3
Transcriptional Activation: Getting a Grip on Condensed Chromatin Craig L. Peterson Chromosomal transcription involves the concerted action of enormous, multisubunit chromatin remod- eling complexes. A recent study, however, suggests a different and surprising viewpoint in which these protein assemblies are quite dynamic and individual subunits play key roles in chromatin remodeling. Biochemical studies over the past ten years have continued to reinforce the view that transcriptional activators control the intitiation of transcription by orchestrating the recruitment of large multisubunit protein complexes (reviewed in [1]). For example, ATP-dependent chromatin remodeling enzymes such as SWI/SNF, histone acetyltransferase complexes such as SAGA, and the RNA polymerase II holoen- zyme are believed to function in vivo as pre-assem- bled protein complexes that have native molecular weights in the millions of Daltons. In a recent issue of Current Biology, Memedula and Belmont [2] describe a powerful in vivo system that directly visualizes the activator-dependent recruitment of multisubunit com- plexes to a target promoter. Surprisingly, they provide compelling evidence that chromatin remodeling com- plexes may not function as pre-assembled units in vivo; rather, their data suggest that individual cat- alytic subunits perform obligatory roles during tran- scriptional activation that cannot be carried out by the 1–2 Mda complexes. In order to visualize how a transcriptional activator functions in the context of a condensed chromatin fiber in vivo, Belmont and colleagues [3] created a novel Chinese hamster ovary (CHO) cell line (A03_1) with an amplified chromosomal region of approxi- mately 90 Mbp containing arrays of lacI repressor binding sites, the gene for dihydrofolate reductase (DHFR) and co-amplified genomic DNA (Figure 1). When imaged in live cells by decoration with a fusion protein consisting of the lac repressor (lacI) linked to the green fluorescent protein (lacI–GFP), this 90 Mbp tract appears as a condensed foci approximately 1 µm in diameter. When these cells are transiently trans- fected with an expression vector that produces a fusion of LacI–GFP to the potent VP16 transcriptional activation domain, the fluorescent foci undergo dra- matic decondensation (Figure 1). This activator-directed decondensation is accompanied by increased acetyla- tion of all four core histones, recruitment of PCAF, Gcn5 and CBP/p300 histone acetyltransferases, and activation of transcription [3]. Inhibition of transcription with α-amanitin does not block the dramatic decon- densation of the lacI array, indicating that these struc- tural changes do not require transcriptional activity. Memedula and Belmont [2] have now followed up these initial studies by monitoring the recruitment kinet- ics of several different chromatin remodeling enzymes. Rather than use transient transfection to introduce the lacI–GFP–VP16 fusion, cells were loaded with glass beads pre-coated with the purified fusion protein. Using this methodology, binding of the lacI–GFP–VP16 fusion throughout the condensed lacI array was detectable by GFP fluorescence immediately following initiation of the bead loading procedure. Likewise, VP16-dependent recruitment of the TRRAP protein, a subunit of multiple histone H3 and H4 acetyltransferase complexes [4,5], was detected within 5 minutes at the lacI array. Sur- prisingly, recruitment of several other histone acetyl- transferase complex subunits, including the Gcn5, PCAF, and CBP/p300 catalytic subunits, was not detectable until 20–30 minutes after binding of the acti- vator. The recruitment of SWI/SNF subunits was also temporally distinct — the catalytic ATPase subunits, Brg1 and hbrm, arrived within 5–10 minutes after acti- vator binding, whereas the BAF155 and BAF170 sub- units were not detectable until at least 20–30 minutes. Thus, although TRRAP, Brg1 and hbrm are generally believed to be obligatory subunits of large complexes, their early arrival at the lacI array — with respect to other histone acetyltransferase or hSWI/SNF subunits — suggests that the VP16 activation domain promotes their recruitment as either isolated polypeptides or as components of smaller subcomplexes. Why should an activator direct recruitment of individual components of multisubunit complexes? One possibility is that the intact, 1–2 MDa complexes may simply be too large to gain access to the interior volume of the condensed lacI locus. If this were the case, however, one might expect to observe recruit- ment of histone acetyltransferase and SWI/SNF complexes to the periphery of the amplified lacI array. This type of localization was clearly not observed [2]. Alternatively, the chromatin structure of the amplified lacI array may block the binding of intact complexes even to nucleosomes positioned on the exterior of the condensed mass. Two recent structural studies of yeast SWI/SNF and RSC remodeling complexes are consistent with this possibility [6,7]. In these studies, single particle recon- structions yielded three-dimensional structures which indicate that mononucleosomes bind within a large cavity of SWI/SNF or RSC that is composed of mul- tiple protein densities. As nucleosomes within con- densed chromatin fibers may have only a limited interaction surface, the ability of SWI/SNF-like com- plexes to ‘envelope’ a nucleosome might only be pos- sible in the context of decondensed, nucleosomal arrays. In contrast, a smaller, isolated Brg1 or hbrm ATPase subunit — which may lie at the base of the Dispatch Current Biology, Vol. 13, R195–R197, March 4, 2003, ©2003 Elsevier Science Ltd. All rights reserved. PII S0960-9822(03)00123-4 Program in Molecular Medicine, University of Massachusetts Medical School, 373 Plantation Street, Biotech 2, Suite 210, Worcester, Massachusetts 01605, USA. E-mail: [email protected]

Transcriptional Activation: Getting a Grip on Condensed Chromatin

Embed Size (px)

Citation preview

Page 1: Transcriptional Activation: Getting a Grip on Condensed Chromatin

Transcriptional Activation: Getting aGrip on Condensed Chromatin

Craig L. Peterson

Chromosomal transcription involves the concertedaction of enormous, multisubunit chromatin remod-eling complexes. A recent study, however, suggestsa different and surprising viewpoint in which theseprotein assemblies are quite dynamic and individualsubunits play key roles in chromatin remodeling.

Biochemical studies over the past ten years havecontinued to reinforce the view that transcriptionalactivators control the intitiation of transcription byorchestrating the recruitment of large multisubunitprotein complexes (reviewed in [1]). For example,ATP-dependent chromatin remodeling enzymes suchas SWI/SNF, histone acetyltransferase complexessuch as SAGA, and the RNA polymerase II holoen-zyme are believed to function in vivo as pre-assem-bled protein complexes that have native molecularweights in the millions of Daltons. In a recent issue ofCurrent Biology, Memedula and Belmont [2] describea powerful in vivo system that directly visualizes theactivator-dependent recruitment of multisubunit com-plexes to a target promoter. Surprisingly, they providecompelling evidence that chromatin remodeling com-plexes may not function as pre-assembled unitsin vivo; rather, their data suggest that individual cat-alytic subunits perform obligatory roles during tran-scriptional activation that cannot be carried out by the1–2 Mda complexes.

In order to visualize how a transcriptional activatorfunctions in the context of a condensed chromatinfiber in vivo, Belmont and colleagues [3] created anovel Chinese hamster ovary (CHO) cell line (A03_1)with an amplified chromosomal region of approxi-mately 90 Mbp containing arrays of lacI repressorbinding sites, the gene for dihydrofolate reductase(DHFR) and co-amplified genomic DNA (Figure 1).When imaged in live cells by decoration with a fusionprotein consisting of the lac repressor (lacI) linked tothe green fluorescent protein (lacI–GFP), this 90 Mbptract appears as a condensed foci approximately 1 µµmin diameter. When these cells are transiently trans-fected with an expression vector that produces afusion of LacI–GFP to the potent VP16 transcriptionalactivation domain, the fluorescent foci undergo dra-matic decondensation (Figure 1). This activator-directeddecondensation is accompanied by increased acetyla-tion of all four core histones, recruitment of PCAF,Gcn5 and CBP/p300 histone acetyltransferases, andactivation of transcription [3]. Inhibition of transcription

with αα-amanitin does not block the dramatic decon-densation of the lacI array, indicating that these struc-tural changes do not require transcriptional activity.

Memedula and Belmont [2] have now followed upthese initial studies by monitoring the recruitment kinet-ics of several different chromatin remodeling enzymes.Rather than use transient transfection to introduce thelacI–GFP–VP16 fusion, cells were loaded with glassbeads pre-coated with the purified fusion protein. Usingthis methodology, binding of the lacI–GFP–VP16 fusionthroughout the condensed lacI array was detectable byGFP fluorescence immediately following initiation of thebead loading procedure. Likewise, VP16-dependentrecruitment of the TRRAP protein, a subunit of multiplehistone H3 and H4 acetyltransferase complexes [4,5],was detected within 5 minutes at the lacI array. Sur-prisingly, recruitment of several other histone acetyl-transferase complex subunits, including the Gcn5,PCAF, and CBP/p300 catalytic subunits, was notdetectable until 20–30 minutes after binding of the acti-vator. The recruitment of SWI/SNF subunits was alsotemporally distinct — the catalytic ATPase subunits,Brg1 and hbrm, arrived within 5–10 minutes after acti-vator binding, whereas the BAF155 and BAF170 sub-units were not detectable until at least 20–30 minutes.Thus, although TRRAP, Brg1 and hbrm are generallybelieved to be obligatory subunits of large complexes,their early arrival at the lacI array — with respect toother histone acetyltransferase or hSWI/SNF subunits— suggests that the VP16 activation domain promotestheir recruitment as either isolated polypeptides or ascomponents of smaller subcomplexes.

Why should an activator direct recruitment ofindividual components of multisubunit complexes?One possibility is that the intact, 1–2 MDa complexesmay simply be too large to gain access to the interiorvolume of the condensed lacI locus. If this were thecase, however, one might expect to observe recruit-ment of histone acetyltransferase and SWI/SNFcomplexes to the periphery of the amplified lacI array.This type of localization was clearly not observed [2].Alternatively, the chromatin structure of the amplifiedlacI array may block the binding of intact complexeseven to nucleosomes positioned on the exterior of thecondensed mass.

Two recent structural studies of yeast SWI/SNF andRSC remodeling complexes are consistent with thispossibility [6,7]. In these studies, single particle recon-structions yielded three-dimensional structures whichindicate that mononucleosomes bind within a largecavity of SWI/SNF or RSC that is composed of mul-tiple protein densities. As nucleosomes within con-densed chromatin fibers may have only a limitedinteraction surface, the ability of SWI/SNF-like com-plexes to ‘envelope’ a nucleosome might only be pos-sible in the context of decondensed, nucleosomalarrays. In contrast, a smaller, isolated Brg1 or hbrmATPase subunit — which may lie at the base of the

Dispatch

Current Biology, Vol. 13, R195–R197, March 4, 2003, ©2003 Elsevier Science Ltd. All rights reserved. PII S0960-9822(03)00123-4

Program in Molecular Medicine, University of MassachusettsMedical School, 373 Plantation Street, Biotech 2, Suite 210,Worcester, Massachusetts 01605, USA.E-mail: [email protected]

Page 2: Transcriptional Activation: Getting a Grip on Condensed Chromatin

SWI/SNF cavity — might be capable of a wider rangeof nucleosomal interactions that are not hindered bychromatin folding. The intrinsic chromatin remodelingactivity of the isolated ATPase subunit may then besufficient to disrupt condensation of the lacI foci,promoting subsequent assembly or recruitment ofremodeling complexes [8].

The recruitment of individual chromatin remodelingsubunits also suggests that the rates of assembly ordisassembly of multisubunit complexes may play akey role in transcriptional control. For instance, theavailable pool of ‘free’ Brg1/hbrm or TRRAP subunitsmay influence the rate or extent of gene induction.This pool is expected to be quite small, as virtually allof the Brg1 detected in whole cell extracts appears tobe complexed with BAF subunits (A.N. Imbalzano,personal communication). Furthermore, studies inyeast have shown that individual SWI/SNF subunitsare unstable and rapidly degraded [9]. Alternatively,stable pools of individual subunits may not exist, butnovel disassembly activities may control the localizedproduction of key subunits or subcomplexes. Like-wise, the completion of complex assembly or the sub-sequent recruitment of a pre-assembled complex mayalso involve novel forms of regulation.

The sequential recruitment of TRRAP, Brg1/hbrmand histone acetyltransferase catalytic subunits mayreflect functional interdependencies among theseisolated subunits (Figure 2). For instance, severalprevious studies have reported that the VP16 activa-tion domain is unable to contact or recruit humanSWI/SNF complexes in purified in vitro systems[10–12]. The rapid VP16-dependent recruitment ofTRRAP in vivo, however, may generate a novel surfacethat is uniquely competent for subsequent recruitmentof Brg1 and hbrm. Similarly, the recruitment ofBrg1/hbrm before the histone acetyltransferase cata-lytic subunits may reflect an obligatory requirement forSWI/SNF-dependent remodeling in histone acetyl-transferase recruitment. This novel order of events maybe determined by the heterochromatin-like state of thelacI array, as in other cases histone acetyltransferase

subunits are clearly recruited to a mammalian targetbefore Brg1 [13].

Interestingly, the VP16-dependent series of factorrecruitments are reminiscent of those detected duringactivation of a large group of yeast genes during latemitosis [14]. In this case, prior SWI/SNF remodelingwas proposed to promote subsequent histone acetyl-transferase binding by driving the release of histoneamino-terminal tail domains from nucleosome–nucle-osome interactions involved in mitotic condensation.Consistent with this type of model, Memedula andBelmont [2] also observe a partial decondensation ofthe lacI array following Brg1/hbrm recruitment. Thus,chromatin unfolding may be the event that triggershistone acetyltransferase recruitment, as well as thesubsequent assembly or recruitment of completeremodeling complexes.

How general is this apparent stepwise assemblyphenomenon? Unfortunately, even the very beststudies that follow in vivo recruitment typically monitoronly one subunit of a putative complex (for examplesee [13]). Recently, Martens and Winston [15] havedescribed what may be a case of stepwise recruit-ment involving the yeast SWI/SNF complex. Theseauthors found that transcriptional repression of SER3required Swi2p/Snf2p, the ATPase subunit of yeastSWI/SNF, but repression was nearly independent ofother SWI/SNF subunits (including the homolog ofBAF155/170, Swi3p). Furthermore, although an intactSWI/SNF complex appeared to be present at the pro-moter, the authors demonstrated that Swi2p could berecruited to the SER3 promoter region even in theabsence of other SWI/SNF subunits. Thus, yeastSwi2p has the capacity to be recruited, and to func-tion, as an individual subunit — or at least as a smallersubcomplex — much like Brg1/hbrm at the amplifiedlacI array in the work of Medulla and Belmont [2].

An even more unexpected example of the dynamicnature of multisubunit complexes comes from arecent study from Misteli and colleagues [16]. Theseinvestigators used an in vivo microscopy method, flu-orescence recovery after photobleaching (FRAP), to

DispatchR196

Figure 1. A transcriptional activatorinduces large-scale changes in chromatinfolding in vivo.

The left panel depicts a repeated array oflacI binding sites organized as an~90 Mbp amplified array. This loci can bevisualized through the binding of alacI–GFP or lacI–GFP–VP16 fusion protein(see text and [2] for details).

LacI–GFP

LacI–GFP–VP16

Current Biology

Page 3: Transcriptional Activation: Getting a Grip on Condensed Chromatin

follow the kinetics for recruitment of RNA polymeraseI subunits at endogenous ribosomal genes. Surpris-ingly, their data indicate that RNA polymerase II maynot function as a pre-assembled complex in vivo, aspreviously believed, but that individual subunits orsmaller subcomplexes assemble in a sequentialmanner at a ribosomal target gene. If true for RNApolymerase I, why not RNA Polymerase II? Indeed, ifthe subunit organization of RNA polymerases isdynamic, then all reports of ‘stable’ multisubunit com-plexes may need to be reinterpreted with a touch ofcaution. Although the transcription field may neverrevert back to a strict ‘step-wise assembly’ model forgene regulation [17], it seems that simplified, holoen-zyme recruitment models are also unlikely to accu-rately describe this process.

References1. Fry, C.J. and Peterson, C.L. (2001). Chromatin remodeling enzymes:

who’s on first? Curr. Biol. 11, R185–R197.2. Memedula, S. and Belmont, A.S. (2003). Sequential recruitment of

HAT and SWI/SNF components to condensed chromatin by VP16.Curr. Biol. 13, 241–246.

3. Tumbar, T., Sudlow, G. and Belmont, A.S. (1999). Large-scale chro-matin unfolding and remodeling induced by VP16 acidic activationdomain. J. Cell Biol. 145, 1341–1354.

4. Brand, M., Yamamoto, K., Staub, A. and Tora, L. (1999). Identifica-tion of TATA-binding protein-free TAFII-containing complex sub-units suggests a role in nucleosome acetylation and signaltransduction. J. Biol. Chem. 274, 18285–18289.

5. Martinez, E., Palhan, V.B., Tjernberg, A., Lymar, E.S., Gamper, A.M.,Kundu, T.K., Chait, B.T. and Roeder, R.G. (2001). Human STAGAcomplex is a chromatin-acetylating transcription coactivator thatinteracts with pre-mRNA splicing and DNA damage-binding factorsin vivo. Mol. Cell. Biol. 21, 6782–6795.

6. Asturias, F.J., Chung, W.H., Kornberg, R.D. and Lorch, Y. (2002).Structural analysis of the RSC chromatin-remodeling complex.Proc. Natl. Acad. Sci. U.S.A. 99, 13477–13480.

7. Smith, C.L., Horowitz-Scherer, R., Flanagan, J.F., Woodcock, C. andPeterson, C.L. (2003). Structural analysis of the yeast SWI/SNFchromatin remodeling complex. Nat. Struct. Biol., in press.

8. Phelan, M.L., Sif, S., Narlikar, G.J. and Kingston, R.E. (1999). Recon-stitution of a core chromatin remodeling complex from SWI/SNFsubunits. Mol. Cell 3, 247–253.

9. Peterson, C.L. and Herskowitz, I. (1992). Characterization of theyeast SWI1, SWI2 and SWI3 genes, which encode a global activa-tor of transcription. Cell 68, 573–583.

10. Sullivan, E.K., Weirich, C.S., Guyon, J.R., Sif, S. and Kingston, R.E.(2001). Transcriptional activation domains of human heat shockfactor 1 recruit human SWI/SNF. Mol. Cell. Biol. 21, 5826–5837.

11. Kadam, S., McAlpine, G.S., Phelan, M.L., Kingston, R.E., Jones, K.Aand Emerson, B.M. (2000). Functional selectivity of recombinantmammalian SWI/SNF subunits. Genes Dev. 14, 2441–2451.

12. Boyer, L.A., Logie, C., Bonte, E., Becker, P.B., Wade, P.A., Wolffe,A.P., Wu, C., Imbalzano, A.N. and Peterson, C.L. (2000). Functionaldelineation of three groups of the ATP-dependent family of chro-matin remodeling enzymes. J. Biol. Chem. 275, 18864–18870.

13. Agalioti, T., Lomvardas, S., Parekh, B., Yie, J., Maniatis, T. andThanos, D. (2000). Ordered recruitment of chromatin modifying andgeneral transcription factors to the IFN-beta promoter. Cell 103,667–678.

14. Krebs, J.E., Fry, C.J., Samuels, M. and Peterson, C.L. (2000). Globalrole for chromatin remodeling enzymes in mitotic gene expression.Cell 102, 587–598.

15. Martens, J.A. and Winston, F. (2002). Evidence that Swi/Snf directlyrepresses transcription in S. cerevisiae. Genes Dev. 16, 2231–2236.

16. Dundr, M., Hoffmann-Rohrer, U., Hu, Q., Grummt, I., Rothblumm,L.I., Phair, R.D. and Misteli, T. (2002). A kinetic framework for amammalian RNA polymerase in vivo. Science 298, 1623–1626.

17. Buratowski, S., Hahn, S., Guarente, L. and Sharp, P.A. (1989). Fiveintermediate complexes in transcription initiation by RNA poly-merase II. Cell 56, 549–561.

Current BiologyR197

Figure 2. Sequential recruitment ofchromatin remodeling subunits at acondensed chromosomal target locus.

The TRRAP subunit of multiple histoneacetyltransferase complexes (red ellipses)associates with the lacI array immediatelyfollowing the binding of the lacI–GFP–VP16fusion (dark blue circles). The brg1 andhbrm ATPase subunits are recruited soonafter (green ovals), followed by associationof histone acetyltransferase (HAT) catalyticsubunits (light blue ovals), histone acetyla-tion and chromosome decondensation.Subsequent events include recruitment ofadditional SWI/SNF subunits and tran-scription.

TRRAP Brg1/hbrm

ATPhydrolysis

SWI/SNF assembly,transcription,additionalde-condensation

HATcatalyticsubunits

TRRAP/HATcomplex

TRRAP/HATcomplex

lacI–GFP–VP16

TRRAP

Brg1/hbrm HAT Catalytic subunits(Gcn5, PCAF, CBP/p300)

Histone amino-terminal tails

Current Biology