPTG
inducer (Fig. 2B). Evidence that skfE and skfF
encode an export pump for the killing factor
came from placing the genes under the control
of an IPTG-inducible promoter and introducing
the construct into a strain that lacked the skf
operon. The mutant cells were mixed with wild-
type cells (tagged with lacZ) and grown in
liquid sporulation medium. The number of mu-
tant cells dropped sharply upon entry into
sporulation when grown in the absence of IPTG
but not when grown in its presence (Fig. 2C).
We next asked whether the operon causes
death in a homogenous population of wild-type
cells. Cultures of cells sporulating in liquid me-
dium show a characteristic drop in optical density
shortly after the start of spore formation. This
drop was associated with a dramatic (⬃70%)
decrease in the number of viable cells, and, of
note, in a manner that was dependent on skf (Fig.
2D and fig. S3). The simplest interpretation of
these results is that the wild type produces a
mixed population in which Spo0A is active (and
directing transcription of skf) in some cells and
not in others [fig. S4 and (13, 14 )]. Cells with
active Spo0A would produce the killing factor
and the pump that exports it. Cells with inactive
Spo0A would produce neither the factor nor the
pump, and they would be killed. Thus, the killing
factor is responsible for killing genetically iden-
tical cells (siblings) in the population. This is
contrary to the traditional paradigm of chemical
warfare among microorganisms in which antibi-
otics are used to kill other, competing species.
Why do colonies of skf mutant cells exhibit
accelerated sporulation (15)? We suggest that the
killing factor causes cells in which Spo0A is
inactive to lyse and release nutrients, which al-
Fig. 1. Mutants of the skf
and sdp operons sporu-
late rapidly. (A) Gene or-
ganization of the skfAB-
CDEFGH and the sdpABC
operons [previously an-
notated as ybcOPST yb-
dABDE and as yvaWXY,
respectively (23)]. The
hairpin symbols repre-
sent transcriptional ter-
minators. (B) Colonies
of skf [⌬(skfABCDEF)::tet;
strain EG168] and sdp
[⌬(sdpABC)::spc; strain
EG407] mutants, as well
as the double skf sdf
mutant (EG523), were
brighter (an indication of
spore formation, see fig.
S2) than those formed by the wild-type strain (PY79) after 14 hours of incubation on solid
sporulation medium. Strains and plasmids used in our experiments are listed in table S2. (C)
Time course of spore formation in solid medium by the wild-type (●), and the skf (Œ), sdp (䡲)
and skf sdp (䉬) mutants. The percentage of heat-resistant, colony-forming units (spores)
versus total viable cells was monitored at the indicated times after inoculation in solid
sporulation medium.
Fig. 2. The skf operon
produces a sporulation
killing factor. (A) skf mu-
tant cells harboring a lacZ
fusion [⌬(skfABCDEF)::tet
amyE::cotD-lacZ; EG169]
and wild-type cells (PY79)
were mixed in equal pro-
portions and grown in liq-
uid sporulation medium.
The ratio of mutant to
wild-type cells was deter-
mined at the indicated
times before and after the
start of sporulation (hour
0, 䡲). As a control, ratios
were determined for wild-
type cells that had been
mixed with wild-type cells
that carried a lacZ fusion
(strain PE29, ●) and for skf
mutant cells (EG168) that
had been mixed with skf
mutant cells that carried a lacZ fusion (EG169, Œ). In these mixed cultures, the total number of viable cells
during the time course was similar to that for individual cultures of the wild type or the skf mutant shown
in (D). (B) Cells harboring the skf operon under the control of an IPTG-inducible promoter (strain EG208) were
spotted on a lawn of wild-type or skf mutant cells growing on a rich (nonsporulation, Luria broth) medium.
The engineered cells produced a halo of growth inhibition (arrow) in the presence (⫹) but not in the absence
(–) of the inducer (1 mM IPTG). (C) Cells lacking the skf operon but containing a copy of skfE and skfF under
the control of an IPTG-inducible promoter (strain EG219) were mixed in equal proportion with wild-type cells
that carried a lacZ fusion (PE29), and the cell mixture was grown in liquid sporulation medium in the absence
(䡲) or in the presence (●) of the inducer (IPTG). The ratio of cells of strain EG219 to the cells of strain PE29
was determined at the indicated times after the start of sporulation. (D) Number of viable cells was measured
in cultures of wild-type cells (●), and cells of skfA (EG165) (䡲) and skfABCDEF (EG168) (Œ) mutants in liquid
sporulation medium at the indicated times after the start of sporulation.
R EPORTS
www.sciencemag.org SCIENCE VOL 301 25 JULY 2003 511
on February 13, 2007 www.sciencemag.orgDownloaded from

low cells in which Spo0A is active but which have
not yet committed
to morphogenesis to keep grow-
ing. An skf mutant, in contrast, does not cause
killing and hence sporulation takes place with-
out delay.
Mutations in a second operon, sdp (Fig.
1A), also caused an accelerated sporulation
phenotype and did so more rapidly than mu-
tations in skf (Fig. 1C and fig. S2). What is
the mechanism by which this three-gene
operon delays spore formation? To answer
this question, we carried out microarray anal-
ysis to identify genes whose transcription was
dependent on the operon (fig. S5 and table
S1). Two genes whose transcription was
strongly dependent on sdp were yvbA (whose
inferred product is similar to the ArsR family
of transcriptional regulators) and yvaZ
(whose product is of unknown function but is
inferred to contain multiple transmembrane
segments). The yvbA and yvaZ genes consti-
tute an apparent operon that is located imme-
diately downstream of, and in convergent
orientation to, the sdp operon itself (Fig. 1A).
The use of lacZ fused to the promoter for
yvbA and yvaZ (P
yvbA yvaZ
-lacZ) confirmed
that transcription of the operon was almost
completely dependent on sdp (Fig. 3A).
Remarkably, this dependence was mediat-
ed by intercellular signaling. Expression of
P
yvbA yvaZ
-lacZ in cells mutant for sdp was
restored when the mutant cells were grown
in close proximity to wild-type cells on
solid medium (Fig. 3B). No restoration of
lacZ expression was observed when the
mutant was grown close to cells mutant for
sdp. Evidently, sdp is responsible for the
production of an extracellular factor that is
capable of inducing the transcription of
yvbA and yvaZ in recipient cells. We puri-
fied from conditioned medium from a cul-
ture of wild-type cells a fraction containing
a ⬃5-kD protein that stimulated ␤-galacto-
sidase synthesis when added to cells of an
sdp mutant that harbored P
yvbA yvaZ
-lacZ
[Fig. 3C and (16 )]. Neither the stimulatory
activity nor the protein was present in con-
ditioned medium from sdp mutant cells. A
protein of similar size was obtained with
cells engineered to express the sdp operon
during growth in response to IPTG. Again,
the fraction containing this protein stimu-
lated ␤-galactosidase production (Fig.3D).
Finally, sequential Edman degradation
(-GLYAV-VAAGYLYVVGVNAALQTAAAV)
(12) revealed that the ⬃5-kD protein origi-
nated from the product of the sdpC gene of
the operon, its N-terminal residue corre-
sponding to residue 141 of the 203residue-
long protein.
Next, we asked whether induction of the
yvbA yvaZ operon, and yvbA in particular,
was responsible for the delay in sporulation
caused by the signaling protein by engineer-
ing cells to express yvbA or yvaZ or both in
response to IPTG. The results show that ar-
tificial induction of yvbA and yvaZ or of yvbA
alone (but not yvaZ alone) was sufficient to
delay sporulation (Fig. 4A). Transcriptional
profiling with cells mutant for the sdp operon
revealed candidates for genes that could be
under the control of the YvbA transcription
factor (above; fig. S5 and table S1). Among
these were the ATP synthase operon (atpIBE-
FHAGDC), which is responsible for ATP
Fig. 3. The sdp operon encodes an
extracellular signaling protein. (A)
Wild-type cells and sdp mutant
cells containing a P
yvbA-yvaZ
-lacZ
fusion (EG381 and EG524, respec-
tively) were grown on solid sporu-
lation medium containing X-gal
(5-bromo-4-chloro-3-indolyl B-
D-
galactopyranoside). (B) Cells of
the sdp mutant harboring P
yvbA-
yvaZ
-lacZ (EG524, white arrows)
were streaked on solid sporulation
medium in the vicinity of a streak
of wild-type (strain PY79, left) or
sdp mutant cells (EG407, right).
(C) Eluates from reversed-phase
chromatography of supernatant
fluids from cultures of wild-type
and sdp mutant (EG407) cells (at
hour 1.5 of sporulation in liquid
medium) were subjected to SDS–
polyacrylamide electrophoresis in
a 4 to 20% gradient gel. (D) Su-
pernatant fluids were collected
from cells of a strain (EG351) in
which the sdp operon was under
the control of an IPTG-inducible
promoter and grown in the ab-
sence (●) or in the presence (Œ)of
the inducer (IPTG). Eluates from
reversed-phase chromatography
of the supernatant fluids from the cultures were added to cells of an sdp mutant harboring P
yvbA-yvaZ
-lacZ
(EG524) growing in minimal medium. Culture samples were collected at the indicated times and assayed for
␤-galactosidase activity.
Fig. 4. The effect of the
sdp-encoded signaling
protein is mediated by the
putative transcription fac-
tor YvbA. (A) Overexpres-
sion of yvbA yvaZ or yvbA
delays sporulation in a
strain lacking the sdp and
the yvbA yvaZ operons.
Constructs were created in
which either yvbA and
yvaZ or yvbA alone or yvaZ
alone were under the con-
trol of an IPTG-inducible
promoter (P
spac-hy
) and in-
troduced into a strain,
EG494, that was mutant
for sdpABC and yvbA yvaZ.
EG494 and its derivatives
were grown on solid
sporulation medium in the
absence and in the pres-
ence of IPTG: 1, EG494; 2,
a derivative of EG494 har-
boring P
spac-hy
-yvbA yvaZ
(EG525); 3, a derivative of
EG494 harboring P
spac-hy
-yvbA (EG526); and 4, a derivative of EG494 harboring P
spac-hy
-yvaZ (EG527).
The wild type was strain PY79. (B) Time course of accumulation of ␤-galactosidase from P
yusLKJ
-lacZ
in a wild-type strain (EG447) (●), and in a strain (EG484) mutant for sdp and yvbA yvaZ and
harboring P
spac-hy
-yvbA. The cells were grown in the absence (Œ) or presence of 1 mM IPTG (䡲).
Culture samples were collected at the indicated times before and after the start of sporulation (hour 0).
(C) Time course of the number of viable cells during sporulation of a strain (EG526) mutant for sdp and
yvbA yvaZ and harboring P
spac-hy
-yvbA (●䡩) and of a derivative of EG526 that was additionally mutant
for sfk (EG528) (䡲▫) grown in the absence (open symbols) and in the presence of 1 mM IPTG (filled symbols).
R EPORTS
25 JULY 2003 VOL 301 SCIENCE www.sciencemag.org512
on February 13, 2007 www.sciencemag.orgDownloaded from

production, and the yusLKJ operon, whose
inferred products are similar to lipid catabo-
lism enzymes (fig. S5 and table S1). Use of
lacZ fused to yusLKJ confirmed that high-
level expression of the operon was dependent
on the signaling protein and on YvbA (Fig.
4B). Also, artificial induction of YvbA syn-
thesis restored the expression of yusLKJ to
cells doubly mutant for the sdp and yvbA
yvaZ operons (Fig. 4B). We propose that the
signaling protein turns on the synthesis of
YvbA, which, in turn, causes an increase in
lipid oxidation and ATP production. The pro-
posed increase in energy production could be
responsible for delaying sporulation, which is
triggered by depletion of energy reserves.
Finally, and coming full circle, we found
that artificial induction of YvbA synthesis
caused a marked drop in cell viability in a
manner that was dependent on the skf operon
(Fig. 4C). Evidently, synthesis of the YvbA
transcription factor causes enhanced sensitiv-
ity to the sporulation killing factor. It could
do so by stimulating the expression of genes
involved in energy production, as metaboli-
cally active cells are more sensitive to anti-
biotics than are quiescent cells (17, 18). Also,
yvbA was previously identified in a screen for
genes that inhibit the expression of the gene
for ␴
W
, a regulatory protein that turns on
genes involved in detoxification and resis-
tance to antibiotics (16, 19, 20). Thus, YvbA-
mediated repression of the gene for ␴
W
could
heighten sensitivity to the killing factor by
suppressing the antibiosis stress response.
We conclude that sporulating cells of B.
subtilis are cannibalistic, feeding on their sib-
lings in order to delay committing to spore
formation. Because sporulation becomes irre-
versible after its earliest stage, delaying spore
formation as long as possible might be ben-
eficial, as a cell that is committed to spore
formation could be at a disadvantage relative
to other cells should nutrient deprivation
prove to be fleeting. Wild (but not laboratory)
strains have been found to assemble into
multicellular structures in which spore forma-
tion preferentially takes place at the apical
tips (21). Perhaps the killing factor and sig-
naling protein influence the timing and local-
ization of spore formation in these fruiting-
body-like structures. Fruiting body formation
by the unrelated spore-forming bacterium
Myxococcus xanthus is reported to involve
lysis of nonsporulating cells (22). Conceiv-
ably, this killing is mediated by cells in the
developing fruiting body that have entered
the pathway to sporulate. It will be interesting
to see whether the killing of genetically iden-
tical siblings is a widespread feature of the
dynamics of bacterial populations.
References and Notes
1. P. J. Piggot, J. G. Coote, Bacteriol. Rev. 40, 908 (1976).
2. P. J. Piggot, R. Losick, in Bacillus subtilis and Its Closest
Relatives: From Genes to Cells, A. L. Sonenshein, J. A.
Hoch, R. Losick, Eds. (ASM Press, Washington, DC,
2002), pp. 473– 481.
3. A. L. Sonenshein, in Prokaryotic Development,Y.V.
Brun, L. J. Shimkets, Eds. (ASM Press, Washington, DC,
2000), pp. 133–150.
4. A. L. Sonenshein, in Bacterial Stress Responses,G.
Storz , R. Hengge-Aronis, Eds. (ASM Press, Washing-
ton, DC, 2000), pp. 199 –215.
5. D. Burbulys, K. A. Trach, J. A. Hoch, Cell 64, 545
(1991).
6. P. Fawcett, P. Eichenberger, R. Losick, P. Youngman,
Proc. Natl. Acad. Sci. U.S.A. 97, 8063 (2000).
7. R. Kolter, F. Moreno, Annu. Rev. Microbiol. 46, 141
(1992).
8. R. J. Siezen, O. P. Kuipers, W. M. de Vos, Antonie Van
Leeuwenhoek 69, 171 (1996).
9. D. Lin, L.-J. Qu, H. Gu, Z. Chen, J. Appl. Microbiol. 91,
1044 (2001).
10. G. Zheng, L. Z. Yan, J. C. Vederas, P. Zuber, J. Bacte-
riol. 181, 7346 (1999).
11. J. Pei, N. V. Grishin, Trends Biochem. Sci. 26, 275
(2001).
12. Single-letter abbreviations for the amino acid resi-
dues are as follows: A, Ala; C, Cys; D, Asp; E, Glu; F,
Phe; G, Gly; H, His; I, Ile; K, Lys; L, Leu; M, Met; N, Asn;
P, Pro; Q, Gln; R, Arg; S, Ser; T, Thr; V, Val; W, Trp; X,
any amino acid; and Y, Tyr.
13. P. Schaeffer, J. Millet, J. P. Aubert, Proc. Natl. Acad.
Sci. U.S.A. 54, 704 (1965).
14. J. D. Chung, G. Stephanopoulos, K. Ireton, A. D. Gross-
man, J. Bacteriol. 176, 1977 (1994).
15. Of particular interest, accelerated sporulation was
observed on solid but not in liquid medium; see
supporting text on Science Online.
16. J. E. Gonza´lez-Pastor, E. Hobbs, R. Losick, data not
shown.
17. M. R. Brown, D. G. Allison, P. Gilbert, J. Antimicrob.
Chemother. 22, 777 (1988).
18. R. M. Cozens et al., Antimicrob. Agents Chemother.
29, 797 (1986).
19. M. S. Turner, J. D. Helmann, J. Bacteriol. 182, 5202
(2000).
20. M. Cao, T. Wang, R. Ye, J. D. Helmann, Mol. Microbiol.
45, 1267 (2002).
21. S. S. Branda, J. E. Gonzalez-Pastor, S. Ben-Yehuda, R.
Losick, R. Kolter, Proc. Natl. Acad. Sci. U.S.A. 98,
11621 (2001).
22. J. W. Wireman, M. Dworkin, Science 189, 516 (1975).
23. F. Kunst et al., Nature 390, 249 (1997).
24. We are grateful to S. Ben-Yehuda for the sdpC mu-
tant and contributing to its characterization; P.
Eichenberger and J. M. Ranz for help with the DNA
microarray experiments; W. Lane for Edman sequenc-
ing; and A. L. Sonenshein, P. Piggot, D. Hartl, D. Haig,
A. Murray, M. Fujita and members of the Losick
laboratory for helpful advice. This work was support-
ed by NIH grant (GM18568) to R.L. J.E.G.-P. was
supported by a Ministerio de Educacio´n y Ciencia
Postdoctoral Fellowship (Spain). E.C.H. was supported
by an NSF Graduate Research Fellowship.
Supporting Online Material
www.sciencemag.org/cgi/content/full/1086462/DC1
Material and Methods
SOM Text
Figs. S1 to S5
Tables S1 and S2
References
5 May 2003; accepted 6 June 2003
Published online 19 June 2003;
10.1126/science.1086462
Include this information when citing this paper.
VDAC2 Inhibits BAK Activation
and Mitochondrial Apoptosis
Emily H.-Y. Cheng,
1
Tatiana V. Sheiko,
2
Jill K. Fisher,
1
William J. Craigen,
2
Stanley J. Korsmeyer
1
*
The multidomain proapoptotic molecules BAK or BAX are required to initiate
the mitochondrial pathway of apoptosis. How cells maintain the potentially
lethal proapoptotic effector BAK in a monomeric inactive conformation at
mitochondria is unknown. In viable cells, we found BAK complexed with
mitochondrial outer-membrane protein VDAC2, a VDAC isoform present in
low abundance that interacts specifically with the inactive conformer of
BAK. Cells deficient in VDAC2, but not cells lacking the more abundant
VDAC1, exhibited enhanced BAK oligomerization and were more susceptible
to apoptotic death. Conversely, overexpression of VDAC2 selectively prevented
BAK activation and inhibited the mitochondrial apoptotic pathway. Death signals
activate “BH3-only” molecules such as tBID, BIM, or BAD, which displace VDAC2
from BAK, enabling homo-oligomerization of BAK and apoptosis. Thus, VDAC2, an
isoform restricted to mammals, regulates the activity of BAK and provides a
connection between mitochondrial physiology and the core apoptotic pathway.
The BCL-2 family of pro- and antiapoptotic
proteins constitutes a critical control point
for apoptosis (1, 2). A combination of ge-
netic and biochemical approaches has
helped to order the components of the mam-
malian cell death pathway. The upstream
“BH3-only” family members respond to se-
lect death signals and subsequently trigger
the activation of the multidomain death
effectors BAX and BAK (3–5). BAX and
BAK constitute an essential gateway to the
intrinsic death pathway operating at the
level of both mitochondria and endoplasmic
reticulum (ER) Ca
2⫹
dynamics (3, 6 ). Acti-
vated homo-oligomerized BAX or BAK re-
sults in the permeabilization of the mitochon-
drial outer membrane (MOM) and the
1
Howard Hughes Medical Institute, Dana-Farber Can-
cer Institute, Harvard Medical School, Boston, MA
02115, USA.
2
Department of Molecular and Human
Genetics, Baylor College of Medicine, Houston, TX
77030–3498, USA.
*To whom correspondence should be addressed. E-
mail: stanley_korsmeyer@dfci.harvard.edu
R EPORTS
www.sciencemag.org SCIENCE VOL 301 25 JULY 2003 513
on February 13, 2007 www.sciencemag.orgDownloaded from