In eukaryotes, DNA is organized into chromatin, a dynamic structure that enables DNA to be accessed for processes such as transcription, replication and repair. To form, maintain or alter this organization according to cellular needs, histones, the main protein component of chromatin, are deposited, replaced, exchanged and post-translationally modified. Histone variants, which exhibit specialized deposition modes in relation to the cell cycle and possibly particular chromatin regions, add an additional level of complexity in the regulation of histone flow. During their metabolism, from their synthesis to their delivery for nucleosome formation, the histones are escorted by proteins called histone chaperones. In the present chapter we summarize our current knowledge concerning histone chaperones and their interaction with particular histones based on key structural properties. From a compilation of selected histone chaperones identified to date, we discuss how they may be placed in a network to regulate histone dynamics. Finally, we provide working models to explain how H3 variants, deposited on to DNA using different histone chaperone machineries, can be transmitted or lost through cell divisions.
Histones, small basic proteins, are the main proteic component of chromatin and are among the most highly conserved proteins in eukaryotes. They are involved in organizing eukaryotic DNA at the level of the fundamental unit of chromatin, the nucleosome . The nucleosome core particle contains an octamer of histones comprising a tetramer of (H3–H4)2 flanked by two dimers of H2A–H2B, around which 147 bp of DNA are wrapped . This organization can have an impact on the regulation of all cellular processes operating on DNA such as transcription, replication, recombination and repair. The structural properties of nucleosomes themselves and of the higher-order chromatin organization can be modulated by a large variety of covalent PTMs (post-translational modifications) of histones, such as acetylation, phosphorylation and methylation . In addition to histone modifications, histone variants that can participate in marking specific chromatin domains should be considered in the context of epigenetic inheritance . Histone variants differ at the level of the primary sequence and regulation of expression. The amino acid differences can range from a few positions to large protein domains and can confer specific properties on nucleosomes . An important aspect of histone variant dynamics relates to their mode of incorporation which can represent a crucial step with major implications for cell fate and stability of expression programmes.
During cell metabolism, histone variant deposition, replacement, exchange or PTM will contribute to preserve or change the chromatin organization according to cellular needs. Although the vast majority of histones present in the cell are associated with DNA, the ‘DNA-free’ fraction is important to consider. Newly synthesized histones have to be translocated to the nucleus and targeted to the required location. Conversely, old or damaged histones have to be discarded and replaced. In addition, cellular processes involving DNA may require transient histone eviction and replacement. One major example of such histone dynamics is observed at the replication fork (Figure 1) . Although in most cases histones are only transiently in a ‘DNA-free’ state, they can persist longer-term in such a state, as illustrated by histone storage during Xenopus laevis oogenesis, which ensures their provision for early development. Since histones are highly basic and positively charged proteins, their presence in excess or unregulated within the cell may lead to deleterious effects through promiscuous interactions and aggregate formation. Thus under most circumstances, when histones are not in association with DNA, they are bound to proteins called histone chaperones . These escort proteins not only avoid unintended interactions of histones with other factors, but also provide means to control histone variant supply and incorporation into chromatin. In this way, histone chaperones and their cognate histone variants are crucial in fundamental processes operating on DNA and in marking specific regions of chromatin.
In the present chapter, we first emphasize the importance of histone chaperones by describing their various roles in histone metabolism and by depicting their interaction properties with histone variants. With a compilation of representative histone chaperones, we will then discuss how mammalian histone H3 variants are deposited by specific histone chaperones and transmitted or lost through cell divisions.
A choice of histone chaperones for histone variants
The term ‘molecular chaperone’ was originally used by Ron Laskey to describe the function of the nuclear protein, nucleoplasmin, which plays several important roles in macromolecular assembly of nucleosomes during early development of the amphibian, Xenopus . Now, by definition, histone chaperones are factors that associate with histones and stimulate a reaction involving histone transfer without being part of the final product. A fundamental property, shared by all histone chaperones, is histone transfer. This can be revealed using a ‘nucleosome-reconstitution’ assay with purified DNA and histones. Notably, all histone chaperones are not necessarily involved in histone deposition on to DNA in vivo and thus they can play other important roles in histone dynamics.
The diversity of histone chaperones
Histone chaperones are involved in various functions such as histone transfer from one chaperone to another, to enzymes using histones as a substrate, on to DNA (deposition) or off DNA (eviction) and in storage of histone pools during development or buffering transient histone overload .
Most histone chaperones are conserved throughout evolution. To list the most representative histone chaperones, we used here three parameters (Table 1): class I are chaperones that alone can bind and transfer histones without necessarily involving additional partners, for example, Asf1 (anti-silencing function 1); class II are complexes consisting of several chaperone subunits, for example, the CAF-1 (chromatin assembly factor-1) complex; and class III are complexes that combine both chaperones and other proteic components providing histone-binding capacity within large enzymatic complexes, for example, Arp4 (actin-related protein-4) in the INO80 chromatin-remodelling complex. Of note, one particular histone chaperone, the human RbAp48 (retinoblastoma-associated protein 48), should be considered as a ‘multiclass chaperone’, since it can fulfil criteria defining all three chaperone categories, depending on the context. Indeed, RbAp48 can function on its own, yet it is also one of the three subunits of the CAF-1 complex, and of several histone-remodelling and histone-modifying enzymatic complexes.
In addition to the interaction with histones, histone chaperones can associate with other key proteins thereby providing an interface between histone metabolism and chromatin dynamics or DNA metabolism. These interactions allow histones to be provided at the right time and the right place when and where they are needed. For example, during replication or repair, the chaperone CAF-1 is targeted to sites of DNA synthesis through its interaction with PCNA (proliferating cell nuclear antigen) . During transcription, Spt6, a histone H3–H4 chaperone involved in transcription elongation, associates with the transcriptional machinery via its interaction with RNA polymerase II . Therefore histone chaperones, through interaction with various proteins, can exhibit a variety of roles in histone metabolism.
Specific chaperones for specific variants
In mammals, a large number of histone variants have been identified for H2A, H3 and H2B, whereas only a single variant is known for H4 until now . Histone variants are separated into two classes depending on their mode of incorporation into chromatin: the canonical (or replicative) histones and the replacement histones. Genes encoding canonical histones are often organized into multicopy clusters, whereas those encoding replacement variants usually occur as single or only a few copies. Canonical histone variants exhibit a peak of expression during S-phase for a major incorporation event during replication in a DSD (DNA-synthesis-dependent) manner. This regulation of expression ensures the provision of new histones to fulfil the requirement for nucleosome assembly on the two daughter strands in the wake of the replication fork. Replacement histone variants do not exhibit the peak of expression during S-phase observed for canonical histones. They can be incorporated in a DSI (DNA-synthesis-independent) manner at various times during the cell cycle depending on the variant.
Histone chaperones can be distinguished on the basis of their histone binding selectivity, which differentiates chaperones that preferentially interact with H3–H4 from those that prefer H2A–H2B. Furthermore, particular chaperones show even further selectivity towards specific histone variants (Table 1). Of note, preferential chaperone–histone interactions do not always involve a strict binding specificity; rather, it may be subject to regulation and should be evaluated in a context-dependent manner. For example, NASP (nuclear autoantigenic sperm protein) has been reported to interact with the linker histone H1 in mammalian cells , however, it does also bind H3–H4 . Similarly, the H2A–H2B chaperone Nap1 (nucleosome assembly protein 1) binds linker histone H1m (B4) in X. laevis eggs, as assessed by immunoprecipitation . How a given chaperone shows selectivity toward specific histones often remains an enigma.
A structural basis for the chaperone–histone interaction
The structural basis underlying a given chaperone–histone interaction, although not sufficient to account for selective interactions with specific histone variants, has proven informative in many ways. Given the basicity of histones, it is not surprising that long acidic stretches are found in numerous histone chaperones, such as nucleoplasmin, yeast Asf1, yeast Nap1, the Spt16 subunit of human FACT and nucleolin . Interaction between acidic chaperones and basic histones thus can lead to charge neutralization, but this does not provide a basis for a specific binding to a distinct histone. Notably, charge neutralization can even be dispensable for histone chaperone activity, as exemplified by mammalian Asf1, which lacks an acidic tail . Therefore, acidic regions may serve to enhance histone interaction and may be an intrinsic feature in one species, for example in a chaperone containing an acidic tail, whereas in other species the interaction is regulated through other features (for example PTM, as in the case of polyglutamylation of Nap1 in Drosophila melanogaster) .
The intimate interaction between chaperones and their cognate histones and the variety of histone metabolic pathways lead us to propose that each histone chaperone is positioned within an escort network regulating histone traffic. Careful evaluation will be needed to determine the extent to which histone chaperones can substitute for one another within this chaperone network and how the chaperones participate in regulating histone traffic in co-ordination with cellular processes involving chromatin reorganization.
How the various histone variants are incorporated into chromatin and which histones chaperones are involved is thus a critical matter. We discuss below recent insights concerning H3 variants in mammals.
Chaperones involved in the deposition of H3 variants
In mammals, five histone H3 variants have been identified up to now: two canonical variants, H3.1 and H3.2, and three replacement variants, H3.3, CENP-A (centromere protein-A) and H3t (Figure 2A). H3.1 is a variant only found in mammals and it exhibits a single amino acid difference with the other canonical variant H3.2 at position 96. Expression of the two canonical variants peak in S-phase and these variants are mainly incorporated during replication. The replacement variant H3.3 has four amino acid differences with H3.2 (at positions 31, 87, 89 and 90) and five with H3.1 (with an additional difference at amino acid 96). H3.3 is expressed in proliferating cells at all stages of the cell cycle and in quiescent cells. Its incorporation has been linked to active transcription  and major chromatin rearrangements during development (see below). H3t is a tissue-specific variant expressed in testis and it exhibits four amino acid differences with H3.1, H3.2 and H3.3 (at positions 24, 71, 98 and 111). CENP-A is a more divergent variant specifically found in centromeric chromatin and its presence is critical for centromere function . How these variants mark specific chromatin states and how they are incorporated into chromatin has been a subject of intensive investigation. The chaperone involved in tissue-specific H3t deposition is not known. To explore how H3.3, H3.1 and CENP-A are assembled into chromatin in human cells, a common successful strategy has been to purify pre-assembled histones H3.3, H3.1 and CENP-A from the nuclear extract fraction of human HeLa cells expressing epitope-tagged H3.3, H3.1 or CENP-A respectively. Isolation of the proteic complexes associated with each variant by immuno-affinity purification and comparative analysis by mass spectrometry enabled identification of histone chaperones that showed specificity for individual H3 variants (Figure 2B). The effective role of the identified chaperones in the deposition of a particular H3 variant has been analysed, and how the presence of a specific variant can be transmitted through cell division is discussed.
CAF-1 and histone variant H3.1, HIRA and histone variant H3.3
Analysis of the complex associated with H3.1 revealed the presence of several histone chaperones for H3–H4: CAF-1 (with its three subunits, p150, p60 and RbAp48), the two isoforms of Asf1 (Asf1a and Asf1b) and NASP . CAF-1 represents the prototype of a chaperone that promotes nucleosome assembly in a DSD pathway during replication and UV repair [6,21]. CAF-1 plays a major role in H3.1 deposition, as demonstrated in vitro using chromatin assembly assays  and in vivo . The deposition of H3.3 into chromatin has been generally associated with regions that are transcriptionally active, suggesting a role for this variant in facilitating transcription. The isolation of H3.3 complexes identified several H3–H4 chaperones: HIRA (histone regulator A), the two isoforms of Asf1 (Asf1a and Asf1b), NASP and RbAp48. HIRA, originally described as a chaperone involved in a DSI nucleosome assembly pathway using the X. laevis egg extract model system in vitro , plays a critical role in H3.3 deposition as described below . In addition to the situation during transcription in somatic cells, H3.3 deposition occurs in other physiological instances . Indeed, chromatin remodelling events, essential for sexual reproduction, involve H3.3 dynamics in Drosophila and mice. Nuclear transfer experiments of Xenopus oocytes have further put forward a potential role for H3.3 in the epigenetic memory of active gene states. Intriguingly, H3.3 localized at interphase telomeres in mouse pluripotent ES (embryonic stem) cells. In this latter case HIRA is not required for H3.3 incorporation and a role for another chaperone to complex DAXX (death-associated protein) ATRX (α-thatassaemia.mental retardation syndrom protein) was proposed to deposit H3.3 at telomeres [25,26].
In contrast with CAF-1 and HIRA that are specific for H3.1 and H3.3 complexes respectively, Asf1 is present in both complexes. Asf1 may participate as an H3.1–H4 or H3.3–H4 histone donor for CAF-1 and HIRA respectively, which in turn deposits histones on to DNA . The function of NASP found in both H3.1 and H3.3 complexes, needs to be elucidated. Thus the histone chaperone CAF-1 appears as a major player in the deposition of the canonical H3 variants in a DSD pathway during replication and repair, whereas HIRA contributes to the deposition of H3.3 in a DSI pathway in the context of transcription.
HJURP (Holliday junction-recognizing protein) and the centromeric H3 variant CENP-A
The histone H3 variant CENP-A marks centromeres and is crucial to ensure proper chromosome segregation . Centromeres form a platform upon which the kinetochore, the multiprotein complex that mediates spindle microtubule attachment during mitosis, is assembled. Intriguingly, the site of centromere formation is not governed by DNA sequence, except in budding yeast where a specific centromeric sequence has been defined. In mammalian cells, the expression of CENP-A peaks in G2-phase and its loading on to centromeres is restricted to a discrete time window during the cell cycle, late telophase-early G1-phase, and therefore proceeds in a replication-independent fashion .
The isolation of soluble CENP-A complexes revealed two known histone chaperones, RbAp48 and Npm1 (nucleophosmin 1). However, only HJURP leads to a defect in CENP-A incorporation and stability at centromeres, indicating a dominant role for HJURP in CENP-A deposition [29,30]. The presence of RbAp48 in H3.3, H3.1 and CENP-A complexes probably reflects its capacity to bind H4; however, its exact role in the assembly line of these variants remains to be clearly elucidated.
Thus HJURP in mammals represents a new histone chaperone specific for CENP-A and its proper incorporation at centromeres at late telophase-early G1-phase.
H3.3 or CENP-A-containing nucleosomes: properties and inheritance
The canonical H3.1 variant is incorporated into chromatin mainly during replication and repair through DSD pathways, whereas the deposition of replacement variants is achieved through DSI pathways, mainly during late telophase-early G1-phase for CENP-A and at any phases of the cell cycle for H3.3. How to envisage the specific marking of a region and its transmission through multiple cell divisions under these circumstances (Figure 3)?
H3.3 associated with transcriptionally active chromatin regions is enriched in active histone marks [31,32]. Furthermore, nucleosomes that contain H3.3 seem to be less stable than those that contain H3.1 , although the two variants exhibit only five amino acid differences at positions that are not predicted as critical in nucleosome structure based on crystal studies . The extent to which this difference of nucleosome stability depends on the differential modification status of the nucleosomes, the presence of other variants, such as H2A.Z , or inherent differences in their structural properties remains to be established. Regardless, in vivo, these properties suggest that H3.3 nucleosomes are more dynamic or amenable to displacement, in particular during transcription.
The variant CENP-A marks the site of centromere identity. Recent studies suggest that CENP-A nucleosomes are unusual and that these peculiarities might provide a means to define centromeric properties . Additional evidence suggests that, like H3.3 nucleosomes, CENP-A nucleosomes are easier to disassemble in vitro than canonical nucleosomes . In Drosophila, Dalal et al.  described a ‘hemisome’ that consists of one molecule each of CenH3, H4, H2A and H2B. Furthermore the potential ability to stabilize positive supercoiling in a CENP-A-containing particle has raised an active debate .
In H3.3 or CENP-A-containing chromatin regions, the deposition of H3.1 during DNA replication results in the dilution of the non-replicative histone variants. Active marks on H3.3 might recruit factors that facilitate the modification of neighbouring H3.1 to ensure the inheritance of an active state in a dominant fashion. Transcription-dependent incorporation promoted by HIRA might counterbalance loss of H3.3. For CENP-A, replication of centromeric DNA might result in the formation of unusual CENP-A-containing nucleosomes in an intermediate state that contains one-half of the amount of CENP-A. These unusual CENP-A-containing nucleosomes will be fully replenished with new CENP-A molecules later in the cell cycle at late telophase-early G1-phase by HJURP.
In the present chapter, we have listed representative histone chaperones on the basis of some of their known structural and functional properties to highlight their diversity. We emphasize that histone chaperones are not necessarily deposition factors and that several other important functions in histone metabolism should be considered. We have underlined the complexity of the choice of histone variants that can mark particular chromatin domains and that are loaded on to DNA through dedicated processes involving histone chaperones. How and why a given chaperone shows selectivity towards specific histones remains a major challenge for our understanding of stability/plasticity of chromatin and associated epigenetic marks.
• The basic repeated unit of chromatin is the nucleosome.
• The nucleosome core particle consists of a histone octamer comprising a tetramer of (H3–H4)2 flanked by two dimers of H2A–H2B around which 147 bp of DNA is wrapped.
• Histone chaperones are defined as factors that associate with histones and stimulate a reaction involving histone transfer without being part of the final product.
• Two types of histone variants are distinguished: the canonical (or replicative) variants and the replacement variants.
• The incorporation of canonical and replacement histone variants is DSD and DSI respectively.
• The deposition of histone variants on to DNA takes advantage of the use of various specific histone chaperones.
We are grateful to A. Cook for critical reading of this manuscript. We apologize to colleagues whose work could not be cited due to a restriction on the number of references. The work in the authors’ laboratory is supported by la Ligue Nationale contre le Cancer (Equipe labellisée Ligue 2010), PIC Programs (‘Rétinoblastome’ and ‘Réplication, Instabilité chromosomique et cancer’), the European Commission Network of Excellence Epigenome ( LSHG-CT-2004-503433), the European Commission ITN FP7-PEOPLE-2007 ‘Image DDR’ and FP7-PEOPLE-2008 ‘Nucleosome 4D’, ACI-2007-Cancéropôle IdF ‘Breast cancer and Epigenetics’, ANR ‘CenRNA’ NT05-4_42267, ANR ‘FaRC’ PCV06_142302 and ANR ‘ECenS’ ANR-09-BLAN-0257-01, INCa ‘GepiG’.
- © The Authors Journal compilation © 2010 Biochemical Society