tionary position. In general, researchers havea good grasp of how to identify those por-tions of a genome that are translated intoproteins,by aligning sequences of messengerRNAs, the precursors of proteins, againstgenomic sequences of interest. One can alsoidentify these ‘coding’ genomic sequences bycomparing the DNA of organisms that areevolutionarily distant. For example, stretch-es of sequence that have been preserved inhumans and fruitflies are likely to be veryimportant for the functioning of the organ-isms. These sequence stretches are calledconserved elements.

However, now that the human genomesequence is essentially finished4, researcherswould like to do more than just identify thesequences that are translated into proteins.They also want to understand all of the regu-latory structures present in a genome —structures that might, for instance,adjust theamount of protein manufactured from aparticular gene. These structures are collec-tively known as functional elements, and thechicken,having diverged from humans morethan 310 million years ago, is considered thebest example so far of an ‘outgroup’ withwhich to identify them. Because enough dif-ferences between the human and chickensequences have accumulated over this period,one can zero in on the precise base pairs thatevolution has left alone for all these years —the base pairs most likely to be functional inthe human genome. By comparison, themouse, which split from humans only 75million years ago, is too similar at the base-pair level, leading to difficulties in identify-ing functional elements5.

The consortium’s initial analysis1

describes 70 million base pairs of sequencethat are highly conserved between chickensand humans. This includes base pairs withingenes, but also base pairs that are betweengenes and therefore relate to potential func-tional elements (interestingly, many of theseseem to be at a considerable distance fromgenes). Questions surrounding what thesestructures are and why evolution has con-strained them over time will only beanswered with targeted experiments, someof which are beginning to get under way6.

Finally, for those who concentrate ongenerating large-scale genomic sequencesand resources, the chicken genome repre-sents another in a series of grand experi-ments to balance two different approaches.Traditional clone-by-clone approaches (see,for example, refs 4, 7) — which involvecloning a genome into bacterial artificialchromosomes (BACs), mapping the clones,then sequencing them and assembling thesequences by using the map — are time-consuming but generally produce an accu-rate representation of all regions of thegenome. Whole-genome shotgun8 (WGS) isquicker, because it involves shattering thewhole genome into pieces, sequencing the

news and views

680 NATURE | VOL 432 | 9 DECEMBER 2004 | www.nature.com/nature

fragments and assembling them by com-puter,but it often fails to represent all regions accurately.

The chicken sequence presented here is ahalfway house: it is not a straight WGSassembly, but has been revised according to a physical map of 180,000 BAC clones,detailed by Wallis et al.3 on page 761. Thismap was crucial in ordering and localizingthe sequence pieces generated by WGS. Thusthe assembly captures an impressive 98% ofthe sequence over most of the genome, withthat number falling slightly in very GC-richregions. The authors were also able to locatepartial or complete sequences of at least 97% of coding genes that were previouslyknown to exist.

However, the genome has received nodirected ‘finishing’ work, and issues do stillexist — there is a distinct lack of continuity in10% of the gene-rich regions, and there areperhaps 1.4 million base pairs of sequencethat are in the wrong position.Recent studies9

suggest that, even with algorithmic improve-ments, WGS assemblies fail to resolve large-scale duplications in vertebrate genomes;even with a BAC map, recently duplicated

sequences in the chicken assembly are poorlyresolved1. And the authors suggest that onereason why they were able to resolve most ofthe WGS sequence was the minimal repetitivecontent of the chicken genome, so the experi-ence will not necessarily translate to all ver-tebrate genomes. As we move forward in thispost-genomic era, we must learn from all past experience, so that we can maintain thehigh quality we have come to expect fromgenome-sequencing projects. ■

Jeremy Schmutz and Jane Grimwood are at theStanford Human Genome Center, 975 CaliforniaAvenue, Palo Alto, California 94304, USA.e-mails: jeremy@shgc.stanford.edujane@shgc.stanford.edu1. International Chicken Genome Sequencing Consortium

Nature 432, 695–716 (2004).

2. International Chicken Polymorphism Map Consortium

Nature 432, 717–722 (2004).

3. Wallis, J. W. et al. Nature 432, 761–764 (2004).

4. International Human Genome Sequencing Consortium

Nature 431, 931–957 (2004).

5. Mouse Genome Sequencing Consortium Nature 420,

520–562 (2002).

6. ENCODE Project Consortium Science 306, 636–640 (2004).

7. International Human Genome Sequencing Consortium

Nature 409, 860–921 (2001).

8. Venter, J. C. et al. Science 291, 1304–1351 (2001).

9. She, X. et al. Nature 431, 927–930 (2004).

Microbiology

Jekyll or hide?George A. O’Toole

Many bacteria can adopt different lifestyles: in a free-living state, theyare virulent and cause disease; in a surface-attached community, theyare less virulent but may go unnoticed. How is this ‘decision’ made?

In the November issue of DevelopmentalCell, Goodman and colleagues1 report theidentification of a regulatory system in

the bacterium Pseudomonas aeruginosa thatdetermines whether it causes disease or lieslow and simply persists. This bacterium isof interest to the medical communitybecause of its ability to infect people whoseimmune system is damaged, who have sustained serious burns or an eye injury, orwho suffer from cystic fibrosis. Goodman et al. found that inactivating a so-calledtwo-component regulatory system in P. aeruginosa results in a strain with a markedly decreased ability to cause disease,but an increased ability to form surface-attached, persistent communities known asbiofilms (Fig. 1).

Although there are many variations onbacterial two-component regulatory sys-tems, their basic job is to constantly samplethe external environment and transmit thisinformation to the bacterial interior. Thisallows the organism to adapt to an ever-changing environment. Goodman et al.1

discovered a new protein component of anew such regulatory system, a componentthat they call RetS.

They also found that a RetS-deficientP. aeruginosa strain was better than a wild-type strain at forming a biofilm on both anabiotic surface, namely glass, and a bioticsurface, cultured hamster cells. The RetS-deficient bacteria were, however, less able todamage the hamster cells they colonized,andto cause disease in a mouse model of pneu-monia. Outside the lab, the ability of P. aeru-ginosa to form biofilms is best known withrespect to abiotic surfaces such as catheters,but it might also be able to produce biofilmson tissues within a host, in diseases such asotitis media (earache) and cystic fibrosis2. Itseems, then, that Goodman et al. might haveidentified a control element that allows thisbacterium to switch between a virulent,disease-causing state and a biofilm state in a mammalian host. The biofilm state,although less virulent, might allow themicrobe to persist for longer.

To understand better how the proteinmight control this pathogenesis/persistenceswitch, the investigators used DNA micro-arrays to identify all the genes in the organismthat are regulated by RetS. They found that,in the RetS-deficient strain, the expression ofgenes required to make a ‘type III secretion

9.12 n&v 679 MH 3/12/04 5:36 pm Page 680

© 2004 Nature Publishing Group

system’ — necessary for P. aeruginosa tocause a short-term infection — was reducedby as much as 25 times. The expression ofother genes associated with virulence wasalso reduced; these genes include those thatproduce surface appendages called pili, aswell as those that encode toxins such as LipAand ToxA.

These data imply that RetS is normallyrequired for the full expression of the factorsrequired to produce an acute infection, andare consistent with the decreased ability ofRetS-deficient P. aeruginosa to cause disease.In contrast, P. aeruginosa genes that areinvolved in the formation of the sugar-richmatrix that encloses a biofilm — the psl andpel genes3–6 — were markedly upregulated inthe mutant bacteria. This suggests that RetSusually turns off the genes needed to make abiofilm.

We make choices every day on the basis ofthe information at hand.Bacteria must do sotoo, and the outcomes of their decisions canhave life-or-death consequences, both to thebacteria and to the host. For bacteria, thesechoices — such as whether or not to form apersistent biofilm — are based in large parton local environmental cues,and are effectedthrough altered gene expression. Forinstance, P. aeruginosa decides to form bio-films on abiotic surfaces, such as catheters orcontact lenses, only when an energy source(such as sugars) and other nutrients (such asiron) are readily available7,8. Otherwise, itremains free-living. Goodman and col-leagues’ findings suggest that P. aeruginosaalso has a decision to make when in the con-text of a mammalian host: does it cause ashort-term infection or does it persist in abiofilm state? An acute infection provides ameans of bacterial propagation, whereas in abiofilm the organism is lying low and is thusless likely to be recognized and attacked bythe immune system.

The very existence of a regulatory systemthat mediates this decision suggests that thechoice is a crucial one for microbes, and onethat they must constantly re-evaluate. Stud-ies of Bordetella bronchiseptica — an organ-ism related to the microbe that causeswhooping cough — provided one of the firstmolecular illustrations of this decision9.Thus, B. bronchiseptica forms biofilms bestwhen the genes required for acute infectionare turned off. However, expression of atoxin required for acute infection can blockbiofilm formation, hinting that the func-tions required to cause disease and thoserequired to make a biofilm might actually be incompatible. Similarly, my own grouphas found that expression of the type IIIsecretion system, required for acute infec-tion in P. aeruginosa, also inhibits biofilmformation10.

We would do well to continue learningabout how bacteria can switch from diseaseto persistence and back again. A better

understanding of this decision could lead tonew strategies for dealing with bacterialinfections. ■

George A. O’Toole is in the Department ofMicrobiology and Immunology, Dartmouth MedicalSchool, Hanover, New Hampshire 03755, USA.e-mail: georgeo@dartmouth.edu

1. Goodman, A. L. et al. Dev. Cell 7, 745–754 (2004).

2. Post, J. C., Stoodley, P. & Hall-Stoodley, L. Curr. Opin.

Otolaryngol. Head Neck Surg. 12, 3185–3190 (2004).

news and views

NATURE | VOL 432 | 9 DECEMBER 2004 | www.nature.com/nature 681

Figure 1 Persistence versus infection. Goodman et al.1 have discovered a regulatory system in thebacterium Pseudomonas aeruginosa that might enable it to choose between two lifestyles in amammalian host: growing as a surface-attached, persistent community (a biofilm), or causing ashort-term infection. a, An electron micrograph of a P. aeruginosa biofilm on a suture. Individualcells can be seen, surrounded by a sugar-rich matrix. (Courtesy Jon Budzik and the Ripple EM facility,Dartmouth College, New Hampshire.) b, An inflamed eye — the effect of an acute P. aeruginosainfection of the cornea. (Courtesy Patrick Saine and Michael E. Zegans, Dartmouth Medical School,New Hampshire.)

a b

Planetary science

Volcanoes on Quaoar?David J. Stevenson

Quaoar, a large body in the Kuiper belt, has crystalline water ice on itssurface, yet conditions there should favour amorphous ice. Does thismean that resurfacing has taken place — perhaps even volcanism?

Our planetary system does not end atPluto. Hundreds of bodies exist inthe Kuiper belt, which extends out-

wards from Pluto, sharing the same plane asthe planetary orbits. The largest known bod-ies in the Kuiper belt are not much smallerthan Pluto, and some have similar dynamicsto that planet. Although the existence of theKuiper belt had long been hypothesized, thefirst Kuiper-belt body was discovered only in1992,by Jewitt and Luu1.These same authorsnow propose2 that Quaoar, the largestknown of these bodies, has crystalline waterice on its surface and possibly also ammonia(see page 731 of this issue). The presence ofcrystalline ice is surprising, because it iswidely believed that its formation requires atemperature of around 100 K or more —substantially higher than the surface tem-perature of these bodies.The precise temper-ature required, however, is not known and may not be the same in laboratory

experiments as it is in space. Yet it might be that we are seeing evidence for ‘planetary’processes such as volcanism within thesebodies.

The discovery and characterization of theKuiper belt is among the most importantdevelopments in planetary science in thepast decade3. As is usual with the discovery of new bodies (inside or outside the SolarSystem), the initial excitement focused onthe dynamical implications: why do theyoccupy these orbits and how did they form?Some orbital migration may occur, but it islikely that these bodies never experiencedmuch higher surface temperatures than theambient conditions provided by the Sun(temperatures of about 50 K or less).Quaoar,discovered in 2002 by Trujillo and Brown, isthe largest body to be found in our systemsince the discovery of Pluto in 1930. It has aradius of about 650 km, roughly half that of Pluto. The composition and nature of the

3. Friedman, L. & Kolter, R. J. Bacteriol. 186, 4457–4465 (2004).

4. Friedman, L. & Kolter, R. Mol. Microbiol. 51, 675–690 (2004).

5. Jackson, K. D., Starkey, M., Kremer, S., Parsek, M. R. & Wozniak,

D. J. J. Bacteriol. 186, 4466–4475 (2004).

6. Matsukawa, M. & Greenberg, E. P. J. Bacteriol. 186, 4449–4456

(2004).

7. Singh, P. K., Parsek, M. R., Greenberg, E. P. & Welsh, M. J.

Nature 417, 552–555 (2002).

8. O’Toole, G. A. & Kolter, R. Mol. Microbiol. 30, 295–304 (1998).

9. Irie, Y., Mattoo, S. & Yuk, M. H. J. Bacteriol. 186, 5692–5698 (2004).

10.Kuchma, S. L., Connolly, J. P. & O’Toole, G. A. J. Bacteriol.

(in the press).

9.12 n&v 679 MH 3/12/04 5:36 pm Page 681

© 2004 Nature Publishing Group