NEWS & VIEWS NATURE|Vol 441|18 May 2006

292

warfare in bacteria8 and male–male competi-tion in lizards9.

This is not the case for Phoenix (Fig. 2d),which can out-compete the original strain andforms spores just fine. Such precipitous recov-eries may turn out to be part of the evolu-tionary process. Extinctions play a key role inthe history of life by removing species that are poorly adapted to persist, a process thatfavours both sexual reproduction and reduced

within-species conflict10,11. However, the studyby Fiegna et al.2 shows that species may alsoescape from the very brink of disaster. There issome sense in this: a virulent cheater thatthreatens a population will necessarily result in strong natural selection for strategies thatcan re-evolve sociality in its wake. It was thiseffect that led to the rapid dominance of thePhoenix mutant, and one can speculate that it might also explain adaptations that policecheating in other societies12. The final twist inthe tale is that Phoenix actually produces morespores than the original strain. Escape from avirulent cheater is not just possible; it can evenimprove things. ■

Kevin R. Foster, currently in the Program forEvolutionary Dynamics, Harvard University, will shortly move to the Bauer Center forGenomics Research, Harvard University,

Cambridge, Massachusetts 02138, USA.e-mail: kfoster@fas.harvard.edu

1. Rankin, D. J. & López-Sepulcre, A. Oikos 111, 616–619(2005).

2. Fiegna, F., Yu, Y.-T. N., Kadam, S. V. & Velicer, G. J. Nature441, 310–314 (2006).

3. Fiegna, F. & Velicer, G. J. Proc. R. Soc. Lond. B 270, 1527–1534(2003).

4. Velicer, G. J., Kroos, L. & Lenski, R. E. Nature 404, 598–601(2000).

5. Velicer, G. J., Kroos, L. & Lenski, R. E. Proc. Natl Acad. Sci.USA 95, 12376–12380 (1998).

6. Velicer, G. J. et al. Proc. Natl Acad. Sci. USA 103, 8107–8112(2006).

7. Robinson, G. E., Grozinger, C. M. & Whitfield, C. W. NatureRev. Genet. 6, 257–270 (2005).

8. Kirkup, B. C. & Riley, M. A. Nature 428, 412 (2004).9. Sinervo, B. & Lively, C. M. Nature 380, 240 (1996).10. Nunney, L. in Levels of Selection in Evolution (ed. Keller, L.)

238–252 (Princeton Univ. Press, 1999).11. Foster, K. R., Shaulsky, G., Strassmann, J. E., Queller, D. C. &

Thompson, C. R. L. Nature 431, 693–696 (2004).12. Frank, S. A. Evolution 57, 693–705 (2003).

EXTRASOLAR PLANETS

A neptunian tripletDavid Charbonneau

Three planets of Neptune mass have been discovered orbiting a Sun-likestar known to have an asteroid belt. Exquisite measurements suggest that the search for habitable planets might be easier than assumed.

Our thirst for knowledge of planets orbitingstars similar to the Sun is tempered by thetechnological challenges of detecting them.We cannot see analogues of the Solar Systemdirectly; rather, the presence of extrasolarplanets is inferred through effects that theyinduce on their parent star. The Dopplermethod, whereby astronomers search for subtle, periodic changes in the apparent speed of a star that result from its gravitational dancewith an unseen planetary companion, hasyielded all but a handful of the more than 180known extrasolar planets1. Heavier planetsproduce larger stellar wobbles, so it is not sur-prising that most of the worlds discovered so far have more in common with the distantgas and ice giants of the Solar System (Jupiter,Saturn, Uranus and Neptune) than with the smaller, closer terrestrial planets fromMercury to Mars.

But as techniques have been refined, soplanets of lower mass have been revealed inincreasing numbers1. The current bestiary ofextrasolar planets is therefore far from com-prehensive. On page 305 of this issue, Lovisand colleagues2 report unprecedentedly pre-cise observations of the nearby, Sun-like starHD 69830. The fruit of their efforts is not one,but three orbiting planets (Fig. 1). The discov-ery is exciting for two reasons. First, theauthors’ technological advance implies thatfurther low-mass planets will be spotted orbit-ing other stars. Second, the architecture of this particular planetary system bears some

intriguing similarities to that of our own SolarSystem.

The newly found planetary system isremarkable in that it possesses three planetslocated on nearly circular orbits within 1 astro-nomical unit of the star (1 AU is the Earth–Sundistance). The same is true of the Solar System.Where the HD 69830 system differs, however,is that the masses of the worlds range from 10to 18 times that of Earth, and so are similar tothat of Neptune. In the Solar System, the divi-sion between the low-mass terrestrial planetsand massive gas giants was determined by the‘ice-line’. This is the distance beyond which thetemperature in the protoplanetary nebula —the reservoir of gas and dust from which theplanets formed — dipped below the freezingpoint of various hydrogen compounds.Beyond this point, much greater amounts ofsolid material, and so planets of much greatermass, were created.

The formation history of the HD 69830 system is thus a puzzle deserving of detailedstudy. Lovis and colleagues present2 a prelimi-nary calculation to show that the inner planetsprobably formed inside the ice line, and thusare likely to be predominantly rock, not gas.Their large masses require that HD 69830’sprotoplanetary nebula contained a largerquantity of solid material than did that of theSolar System. That inference is at odds withthe observation that the star itself actually hasa lower abundance of heavier elements, thestuff of planets.

Preadapted resistance to cheatinga

b

c

d

Abu

ndan

ce

Extinction

Abu

ndan

ceA

bund

ance

Unstable recovery from cheating

Abu

ndan

ce

Stable recovery from cheating

Time

New cooperative strain

Cooperative strainCheater

Figure 2 | Four possible outcomes when a cheaterevolves in a social species. A cheater is anorganism that exploits a cooperative adaptationfor selfish gain. a, Preadapted resistance tocheating. It is typical to assume that socialsystems arise in such a way that cheaters can have only a limited impact (as shown), or do not succeed at all. Examples of preadaptationsinclude high relatedness and pre-existingconstraints that link cheating to a cost to thecheater11. Policing and enforcement systems may evolve later to further constrain cheaters12. b, Extinction. The cheater causes extinction of the social trait, or species (evolutionarysuicide1). This selects for species preadapted toresist cheating10. c, Unstable recovery. A socialstrategy arises that resists the cheater but cannotout-compete the original strategy. The originalstrategy may reinvade and perpetuate a cycle ofreinvasions in a rock–paper–scissors dynamic8,9.d, Stable recovery. Sociality is restored by astrategy that out-competes both the cheater and the original strategy, as occurred with thePhoenix mutant2. The result is a stable adaptationthat protects the social system from the cheater.This process may be behind the policing andenforcement systems in other social species12.

18.5 N&V 291 MH 12/5/06 5:27 PM Page 292

Nature Publishing Group ©2006

NATURE|Vol 441|18 May 2006 NEWS & VIEWS

293

Doppler wobble. Second, the reduced stellarbrightness shifts the habitable zone inwards,reducing the orbital period of a habitableplanet and further increasing the Dopplerwobble. Should astronomers succeed in trans-ferring the precision achieved by Lovis et al.2

to low-mass stars, they might just turn up afew planets akin to our own.

What the Doppler method will, unfortu-nately, never reveal is the composition of theplanets it detects. An exciting synergy there-fore surely exists between such high-precisionDoppler measurements and measurements tobe made by two planned satellite missions, the COROT mission4 led by the French spaceagency CNES, and NASA’s Kepler mission5.Both spacecraft will seek to identify rockyplanets by the complementary method of tran-sits6, in which a planet is revealed as it crossesthe face of its parent star by a minute dimmingof that star’s light. By observing both Dopplersignals and transits, both the mass and physicalsize of a planet can be estimated. That in turnyields a density and (by inference) a composi-tion. Previous limits on the Doppler methodimplied that any terrestrial planets detected byCOROT and Kepler would not receive corrob-orative Doppler measurements. Excitingly,Lovis and colleagues2 have given us cause torethink this pessimistic assumption. ■

David Charbonneau is in the Department ofAstronomy, Harvard University, 60 GardenStreet, Cambridge, Massachusetts 02138, USA.e-mail: dcharbonneau@cfa.harvard.edu

1. Udry, S., Fischer, D. A. & Queloz, D. in Protostars and PlanetsV (eds Reipurth, B., Jewitt, D. & Keil, K.) (Univ. ArizonaPress, Tucson, in the press).

2. Lovis, C. et al. Nature 441, 305–309 (2006).3. Beichman, C. A. et al. Astrophys. J. 626, 1061–1069

(2005).4. Bordé, P., Rouan, D. & Léger, A. Astron. Astrophys. 405,

1137–1144 (2003).5. Basri, G., Borucki, W. J. & Koch, D. New Astron. Rev. 49,

478–485 (2005).6. Charbonneau, D., Brown, T. M., Burrows, A. & Laughlin, G.

in Protostars and Planets V (eds Reipurth, B., Jewitt, D. &Keil, K.) (Univ. Arizona Press, Tucson, in the press).

HD 69830 is no stranger to the spotlight.Last year, researchers using the NASA SpitzerSpace Telescope announced that it probablypossesses an asteroid belt3 — the only star sim-ilar in mass and age to the Sun that is known tohave one. This conclusion came from theobservation that the star was brighter thanexpected at infrared wavelengths, suggestingthe presence of grains of dust that were smalland were located within 1 AU of the star. At thatdistance, small grains could not survive verylong, as the star’s light would cause them eitherto fall inwards towards the star, or be blownoutwards. The researchers therefore inferredthat the grains must be constantly replenishedby material spun off in collisions of large bod-ies in an asteroid belt. Notably, the impliedmass of HD 69830’s asteroid belt is roughly 25times that of our own, seemingly in line withthe beefed-up values of the planetary masses.

Intriguingly, these researchers also posited3

the existence of an unseen planet that quietlyshepherded the asteroid belt in its orbit. Itwould seem that Lovis et al. have found the(plural) shepherds. Working with the converselogic, they consider2 the gravitational influ-ence of their three planets on an asteroid belt,and find that its position must be constrainedto be either close to the star (0.3–0.5 AU) or farfrom it (beyond 0.8 AU). The earlier infraredobservations favour the former location,which places the belt between the orbits of thecentre and the outermost planet, but whetherthis location is truly stable and consistent withthe infrared observations remains an openquestion. In the Solar System, the asteroid beltlies near 2.6 AU, between the orbits of Marsand Jupiter. The difference in its location in theHD 69830 system is surely a clue to differencesin that star’s planet-formation history.

One of the great quests of astronomy is todiscover a small, rocky, Earth-like planetorbiting within the ‘habitable zone’ — therange of distances from a star for which theambient temperature would permit a planetwith liquid water and, perhaps, life as we knowit. For the Sun, which is comparatively largeand hot for our Galaxy, this orbital distance is great enough that the Doppler wobbleinduced by Earth would be a measly 9 cm s�1,nearly an order of magnitude below even theexquisite measurement precision establishedby Lovis and colleagues. But most neighbour-ing stars are significantly less massive andcooler than the Sun, so a search for habitableplanets using the Doppler method is feasiblefor two reasons. First, the lower stellar massmeans an Earth-mass planet will cause a larger

Figure 1 | Three’s company. How the planetary system of the Sun-like star HD 69830 might look,according to Lovis and colleagues2. Three Neptune-mass planets orbit the star on near-circular orbits ataround 0.08 AU, 0.19 AU and 0.63 AU (where 1 AU is the distance from Earth to the Sun). Considerationsof the gravitational influence of the three planets puts the most likely position for an asteroid belt, theexistence of which has been inferred by measurements of infrared radiation3, at between 0.3 and 0.5 AU.

MICROBIOLOGY

Antibiotic stops ‘ping-pong’ matchEric D. Brown

As bacteria become resistant to existing drugs, there is a need forantibiotics with new modes of action. Such a compound has been found,and it works by binding to an intermediate in the catalytic cycle of its target.

Pathogenic bacteria have developed strainsthat are resistant to almost all antibiotics in usetoday. Particularly worrisome are infections bya large group of bacteria classified as beingGram-positive, such as staphylococci and ente-rococci, which cause pneumonia and other,often fatal, infections. The problem is high-lighted by the emergence of multiply-drug-

resistant strains of these organisms — so-calledsuperbugs — that are resistant to vancomycin1,a drug widely recognized as the last line ofdefence in many Gram-positive bacterial infec-tions. On page 358 of this issue2, Wang and col-leagues report the discovery of a new antibiotic,platensimycin, that has potent antibacterialactivity against these Gram-positive pathogens.

18.5 N&V 291 MH 12/5/06 5:27 PM Page 293

Nature Publishing Group ©2006