Assembly of Variant Histones into Chromatin Steven Henikoff 1 and Kami Ahmad 2 1 Howard Hughes Medical Institute, Fred Hutchinson Cancer Research Center, Seattle, Washington 98109; email: [email protected]2 Department of Biological Chemistry and Molecular Pharmacology, Harvard Medical School, Boston, Massachusetts 02115; email: kami − [email protected]Annu. Rev. Cell Dev. Biol. 2005. 21:133–53 First published online as a Review in Advance on July 18, 2005 The Annual Review of Cell and Developmental Biology is online at http://cellbio.annualreviews.org doi: 10.1146/ annurev.cellbio.21.012704.133518 Copyright c 2005 by Annual Reviews. All rights reserved 1081-0706/05/1110- 0133$20.00 Key Words nucleosome, chromatin remodeling, histone replacement, epigenetics, centromeric chromatin Abstract Chromatin can be differentiated by the deposition of variant his- tones at centromeres, active genes, and silent loci. Variant histones are assembled into nucleosomes in a replication-independent man- ner, in contrast to assembly of bulk chromatin that is coupled to replication. Recent in vitro studies have provided the first glimpses of protein machines dedicated to building and replacing alternative nucleosomes. They deposit variant H2A and H3 histones and are targeted to particular functional sites in the genome. Differences between variant and canonical histones can have profound conse- quences, either for delivery of the histones to sites of assembly or for their function after incorporation into chromatin. Recent studies have also revealed connections between assembly of variant nucleo- somes, chromatin remodeling, and histone post-translational mod- ification. Taken together, these findings indicate that chromosome architecture can be highly dynamic at the most fundamental level, with epigenetic consequences. 133 Annu. Rev. Cell Dev. Biol. 2005.21:133-153. Downloaded from www.annualreviews.org by University of Southern California (USC) on 04/03/14. For personal use only.
Transcript
ANRV255-CB21-06 ARI 1 September 2005 11:43
Assembly of VariantHistones into ChromatinSteven Henikoff1 and Kami Ahmad2
1Howard Hughes Medical Institute, Fred Hutchinson Cancer Research Center,Seattle, Washington 98109; email: [email protected] of Biological Chemistry and Molecular Pharmacology, Harvard MedicalSchool, Boston, Massachusetts 02115; email: kami−[email protected]
Annu. Rev. Cell Dev. Biol.2005. 21:133–53
First published online as aReview in Advance onJuly 18, 2005
The Annual Review ofCell and DevelopmentalBiology is online athttp://cellbio.annualreviews.org
AbstractChromatin can be differentiated by the deposition of variant his-tones at centromeres, active genes, and silent loci. Variant histonesare assembled into nucleosomes in a replication-independent man-ner, in contrast to assembly of bulk chromatin that is coupled toreplication. Recent in vitro studies have provided the first glimpsesof protein machines dedicated to building and replacing alternativenucleosomes. They deposit variant H2A and H3 histones and aretargeted to particular functional sites in the genome. Differencesbetween variant and canonical histones can have profound conse-quences, either for delivery of the histones to sites of assembly orfor their function after incorporation into chromatin. Recent studieshave also revealed connections between assembly of variant nucleo-somes, chromatin remodeling, and histone post-translational mod-ification. Taken together, these findings indicate that chromosomearchitecture can be highly dynamic at the most fundamental level,with epigenetic consequences.
Over the past decade we have witnessed arenaissance of interest in core histones withthe general realization that these four sim-ple components of nucleosomal octamers,histones H2A, H2B, H3 and H4, are alsokey players in basic nuclear processes. Thisis especially true for post-translational mod-ifications of histones, which have been im-plicated both in modulating chromatin archi-tecture, and in the regulation of transcription(Brownell et al. 1996, Jenuwein & Allis 2001,Turner et al. 1992). The realization that ad-dition or removal of these modifications canfacilitate gene activation or silencing has fu-eled considerable excitement in the chromatinfield.
Much less attention has been paid to thedifferentiation of chromatin by the incorpo-ration of variant histones (Table 1). Theseare separately encoded forms of canonicalhistones that are distinguished by sequencedifferences (Malik & Henikoff 2003). Many
variant histones are simply polymorphic ver-sions of the major canonical forms that areassembled into bulk chromatin behind thereplication fork. However, other variants arefound to have profound differences that dis-tinguish them from canonical forms, either inthe way that they are deposited or the waythat they function after deposition, or both.In the past few years, we have come to realizethat this latter class of histone variants and thespecial machineries that deposit them play im-portant roles in chromatin differentiation andepigenetic maintenance. The recent study ofvariants and their assembly has begun to re-veal a highly dynamic picture of chromatin, inwhich processes of post-transcriptional mod-ification appear to be coupled to processes ofhistone replacement.
In our review, we examine the basis for thismore dynamic view of chromatin by consider-ing recent studies on particular core histonevariants and chromatin assembly complexes.We explore the possibility that the processesthat replace histones at active genes and thatpropagate chromatin states when DNA repli-cates involve the concerted action of nucle-osome remodeling and histone modificationactivities. Although the study of chromatindynamics is technically demanding, we expectthat the rapid improvements in molecular bi-ology, cytogenetics, and genomics technolo-gies mean that this area of research is still inits infancy.
Centromeric Chromatin is Identifiedby a Special Histone H3 Variant
Every eukaryotic chromosome requires a cen-tromere for it to segregate at mitosis, andthe uniqueness of this structure has facilitatedthe cytological identification of centromere-specific protein components. The first suchcomponents were identified as epitopesof autoimmune antibodies (Earnshaw &Rothfield 1985), and one of these, CENP-A,was found to be a histone H3 homolog thatcopurifies with nucleosomes (Palmer et al.1991, Palmer et al. 1987). Genetic studies
134 Henikoff · Ahmad
Ann
u. R
ev. C
ell D
ev. B
iol.
2005
.21:
133-
153.
Dow
nloa
ded
from
ww
w.a
nnua
lrev
iew
s.or
gby
Uni
vers
ity o
f So
uthe
rn C
alif
orni
a (U
SC)
on 0
4/03
/14.
For
per
sona
l use
onl
y.
ANRV255-CB21-06 ARI 1 September 2005 11:43
Table 1 Histone variants and associated chromatin assembly complexes
Histones Features Assembled by (organism)Archaeal histones Ancestral histone fold proteins without tails found in singly wrapped
tetrameric units that comprise nucleosome particles.Unknown
H2A, H2B Canonical core histones encoded by replication-coupled genes. FACT (yeast, Drosophila)H2AZ H2A variant found in nearly all eukaryotes that has a diverged
H3 variant that replaces H3 and differs at position 31 and at a few residueson helix 2 that allow deposition outside of replication.
HIRA (mammals)
Packaginghistones
Core and linker histone variants adapted for tight packaging of DNA insperm and pollen in some organisms.
reveal that mammalian CENP-A and its coun-terparts in other eukaryotes (generically re-ferred to as CenH3s) are absolutely requiredfor assembly of the proteinaceous kinetochoreto which the spindle microtubules attach atmitosis and meiosis (Blower & Karpen 2001,Buchwitz et al. 1999, Howman et al. 2000,Stoler et al. 1995). Antibodies against CenH3sfrom both plants and animals have been usedto map centromeres (Alonso et al. 2003, Loet al. 2001, Nagaki et al. 2004), because inthese organisms the centromere is not de-termined by DNA sequence but rather bythe presence of centromeric chromatin. Infact, human neocentromeres that show no re-semblance in DNA sequence to native alpha-satellite-containing centromeres are nev-ertheless packaged in CENP-A-containingnucleosomes (Alonso et al. 2003, Amor et al.2004). These observations imply that the con-stant location of the centromere in all cellsof an organism through millions of years ofevolution is maintained by the faithful assem-bly of CenH3-containing chromatin. Appar-ently centromeres are maintained indefinitelyby the action of a chromatin assembly process.
What features of CenH3s are recognizedfor assembly into chromatin? Swaps between
CENP-A or Cse4p and H3 identified the coreregion as being crucial (Keith et al. 1999,Shelby et al. 1997), although further infer-ences were complicated by the possibility thatsome H3-specific residues might be incom-patible with CenH3 function. A more refinedapproach is to use heterologous CenH3s;in the case of Drosophila CenH3 (Cid), thisled to the identification of Loop I as beingboth necessary and sufficient for Cid local-ization to centromeres (Vermaak et al. 2002)(Figure 1). Thus Loop I of human CENP-A is included within a region inferred to bemore compact than the corresponding re-gions of H3 (Black et al. 2004). The pos-sibility that contacts between Loop I andcentromeric DNA are important for correctassembly of centromeric nucleosomes is im-plied by the evidence for adaptive evolution ofLoop I in Drosophila and Arabidopsis (Cooper& Henikoff 2004, Malik & Henikoff 2001).However, mammalian CENP-A is differentin that no adaptive evolution is seen (Talbertet al. 2004), and heterologous CenH3s canlocalize to human centromeres (Henikoffet al. 2000, Wieland et al. 2004). In fact,yeast Cse4p can even functionally replaceCENP-A (Wieland et al. 2004), unlike the
www.annualreviews.org • Histone Variants in Chromatin 135
Ann
u. R
ev. C
ell D
ev. B
iol.
2005
.21:
133-
153.
Dow
nloa
ded
from
ww
w.a
nnua
lrev
iew
s.or
gby
Uni
vers
ity o
f So
uthe
rn C
alif
orni
a (U
SC)
on 0
4/03
/14.
For
per
sona
l use
onl
y.
ANRV255-CB21-06 ARI 1 September 2005 11:43
Figure 1Regions of H3 and H2A variants responsible for chromatin differentiation. Surface-accessible residues(shaded blue) were annotated from the 1.9 A X-ray crystal structure (Davey et al. 2002). N-terminal tailsthat are not included in nucleosome models but are presumed to be accessible are also indicated (shadedgreen). Unshaded residues are occluded by DNA or by other histones within the octamer. Schematics forthe secondary structure of each histone are indicated, and regions of the histones and their variants withthe functions described in the text are indicated.
situation for Drosophila bipectinata Cid, whichrequires Drosophila melanogaster Loop I tolocalize to D. melanogaster centromeres(Vermaak et al. 2002). These differences be-tween organisms might be attributable to theother centromere-specific DNA-binding pro-tein, CENP-C, which is adaptively evolving inmammals and plants, but which has not beenidentified in Drosophila (Talbert et al. 2004). Ithas been proposed that CenH3s and CENP-Cs adapt the rapidly evolving centromericsatellites to the conserved kinetochore ma-chinery (Malik & Henikoff 2001), in whichcase proteins that are not adaptively evolving,such as CENP-A and Cse4p, would not re-quire species-specific interactions to packagecentromeric chromatin.
Assembly of CenH3-containing nucleo-somes is independent of replication (Ahmad
& Henikoff 2001a, Shelby et al. 2000). Theprocess presumably initiates with interac-tions between DNA and CenH3 Loop I ofCenH3•H4 units, and when the full core isassembled, the CenH3 N-terminal tail wouldinteract with linker DNA (Vermaak et al.2002) (Figure 2). In the canonical nucleo-some, the H3 N-terminal tail exits betweenthe DNA helices and contacts the DNA mi-nor groove where the DNA leaves the nucle-osome core (Luger et al. 1997). In contrastto the nearly invariant tail of canonical H3,which is constrained by the density of post-translational modification sites, the CenH3N-terminal tails are extraordinarily diverse,differing in length and sequence to such anextent that they cannot be aligned betweendistant species (Malik & Henikoff 2003). Mi-nor groove-binding motifs have been detected
136 Henikoff · Ahmad
Ann
u. R
ev. C
ell D
ev. B
iol.
2005
.21:
133-
153.
Dow
nloa
ded
from
ww
w.a
nnua
lrev
iew
s.or
gby
Uni
vers
ity o
f So
uthe
rn C
alif
orni
a (U
SC)
on 0
4/03
/14.
For
per
sona
l use
onl
y.
ANRV255-CB21-06 ARI 1 September 2005 11:43
Figure 2A model for assembly of centromeric chromatin.
in CenH3 N-terminal tails (Malik et al.2002), suggesting that contacts with the mi-nor groove stabilize centromeric nucleosomeassembly. In support of this possibility, the N-terminal tails of some CenH3s are adaptivelyevolving (Malik & Henikoff 2001, Talbert
RI: replication-independent
et al. 2002). In Drosophila Cid, regions of adap-tive evolution are interspersed with patches ofsequence conservation that might correspondto sites of interaction with the conserved kine-tochore machinery (Keith et al. 1999, Maliket al. 2002). Centromeric chromatin wouldthen mature with the recruitment of CENP-C, a DNA-binding protein that neverthelessdepends upon CenH3 for centromeric local-ization (Moore & Roth 2001, Sugimoto et al.1994, Yang et al. 1996).
It is striking that centromeric chro-matin is organized as interspersed stretchesof CenH3- and H3-containing nucleosomes(Blower et al. 2002). It is difficult to envisionprocessive assembly leading to an interspersedorganization, which would require switch-ing back and forth between substrates. Al-ternatively, CenH3- and H3-containing nu-cleosomes might be assembled at differenttimes in the cell cycle (Ahmad & Henikoff2002b): Canonical H3 incorporates behindthe replication fork, whereas both humanand fly CenH3s deposit in an RI manner.Just how CenH3 assembly is maintained atsites of pre-existing CenH3 in the absence ofDNA sequence determinants is a major unan-swered question. A speculative possibility isthat anaphase tension on CenH3-containingchromatin causes adjacent nucleosomes to un-ravel and subsequent chromatin repair in-corporates new CenH3 (Ahmad & Henikoff2002b, Mellone & Allshire 2003). It is alsopossible that unique patterns of H3 mod-ifications found in interspersed and flank-ing stretches of nucleosomes predispose cen-tromeres for deposition of CenH3 (Sullivan& Karpen 2004). How such patterns arisemight depend on modes of assembly of H3-containing nucleosomes, which we turn tonext.
Active Genes are Sites of H3.3Replacement
Another universal H3 variant, H3.3, is sosimilar to canonical H3 that its distinctiveproperties were not realized until recently. In
www.annualreviews.org • Histone Variants in Chromatin 137
Ann
u. R
ev. C
ell D
ev. B
iol.
2005
.21:
133-
153.
Dow
nloa
ded
from
ww
w.a
nnua
lrev
iew
s.or
gby
Uni
vers
ity o
f So
uthe
rn C
alif
orni
a (U
SC)
on 0
4/03
/14.
For
per
sona
l use
onl
y.
ANRV255-CB21-06 ARI 1 September 2005 11:43
RC:replication-coupled
animals, H3.3 differs from H3 at only fouramino acid positions (Figure 1), consistentwith the view that it is essentially a constitu-tive version of canonical H3 (Yu & Gorovsky1997). Indeed, H3.3 is the dominant H3-subtype in nondividing differentiated cells invertebrates (Pina & Suau 1987, Urban &Zweidler 1983) and in endoreplicating cellsduring development of a chordate (Chiodaet al. 2004). Nevertheless, we found thatthree of the four differences between H3 andH3.3 determine nucleosome assembly behav-ior: Changes from the H3 to the H3.3 formallowed RI assembly (Ahmad & Henikoff2002c). This difference between H3 and H3.3defines a pathway for RI assembly that is dis-tinct from the RC pathway whereby canon-ical histones are assembled into bulk DNA.Another difference is that RC assembly of ei-ther H3 or H3.3 requires the N-terminal tail,whereas RI does not.
Confirmation that alternative RC and RIassembly pathways exist came from the purifi-cation of H3- and H3.3-containing complexes(Tagami et al. 2004). The well-studied RC as-sembly complex, CAF-1, copurifies with H3,whereas the replication-independent histonechaperone, HIRA, copurifies with H3.3. Bothcomplexes include a common H4-bindingcomponent, RbAp48, a homolog of which isalso essential for the assembly of Schizosaccha-romyces pombe CenH3 (Hayashi et al. 2004).Interestingly, the assembly form of histonesin both human CAF-1 and HIRA complexeswas an H3•H4 (or H3.3•H4) dimer (Tagamiet al. 2004), not a tetramer as might havebeen expected from the existence of (H3•H4)2
tetramers in solution (Wolffe 1992). Thisimplies that histone heterodimers are sub-strates in all known assembly complexes(Figure 3), consistent with a common ori-gin of eukaryotic nucleosome assembly mech-anisms. Core eukaryotic histones evolvedfrom structurally similar archaeal histonesthat assemble to form tetrameric nucleosomesclosely resembling (H3•H4)2 tetramers pro-duced in vitro (Pereira & Reeve 1998), and itwill be interesting to determine whether there
are biochemical similarities in assembly aswell.
H3.3 is enriched in active chromatin(Hendzel & Davie 1990), and the basis forenrichment has been elucidated by cytologicalstudies using epitope-tagged versions of H3.3.We showed that RI assembly of H3.3 local-izes to active, but not inactive, rDNA arraysand to euchromatin, but not heterochromatin,in Drosophila (Ahmad & Henikoff 2002c). Inhuman cells, H3.3 incorporates at a trans-gene array that has been induced to transcribe(Janicki et al. 2004). The concomitant loss ofheterochromatic markers in both Drosophilaand human cells demonstrates that the pro-cess of H3.3 deposition at active genes accom-panies replacement of pre-existing histones.The replacement process can be rapid, occur-ring on the order of an hour at a transgenearray observed in living cells (Janicki et al.2004).
RI assembly of H3.3•H4 provides an at-tractive mechanism for the resetting and per-petuation of histone modifications (Ahmad &Henikoff 2002c). RC assembly leaves a mix-ture of old and new nucleosomes on daughterstrands and, at active genes, this would be amixture of H3.3- and H3-containing nucleo-somes. If H3.3 is post-translationally modifiedin a way that suits transcriptional activity, thenthe mixture of histones after replication wouldcontinue to promote transcription (Henikoffet al. 2004). Continued transcriptional ac-tivity would then result in nucleosome re-placement over the body of genes, leadingto differentiation of chromatin whereby ac-tive regions are packaged in H3.3-containingnucleosomes. Indeed, there is enough H3.3in Drosophila Kc cell chromatin to denselypackage transcribed regions (McKittrick et al.2004). Furthermore, the enrichment in bulkH3.3 of lysine modifications that have beenfound to correlate with active transcription(McKittrick et al. 2004, Waterborg 1990) pro-vides a connection between histone replace-ment via the RI pathway and changes in prop-erties of chromatin by post-transcriptionalmodification.
138 Henikoff · Ahmad
Ann
u. R
ev. C
ell D
ev. B
iol.
2005
.21:
133-
153.
Dow
nloa
ded
from
ww
w.a
nnua
lrev
iew
s.or
gby
Uni
vers
ity o
f So
uthe
rn C
alif
orni
a (U
SC)
on 0
4/03
/14.
For
per
sona
l use
onl
y.
ANRV255-CB21-06 ARI 1 September 2005 11:43
It remains to be determined whether theform of H3.3•H4 that is assembled at ac-tive genes is post-translationally modifiedprior to or during assembly. Nevertheless,the strong correlations among lysine mod-ifications found in active chromatin of fliesand yeast are most easily understood if as-sembly and modification are concerted pro-cesses (Workman & Abmayr 2004). Associ-ations between histone-modifying enzymesand elongating RNA polymerases are alsoconsistent with concerted assembly and mod-ification. In yeast, histone replacement canoccur within a minute at sites of active tran-scription (Schwabish & Struhl 2004), whichdelimits the processes of histone modificationand nucleosome assembly to the same short-time interval.
Nucleosomes are Disassembled atPromoters
Nucleosome replacement during transcrip-tional elongation might also play a role in generegulation during development. Transcrip-tion within the mouse beta-globin locus con-trol region has been proposed to potentiateactivation (Gribnau et al. 2000), and transcrip-tion through a Drosophila Polycomb responseelement derepresses expression of genes incis (Bender & Fitzgerald 2002, Drewell et al.2002, Hogga & Karch 2002, Rank et al. 2002).Inactive chromatin can obstruct the binding oftranscription factors to their target sites, butonce bound, the result is a heritable state ofactivity (Ahmad & Henikoff 2001b). Herita-ble activation of previously silent chromatinby transcription-coupled replacement of hi-stones is one way that the RI assembly pro-cess might lead to inheritance of an epigeneticstate (Ahmad & Henikoff 2002a).
Although histone modification is the best-established change in chromatin states uponactivation of a gene, recent studies have shownthat the distribution of nucleosomes canchange as well. Activation of the yeast PHO5gene leads to loss of the nucleosome at thePHO5 promoter (Boeger et al. 2003, Reinke
Figure 3A model for RI replacement with histone variants (Henikoff et al. 2004).Parallels between H2A•H2B and H3•H4 replacement processes suggest acommon underlying mechanism, where a large molecular machine (eitherRNA polymerase or a SWI/SNF remodeler) partially or completelyunravels a nucleosome during transit. The result is either retention ofheterodimeric subunits, such as the FACT-facilitated transfer of H2A•H2Bfrom in front of RNA polymerase to behind (Belotserkovskaya et al. 2003,Formosa et al. 2002), or loss of a heterodimer. In the latter case, chromatinrepair replaces the lost heterodimer with either H3.3•H4 (top) orH2AZ•H2B (bottom). Failure to repair will result in nucleosome evictionand reduced nucleosome densities, such as has been observed at promotersand in the body of highly transcribed genes in yeast.
& Horz 2003). The nucleosome is not sim-ply moved aside; rather it unravels (Figure 3),a process that is facilitated by the ASF1 his-tone chaperone (Adkins et al. 2004, Boegeret al. 2004). This process is evidently not lim-ited to PHO5 because depletion of nucleo-somes at promoters has been found to occurgenome-wide in yeast (Bernstein et al. 2004,Lee et al. 2004). In addition, transiting RNApolymerases displace nucleosomes, leading tovariation in nucleosome occupancy over the
www.annualreviews.org • Histone Variants in Chromatin 139
Ann
u. R
ev. C
ell D
ev. B
iol.
2005
.21:
133-
153.
Dow
nloa
ded
from
ww
w.a
nnua
lrev
iew
s.or
gby
Uni
vers
ity o
f So
uthe
rn C
alif
orni
a (U
SC)
on 0
4/03
/14.
For
per
sona
l use
onl
y.
ANRV255-CB21-06 ARI 1 September 2005 11:43
HP1:heterochromatin-associatedprotein 1
body of genes (Kristjuhan & Svejstrup 2004,Schwabish & Struhl 2004). These findingsraise questions about the implicit assumptionthat the distribution of histone modificationsusing chromatin immunoprecipitation (ChIP)accurately reflects their relative density alongthe DNA (Hanlon & Lieb 2004). The actionof chromatin remodeling machines, DNA-binding proteins, and RNA polymerases canaffect local nucleosome densities such that thisassumption is valid, at best, as a first approx-imation. Thus transcription-coupled replace-ment or loss of histones could underlie someof the patterns of histone modifications thathave been reported. This interpretation mightextend to the mapping of histone-modifyingenzymes, which are cross-linked to their tar-get sites by formaldehyde, a reagent that pri-marily cross-links primary amines (Nagy et al.2003, Solomon & Varshavsky 1985), whichare especially abundant and accessible on thehistones themselves.
Chromatin Remodeling MachinesReplace H2A Variants
Structural alignments of nucleosomal sub-units reveal ancestral homology betweenH3•H4, H2A•H2B, and archaeal dimericunits, where H3 aligns with H2A and H4 withH2B (Pereira & Reeve 1998). This structuralequivalence might underlie the fact that H3and H2A have diverse variant forms, whereasH4 and H2B have (almost) none (Malik &Henikoff 2003). Of the H2A variants, H2AZis conspicuous in having a single evolution-ary origin very early in eukaryotic evolu-tion. Although H2AZ is essential in animals(Faast et al. 2001, van Daal & Elgin 1992), itis nonessential in budding yeast (Dhillon &Kamakaka 2000, Jackson & Gorovsky 2000),and this has facilitated its in vivo study. Thuswe know that H2AZ can act as a transcrip-tional activator and an antisilencer at differ-ent loci (Meneghini et al. 2003, Santistebanet al. 2000). Whether these functions ofH2AZ generalize to plants and animals isnot known. Surprisingly, mammalian H2AZ
shows a heterochromatic distribution and in-teracts with HP1 (Fan et al. 2004, Rangasamyet al. 2003), making it difficult to draw firmconclusions about conserved roles for H2AZin chromatin function.
Whereas the mechanism whereby H2AZaffects chromatin remains uncertain, muchhas been learned about how it is depositedinto chromatin, thanks to a combination ofin vivo and in vitro studies in yeast. H2AZ isassembled by the SWR1 complex, and muta-tions in the swr1 gene, which encodes a keycomponent of this complex, show gene ex-pression phenotypes similar to those of H2AZ(htz1) mutants (Kobor et al. 2004, Kroganet al. 2003, Mizuguchi et al. 2004). In vitro,SWR1 replaces H2A•H2B dimers in nucle-osomes with H2AZ•H2B dimers (Mizuguchiet al. 2004). The fate of the leaving H2A•H2Bdimer is unknown, so that it is prematureto refer to this process as an exchange, im-plying a reciprocal event, as opposed to re-placement, which does not. The actual mech-anism of H2A•H2B replacement might bevery similar to that for H3•H4 replacement,in which chromatin is perhaps repaired afterloss of a heterodimeric subunit during transitof a large complex such as RNA polymeraseor an ATP-dependent remodeling complex(Figure 3).
Replacement of H2A•H2B withH2AZ•H2B requires ATP, as expectedfrom the fact that the SWR1 subunit is amember of the SWI/SNF family of ATP-dependent chromatin remodelers. Thisfinding has important implications both forunderstanding the assembly of histone vari-ants, and for the understanding of chromatinremodeling. It remains uncertain as to justhow actions of various SWI/SNF familymembers observed in vitro relate to theirfunctions in vivo; but in the case of SWR1,the in vitro replacement of H2A with H2AZand the supporting in vivo data provide un-equivocal evidence for this specific function.We look forward to other examples of specificroles played by chromatin remodelers innucleosome assembly or disassembly.
140 Henikoff · Ahmad
Ann
u. R
ev. C
ell D
ev. B
iol.
2005
.21:
133-
153.
Dow
nloa
ded
from
ww
w.a
nnua
lrev
iew
s.or
gby
Uni
vers
ity o
f So
uthe
rn C
alif
orni
a (U
SC)
on 0
4/03
/14.
For
per
sona
l use
onl
y.
ANRV255-CB21-06 ARI 1 September 2005 11:43
The Drosophila SWR1 ortholog performsan analogous H2AZ replacement reaction aspart of the Tip60 complex (Kusch et al. 2004).An intriguing twist arises from the role ofDrosophila H2AZ in repair of double-strandedbreaks (Madigan et al. 2002). In most eukary-otes, this role is played by the H2AX vari-ant, which is otherwise similar to canonicalH2A except for the presence of a four-aminoacid C-terminal motif (SQ[D/E]Ø, where Ørepresents a hydrophobic amino acid). TheDrosophila version of H2AZ (H2AvD) hasevolved an H2AX-like motif and as a resultfunctions similarly in double-strand break re-pair. In diverse eukaryotes, H2AX is phospho-rylated on the serine of the H2AX C-terminalmotif at sites of double-strand breaks, andphosphorylation spreads rapidly to otherH2AXs along the chromosome, an event thatis important for recruitment of break repairmachinery (Fernandez-Capetillo et al. 2004,Rogakou et al. 1998). In vitro, Tip60 specifi-cally binds phosphorylated H2AvD, acetylatesit at Lys5 and replaces it with an unphosphory-lated H2AvD (Kusch et al. 2004). This com-bination of activities suggests that the func-tion of Tip60 is to remodel chromatin at sitesof double-strand breaks, while restoring theground state by effectively erasing the phos-phorylation mark.
It is likely that a similar process of ATPase-catalyzed replacement occurs in yeast, wherethe INO80 chromatin remodeling complexis recruited to H2AX when it is phos-phorylated following a double-strand break(Morrison et al. 2004, van Attikum et al.2004). INO80 is an ATPase distinct fromSWR1 and is consistent with yeast H2AX,which is actually the canonical version ofH2A, distinct from H2AZ. Therefore, twodifferent remodeling machines with distinctH2A substrates appear to have evolved in dif-ferent organisms to assume similar roles inDNA repair.
Other H2A variants are lineage-specific.The macroH2A histone is a vertebrate-specific variant unique among histones inhaving an additional globular domain. This
H2ABbd: H2A-Barr-body-deficient
C-terminal 200-amino acid domain is ho-mologous to a broad class of polynucleotideand peptide hydrolases, raising the possibilitythat macroH2A alters chromatin via the ac-tion of a tethered enzyme (Allen et al. 2003).macroH2A is enriched in regions of the mam-malian inactive X chromosome that are asso-ciated with determinants of facultative silenc-ing, including Xist RNA and H3 trimethyllysine-27 (Chadwick & Willard 2004). Incontrast, another vertebrate-specific variant,H2ABbd, shows a cytological distribution thatindicates an association with active chromatin(Chadwick & Willard 2001). These patternsare suggestive of functional differentiation ofvariant-containing chromatin.
How Is Variant Structure Related toFunction in Chromatin?
As we have seen, histone variants are dis-tinguished from canonical histones both bytheir mode of assembly into nucleosomes andby their properties in chromatin. Ultimately,processes such as transcription and replicationmust alter the structure of nucleosomes to ex-pose DNA, and this must involve regulatingDNA-histone affinities. The structure of thenucleosome suggests ways that exposure canhappen. The protein octamer can be dividedinto three surfaces, each with distinctive roles:(a) a perimeter ramp that underlies the super-helix of wrapped DNA, (b) the two exposedfaces of the disk, and (c) flexible N- and C-terminal tails of the histones that extend outof the nucleosome. The roles of these surfacesare becoming apparent, as are the ways thathistone modifications and variants alter acces-sibility of specific nucleosomes. For example,acetylation of N-terminal lysines may neutral-ize DNA-histone interactions (Turner et al.1992). An excellent case has been made thatmodifications decorating the protein super-helical ramp sterically disrupt DNA-histonecontacts (Cosgrove et al. 2004, Freitas et al.2004). A second mode-of-action for modifi-cations is as binding sites for proteins that re-model or restrict nucleosomes. Modifications
www.annualreviews.org • Histone Variants in Chromatin 141
Ann
u. R
ev. C
ell D
ev. B
iol.
2005
.21:
133-
153.
Dow
nloa
ded
from
ww
w.a
nnua
lrev
iew
s.or
gby
Uni
vers
ity o
f So
uthe
rn C
alif
orni
a (U
SC)
on 0
4/03
/14.
For
per
sona
l use
onl
y.
ANRV255-CB21-06 ARI 1 September 2005 11:43
Figure 4Location of differences between canonical and variant histones H3 (blue)and H2A (brown) shown on their three-dimensional structures. Segmentsthat differ are highlighted in yellow.
on the histone tails and the exposed disk facesare clearly suitable for display to incoming fac-tors, and both repressive partners (e.g., HP1)and activating ones (e.g., SWI/SNF nucle-osome remodelers) are known (Jenuwein &Allis 2001).
Studies of the conserved histone variantH2AZ support these paradigms. The regionof Drosophila H2AZ homolog (H2AvD) thatis essential for development lies in the dock-ing domain (Figure 4), where H2A interactswith the H3•H4 dimer within the nucleosome(Clarkson et al. 1999). This specialized dock-ing domain of H2AZ presents a binding sitefor HP1 exposed on the face of the nucleo-some (Fan et al. 2004). Thus the primary se-quence differences in the variant create a newbinding site in nucleosomes. Additionally, thissame region shifts the underlying H3 αN he-lix (Suto et al. 2000). The H3 αN helix alsocontacts DNA, and the changes in the H2AZdocking domain appear to alter these DNAcontacts, thus subtly destabilizing the nucleo-some. Other differences in the Loop 1 regionof H2AZ appear to configure the compositionof nucleosomes. Comparison of crystal struc-tures shows that the H2A Loop I will clashwith H2AZ Loop I within the nucleosome(Figure 4). Thus nucleosomes homotypic forH2A or for H2AZ are structurally preferable.This incompatibility between H2A and H2AZimplies that the SWR1-catalyzed replacementof one H2A•H2B dimer by H2AZ•H2B willfacilitate replacement of the other H2A•H2Bdimer.
Despite these insights gleaned from struc-tures, it is not yet clear how altered struc-tural features of H2AZ-containing nucleo-somes lead to their diverse roles inferred fromin vivo studies. Comparison of variant andcanonical nucleosomes based on their physicalproperties has led to different conclusions bydifferent laboratories: Some observations areconsistent with a destabilizing role for H2AZ(Abbott et al. 2001), whereas others are consis-tent with greater stabilization (Fan et al. 2002,Park et al. 2004). It is possible that higherorder interactions are more important than
142 Henikoff · Ahmad
Ann
u. R
ev. C
ell D
ev. B
iol.
2005
.21:
133-
153.
Dow
nloa
ded
from
ww
w.a
nnua
lrev
iew
s.or
gby
Uni
vers
ity o
f So
uthe
rn C
alif
orni
a (U
SC)
on 0
4/03
/14.
For
per
sona
l use
onl
y.
ANRV255-CB21-06 ARI 1 September 2005 11:43
particle stabilization for in vivo behavior ofH2AZ-containing nucleosomes, as suggestedby the interaction with HP1 (Fan et al. 2004).Differences in higher order interactions in-volving H2AZ could help rationalize the dif-ferent in vivo behaviors implied from studiesin yeast and animals.
There is more limited evidence for howother variants of H2A affect function. TheH2AX variant is identical to H2A through-out most of the protein, with primary se-quence differences in the C-terminal includ-ing the defining four-amino acid motif. Thusthis variant functions by providing a newphosphorylation site. The macroH2A variantcarries a large, globular C-terminal domainthat impedes transcription factor binding invitro, and its histone fold domain interfereswith ATP-dependent remodeling (Angelovet al. 2003). Both attributes are consistentwith macroH2A playing a role in facultativesilencing of the inactivated X chromosome(Costanzi & Pehrson 1998). In contrast, nu-cleosomes containing the H2ABbd variant aremore accessible than canonical nucleosomes(Angelov et al. 2004), consistent with the strik-ing depletion of H2ABbd on the inactive Xchromosome (Chadwick & Willard 2001).
Some variants of H3 discussed in this re-view also show evidence of structural dif-ferentiation. CenH3s are the most extreme,where positive selection is thought to actas an adaptor between the rapidly evolv-ing centromeric satellite sequences and theconserved kinetochore apparatus (Malik &Henikoff 2001). Rapid substitutions are fo-cused on sites of DNA-histone contacts andmay alter the affinity of centromeric nucleo-somes for underlying satellite sequences. TheCid Loop 1 that is required for targeting tothe centromere is on the disk face, consis-tent with a role as an exposed binding sur-face (Vermaak et al. 2002), and the divergedN-terminal tails of all CenH3 histones arethought to be platforms for binding kineto-chore components (Malik & Henikoff 2003).In contrast, H3.3 is an example of a variantthat appears to be structurally almost identical
to its canonical counterpart (Figure 4). Thecluster of core residues that specify assem-bly pathway are located behind the sheathof water residues that are structured by theDNA double helices (Davey et al. 2002) andthus are not accessible in the complete nu-cleosome. The single remaining difference,Ala31 in H3 versus Ser/Thr31 in H3.3, isnearly universal, suggestive of phosphoreg-ulation of H3.3 in as-yet unidentified pro-cesses. However, the interchangeability of RCand RI forms in Tetrahymena (Yu & Gorovsky1997), the presence of only H3.3 in yeastsand molds (Malik & Henikoff 2003), andthe dominance of H3.3 in certain cell lin-eages (Chioda et al. 2004, Pina & Suau 1987,Urban & Zweidler 1983) suggest minimalstructural differentiation between H3 andH3.3. Rather, a consistent requirement isseen for an RI H3-subtype in all eukaryotes,with the addition of a distinct RC-only formin multicellular eukaryotes. This phyleticpattern—and the evidence that H3.3 assem-bly is coupled to transcription—suggests thatorganisms with large, mostly silent genomeshave evolved the RC-only form to pack-age silent chromatin (Ahmad & Henikoff2002c).
However, the functional distinctions be-tween the two subtypes do not seem to be be-cause of alterations of exposed surfaces in astatic model of the nucleosome. We will needto consider the dynamics of nucleosome as-sembly to understand the role of the H3.3variant.
Functions During NucleosomeAssembly
A number of protein-binding sites on histonesare inaccessible in the complete nucleosomestructure, implying that the sites are func-tional only in structural intermediates. Oneexample of this is the RbAp48 subunit of CAF-1, which binds to the αN helix of histone H4(Vermaak et al. 1999). This interaction prob-ably occurs as histones are being delivered fordeposition because the helix will be buried in
www.annualreviews.org • Histone Variants in Chromatin 143
Ann
u. R
ev. C
ell D
ev. B
iol.
2005
.21:
133-
153.
Dow
nloa
ded
from
ww
w.a
nnua
lrev
iew
s.or
gby
Uni
vers
ity o
f So
uthe
rn C
alif
orni
a (U
SC)
on 0
4/03
/14.
For
per
sona
l use
onl
y.
ANRV255-CB21-06 ARI 1 September 2005 11:43
CAFs: chromatinassembly factors
the completely assembled nucleosome. Sur-prisingly, a number of complexes that regu-late transcription and operate on nucleosomaltemplates also contain RbAp48. This suggeststhat nucleosomal intermediates with exposedinternal portions may be regulatory targets(Vermaak et al. 2003).
Similarly, the cluster of core residues thatare not accessible in the complete nucleosomehave been proposed to form a binding surfacein predeposition complexes that delivers H3.3to DNA because they are essential for RI de-position of the histone (Ahmad & Henikoff2002c). However, the inaccessibility of theseresidues in H3.3 in the nucleosome leaves lit-tle to distinguish it from the canonical H3histone. The observation that H3.3 can de-posit in actively transcribed chromatin at anytime implies that these regions are structurallyand compositionally dynamic, with nucleo-somes continually disassembled and then re-assembled (Figure 3). If this leaves little timeto complete nucleosome assembly before anew round of replacement begins, transcribedchromatin would remain in a partially assem-bled state. Thus the repetition of RI assemblyalone would be sufficient to confer high DNAaccessibility. This function does not requirethat the H3.3 variant generates a specializednucleosome structure. In this view, the lack ofstructural specialization in H3.3 is a key fea-ture that results from the need for H3.3 to per-form the same role as the major histone H3,i.e., to package DNA. The sequence identitybetween the two H3-subtypes would be espe-cially important as genes become repressed,because transcription will have enriched theH3.3 variant in chromatin that must now bepackaged in a silent configuration.
A number of observations suggest that nu-cleosomal intermediates may be critical for ac-tive chromatin. As mentioned above, the veryprocess of targeted RI assembly suggests thatintermediates will be common in transcribedchromatin. Structures consistent with splitnucleosomes are indeed observed at highlytranscribed genes (Lee & Garrard 1991), andin vivo cross-linking studies also imply that
buried nucleosomal surfaces are exposed in ac-tive chromatin (Jackson 1978). Finally, a num-ber of transcription-promoting factors havesubunits with high histone-binding affinities,which are thought to assist chromatin bind-ing. However, similar histone affinities arefound in the protein chaperones that deliverhistones for nucleosome assembly (Akey &Luger 2003). Indeed, some transcription fac-tor complexes contain free histones (Keener1997), and thus do not appear to be usingtheir histone-binding subunits to bind chro-matin. Instead, they might act by assisting nu-cleosome assembly. A nucleosome disassem-bly role has been reported for one of thesefactors, ASF1 (Adkins et al. 2004). These ob-servations are consistent with the idea that nu-cleosomal intermediates are prevalent in ac-tive chromatin and distinguish it from inactiveregions.
What Are the FunctionalConsequences of MultipleNucleosome Assembly Pathways?
The bulk of nucleosome assembly occurs dur-ing DNA replication, and experiments withextracts defined a conserved set of CAFsthat support histone deposition specifically onreplicating DNA (Verreault et al. 1996). Theexpectation from these studies was that DNAreplication without RC nucleosome assem-bly would be lethal. Thus it was surprisingto find that null mutations for CAF compo-nents in budding yeast are viable (Enomoto &Berman 1998, Kaufman et al. 1997). CAF mu-tants show defects in telomere silencing andin DNA damage repair, but have normal pack-aging of chromatin. Thus other chromatin as-sembly pathways must compensate when RCassembly is defective. Indeed, other RC andRI nucleosome assembly activities have beenidentified. For example, the RCAF complexalso supports RC assembly and enhances theactivity of CAF. The lack of lethal phenotypesfor chromatin assembly mutants suggests thatany gaps in chromatin are filled by these otherpathways. As most of the yeast genome is
144 Henikoff · Ahmad
Ann
u. R
ev. C
ell D
ev. B
iol.
2005
.21:
133-
153.
Dow
nloa
ded
from
ww
w.a
nnua
lrev
iew
s.or
gby
Uni
vers
ity o
f So
uthe
rn C
alif
orni
a (U
SC)
on 0
4/03
/14.
For
per
sona
l use
onl
y.
ANRV255-CB21-06 ARI 1 September 2005 11:43
transcriptionally active, gaps left after replica-tion can be filled by transcription-coupled RIassembly. However, compensation also occursin organisms with more complex genomes.FASCIATA mutations eliminate CAF in Ara-bidopsis and cause some meristematic defects,but plants remain viable, indicating functionalredundancy (Kaya et al. 2001). Therefore,even transcriptionally silent heterochromatinis being duplicated by alternative pathways inthis mutant. A simple model is that alternativeassembly pathways are capable of working onany gapped templates, but normally do notact on regions where CAF rapidly completesnucleosome assembly in S phase. BlockingCAF function in mammalian cells stimulatesa DNA damage checkpoint (Hoek & Stillman2003, Nabatiyan & Krude 2004), consistentwith the idea that unpackaged DNA is nottolerated or is easily damaged.
Compensation between nucleosome as-sembly pathways can also explain why defectsin the CAF and HIR nucleosome assemblyactivities result in mis-targeting of the cen-tromeric Cse4p histone (Sharp et al. 2002).A compensation effect is also consistent withthe results of altering the expression of his-tone variants. Overexpression of centromerichistones in yeast, mammals, or Drosophila re-sults in its deposition in euchromatin, as ifextra CenH3 fills gaps in transcriptionally ac-tive chromatin (Ahmad & Henikoff 2002b,Collins et al. 2004, Shelby et al. 1997). Thissuggests that there is a balance between allnucleosome assembly pathways in these or-ganisms, even though they normally act ondistinct parts of the genome.
CAF mutants in both yeast and Arabidopsispackage DNA into chromatin and are viable;however, they show defective telomeric si-lencing and unstable developmental fates. Re-duction of an RbAp48 component of CAF inArabidopsis also causes spectacular epigeneticdefects (Hennig et al. 2003), although nulls forsome of these are lethal (Kohler et al. 2003).These phenotypes point to critical links be-tween the mode of nucleosome assembly andthe inheritance of epigenetic states. For exam-
ple, if each chromatin assembly complex re-cruits specific chromatin-modifying enzymes,epigenetic patterns would not be preservedwhen alternate assembly pathways duplicatea chromatin region.
Indeed, there is evidence that the propa-gation of heterochromatin and DNA repli-cation are linked in this way. Proper hete-rochromatic localization of HP1 depends inpart on binding to CAF-1 at heterochromaticreplication forks, as if HP1 is recruited forloading onto H3K9-methylated nucleosomes(Quivy et al. 2004). Moreover, CAF-1 de-livers histone H3 pre-methylated at lysine-9 to replication forks at sites of methylatedDNA (Sarraf & Stancheva 2004). Both mech-anisms for perpetuating heterochromatin re-quire that CAF-1 be used to duplicate thechromatin during replication. These consid-erations predict that CAF mutations in plantsare accumulating replacement H3.2 histonesin heterochromatin because RC assembly failsand gap-filling, using RI pathways, compen-sates. Whereas the phenotypes for elimina-tion of CAF-1 are more severe in mammaliancells, this may reflect its more critical role inepigenetic control of essential processes.
CONCLUSIONS
The differentiation of nucleosomes by incor-poration of variant histones must be as an-cient as the eukaryotes themselves, insofar asa CenH3-containing centromere is a definingfeature of eukaryotic chromosomes. The factthat centromeric chromatin is maintained inthe same chromosomal position for millionsof years, yet can shift spontaneously to anunrelated DNA sequence, is the most ex-treme example imaginable of faithful epi-genetic inheritance. Nevertheless, the rapidevolution of both the highly repetitive cen-tromeric satellite DNA and the CenH3 vari-ant that evidently adapts to it implies that cen-tromeric stability is dynamically maintained.Yet little is known about how centromericnucleosomes are assembled and propagatedthrough the cell cycle.
www.annualreviews.org • Histone Variants in Chromatin 145
Ann
u. R
ev. C
ell D
ev. B
iol.
2005
.21:
133-
153.
Dow
nloa
ded
from
ww
w.a
nnua
lrev
iew
s.or
gby
Uni
vers
ity o
f So
uthe
rn C
alif
orni
a (U
SC)
on 0
4/03
/14.
For
per
sona
l use
onl
y.
ANRV255-CB21-06 ARI 1 September 2005 11:43
Variant histones, such as CenH3s, are de-posited independently of replication, in con-trast to the bulk of chromatin, which isassembled by RC assembly of canonical hi-stones. The use of distinct assembly path-ways provides a simple means of differenti-ating chromatin, by targeting replacement ofnucleosomes throughout the cell cycle. Tran-scription itself appears to catalyze the tar-geting of the H3.3 replacement variant, andthe biochemistry of this process has begunto reveal insights into both variant deposi-tion and the possible maintenance of associ-ated active histone modifications. Disruptionof nucleosomes is not limited to transcription-coupled replacement, because nucleosomesare “evicted” from promoters upon geneactivation.
These dynamic processes of nucleosomereplacement and eviction can help accountfor the abundance and diversity of ATP-dependent chromatin remodeling complexes.Indeed, replacement of the H2AZ variant iscatalyzed by one such complex, and it seemslikely that other H2A variants are associatedwith dedicated members of the SWI/SNFfamily of ATPases. In this way, nucleosomeassembly and remodeling, formerly assumedfrom in vitro work to be distinct processes,
are now seen to be aspects of the same in vivoprocess.
Once incorporated into chromatin, mosthistone variants have distinct structural prop-erties that are likely to profoundly alter chro-matin. Differences between H2A and H2AZin the docking domain can affect nucleosomeintegrity, and the large globular domain ofmacroH2A is suspected to have enzymaticfunction. H3.3 is the exception, because theonly difference from H3 that is exposed in thenucleosome is a single tail residue of uncer-tain significance; rather, differences in post-translational modifications that are found todistinguish H3 from H3.3 are more likely toaffect nucleosome properties.
The study of histone variants and the mul-tiple biochemical processes that deposit theminto chromatin has led to new insights intochromatin dynamics. The effects of these pro-cesses on gene regulation and chromosomebehavior have yet to be elucidated, but theavailability of powerful new tools promises tochange that. Most importantly, the excitementgenerated by these new insights has fueled aresurgence of interest in histone variants afterdecades of relative neglect. We look forwardto the deeper insights into eukaryotic biologythat now appear to be just around the corner.
SUMMARY POINTS
1. Variants of histones H3 and H2A differentiate chromatin at centromeres, active genes,and heterochromatin.
2. Nucleosomes characterized by a special H3 variant identify the centromeres of everyeukaryotic chromosome.
3. The replacement histone, H3.3, marks actively transcribed loci by replication-independent nucleosome assembly.
4. Gene activation is accompanied by disassembly of a nucleosome at the promoter.
5. A chromatin remodeling machine replaces the conserved histone variant, H2AZ.
6. Epigenetically silenced chromatin is enriched or depleted in abundance of diverseH2A variants.
146 Henikoff · Ahmad
Ann
u. R
ev. C
ell D
ev. B
iol.
2005
.21:
133-
153.
Dow
nloa
ded
from
ww
w.a
nnua
lrev
iew
s.or
gby
Uni
vers
ity o
f So
uthe
rn C
alif
orni
a (U
SC)
on 0
4/03
/14.
For
per
sona
l use
onl
y.
ANRV255-CB21-06 ARI 1 September 2005 11:43
7. Variant structure can affect properties of chromatin.
8. The operation of multiple nucleosome assembly pathways has important implicationsfor nucleosome dynamics.
LITERATURE CITED
Abbott DW, Ivanova VS, Wang X, Bonner WM, Ausio J. 2001. Characterization of the stabilityand folding of H2A.Z chromatin particles: implications for transcriptional activation. J.Biol. Chem. 276:41945–49
Adkins MW, Howar SR, Tyler JK. 2004. Chromatin disassembly mediated by the histonechaperone Asf1 is essential for transcriptional activation of the yeast PHO5 and PHO8genes. Mol. Cell. 14:657–66
Ahmad K, Henikoff S. 2001a. Centromeres are specialized replication domains in heterochro-matin. J. Cell Biol. 153:101–10
Ahmad K, Henikoff S. 2001b. Modulation of a transcription factor counteracts heterochromaticgene silencing in Drosophila. Cell 104:839–47
Ahmad K, Henikoff S. 2002a. Epigenetic consequences of nucleosome dynamics. Cell 111:281–84
Ahmad K, Henikoff S. 2002b. Histone H3 variants specify modes of chromatin assembly. Proc.Natl. Acad. Sci. USA 99(Suppl.) 4:16477–84
Ahmad K, Henikoff S. 2002c. The histone variant H3.3 marks active chromatin by replication-independent nucleosome assembly. Mol. Cell 9:1191–200
Akey CW, Luger K. 2003. Histone chaperones and nucleosome assembly. Curr. Opin. Struct.Biol. 13:6–14
Allen MD, Buckle AM, Cordell SC, Lowe J, Bycroft M. 2003. The crystal structure ofAF1521 a protein from Archaeoglobus fulgidus with homology to the non-histone domainof macroH2A. J. Mol. Biol. 330:503–11
Alonso A, Mahmood R, Li S, Cheung F, Yoda K, Warburton PE. 2003. Genomic microar-ray analysis reveals distinct locations for the CENP-A binding domains in three humanchromosome 13q32 neocentromeres. Hum. Mol. Genet. 12:2711–21
Amor DJ, Bentley K, Ryan J, Perry J, Wong L, et al. 2004. Human centromere repositioning“in progress.” Proc. Natl. Acad. Sci. USA 101:6542–47
Angelov D, Molla A, Perche PY, Hans F, Cote J, et al. 2003. The histone variant macroH2Ainterferes with transcription factor binding and SWI/SNF nucleosome remodeling. Mol.Cell 11:1033–41
Angelov D, Verdel A, An W, Bondarenko V, Hans F, et al. 2004. SWI/SNF remodeling andp300-dependent transcription of histone variant H2ABbd nucleosomal arrays. EMBO J.23:3815–24
Belotserkovskaya R, Oh S, Bondarenko VA, Orphanides G, Studitsky VM, Reinberg D. 2003.FACT facilitates transcription-dependent nucleosome alteration. Science 301:1090–93
Bender W, Fitzgerald DP. 2002. Transcription activates repressed domains in the Drosophilabithorax complex. Development 129:4923–30
Bernstein BE, Liu CL, Humphrey EL, Perlstein EO, Schreiber SL. 2004. Global nucleosomeoccupancy in yeast. Genome Biol. 5:R62
www.annualreviews.org • Histone Variants in Chromatin 147
Ann
u. R
ev. C
ell D
ev. B
iol.
2005
.21:
133-
153.
Dow
nloa
ded
from
ww
w.a
nnua
lrev
iew
s.or
gby
Uni
vers
ity o
f So
uthe
rn C
alif
orni
a (U
SC)
on 0
4/03
/14.
For
per
sona
l use
onl
y.
ANRV255-CB21-06 ARI 1 September 2005 11:43
Black BE, Foltz DR, Chakravarthy S, Luger K, Woods VL Jr, Cleveland DW. 2004. Structuraldeterminants for generating centromeric chromatin. Nature 430:578–82
Blower MD, Karpen GH. 2001. The role of Drosophila CID in kinetochore formation, cell-cycle progression and heterochromatin interactions. Nat. Cell Biol. 3:730–39
Blower MD, Sullivan BA, Karpen GH. 2002. Conserved organization of centromeric chromatinin flies and humans. Developmental Cell 2:319–30
Boeger H, Griesenbeck J, Strattan JS, Kornberg RD. 2003. Nucleosomes unfold completelyat a transcriptionally active promoter. Mol. Cell. 11:1587–98
Boeger H, Griesenbeck J, Strattan JS, Kornberg RD. 2004. Removal of promoter nucleosomesby disassembly rather than sliding in vivo. Mol. Cell. 14:667–73
Brownell JE, Zhou J, Ranalli T, Kobayashi R, Edmondson DG, et al. 1996. Tetrahymenahistone acetyltransferase A: a homolog to yeast Gcn5p linking histone acetylation to geneactivation. Cell 84:843–51
Buchwitz BJ, Ahmad K, Moore LL, Roth MB, Henikoff S. 1999. A histone-H3-like protein inC. elegans. Nature 401:547–48
Chadwick BP, Willard HF. 2001. A novel chromatin protein, distantly related to histone H2A,is largely excluded from the inactive X chromosome. J. Cell Biol. 152:375–84
Chadwick BP, Willard HF. 2004. Multiple spatially distinct types of facultative heterochro-matin on the human inactive X chromosome. Proc. Natl. Acad. Sci. USA 101:17450–55
Chioda M, Spada F, Eskeland R, Thompson EM. 2004. Histone mRNAs do not accumulateduring S phase of either mitotic or endoreduplicative cycles in the chordate Oikopleuradioica. Mol. Cell. Biol. 24:5391–403
Clarkson MJ, Wells JR, Gibson F, Saint R, Tremethick DJ. 1999. Regions of variant histoneHis2AvD required for Drosophila development. Nature 399:694–97
Collins KA, Furuyama S, Biggins S. 2004. Proteolysis contributes to the exclusive centromerelocalization of the yeast Cse4/CENP-A histone H3 variant. Curr. Biol. 14:1968–72
Cooper JL, Henikoff S. 2004. Adaptive evolution of the histone fold domain in centromerichistones. Mol. Biol. Evol. 21(9):1712–18
Cosgrove MS, Boeke JD, Wolberger C. 2004. Regulated nucleosome mobility and the histonecode. Nat. Struct. Mol. Biol. 11:1037–43
Costanzi C, Pehrson JR. 1998. Histone macroH2A1 is concentrated in the inactive X chro-mosome of female mammals. Nature 393:599–601
Davey CA, Sargent DF, Luger K, Maeder AW, Richmond TJ. 2002. Solvent mediated inter-actions in the structure of the nucleosome core particle at 1.9 A resolution. J. Mol. Biol.319:1097–113
Dhillon N, Kamakaka RT. 2000. A histone variant, Htz1p, and a Sir1p-like protein, Esc2p,mediate silencing at HMR. Mol. Cell. 6:769–80
Drewell RA, Bae E, Burr J, Lewis EB. 2002. Transcription defines the embryonic domainsof cis-regulatory activity at the Drosophila bithorax complex. Proc. Natl. Acad. Sci. USA99:16853–58
Earnshaw WC, Rothfield N. 1985. Identification of a family of human centromere proteinsusing autoimmune sera from patients with schleroderma. Chromosoma 91:313–21
Enomoto S, Berman J. 1998. Chromatin assembly factor I contributes to the maintenance, butnot the re-establishment, of silencing at the yeast silent mating loci. Genes Dev. 12:219–32
Faast R, Thonglairoam V, Schulz TC, Beall J, Wells JR, et al. 2001. Histone variant H2A.Z isrequired for early mammalian development. Curr. Biol. 11:1183–87
Fan JY, Gordon F, Luger K, Hansen JC, Tremethick DJ. 2002. The essential histone variantH2A.Z regulates the equilibrium between different chromatin conformational states. Nat.Struct. Biol. 9:172–76
Fernandez-Capetillo O, Lee A, Nussenzweig M, Nussenzweig A. 2004. H2AX: the histoneguardian of the genome. DNA Repair 3:959–67
Formosa T, Ruone S, Adams MD, Olsen AE, Eriksson P, et al. 2002. Defects in SPT16 orPOB3 (yFACT) in Saccharomyces cerevisiae cause dependence on the Hir/Hpc pathway:polymerase passage may degrade chromatin structure. Genetics 162:1557–71
Freitas MA, Sklenar AR, Parthun MR. 2004. Application of mass spectrometry to the iden-tification and quantification of histone post-translational modifications. J. Cell Biochem.92:691–700
Gribnau J, Diderich K, Pruzina S, Calzolari R, Fraser P. 2000. Intergenic transcription anddevelopmental remodeling of chromatin subdomains in the human β-globin locus. Mol.Cell 5:377–86
Hanlon SE, Lieb JD. 2004. Progress and challenges in profiling the dynamics of chromatin andtranscription factor binding with DNA microarrays. Curr. Opin. Genet. Dev. 14:697–705
Hayashi T, Fujita Y, Iwasaki O, Adachi Y, Takahashi K, Yanagida M. 2004. Mis16 and Mis18 arerequired for CENP-A loading and histone deacetylation at centromeres. Cell 118:715–29
Hendzel MJ, Davie JR. 1990. Nucleosomal histones of transcriptionally active/competent chro-matin preferentially exchange with newly synthesized histones in quiescent chicken ery-throcytes. Biochem. J. 271:67–73
Henikoff S, Ahmad K, Platero JS, van Steensel B. 2000. Heterochromatic deposition of cen-tromeric histone H3-like proteins. Proc. Natl. Acad. Sci. USA 97:716–21
Henikoff S, Furuyama T, Ahmad A. 2004. Histone variants, nucleosome assembly and epige-netic inheritance. Trends Genet. 20:320–26
Hennig L, Taranto P, Walser M, Schonrock N, Gruissem W. 2003. Arabidopsis MSl1 is requiredfor epigenetic maintenance of reproductive development. Development 130:2555–65
Hoek M, Stillman B. 2003. Chromatin assembly factor 1 is essential and couples chromatinassembly to DNA replication in vivo. Proc. Natl. Acad. Sci. USA 100:12183–88
Hogga I, Karch F. 2002. Transcription through the iab-7 cis-regulatory domain of the bitho-rax complex interferes with maintenance of Polycomb-mediated silencing. Development129:4915–22
Howman EV, Fowler KJ, Newson AJ, Redward S, MacDonald AC, et al. 2000. Early disruptionof centromeric chromatin organization in centromere protein A (CenpA) null mice. Proc.Natl. Acad. Sci. USA 97:1148–53
Jackson JD, Gorovsky MA. 2000. Histone H2A.Z has a conserved function that is distinct fromthat of the major H2A sequence variants. Nucleic Acids Res. 28:3811–16
Jackson V. 1978. Studies on histone organization in the nucleosome using formaldehyde as areversible cross-linking agent. Cell 15:945–54
Janicki SM, Tsukamoto T, Salghetti SE, Tansey WP, Sachidanandam R, et al. 2004. Fromsilencing to gene expression: real-time analysis in single cells. Cell 116:683–98
Jenuwein T, Allis CD. 2001. Translating the histone code. Science 293:1074–80Kaufman PD, Kobayashi R, Stillman B. 1997. Ultraviolet radiation sensitivity and reduction
of telomeric silencing in Saccharomyces cerevisiae cells lacking chromatin assembly factor-I.Genes Dev. 11:345–57
Kaya H, Shibahara KI, Taoka KI, Iwabuchi M, Stillman B, Araki T. 2001. FASCIATA genesfor chromatin assembly factor-1 in Arabidopsis maintain the cellular organization of apicalmeristems. Cell 104:131–42
www.annualreviews.org • Histone Variants in Chromatin 149
Ann
u. R
ev. C
ell D
ev. B
iol.
2005
.21:
133-
153.
Dow
nloa
ded
from
ww
w.a
nnua
lrev
iew
s.or
gby
Uni
vers
ity o
f So
uthe
rn C
alif
orni
a (U
SC)
on 0
4/03
/14.
For
per
sona
l use
onl
y.
ANRV255-CB21-06 ARI 1 September 2005 11:43
Keener J, Dodd JA, Lalo D, Nomura M. 1997. Histones H3 and H4 are components ofupstream activation factor required for the high-level transcription of yeast rDNA byRNA polymerase I. Proc. Natl. Acad. Sci. USA 94:13458–62
Keith KC, Baker RE, Chen Y, Harris K, Stoler S, Fitzgerald-Hayes M. 1999. Analysis of pri-mary structural determinants that distinguish the centromere-specific function of histonevariant Cse4p from histone H3. Mol. Cell. Biol. 19:6130–39
Kobor MS, Venkatasubrahmanyam S, Meneghini MD, Gin JW, Jennings JL, et al. 2004.A protein complex containing the conserved Swi2/Snf2-related ATPase Swr1p depositshistone variant H2A.Z into euchromatin. PLoS Biol. 2:E131
Kohler C, Hennig L, Bouveret R, Gheyselinck J, Grossniklaus U, Gruissem W. 2003. Ara-bidopsis MSI1 is a component of the MEA/FIE Polycomb group complex and required forseed development. EMBO J. 22:4804–14
Kristjuhan A, Svejstrup JQ. 2004. Evidence for distinct mechanisms facilitating transcriptelongation through chromatin in vivo. EMBO J. 23:4243–52
Krogan NJ, Keogh MC, Datta N, Sawa C, Ryan OW, et al. 2003. A Snf2 family ATPasecomplex required for recruitment of the histone H2A variant Htz1. Mol. Cell 12:1565–76
Kusch T, Florens L, Macdonald WH, Swanson SK, Glaser RL, et al. 2004. Acetylation by Tip60is required for selective histone variant exchange at DNA lesions. Science 306:2084–87
Lee CK, Shibata Y, Rao B, Strahl BD, Lieb JD. 2004. Evidence for nucleosome depletion atactive regulatory regions genome-wide. Nat. Genet. 36:900–5
Lee MS, Garrard WT. 1991. Transcription-induced nucleosome ‘splitting’: an underlyingstructure for DNase I sensitive chromatin. EMBO J. 10:607–15
Lo AW, Magliano DJ, Sibson MC, Kalitsis P, Craig JM, Choo KHA. 2001. A novel chromatinimmunoprecipitation and array (CIA) analysis identifies a 460-kb CENP-A-binding neo-centromeric DNA. Genome Res. 11:448–57
Luger K, Mader AW, Richmond RK, Sargent DF, Richmond TJ. 1997. Crystal structure ofthe nucleosome core particle at 2.8 A resolution. Nature 389:251–60
Malik HS, Henikoff S. 2001. Adaptive evolution of Cid, a centromere-specific histone inDrosophila. Genetics 157:1293–98
Malik HS, Henikoff S. 2003. Phylogenomics of the nucleosome. Nat. Struct. Biol. 10:882–91Malik HS, Vermaak D, Henikoff S. 2002. Recurrent evolution of DNA-binding motifs in the
Drosophila centromeric histone. Proc. Natl. Acad. Sci. USA 99:1449–54McKittrick E, Gafken PR, Ahmad K, Henikoff S. 2004. Histone H3.3 is enriched in covalent
modifications associated with active chromatin. Proc. Natl. Acad. Sci. USA 101:1525–30Mellone BG, Allshire RC. 2003. Stretching it: putting the CEN(P-A) in centromere. Curr.
chromatin from the ectopic spread of silent chromatin. Cell 112:725–36Mizuguchi G, Shen X, Landry J, Wu WH, Sen S, Wu C. 2004. ATP-driven exchange of histone
H2AZ variant catalyzed by SWR1 chromatin remodeling complex. Science 303:343–48Monson EK, de Bruin D, Zakian VA. 1997. The yeast Cac1 protein is required for the stable
inheritance of transcriptionally repressed chromatin at telomeres. Proc. Natl. Acad. Sci.USA 94:13081–86
Moore LL, Roth MB. 2001. HCP-4, a CENP-C-like protein in Caenorhabditis elegans, is re-quired for resolution of sister centromeres. J. Cell Biol. 153:1199–208
150 Henikoff · Ahmad
Ann
u. R
ev. C
ell D
ev. B
iol.
2005
.21:
133-
153.
Dow
nloa
ded
from
ww
w.a
nnua
lrev
iew
s.or
gby
Uni
vers
ity o
f So
uthe
rn C
alif
orni
a (U
SC)
on 0
4/03
/14.
For
per
sona
l use
onl
y.
ANRV255-CB21-06 ARI 1 September 2005 11:43
Morrison AJ, Highland J, Krogan NJ, Arbel-Eden A, Greenblatt JF, et al. 2004. INO80 andgamma-H2AX interaction links ATP-dependent chromatin remodeling to DNA damagerepair. Cell 119:767–75
Nabatiyan A, Krude T. 2004. Silencing of chromatin assembly factor 1 in human cells leads tocell death and loss of chromatin assembly during DNA synthesis. Mol. Cell. Biol. 24:2853–62
Nagaki K, Cheng Z, Ouyang S, Talbert PB, Kim M, et al. 2004. Sequencing of a rice centromereuncovers active genes. Nat. Genet. 36:138–45
Nagy PL, Cleary ML, Brown PO, Lieb JD. 2003. Genomewide demarcation of RNA poly-merase II transcription units revealed by physical fractionation of chromatin. Proc. Natl.Acad. Sci. USA 100:6364–69
Palmer DK, O’Day K, Trong HL, Charbonneau H, Margolis RL. 1991. Purification of thecentromere-specific protein CENP-A and demonstration that it is a distinctive histone.Proc. Natl. Acad. Sci. USA 88:3734–38
Palmer DK, O’Day K, Wener MH, Andrews BS, Margolis RL. 1987. A 17-kD centromereprotein (CENP-A) copurifies with nucleosome core particles and with histones. J. CellBiol. 104:805–15
Park YJ, Dyer PN, Tremethick DJ, Luger K. 2004. A new fluorescence resonance energy trans-fer approach demonstrates that the histone variant H2AZ stabilizes the histone octamerwithin the nucleosome. J. Biol. Chem. 279:24274–82
Pereira SL, Reeve JN. 1998. Histones and nucleosomes in Archaea and Eukarya: a comparativeanalysis. Extremophiles 2:141–48
Pina B, Suau P. 1987. Changes in histones H2A and H3 variant composition in differentiatingand mature rat brain cortical neurons. Dev. Biol. 123:51–58
Quivy JP, Roche D, Kirschner D, Tagami H, Nakatani Y, Almouzni G. 2004. A CAF-1 de-pendent pool of HP1 during heterochromatin duplication. EMBO J. 23:3516–26
Rangasamy D, Berven L, Ridgway P, Tremethick DJ. 2003. Pericentric heterochromatin be-comes enriched with H2A.Z during early mammalian development. EMBO. J. 22:1599–607
Rank G, Prestel M, Paro R. 2002. Transcription through intergenic chromosomal memoryelements of the Drosophila bithorax complex correlates with an epigenetic switch. Mol.Cell. Biol. 22:8026–34
Reinke H, Horz W. 2003. Histones are first hyperacetylated and then lose contact with theactivated PHO5 promoter. Mol. Cell 23:1599–607
Rogakou EP, Pilch DR, Orr AH, Ivanova VS, Bonner WM. 1998. DNA double-strandedbreaks induce histone H2AX phosphorylation on serine 139. J. Biol. Chem. 273:5858–68
Santisteban MS, Kalashnikova T, Smith MM. 2000. Histone H2A.Z regulates transcriptionand is partially redundant with nucleosome remodeling complexes. Cell 103:411–22
Sarraf SA, Stancheva I. 2004. Methyl-CpG binding protein MBD1 couples histone H3 methy-lation at lysine 9 by SETDB1 to DNA replication and chromatin assembly. Mol. Cell.15:595–605
Schwabish MA, Struhl K. 2004. Evidence for eviction and rapid deposition of histones upontranscriptional elongation by RNA polymerase II. Mol. Cell. Biol. 24:10111–17
Sharp JA, Franco AA, Osley MA, Kaufman PD. 2002. Chromatin assembly factor I and Hirproteins contribute to building functional kinetochores in S. cerevisiae. Genes Dev. 16:85–100
Shelby RD, Monier K, Sullivan KF. 2000. Chromatin assembly at kinetochores is uncoupledfrom DNA replication. J. Cell Biol. 151:1113–18
www.annualreviews.org • Histone Variants in Chromatin 151
Ann
u. R
ev. C
ell D
ev. B
iol.
2005
.21:
133-
153.
Dow
nloa
ded
from
ww
w.a
nnua
lrev
iew
s.or
gby
Uni
vers
ity o
f So
uthe
rn C
alif
orni
a (U
SC)
on 0
4/03
/14.
For
per
sona
l use
onl
y.
ANRV255-CB21-06 ARI 1 September 2005 11:43
Shelby RD, Vafa O, Sullivan KF. 1997. Assembly of CENP-A into centromeric chromatinrequires a cooperative array of nucleosomal DNA contact sites. J. Cell Biol. 136:501–13
Solomon MJ, Varshavsky A. 1985. Formaldehyde-mediated DNA-protein crosslinking: a probefor in vivo chromatin structures. Proc. Natl. Acad. Sci. USA 82:6470–74
Stoler S, Keith KC, Curnick KE, Fitzgerald-Hayes M. 1995. A mutation in CSE4, an es-sential gene encoding a novel chromatin-associated protein in yeast, causes chromosomenondisjunction and cell cycle arrest at mitosis. Genes Dev. 9:573–86
Sugimoto K, Yata H, Muro Y, Himeno M. 1994. Human centromere protein C (CENP-C) is aDNA-binding protein which possesses a novel DNA-binding motif. J. Biochem. 116:877–81
Sullivan BA, Karpen GH. 2004. Centromeric chromatin exhibits a histone modification pat-tern that is distinct from both euchromatin and heterochromatin. Nat. Struct. Mol. Biol.11:1076–83
Suto RK, Clarkson MJ, Tremethick DJ, Luger K. 2000. Crystal structure of a nucleosome coreparticle containing the variant histone H2A.Z. Nat. Struct. Biol. 7:1121–24
Tagami H, Ray-Gallet D, Almouzni G, Nakatani Y. 2004. Histone H3.1 and H3.3 complexesmediate nucleosome assembly pathways dependent or independent of DNA synthesis. Cell116:51–61
Talbert PB, Bryson TD, Henikoff S. 2004. Adaptive evolution of centromere proteins in plantsand animals. J. Biol. 3:18
Talbert PB, Masuelli R, Tyagi AP, Comai L, Henikoff S. 2002. Centromeric localization andadaptive evolution of an Arabidopsis histone H3 variant. Plant Cell 14:1053–66
Turner BM, Birley AJ, Lavendar J. 1992. Histone H4 isoforms acetylated at specific lysineresidues define individual chromosomes and chromatin domains in Drosophila polytenenuclei. Cell 69:375–84
Urban MK, Zweidler A. 1983. Changes in nucleosomal core histone variants during chickendevelopment and maturation. Dev. Biol. 95:421–28
van Attikum H, Fritsch O, Hohn B, Gasser SM. 2004. Recruitment of the INO80 complex byH2A phosphorylation links ATP-dependent chromatin remodeling with DNA double-strand break repair. Cell 119:777–88
van Daal A, Elgin SC. 1992. A histone variant, H2AvD, is essential in Drosophila melanogaster.Mol. Biol. Cell 3:593–602
Vermaak D, Ahmad K, Henikoff S. 2003. Maintenance of chromatin states: an open-and-shutcase. Curr. Opin. Cell Biol. 15:266–74
Vermaak D, Hayden HS, Henikoff S. 2002. Centromere targeting element within the histonefold domain of Cid. Mol. Cell. Biol. 22:7553–61
Vermaak D, Wade PA, Jones PL, Shi YB, Wolffe AP. 1999. Functional analysis of the SIN3-histone deacetylase RPD3-RbAp48-histone H4 connection in the Xenopus oocyte. Mol.Cell. Biol. 19:5847–60
Verreault A, Kaufman PD, Kobayashi R, Stillman B. 1996. Nucleosome assembly by a complexof CAF-1 and acetylated histones H3/H4. Cell 87:95–104
Waterborg JH. 1990. Sequence analysis of acetylation and methylation in two histone H3variants of alfalfa. J. Biol. Chem. 265:17157–61
Wieland G, Orthaus S, Ohndorf S, Diekmann S, Hemmerich P. 2004. Functional complemen-tation of human centromere protein A (CENP-A) by Cse4p from Saccharomyces cerevisiae.Mol. Cell. Biol. 24:6620–30
Wolffe AP. 1992. Chromatin: Structure and Function. San Diego: AcademicWorkman JL, Abmayr SM. 2004. Histone H3 variants and modifications on transcribed genes.
Proc. Natl. Acad. Sci. USA 101:1429–30
152 Henikoff · Ahmad
Ann
u. R
ev. C
ell D
ev. B
iol.
2005
.21:
133-
153.
Dow
nloa
ded
from
ww
w.a
nnua
lrev
iew
s.or
gby
Uni
vers
ity o
f So
uthe
rn C
alif
orni
a (U
SC)
on 0
4/03
/14.
For
per
sona
l use
onl
y.
ANRV255-CB21-06 ARI 1 September 2005 11:43
Yang CH, Tomkiel J, Saitoh H, Johnson DH, Earnshaw WC. 1996. Identification of overlap-ping DNA-binding and centromere-targeting domains in the human kinetochore proteinCENP-C. Mol. Cell. Biol. 16:3576–86
Yu L, Gorovsky MA. 1997. Constitutive expression, not a particular primary sequence, is theimportant feature of the H3 replacement variant hv2 in Tetrahymena thermophila. Mol. Cell.Biol. 17:6303–10
www.annualreviews.org • Histone Variants in Chromatin 153
Protein Translocation by the Sec61/SecY ChannelAndrew R. Osborne, Tom A. Rapoport, and Bert van den Berg � � � � � � � � � � � � � � � � � � � � � � � � � � � 529
