seen in the classic fungus-farming symbioses2 and in the more recently discovered incipi-ent practices of fungus farming by Littoraria snails9 and red-alga farming by Stegastes damselfish10.

The damselfish mutualism has led to at least one case of monoculture in which the crops no longer have free-living relatives. The fact that slime mould husbandry has not achieved this status underlines the point that parent–offspring transmission is insufficient to install absolute co-dependency in a mutualism11. Some form of monoculture farming is appar-ently essential to make symbionts put all their eggs in a host’s basket, because being eaten is profitable only when it benefits clone mates that are nursed and dispersed. Monocultural commitment makes these kin-selected ben-efits consistent, paving the way for mutual coadaptation, irrespective of symbionts being acquired from the environment rather than being inherited12. The limitations of bacterial husbandry in slime moulds3 therefore clarify a major cornerstone of our general understand-ing of mutualistic interactions. They also invite further study to unravel the molecular mech-anisms that allow or prevent bacterial trans-mission, and to establish the dynamics of food transmission in slugs that are genetic mixtures of several strains6.

The Dictyostelium symbiosis presents an interesting analogy to culturally adjustable human subsistence farming in its various contemporary and historical combinations with hunter–gatherer strategies1. In both slime moulds and humans, farmers did not become reproductively isolated from non-farmers, nor did crops or livestock lose the possibil-ity of hybridizing with wild relatives, as has happened in the specialized insect fungus-farming symbioses2,12. The slime moulds may have insufficient multicellular complexity to evolve specialized nurturing traits for par-ticular crops, whereas our own species lacked evolutionary time and consistent selection for extreme crop specialization. Neither of these constraints applied to the farming societies of ants and termites.

Although Dictyostelium do not actively rear their crops, they may well possess unknown adaptations that, if revealed, would illuminate fundamental questions of conflict and cooper-ation across species boundaries. The ancestors of these slime moulds were among the earliest colonizers of terrestrial habitats, so the history of this bacterial husbandry symbiosis may go back further than any other farming system. Mapping farming practices on a large-scale evolutionary tree of the slime moulds would therefore be a worthy objective. If this farm-ing symbiosis turns out to be ancient, our new understanding of Dictyostelium biology could be summed up in a lyric of the rock band Metallica. To paraphrase: wherever they roamed, they redefined the unknown, by themselves but not alone. ■

Jacobus J. (Koos) Boomsma is in the Centre for Social Evolution, Department of Biology, University of Copenhagen, 2100 Copenhagen, Denmark. e-mail: jjboomsma@bio.ku.dk

1. Diamond, J. Nature 418, 700–707 (2002).2. Mueller, U. G. et al. Annu. Rev. Ecol. Evol. Syst. 36,

563–595 (2005). 3. Brock, D. A., Douglas, t. e., Queller, D. C. &

Strassmann, J. e. Nature 469, 393–396 (2011).4. Bloomfield, G., Skelton, J., ivens, A., tanaka, y. &

Kay, R. R. Science 330, 1533–1536 (2010).5. Kessin, R. H. Dictyostelium: Evolution, Cell Biology,

and the Development of Multicellularity (Cambridge Univ. Press, 2001).

6. Gilbert, o. M. et al. Proc. Natl Acad. Sci. USA 104, 8913–8917 (2007).

7. Maynard Smith, J. Proc. R. Soc. Lond. B 205, 475–488 (1979).

8. Baldauf, S. l. et al. Science 290, 972–977 (2000).

9. Silliman, B. R. & newell, S. y. Proc. Natl Acad. Sci. USA 100, 15643–15648 (2003).

10. Hata, H., Watanabe, K. & Kato, M. BMC Evol. Biol. 10, 185 (2010).

11. Herre, e. A., Knowlton, n., Mueller, U. G. & Rehner, S. A. Trends Ecol. Evol. 14, 49–53 (1999).

12. Aanen, D. K. et al. Science 326, 1103–1106 (2009).

i m m u N o l o G Y

Peptide gets in shape for self-defence The transformation of tadpole to frog and of caterpillar to butterfly are two of the more obvious examples of metamorphosis. But molecular shape-shifting may occur in each of us as part of our innate antibacterial defence system. See Letter p.419

R o B e R T i . l e h R e R

Among the immune mediators that fight microorganisms within us, one is human β-defensin 1 (hBD-1). This

peptide, which was first described1 in 1995, is continually expressed in skin and epithelial cells throughout the body2. But because its direct antibacterial properties are modest, the reason it stands guard at interfaces between microbe-laden environments — such as the colon or skin — and their adjacent, normally sterile tissues has remained enigmatic. On page 419 of this issue, Schroeder et al.3 show that the mild antibacterial activity of hBD-1 changes drastically after it undergoes a chemi-cally induced change in shape.

The peptide backbone of hBD-1 is folded into a well-defined structure that is held together by three internal disulphide bonds4. Schroeder et al. find that an enzyme called thioredoxin reductase can sever these disul-phide bonds in a reduction reaction. The reduced hBD-1 molecule undergoes a pro-found change in shape (Fig. 1, overleaf) that allows it to kill some Gram-positive bacteria, against which its normally oxidized form is powerless. The authors’ electron micrographs of the slain bacteria show changes that might lead a forensically inclined microbiologist to wonder whether reduced hBD-1 induced the bacteria to self-destruct, by triggering their latent autolysis (self-breakdown) systems5.

Schroeder and co-workers’ observations3 are unlikely to be merely a ‘test-tube’ phenom-enon. They demonstrate that in human epithe-lia, oxidized hBD-1 and thioredoxin reductase colocalize with reduced hBD-1.

If shape change alone imparted the expanded antimicrobial range of hBD-1, then analogues of this peptide in which cysteine residues are replaced by other amino acids should also show enhanced function, because the cysteine-free peptides would be unable to form shape-restraining disulphide bonds. But Schroeder et al. report that such analogues are ineffective antibiotics. Evidently, there is something spe-cial about the cysteine residues of hBD-1, but exactly what remains unknown.

There is also something special about having a positive charge. Full-length hBD-1 has 36 amino-acid residues, including six cysteines, one negatively charged aspartic acid and five positively charged residues (four lysines and one arginine). If the partial charge of its single histidine residue is ignored, hBD-1 has a net charge of +4. This charge is concentrated in its carboxy-terminal octapeptide — arginine- glycine-lysine-alanine-lysine-cysteine-cysteine-lysine. A truncated variant of hBD-1 lacking the last seven of these residues shows no activity, but a seven-residue peptide lacking only the initial arginine of the octapeptide does.

It is ironic that the bacteria that Schroeder et al. find to be susceptible to reduced hBD-1 belong to either the lactobacilli or bifido-bacteria genera — organisms that are generally considered6 to be health-promoting pro biotics rather than potential pathogens. This observa-tion, however, should be considered a proof of concept, rather than a serious colonic conundrum.

At least 1,000 species of bacteria reside in the colon of a healthy adult, and a single gram of faeces may contain up to 1012 bacteria. Unlike bifidobacteria, most bacterial species residing

2 0 J A N U A R y 2 0 1 1 | V O L 4 6 9 | N A T U R E | 3 0 9

NEWS & VIEWS RESEaRch

© 2011 Macmillan Publishers Limited. All rights reserved

C h e m i C a l B i o l o G Y

Catalytic detoxificationProtein engineering of an enzyme that catalytically detoxifies organophosphate compounds in the body opens up fresh opportunities in the search for therapeutic protection against nerve agents used in chemical warfare.

F R a N K m . R a u s h e l

Organophosphates are among the most toxic compounds that have been chem-ically synthesized. Since the discovery

of their biological activity in the 1930s, these compounds have found use as broad-spectrum insecticides for agricultural and domestic appli-cations. But organophosphates have also been developed as chemical-warfare agents, includ-ing VX and the ‘G-agents’ (such as sarin, soman and cyclosarin). Because these compounds are relatively easy to synthesize, their use by inter-national terrorist groups is a serious threat. Current protocols for the prevention and treat-ment of organophosphate poisoning are largely ineffective, and so new strategies are desper-ately needed. Reporting in Nature Chemical Biology, Gupta et al.1 describe an approach that

might one day find use in preventing organo-phosphate poisoning.

Organophosphates are highly toxic because they rapidly inactivate acetylcholinesterase (AChE), an enzyme required for nerve func-tion (Fig. 1). AChE breaks down (hydrolyses) acetylcholine, a neurotransmitter that relays nerve impulses to muscles and other organs. Organophosphates form a covalent bond to a serine amino-acid residue in the active site of AChE, stopping the enzyme from function-ing. The subsequent build-up of acetylcholine blocks cholinergic nerve impulses, leading to paralysis, suffocation and death.

Various prophylactic approaches have been developed to diminish the toxic effect of organophosphates. Atropine, for example, is a competitive antagonist for muscarinic acetylcholine receptors — it blocks the action

in the colon cannot be grown in culture, and their presence can only be disclosed using various ‘gene-sniffing’ techniques. It could be, therefore, that the effects of reduced hBD-1 on probiotic bacteria are simply collateral damage on these harmless bystanders by a defence system that also targets less-well-intentioned intestinal residents or transients. Alternatively, even probiotics may require surveillance to keep them from overstepping their boundaries.

Ideally, the activity of an antibiotic should be examined in a defined medium, the composi-tion of which closely resembles, or precisely replicates, the in vivo environment. With pre-cise simulation of the colonic content being too challenging to contemplate, it would be informative to learn how defined factors — such as pH and, in particular, salinity — affect the activity of reduced hBD-1 in vitro. Another interesting experiment would be to test the antibacterial activity of mixtures of reduced and oxidized hBD-1, because clearly such mixtures occur in vivo.

It remains unknown whether Schroeder and colleagues’ results are unique to hBD-1 or whether they are also true for other defensin peptides. Defensins and defensin-like peptides are fairly universal participants in host defence against infection7 : they occur in plants, fungi, invertebrates and vertebrates. Vertebrates have three subfamilies of defensins (designated α, β and θ)8, the members of which consist exclu-sively of cationic peptides with six cysteines and three disulphide bonds, which also provide resistance to premature proteolytic digestion.

Although more than 20 genes have been identified9 that encode hBDs, only hBDs 1–4 have received extensive attention. The net

positive charge of these four peptides varies from +4 for hBD-1 to an astounding +11 for hBD-3, whose eight carboxy-terminal residues alone carry a net charge of +6. From previous work10 on hBD-3, its high net positive charge contributes substantially to the peptide’s ability

Figure 1 | Computer-generated structures of human β-defensin 1. a, In its oxidized form, the peptide human β-defensin 1 (hBD-1) contains three disulphide bonds. The peptide backbone is shown in green except for cysteine residues (yellow) and the six non-cysteine carboxy-terminal residues (pink). The broad arrows represent β-sheet components found only in the oxidized form. b, The reduced hBD-1 structure was generated from the lowest-energy conformation of oxidized hBD-1, after breaking its disulphide bonds and causing it to assume a random conformation. (Structure generated by Alan J. Waring, Univ. California, Los Angeles.)

a Oxidized hBD-1 b Reduced hBD-1

to kill bacteria or fungi such as Candida albi-cans. This is especially true when the assays are performed in media of low ionic strength.

Given the extreme cationicity and high intrinsic activity of oxidized hBD-3, it is not surprising that when Schroeder et al.3 removed its disulphide bonds, they did not detect improved activity of this peptide against bifidobacteria. After all, a bacterium can only be killed once. For thioredoxin reductase to empower hBD-3 to do so twice would be a reductio ad absurdum. ■

Robert I. Lehrer is in the Department of Medicine, David Geffen School of Medicine, University of California, Los Angeles, Los Angeles, California 90045, USA. e-mail: rlehrer@mednet.ucla.edu

1. Bensch, K. W., Raida, M., Mägert, H.-J., Schulz-Knappe, P. & Forssmann, W.-G. FEBS Lett. 368, 331–335 (1995).

2. Zhao, C., Wang, i. & lehrer, R. i. FEBS Lett. 396, 319–322 (1996).

3. Schroeder, B. o. et al. Nature 469, 419–423 (2011).4. Schibli, D. J. et al. J. Biol. Chem. 277, 8279–8289

(2002).5. Sahl, H.-G. et al. J. Leukoc. Biol. 77, 466–475 (2005).6. Kleerebezem, M. & Vaughan, e. e. Annu. Rev.

Microbiol. 63, 269–290 (2009).7. Wong, J. H., Xia, l. & ng, t. B. Curr. Protein Pept. Sci.

8, 446–459 (2007).8. Selsted, M. e. & ouellette, A. J. Nature Immunol. 6,

551–557 (2005).9. Schutte, B. C. et al. Proc. Natl Acad. Sci. USA 99,

2129–2133 (2002).10. Hoover, D. M. et al. Antimicrob. Agents Chemother.

47, 2804–2809 (2003).

3 1 0 | N A T U R E | V O L 4 6 9 | 2 0 J A N U A R y 2 0 1 1

NEWS & VIEWSRESEaRch

© 2011 Macmillan Publishers Limited. All rights reserved