meeting report

NATURE CELL BIOLOGY VOL 5 MAY 2003 www.naturecellbiology.com 395

Chromatin and transcription: histones continue to make their marks

Mariela Jaskelioff and Craig L. Peterson

Chromatin architecture is modulated by a large number of enzymes, resulting in the proper regulation of transcrip-tion, replication, cell cycle progression, DNA repair, recombination and chromosome segregation. The structure,regulation and coordination of these enzymatic activities were the main topics of discussion at The Enzymology ofChromatin and Transcription Keystone Symposium held in Santa Fe, NM (March 10–16, 2003).

Chromatin fibres are the natural sub-strates for all DNA-mediated process-es, and thus it is not too surprising

that a variety of evolutionarily conservedenzymes modulate the architecture of chro-matin, allowing the precise temporal andspatial regulation of genetic processes.These enzymes, the focal point of thisKeystone Symposium organized by ShellyL. Berger (Wistar Institute, Philadelphia,PA) and Jerry L. Workman (PennsylvaniaState University, State College, PA), can bedivided in two major classes: first, thosethat introduce covalent modifications onhistone amino- or carboxy-terminal ‘tail’domains; second, those that use the freeenergy derived from ATP hydrolysis toactively disrupt nucleosomal structure.

With the exception of one plenary ses-sion devoted to the mechanism of action ofthe ATP-dependent remodelling enzymes,many of the presentations centred on thecoordinated action of histone-modifyingenzymes during the transcription processand the various roles that the subsequenthistone ‘marks’ play in establishing on or offstates of gene expression.

Histone methylation comes to centre stageIn his keynote address, C. David Allis(University of Virginia, Charlottesville, VA)proposed that histones are major carriers ofepigenetic information and that covalentmodifications on the histone N-terminaltails function as master on/off switchesthat determine whether a gene is active orinactive. Histone tails are subject to a widerange of post-translational modifications(Fig. 1). Although in past years, histoneacetylation on lysine residues has receivedthe lion’s share of attention, remarkably,histone acetylation was not a major topicof discussion at this meeting. In its place,histone methylation permeated nearlyevery session. Methylation can have mul-tiple effects on chromatin function,depending on the specific lysine and thelevel of modification (that is, mono-, di-, ortri-methylation of a single lysine). For

instance, H3-K9 di-methylation and H3-K27 tri-methylation are both largely associ-ated with gene silencing and heterochro-matin formation1, whereas methylation ofH3-K4, H3-K36, or H3-K79 is associatedwith active chromatin. Tony Kouzarides(University of Cambridge, Cambridge, UK)also presented evidence that the number ofmethyl groups on a single lysine is associat-ed with distinct chromatin states. Whereasseveral inactive genes are marked by H3-K4di-methylation, the transition to transcrip-tional competence correlated with H3-K4tri-methylation2. Thus, it is clear that thesite specificity of lysine methylation, as wellas the number of methyl groups attached toa particular lysine, can have distinct conse-quences for transcription.

With few exceptions, histone methyla-tion is catalysed by proteins that contain aconserved SET domain flanked by cysteine-rich pre-SET and post-SET domains. SteveGamblin (Institute for Medical Research,London, UK), Raymond Trievel (NIH,Bethesda, MD) and Xing Zhang (EmoryUniversity, Atlanta, GA) each presentedcrystallographic structures of the catalyticdomains from three different methyltrans-ferases — SET7/9, LSRT and DIM-5,respectively — in complex with their sub-strate. A hallmark of each structure was anarrow, doughnut-like hole at the catalyticcore where methyl transfer occurs. Thepeptide substrate binds on one side of thechannel and the methyl donor occupies theopposite face. Once the lysine is docked into

H2A - S G R G K Q G G K A R A...A V L L P K K T E S H H K 1 5 119

NH3

P AC Ub

H2B - P E P V K S A P V P K K G S K K A I N K ...V K Y T S S K 5 12 15 120 (123 in yeast)

NH3

AC AC AC Ub

H4 - S G R G K G G K G L G K G G A K R H R K V L R D N I Q G I T 1 3 2016 12 8 5

NH3

P Me MeAC AC AC AC

Me MeMe Me

H3 - A R T K Q T A R K S T G G K A P R K Q L A S K A A R K S A ... G V K K... E F K T D 2 3 4 9 10 14 1718 23 26 27 28 36 79

NH3

P P PMe

MeMe MeAC AC ACACAC

Me

AC

Figure 1 Potential sites of post-translational modification on nucleosomal histones. Manymodification patterns have been closely linked to unique biological outcomes. Ac,acetylation; Me, methylation; P, phosporylation; Ub, ubiquitination.

© 2003 Nature Publishing Group

meeting report

NATURE CELL BIOLOGY VOL 5 MAY 2003 www.naturecellbiology.com396

the active site, one hydrogen from the ε-NH

3protrudes through the hole to the

opposite face where the transfer reactiontakes place. In the case of a mono-methy-lase, SET7/9, the target ε-NH

3group is

rigidly bound in the active site such thatonly a single methyl-transfer event canoccur3. In contrast, the ε-NH

3group has

greater freedom of rotation in the active siteof di- or tri-methyltransferases (LSMT,DIM-5), which allows for additionalmethylation events4,5. Remarkably, XingZhang (Emory University) demonstratedthat a tri-methylase or a mono-methylasecould be converted to a di-methylase simplyby changing one of the amino acids thatcontrols the hydrogen-bonding environ-ment of the ε-NH

3group.

Combinatorial histone modifications: regulating the marksTo date, there are still no reports of a bonafide ‘demethylase’, and thus it seems thathistone methylation is a relatively stablechromatin mark that can only be lost bysuccessive rounds of DNA replication or byreplication-independent histone replace-ment. David Allis suggested that cells mayuse reversible serine/threonine phosphory-lation to regulate the biological outcomes oflysine methylation1. His ‘methyl/phosphoswitch’ hypothesis argues that a phosphory-lation event on a histone tail would regulatethe binding of an effector protein to anadjacent methylated residue. In support ofthis model, he presented numerous exam-ples in which a serine or threonine is foundadjacent to a methylation site (lysine)throughout the sequence of histone pro-teins. As a proof of principle, David Allispresented evidence that phosphorylation ofSer 10 of a histone H3 peptide blocks theability of heterochromatin protein 1 (HP1)to bind its cognate substrate, di-methylatedH3-K9.

Whereas the control of histone methyla-tion by adjacent phosphorylation is a newlyemerging paradigm, other types of combi-natorial histone modifications are betterestablished. For instance, several presenta-tions, including those from Mary AnnOsley (University of New Mexico,Albuquerque, NM), Karl Henry (WistarInstitute, Philadelphia, PA), Brad Bernstein(Harvard Medical School, Boston, MA) andAli Shilatifard (St. Louis University, St.Louis, MO) demonstrated that ubiquitina-tion of H2B-K123 is required for subse-quent methylation of H3-K4 and H3-K79(refs 6–8). All three of these marks areimportant for transcriptional activity.Similarly, Robert Roeder (RockefellerUniversity, New York, NY) presented evi-dence from in vitro studies for a multistepmodel for p53-mediated in vitro transcrip-tion, in which methylation of H4-R3 byprotein arginine methyltransferase 1

Bre1

Ub

Rad6 Bre1

Rad6Bre1

Rad6

Activator binding

Transcription initiation

Transcription termination

Late elongation

Early elongation

AD

AD

Pol II

Phos-Ser5

GTFsPol IIKin28

Pol II

Pol II

Phos-Ser5

Phos-Ser2

Paf1 complex

Polyadenylation,termination factors

Elongation factors

mRNA transcript

Paf1

Paf1

Chd1

Chd1

TFIIS

TFIIS

K36CH3

ElonginB C

ElonginB C

K4CH3

K4CH3

K4CH3

K4CH3

K36CH3

ISW1

ISW1

Set2

PolyA signal

FACT

FACT

RNA15

AAAAAA

SAM

SAH

COMPASS

Phos-Ser5

ElonginB C

Set2

FACT

COMPASS

Ctk1

K4CH3

K4CH3

K4CH3

ISW1

Pol II

Paf1

Chd1

TFIISCOMPASS

© 2003 Nature Publishing Group

meeting report

NATURE CELL BIOLOGY VOL 5 MAY 2003 www.naturecellbiology.com 397

(PRMT1) stimulates CBP/p300 acetylationof H4-K5, K8, K12 and K16, which in turnpromotes the methylation of H3-R2, R17and R26 by another PRMT family member,CARM1. The integrated action of these his-tone-modifying enzymes correlates withstrong transcriptional activation of thep53-dependent gene. Cellular chromatinimmunoprecipitation (ChIP) assays con-firmed the stepwise recruitment of thesecofactors and accompanying histone modi-fications in response to activation of a p53-dependent gene by ultraviolet irradiation.Similarly, Michael Stallcup (University ofSouthern California, Los Angeles, CA)showed that oestrogen-dependent tran-scription also required the concerted actionof PRMT1/CARM1/p300, recruited to theoestrogen-responsive promoter by GRIP-1/ER-activating complex9. Finally, RonenMarmorstein (Wistar Institute,Philadelphia, PA) presented the structuralbasis for why phosphorylation of H3-S10enhances the ability of Gcn5p to acetylateH3-K14 in vitro and at some sites in vivo10.He found that the phosphorylated serineinduces H3 adjustments both local and dis-tal to H3-S10, resulting in a more optimalanchoring of H3 for K14 acetylation.

How do histone modificationsexert their effects? A large number of presentations focused onhow histone marks exert their biologicaleffects. In theory, site-specific histone mod-ifications might influence chromatin func-tion through two mechanisms: first, byaffecting nucleosome–nucleosome interac-tions that control the folding of nucleoso-mal arrays; second, by promoting or dis-rupting the chromatin binding of non-his-tone proteins that link the covalent modifi-cation to a biological outcome. Addressing

the first potential mechanism has beenproblematic because of the technical inabil-ity to reconstitute nucleosomal arrays thatharbour homogeneous and site-specificmodifications. In response to this problem,we presented our recent success at develop-ing a native chemical ligation strategy per-mitting the reconstitution of nucleosomalarrays that harbour a wide range of site-specific histone modifications, includingserine phosphorylation, lysine acetylationand lysine methylation11.

A great deal of progress is being made inthe identification of proteins that bind tohistone tails harbouring specific marks.Several talks discussed the binding of theheterochromatin component HP1 to theH3 tail dimethylated at K9. Similarly, YiZhang (UNC Chapel Hill, NC) presented invivo and in vitro studies demonstrating thatthe Drosophila melanogaster and mam-malian Polycomb complexes interact withhistone H3 tri-methylated at K27. In bothof these cases, the binding of effector pro-teins to methylated H3-K9 or H3-K27 pro-moted gene silencing12. Chromatin marksthat are associated with transcriptionallyactive genes may also exert their effects byinfluencing the binding of chromatinremodelling factors. Although not dis-cussed at this meeting, acetylated lysinescan interact with bromodomains foundwithin subunits of SWI/SNF-like chro-matin remodelling complexes13.Additionally, Tony Kouzarides presentedcompelling evidence that methylation ofH3-K4 promotes the binding of the yeastIsw1p ATPase. Thus, histone modificationprovides a means to recruit enzymes thatdisrupt chromatin structure (for example,Isw1p) or to recruit proteins that furtherpackage the chromatin fibre into inaccessi-ble states (for example, HP1).

Heterochromatin formation and silencingThe establishment and maintenance oftranscriptional silencing by heterochro-matic structures was discussed at length.Michael Grunstein (UCLA, Los Angeles,CA) provided evidence supporting a step-wise model for telomeric heterochro-matin assembly in budding yeast. Hisresults indicated that the Rap1p protein,which is localized to the T

1–3telomeric

repeats, recruits Sir4p, which in turnrecruits the histone-binding protein Sir3pand the histone deacetylase Sir2p. TheNAD+-dependent deacetylation of H4-K16 by Sir2p then allows the spreading ofthese three Sir proteins from the telom-ere14. Work presented by both MichaelGrunstein and Masami Horikoshi(University of Tokyo, Tokyo, Japan) indi-cated that the advancement of this silenc-ing structure into euchromatic regions isblocked by the action of Sas2p, a histone

acetyltransferase that targets H4-K16 andcounteracts the action of Sir2p15,16. Infact, Sas2p seems to have a global role inthe maintenance of H4-K16 acetylation.Ann Ehrenhofer-Murray (Max-PlanckInstitute, Berlin, Germany) reported thatSas2p is part of the yeast SAS-I complex,which interacts with the chromatinassembly factor CAF-1 and is recruited toDNA replication forks to re-establish H4-K16 acetylation on newly assembledchromatin17.

A number of speakers discussed theroles of HP1 in gene silencing, heterochro-matin formation and chromosome segrega-tion. In Drosophila, HP1 is associated withpericentric and telomeric regions, a regionalong the arm of the 4th chromosome andapproximately 200 other euchromatic sites.In vitro, HP1 prefers to bind to a histone H3tail dimethylated at K9, and in vivo HP1generally (but not always) colocalizes withthis histone mark. Lori Wallrath (Universityof Iowa, Iowa City, IA) demonstrated thatthe artificial tethering of a LacI–HP1 fusionprotein to an ectopic location could causethe long range silencing of reporter genes18.Remarkably, the tethered HP1 did notresult in H3-K9 methylation. Sally Elgin(Washington University, St. Louis) then dis-cussed the fact that HP1 requires other geneproducts to effect heterochromatin-basedgene silencing, and that one of these factors,HP2, colocalizes with HP1 at many loci19.This HP1-based silencing system is notused solely for establishment of heterochro-matin domains, as Frank Rauscher III(Wistar Institute, Philadelphia, PA) provid-ed an example of HP1 recruitment to spe-cific promoters by the KAP-1 corepressor.KAP-1 coordinates histone deacetylation,histone methylation and HP1 deposition toeffect silencing by the KRAB-ZFP super-family of transcriptional repressors.Interestingly, this highly localized silencingstructure was mitotically stable and wasmaintained for more than 50 populationdoublings, even in the absence of the DNA-bound KRAB repressor20.

Perhaps the most highly anticipatedtalks of the symposium were those present-ed by Shiv Grewal (CSHL, Cold SpringHarbor, NY) and Robin Allshire (WellcomeTrust Centre for Cell Biology, University ofEdinburgh), who discussed their recentstudies on the role of the RNA interference(RNAi) machinery in establishing andmaintaining heterochromatin structures atthe fission yeast mating-type locus and cen-tromeres21,22. Both groups demonstratedthat a short piece of centromeric repetitiveDNA was sufficient to establish RNAi-dependent, heterochromatin-based genesilencing at an ectopic location.Furthermore, RNAi was shown to berequired to recruit the histone H3-K9methyltransferase, Clr4, the fission yeasthomologue of HP1, Swi6 and cohesion pro-

Figure 2 Transcriptional elongation byRNAPII on chromatin substrates.Rad6p/Bre1p are recruited by transcrip-tional activators to ubiquitinate H2B. TheRNAPII CTD repeats are phosphorylated onSer 5 by Kin28p. As the RNAPII complexprogresses, Ser 5 is dephosphorylated andSer 2 is phosphorylated by Ctk1p. The H3K36-specific methyltransferase Set2passociates with the ongoing complex, pro-moting late elongation. When a polyadeny-lation site is reached, Ser 2 is dephospho-rylated by Fcp1p and most elongation fac-tors dissociate from the RNAPII complex.Polyadenylation and termination factorsassociate with the RNAPII complex, result-ing in the release of the polyadenylatedmRNA transcript. We thank Ali ‘the fisher-man’ Shilatifard for help with this figure.

�

© 2003 Nature Publishing Group

meeting report

teins. Shiv Grewal then described recentwork indicating that the fission yeasthomologue of the Sir2 deacetylase is alsorequired for establishing heterochromatin-based gene silencing at centromeres and themating-type locus. Robin Allshire also pre-sented data indicating that repression ofsome euchromatic genes may be regulatedby the RNAi machinery/Clr4/Swi6. Thus, itseems that the RNAi-dependent establish-ment of heterochromatin may represent amore general means for creating silentstates at euchromatic loci.

The transcriptional cycleIn contrast to previous meetings of thiskind, very few presentations focused on theregulated formation of the transcriptionpre-initiation complex, but rather tran-scriptional elongation by RNA polymeraseII was the centrepiece of discussion. Workfrom a number of groups indicated thattranscriptional elongation involves anenormous number of regulatory factors(Fig. 2). As expected, several talks focusedon the how RNAPII contends with thenucleosomal barrier to elongation. DannyReinberg (UMDNJ, Piscataway, NJ) pre-sented compelling evidence for associationof the FACT complex with elongating poly-merase in Drosophila (in collaboration withJ. Lis) and yeast cells. Mechanistic studies byhis group (and in collaboration with V.Studitsky) indicated that FACT facilitatesnucleosomal elongation by destabilizingone of the histone H2A/H2B dimers.Robert Kingston (Harvard Medical School,Boston, MA) also presented in vivo evi-dence that the human SWI/SNF ATP-dependent remodelling complex is recruit-ed to the hsp70 promoter by heat shocktranscription factor, where SWI/SNF isthen central to transcriptional elongation.

Several presentations focused on therole of multisite phosphorylation of theheptapeptide repeats in the C-terminaldomain (CTD) of the largest subunit ofRNAPII as a means for recruiting variouselongation factors. For instance, StephenBuratowski (Harvard Medical School,Boston, MA) discussed how phosphoryla-tion of Ser 5 within the CTD of early elon-gation complexes results in recruitment ofthe mRNA-capping enzyme and the Set1p-containing histone H3-K4 methyltrans-ferase complex, COMPASS (Fig. 2).Similarly, Ali Shilatifard showed that recruit-ment of COMPASS to this elongating formof RNAPII involved its interaction with thePAF1 complex, which is known to associatewith the elongating form of RNAPII. He alsoshowed that the PAF1 complex controls thehistone methyltransferase activity of Dot1p,thereby linking methylation of H3-K79 totranscriptional elongation23 (See News andViews by Bryan Turner on page 390 of thisissue). Work presented by Tony Kouzarides

and Antonin Morillon (University ofOxford, Oxford, UK) indicated that theelongation role of the COMPASS methyl-transferase complex may involve recruit-ment of the Isw1p ATP-dependent chro-matin remodelling enzyme, as Isw1p bindsto an H3 tail that is di- or tri-methylated atK4, and ISW1 is required in vivo for a nor-mal distribution of RNAPII on codingregions. As elongation by RNAPII progress-es, Ser 2 on the CTD is phosphorylated bythe Ctk1p kinase. Results presented by bothBrian Strahl (UNC, Chapel Hill, NC) andAli Shilatifard indicated that phosphoryla-tion of Ser 2 may release the COMPASScomplex and result in recruitment of Set2p,which methylates H3-K36. Why H3 methy-lation sites change during the stages ofRNAPII elongation is not known.

Ubiquitin-dependent proteolysisand transcriptionSeveral presentations described examples oftranscriptional regulation by ubiquitin-dependent recruitment of proteosomalcomponents. Talks by Thomas Kodadek(UT Southwestern, Dallas, TX) and MichaelHubner (EMBL, Heidelberg, Germany)described how the functioning of eitheryeast Gal4p or the mammalian oestrogenreceptor are controlled by ubiquitin-dependent recruitment of the 19S APISproteasomal subcomplex. In the case ofGal4p, Kodadek presented evidence thatGal4p may be ubiquitinated when bound atthe GAL1 promoter, and that subsequentrecruitment of a 19S APIS proteasomalsubcomplex might regulate Gal4p DNAbinding or control Gal4-RNAPII holoen-zyme interactions24. Hubner presented anintegrated model for the cyclic turnover ofoestrogen receptor α on responsive pro-moters. In this model, both unliganded andliganded oestrogen receptor α, as well as E3ubiquitin ligases and the 19S APIS complex,cycle on and off an oestrogen response ele-ment in vivo. Subsequent ubiquitinationand proteasome-mediated removal ofoestrogen receptor α from the promoter areinstrumental for maintaining oestrogensignalling. Consequently, inhibition of theproteasome results in the loss of oestrogenreceptor α-mediated transcription, whereastranscriptional inhibition prevents oestro-gen receptor α degradation by the protea-some25. Thus, both of these studies suggestthat dynamic regulation of activator degra-dation or function may control transcrip-tional activation.

Transcriptional regulation by ubiqui-tin-dependent events does not seem to berestricted to the gene-specific activators.Jesper Svejstrup (Cancer Research UKLondon Research Institute, London, UK)discussed their identification of a yeastfactor called DEF1, which seems to be anovel elongation factor that is required for

ubiquitin-dependent degradation ofRNAPII in response to DNA damage26. Ona related note, Joan Conaway (StowerInstitute, Kansas City, MO) described theirongoing studies characterizing the ElonginB/C ubiquitin ligase complex. Shedescribed yeast two-hybrid and biochemi-cal studies, which show that Elongin B/Cinteracts directly with the Med8p subunit ofthe mammalian mediator complex. In fact,she showed that purified preparations ofmammalian mediator contain sub-stoi-chiometric levels of Elogin B/C and thatmediator has ubiquitin ligase activity27.What the targets for this activity might bein the transcription initiation or elongationcomplex are not known.

The futureAs researchers continue to pry their wayinto the mysteries of chromatin and tran-scriptional control, the possibilities forregulation seem endless. A multitude ofhistone modifications control recruitmentof chromatin remodelling enzymes, as wellas proteins that influence the higher-orderfolding of chromatin fibres. It now seemsthat millions of Daltons of protein are notonly bound at the transcription initiationsite, but that even larger protein assembliestravel down the gene with the elongatingpolymerase. And yet, although the meetingdescribed many new factors and new regu-latory paradigms, many old topics wentvirtually unmentioned. For instance, link-er histones and histone variants are gener-al features of the intrinsically heteroge-neous chromatin fibre, but almost nomention was made of how these factorscontrol the solution state behaviour andfunction of chromatin fibres. Recent invivo studies from Belmont and colleaguesindicate that transcription actually occurson enormous, 100–400-nm thick chro-matin fibres28. How can this fibre accom-modate the many proteins involved in theinitiation and elongation of transcription?And finally, as we noted at the start of thisreport, chromatin structure and histonemodifications are commonly thought toimprint an epigenetic code on transcrip-tional patterns. However, we still have littleknowledge of how chromatin states arepropagated after passage of a replicationfork. It is clear that we still have lots ofwork to do before all of the mysteries aresolved.Mariela Jaskelioff and Craig L. Peterson are in theProgram in Molecular Medicine and theInterdisciplinary Graduate Program, University ofMassachusetts Medical School, 373 Plantation St,Worcester, MA 01605, USAe-mail: Craig.Peterson@umassmed.edu

1. Fischle, W., Wang, Y. & Allis, C. D. Curr. Opin. Cell Biol. 15,172–183 (2003).

2. Santos-Rosa, H. et al. Nature 419, 407–411 (2002).

3. Xiao, B. et al. Nature 421, 652–656 (2003).

NATURE CELL BIOLOGY VOL 5 MAY 2003 www.naturecellbiology.com398

© 2003 Nature Publishing Group

4. Trievel, R. C. et al. Cell 111, 91–103 (2002).

5. Zhang, X. et al. Cell 111, 117–127 (2002).

6. Henry, K. W. & Berger, S. L. Nature Struct. Biol. 9, 565–566

(2002).

7. Bernstein, B. E. & Schreiber, S. L. Chem. Biol. 9, 1167–1173

(2002).

8. Dover, J. et al. J. Biol. Chem. 277, 28368–28371 (2002).

9. Lee, Y. H. et al. Mol. Cell Biol. 22, 3621–3632 (2002).

10. Lo, W. S. et al. Mol Cell 5, 917–926 (2000).

11. Shogren-Knaak, M. A., Fry, C. J. & Peterson, C. L. J. Biol. Chem.

PMID: 12595522 (2003).

12. Cao, R. et al. Science 298, 1039–1043 (2002).

13. Hassan, A. H. et al. Cell 111, 369–379 (2002).

14. Luo, K., Vega-Palas, M. A. & Grunstein, M. Genes Dev. 16,1528–1539 (2002).

15. Suka, N., Luo, K. & Grunstein, M. Nature Genet. 32, 378–383

(2002).

16. Kimura, A., Umehara, T. & Horikoshi, M. Nature Genet. 32,370–377 (2002).

17. Meijsing, S. H. & Ehrenhofer-Murray, A. E. Genes Dev. 15,3169–3182 (2002).

18. Li, Y. et al. Development 130, 1817–1824 (2003).

19. Shaffer, C. D. et al. Proc. Natl Acad. Sci USA 99, 14332–14337

(2002).

20. Schultz, D. C. et al. Genes Dev 16, 919–932 (2002).

21. Hall, I. M., Noma, K. & Grewal, S. I. Proc. Natl Acad. Sci. USA

100, 193–198 (2003).

22. Volpe, T. et al. Chromosome Res. 11, 137–146 (2003).

23. Krogan, N. J. et al. Mol. Cell 11, 721–729 (2003).

24. Gonzalez, F., Delahodde, A., Kodadek, T. & Johnston, S. A.

Science 296, 548–550 (2002).

25. Reid, G.et al. Mol. Cell 11, 695–707 (2003).

26. Woudstra, E. C. et al. Nature 415, 929–933 (2002).

27. Brower, C. S. et al. Proc. Natl Acad. Sci. USA 99, 10353–10358

(2002).

28. Memedula, S. & Belmont, A. S. Curr. Biol. 13, 241–246 (2003).

meeting report

NATURE CELL BIOLOGY VOL 5 MAY 2003 www.naturecellbiology.com 399

© 2003 Nature Publishing Group