DOI: 10.1126/science.1093990, 324 (2004);303 Science

Tom Owen-Hughes and Michael BrunoBreaking the Silence

This copy is for your personal, non-commercial use only.

clicking here.colleagues, clients, or customers by , you can order high-quality copies for yourIf you wish to distribute this article to others

  here.following the guidelines

can be obtained byPermission to republish or repurpose articles or portions of articles

  ): May 20, 2014 www.sciencemag.org (this information is current as of

The following resources related to this article are available online at

http://www.sciencemag.org/content/303/5656/324.full.htmlversion of this article at:

including high-resolution figures, can be found in the onlineUpdated information and services,

http://www.sciencemag.org/content/303/5656/324.full.html#relatedfound at:

can berelated to this article A list of selected additional articles on the Science Web sites

http://www.sciencemag.org/content/303/5656/324.full.html#ref-list-1, 5 of which can be accessed free:cites 11 articlesThis article

11 article(s) on the ISI Web of Sciencecited by This article has been

http://www.sciencemag.org/cgi/collection/molec_biolMolecular Biology

subject collections:This article appears in the following

registered trademark of AAAS. is aScience2004 by the American Association for the Advancement of Science; all rights reserved. The title

CopyrightAmerican Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005. (print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by theScience

on

May

20,

201

4w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

o

n M

ay 2

0, 2

014

ww

w.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om

on

May

20,

201

4w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

324

K) and T dwarfs (800 to 1200 K) are shiftedto the near-infrared (1 to 2 µm), which alsomakes brown dwarfs intrinsically faint.Moreover, their outer atmospheres areloaded with molecules such as water, carbonmonoxide, methane, and ammonia that ab-sorb light emitted from the interior, furtherdimming the flux of light. These considera-tions explain why finding brown dwarfs wassuch an observational challenge.

The dense, cool atmospheres of browndwarfs are an ideal environment for pro-ducing molecules, rather than themonatomic ions and neutral atoms soprominent in normal stars. Which gases arepresent at a given depth in brown dwarf at-mospheres depends on temperature, pres-sure, and overall elemental composition.Like Jupiter and the Sun, brown dwarfsmainly consist of hydrogen and helium, butmolecules made of less abundant elementsmake all the difference. Water and carbonmonoxide absorption shape the near-in-frared spectra of L dwarfs, whereas thenear-infrared spectra of T dwarfs are dom-inated by methane absorption (1, 5–8).Indeed, methane absorption was the proofof the brown dwarf status of Gl229 B.Methane is more abundant than CO onlybelow 1200 to 1500 K and at total pres-sures characteristic of brown dwarf atmos-pheres (11). Such low temperatures are on-ly reached at substellar masses.

Perhaps the most unusual aspect ofbrown dwarf atmospheres is the presence ofclouds. Refractory oxides, silicates, and liq-uid iron metal can form cloud layers even inthe hottest L dwarfs (11, 12). Cloud forma-tion on brown dwarfs is similar to that on gi-ant planets and did not come as a surprise toplanetary scientists, who have modeledcloud layer formation at different atmo-

spheric levels by gravitational settling of con-densates from overlying cooler atmosphericregions (13–16). However, what distinguish-es giant planets, T dwarfs, and L dwarfs is thenumber of different cloud layers.

Jupiter has top-level clouds of water,ammonia hydrogen sulfide (NH4HS), andammonia (see the figure). Deeper inside,there are alkali halide and sulfide cloudsfollowed by silicate and iron cloud layers.The deepest cloud layer on Jupiter is madeof refractory ceramics such as corundumand perovskite. In the hotter T and Ldwarfs, the cloud layer structure shown forJupiter is successively stripped at the top.In the hottest L dwarfs, only the layer withthe most refractory condensates may be left(see the figure). Thus, looking at hotterbrown dwarfs is like looking at deeper anddeeper regions of Jupiter’s atmosphere.

The atmosphere above a cloud is deplet-ed in the gases that contain the elementstrapped in the cloud. For example, molecu-lar absorption of TiO and VO is the trade-mark of the coolest M dwarfs but disap-pears in L dwarfs (1, 5–8), where both mol-ecules enter a perovskite cloud. Calciumand aluminum lines and CaH bands disap-pear near the transition from L to Mdwarfs, because clouds made of corundumand calcium aluminates take up Ca and Al.FeH absorption fades away in cooler Ldwarfs when metallic iron clouds removeiron gases from the observable atmosphere.Atomic alkali lines are prominent in all Ldwarfs and detectable in the hottest Tdwarfs, but not in cooler T dwarfs, wherethe alkali atoms form halide gases that con-dense into clouds.

Drastic compositional changes in theiratmospheres thus alter the observable spec-tral characteristics of brown dwarfs. In ad-

dition, clouds can block light emitted bythe brown dwarf, depending on how closethey are to the observable atmosphere. Theeffects of clouds on brown dwarf spectraare just beginning to be understood(17–20).

Despite the complexity of the browndwarf spectra, the underlying chemistrycan be identified and used as a “thermome-ter” to sort brown dwarfs (1, 16). No directobservations have yet been made of objectswith even lower mass and temperature thanT8 dwarfs (effective temperature ~800 K)to complete the bridge to Jupiter (~125 K).But with the Spitzer Space Telescope in or-bit ready to gather mid-infrared spectra,such ultralight brown dwarfs may finallycome into view.

References and Notes1. I. S. McLean et al., Astrophys. J. 596, 561 (2003).2. R. Jayawardhana, Science 303, 322 (2004).3. T. Nakajima et al., Nature 378, 463 (1995).4. B. R. Oppenheimer, S. R. Kulkarni, K. Matthews, T.

Nakajima, Science 270, 1478 (1995).5. J. D. Kirkpatrick et al., Astrophys. J. 519, 802 (1999).6. E. L. Martin et al., Astron. J. 118, 2466 (1999).7. A. J. Burgasser et al., Astrophys. J. 564, 421 (2002).8. T. R. Geballe et al., Astrophys. J. 564, 466 (2002).9. For lists of known L and T dwarfs, see

http://spider.ipac.caltech.edu/staff/davy/ARCHIVEand www.astro.ucla.edu/~adam/homepage/research/tdwarf.

10. Mass limits for hydrogen and deuterium burning referto objects of solar composition. Brown dwarf spectraand evolution are reviewed in (21).

11. B. Fegley, K. Lodders, Astrophys. J. 472, L37 (1996).12. T. Tsuji, K. Ohnaka, W. Aoki, Astron. Astrophys. 305, L1

(1996).13. J. S. Lewis, Icarus 10, 393 (1969).14. B. Fegley, K. Lodders, Icarus 110, 117 (1994).15. K. Lodders, Astrophys. J. 519, 793 (1999).16. K. Lodders, Astrophys. J. 577, 974 (2002).17. A. S.Ackerman, M. S. Marley, Astrophys. J. 556, 872 (2001).18. A. Burrows et al., Astrophys. J. 573, 394 (2002).19. M. S. Marley et al., Astrophys. J. 568, 335 (2002).20. T. Tsuji, Astrophys. J. 575, 264 (2002).21. A. Burrows, W. B. Hubbard, J. I. Lunine, J. Liebert, Rev.

Mod. Phys. 73, 719 (2001).

The genomes of eukaryotic cells arenot composed of free DNA but existas chromatin in which the DNA is as-

sociated with octamers of histone proteinscalled nucleosomes. The structure of chro-matin varies throughout the genome, pro-viding a means to regulate access of thetranscriptional machinery to the underly-ing genes. In an extreme case, chromatinadopts a condensed structure termed hete-

rochromatin in which genes are less acces-sible and frequently transcriptionallysilent. Uncondensed chromatin (euchro-matin) is much more accessible than hete-rochromatin and contains the majority ofactively expressed genes. To regulate geneexpression, cells adopt several differentstrategies to alter chromatin structure.These include the posttranslational modifi-cation of histones (for example, by acetyla-tion), the incorporation of variant histoneproteins into nucleosomes, and adenosinetriphosphate (ATP)–dependent chromatinremodeling by protein complexes related tothe yeast Swi2/Snf2 ATPase. Three recent

reports, including one by Mizuguchi et al.(1) on page 343 of this issue, indicate thatthese different strategies may be more inti-mately related than previously appreciated.

Mizuguchi and co-workers report theirpurification from yeast extracts of a pro-tein complex containing Swr1, a previous-ly uncharacterized Swi2/Snf2-relatedATPase (1). The authors identified 12 pro-teins in the Swr1 complex, some of whichare shared by other chromatin-associatedcomplexes. The same 12 proteins wereidentified in two complementary reportspublished elsewhere (2, 3). Histones in-cluding a significant proportion of Htz1,the yeast histone H2A.Z variant, werefound to associate with the Swr1 complex(1–3). The investigators then compared thegenome-wide transcription profiles ofswr1 and htz1 yeast mutants. They foundthat many genes that are either activated or

M O L E C U L A R B I O L O G Y

Breaking the SilenceTom Owen-Hughes and Michael Bruno

The authors are in the Division of Gene Regulation andExpression, School of Life Sciences, University of Dundee,Dundee DD1 5EH, UK. E-mail: t.a.owenhughes@dundee.ac.uk

P E R S P E C T I V E S

16 JANUARY 2004 VOL 303 SCIENCE www.sciencemag.org

Silenced chromatin Euchromatin

H2A.Z/H2B H2A/H2B

Swr1complex

H2A.Z/H2BBHnucleosomen

H2A/H2BHnucleosome

Sir2, Sir3Sir4

325

repressed by swr1 are similarly affected byhtz1, suggesting that the association ofSwr1 and Htz1 is functional as well asphysical (1–3). Remarkably, all threegroups discovered that Swr1 is needed forthe incorporation of the H2A.Z histonevariant into nucleosomes (replacing his-tone H2A) at approximately 25 chromoso-mal locations scattered throughout theyeast genome (1–3).

In budding yeast, genes activated by theincorporation of Htz1 into nucleosomesare found at the ends of chromosomes(telomeres), which are composed of chro-matin maintained in a transcriptionallysilent state through the action of Sir pro-teins. In htz1 mutant strains, silenced chro-matin spreads into adjacent regions of un-silenced chromatin (4). Thus, the incorpo-ration of Htz1 into nucleosomes mediatedby Swr1 may act as a buffer preventing thespread of chromatin silencing (see the fig-ure). A caveat to this is that in some casesthe H2A.Z histone variant appears to colo-calize with components of yeast hete-rochromatin (3). This raises the possibilitythat another factor may modulate the activ-ity of H2A.Z in order to prevent the spreadof chromatin silencing. An attractive op-tion would be that either H2A.Z or othernucleosome histones could be modified by,for example, the addition of acetyl groups.In the protozoan Tetrahymena, acetylationof the H2A.Z amino terminus is essentialfor survival (5). Intriguingly, the Swr1complex shares four subunits with theNuA4 histone acetyltransferase complex,which is also involved in the silencing oftelomeric heterochromatin (6, 7). Indeed, ahuman H2A.Z complex, equivalent to theyeast Swrl complex, has histone acetyl-transferase activity (8). In one of the new

studies, Krogan et al. (3) report that thebromodomain-containing protein Bdf1 isalso associated with certain Swr1 complex-es. This protein binds to acetylated histonesH3 and H4 and has anti-silencing proper-ties (9, 10). It is possible that Bdf1 may beinvolved in the recruitment of Swr1 com-plexes to chromatin.

How then might the Swr1 complex directthe incorporation of H2A.Z into nucleo-somes? Mizuguchi et al. show that the Swr1complex directly catalyzes the ATP-depend-ent replacement of histone H2A/H2Bdimers with H2A.Z/H2B dimers with highefficiency (1) (see the figure). Histone re-placement represents a dramatic new addi-tion to the arsenal of alterations to nucleo-some structure that are driven by ATP-de-pendent remodeling enzymes. Previously, ithad been assumed that ATP-dependent re-modeling enzymes only altered the interac-tion of DNA with the histone octamer ratherthan altering the structure or composition ofhistones within the nucleosomes. The abili-ty to direct the incorporation of histonedimers is not restricted to the Swr1 com-plex. The yeast RSC and SWI/SNF com-plexes also direct the exchange of histonedimers, albeit with reduced efficiency (1,11). However, unlike the Swr1 complex,these complexes show no preference for theH2A.Z histone variant. Remodeling en-zymes do not just direct the incorporation offree histone dimers into chromatin: TheATP-dependent exchange of histones be-tween chromatin fragments has also beenreported (11). These observations raise theintriguing possibility that some chromatinremodeling enzymes may have as yetunidentified specificities for directing theenrichment or removal of other histone vari-ants or histone modifications.

The importance of the Swr1 ATPase forhistone dimer exchange is unequivocal, butis not so clear-cut for the RSC andSWI/SNF complexes. Both of these com-plexes generate more accessible chromatin,and the generation of nucleosomes lackingone histone dimer may contribute to this.In support of this possibility, depletion ofH2A and H2B in vivo (12) or mutation ofthe dimer-tetramer interface of nucleo-somes (13) suppresses the requirement forthe SWI/SNF complex in transcriptionalactivation.

The lead that drew Krogan et al. (3) in-to the Swr1-H2A.Z story was their detec-tion of interactions between both Swr1and Htz1 and the genes involved in theelongation of transcripts during transcrip-tion. This suggests that both H2A.Z andSwr1 are involved in transcript elongationin a way that we do not yet fully under-stand. Another complex, FACT, has re-cently been reported to assist in the shut-tling of histone dimers around the RNApolymerase during transcription, in anATP-independent process (14). There area number of cases where SWI/SNF-relat-ed remodeling enzymes have been linkedto transcription elongation; in some ofthese cases it is tempting to speculate thatdestabilization of histone dimers may beinvolved.

The three new studies have broken freshground on several fronts. The idea that thehistone octamer remains unchanged duringthe course of chromatin remodeling hasbeen shown not to hold true in all cases.The new work establishes the importanceof ATP-dependent chromatin remodelingfor directing the incorporation of varianthistone proteins into nucleosomes and pro-vides fresh insights into how the spread ofheterochromatin is halted.

References and Notes1. G. Mizuguchi et al., Science 303, 343 (2004); pub-

lished online 26 November 2003 (10.1126/science.1090701).

2. M. S. Kobor et al., PLoS Biol., in press.3. N. J. Krogan et al., Mol. Cell 12, 1565 (2003).4. M. D. Meneghini, M. Wu, H. D. Madhani, Cell 112, 725

(2003).5. Q. Ren, M. A. Gorovsky, Mol. Cell 7, 1329 (2001).6. A. A. Boudreault et al., Genes Dev. 17, 1415 (2003).7. J. Côté, personal communication.8. P. Nakatani, personal communication.9. O. Matangkasombut, S. Buratowski, Mol. Cell 11, 353

(2003).10. A. G. Ladurner, C. Inouye, R. Jain, R. Tjian, Mol. Cell 11,

365 (2003).11. M. Bruno et al., Mol. Cell 12, 1599 (2003).12. J. N. Hirschhorn, S. A. Brown, C. D. Clark, F. Winston,

Genes Dev. 6, 2288 (1992).13. M. S. Santisteban, G. Arents, E. N. Moudrianakis, M. M.

Smith, EMBO J. 16, 2493 (1997).14. R. Belotserkovskaya et al., Science 301, 1090 (2003).15. We thank J. Côté, A. Flaus, and C. Stockdale for help-

ful discussions, and C. Wu, J. Rine, J. Côté, J.Greenblatt, and P. Nakatani for communicating re-sults before publication. Supported by a WellcomeTrust Senior Fellowship (TOH).C

RED

IT:K

AT

HA

RIN

E S

UT

LIFF

/SC

IEN

CE

Octamer interrupted. In yeast, the transcriptionally silenced chromatin of telomeres is formed by

the association of the Sir2, Sir3, and Sir4 proteins (dark blue) with nucleosomes. Replacement of

H2A/H2B histone dimers (yellow) with H2A.Z/H2B dimers (red) containing the variant histone pro-

tein H2A.Z prevents the spread of Sir proteins into adjacent regions of unsilenced chromatin. This

histone replacement acts as a buffer halting the spread of chromatin silencing. The Swr1 complex

mediates the removal of H2A/H2B dimers from nucleosomes and their replacement with

H2A.Z/H2B dimers.

P E R S P E C T I V E S

www.sciencemag.org SCIENCE VOL 303 16 JANUARY 2004